Citation
Placebo psychotherapy and change in anxiety, mood and adjustment.

Material Information

Title:
Placebo psychotherapy and change in anxiety, mood and adjustment.
Creator:
Goldstein, Herbert, 1936-
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Adjectives ( jstor )
Anxiety ( jstor )
Control groups ( jstor )
Covariance ( jstor )
Mathematical dependent variables ( jstor )
Placebo effect ( jstor )
Placebos ( jstor )
Pretreatment ( jstor )
Psychotherapists ( jstor )
Psychotherapy ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
This item is presumed in the public domain according to the terms of the Retrospective Dissertation Scanning (RDS) policy, which may be viewed at http://ufdc.ufl.edu/AA00007596/00001. The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
22567387 ( ALEPH )
13868254 ( OCLC )

Downloads

This item has the following downloads:


Full Text









PLACEBO PSYCHOTHERAPY AND CHANGE IN

ANXIETY, MOOD AND ADJUSTMENT



















By

HERBERT GOLDSTEIN










A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY












UNIVERSITY OF FLORIDA April, 1965






/













ACKNOeLEDGMENTS


This study would not have been completed had it not been for the foresight, wi sdom and encouragement of many people to whom the wrl ter is indebted aend grateful.

A special measure of gratitude Is due Dr. Audrey S. Schumacher, chairman of the supervisory committee, for her genuine interest, her untiring patience and her constant support and guidance. The assistance and encouragement of the remainder of the sujervisory committee, Dr. R. J. Anderson, Dr. B. Barger, Dr. J. J. Wright and Dr. V. A. Hines, is warmly accepted and greatly appreciated.

The writer wishes to acknowledge the staff and office workers of the Gulfport, Mississippi Veterans' Administration Hospital who gave unsparingly of their time, their cooperation and their resources. A particular note of gratitude is extended to three people who specifically asked to remain unmentikrned, but without whom the progress of this study would certainly have been impeded.

The writer wishes to thank his parents for their faith, for their investment and for wanting someday to say: "there's my son, the Doctor."

Most of all, the writer thanks his wife, Alene, for her understanding, her support, her guidance and her remarkable patience which sustained the writer in his efforts to complete this study.






iI













TABLE OF CONTENTS


Page

ACKNOWLEDGMENTS . . . . . . . . .. . 1

LIST OF TABLES. . . . . . . . .. . Iv

INTRODUCTION. . . . .. . . . . . . . .

The Present Study. ... ... . ... . .... .... 4


METHOD ........ . . .... ..... . . ... 13

Subjects . . . . . . ........ . . . 14
Means of Measurement .................. 16
Procedure ................. .... ..... 24


RESULTS ...... . .. .... ........ ..... . 2

DISCUSSION. ............ ..... .. ..... 62

APPENDICES. ................. . . . . ....... 75

A INSTRUCTIONS TO SUBJECTS. ....... ..... 76
8 SAMPLE ANXIETY ADJECTIVE CHECK LIST ......... 80 C SAMPLE MOOD ADJECTIVE CHECK LIST. ....... .. ... 82
S SAMPLE ADJUSTMENT SCORE. ..... . . . . . . 85
E SAMPLE OF THE MODIFICATION OF THE EXPECTATION
Q-SORT . . ............. . ....... . . 90
F SAMPLE PERSONAL INFORMATION QUESTIONNAIRE ..... 92
G SAMPLE LETTER SENT TO ALL SUBJECTS WHO SPOKE
INTO THE TAPE RECORDER ............. 94
H SUBJECT RAW DATA .................. 96

BIBLIOGRAPHY ........... . ....... ...... . 103













LIST OF TABLES


Tabl e Page

1 T ;L EXPERIMENTAL DESIGN ........ . . . . 13

2 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE
DEMONSTRATING THE RANDOM DISTRIBUTION OF
AGE ...... . . .. ..... .... 5... 15

3 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE
DEMONSTRATING THE RANDOM DISTRIBUTION OF
I.Q. . . . . . ........................ 16

4 THE TELVE MOOD FACTORS OF THE MACL AND
THE ADJECTIVES CONSTITUTING EACH MOOD
FACTOR. . . . . . . . . . . ... 19

5 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE ZUCKERMAN ADJECTIVE CHECK LIST. . .. . . 30

5A PRETREATMENT MEANS (Xi), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE
ZUCKERMAN ADJECTIVE CHECK LIST ... . ... ... . . 30

6 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL AGGRESSION MOOD. ..... . .... 31

6A P.&ETREATMENT MEANS (i ), POSTTREATMENT MEANS
(x2) AND MEAN CHANGE SCORES (O) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE
MACL AGGRESSION MOOD. ................ 31

7 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL CONCENTRATION MOOD ..... ... . 32

7A PRETREATMENT MEANS (), POSTTREATMENT MEANS
(2) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE
MACL CONCENTRATION MOOD ................ 32

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL DEACTIVATION MOOD .......... ... 33



Iv










LIST OF TABLES--Contnued


Table Page

8A PRETREATMENT MEANS (X), POSTTREAThENT MEANS
(" ) AND MEAN CHANGE SCORES (D) FOR EACH TRATHENT CONDITION OF THE DESIGN ON THE
MACL DEACTIVATION MOOD. ............. . . 33

9 SUMMARY TALE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL SOCIAL AFFECTION MOOD ........... 34

9A PRETREATMENT MEANS (XI), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE
MACL SOCIAL AFFECTION MOOD ............... 34

10 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL ANXIETY MOOD ............... 35

IOA PRETREATMENT MEANS (), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE
MACL ANXIETY MOOD ............... . . 35

II SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE rOR THE MACL DEPRESSION MOOD. .... ........ . . 36

I IA PRETREATMENT MEANS (il), POSTTREATMENT MEANS
(x2) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE
MACL DEPRESSION MOO .................. 36

12 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL EGOTISM MOOD .............. 37

12A PRETREATMENT MEANS (XI), POSTTREATMENT MEANS
(x2) AND MEANS CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL EGOTISM MOOD ..... .......... . . 37

13 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL PLEASANTNESS MOOD ............ 38

13A PRETREATMENT MEANS (X1), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE
MACL PLEASANTNESS MOOD. .......... ....... 38










LIST OF TABLES--ontinued


Table Page

14 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL ACTIVATION MOOD.. .. .... ....... 39

1L4A PRETREATMENT MEANS ( X), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE
MACL ACTIVATION MOO ... ............ .. . 39

15 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE.
FOR THE MACL NONCHALANCE MOOD ....... ..... 40

15A PRETREAThENT MEANS (X), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE
MACL NONCHALANCE MOOD ............... 40

to SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL SKEPTICISM MOOD . ...... . ... 41

16A PRETREATMENT MEANS (XI), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH TREAThENT CONDITION OF THE DESIGN ON THE
MACL SKEPTICISM MOOD. ........... ... 41

17 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL STARTLE MOOD ...... ...... . 42

17A PRETREATMENT MEANS (Xi), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE
MACL STARTLE MOOD .................. 42

18 SUMMAY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE ADJUSTMENT SCORE ............ . . 43

18A PRETREAThENT MEANS (X), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (O) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE
ADJUSTMENT SCORE ................... 43

19 SUMMARY OF THE DEPENDENT VARIABLES AND THE SIGNIFICANCE LEVELS AT WHICH THEY
DISTINGUISHED BETWEEN THE INDEPENDENT
VARIABLES ... .............. . .. . 4


vi











LIST OF TABLES--Continued


Table Page

20 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANDOM DISTRIBUTION OF
MPI INTROVERSION SCORES ................ 50

21 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANDOM DISTRIBUTION OF
MPI NEUROTICIS 4 SCORES. ..... .. ........ 51

22 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EIGHT OF THE DEPENDENT
VARIABLES AND SCORES ON THE BETA INTELLIGENCE TEST. THE DATA ARE PRESENTED FOR EACH OF THE
EIGHT TREATMENT CONDITIONS. ............. 53

23 PEARSON PRODUCT-hOMENT CORRELATIONS BETWEEN SUBJECTS- CHANGE SCORES FROM PRE- TO POSTTREAThENT TESTING ON EACH OF THE DEPENDENT
VARIABLES AND SCORES ON THE MPI INTROVERSION (LOW SCORES)-EXTROVERSION (HIGH SCORES). THE
DATA ARE PRESENTED FOR EACH OF THE EIGHT
TREATMENT CONDITIONS. .............. . 55

24 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND SUBJECTS' AGES. THE DATA ARE
PPESENTED FOR EACH OF THE EIGHT TREATMENT
CONDITIONS .. . . ........ . .. ... 56

25 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT
VARIABLES AND SCORES ON THE MPI NEUROTICISM
SCALE. THE DATA ARE PRESENTED FOR EACH OF
THE EIGHT TREATMENT CONDITIONS. ... .. ....... 5

26 MULTIPLE REGRESSION COEFFICIENTS BETWEEN AMOUNT OF SPEECH AND DESIRE FOR PSYCHOTHERAPY PRESENTED
FOR EACH INDEPENDENT VARIABLE (EXPERIMENTAL
DATA ONLY) .. . ............... 59





vil










LIST OF TAbLES--ContinueJ


Table Page

27 MULTIPLE REGRESSION COEFFICIENTS BETWEEN THE
VARIABLES LISTED 01 THE VERTICAL AXIS AND FIVE
DEPENDENT VARIABLE CHANGE SCORES FOR THE EXPERIMENTAL GROUP, THE DATA ARE PRESENTED
SEPARATELY FOR EACH OF THE INDEPENDENT VARIABLES IN ADDITIV4 TO THE TOTAL EXPERIMENTAL
GROUP . . . . . . ................... 61









































vilii














I NTRODUCTION


The extensive literature dealing with processes and systems of psychotherapy generally makes the explicit assumption that behavioral changes following psychotherapeutic encounter are a function of the Intended efforts of the psychotherapist. That is, behavioral modification results from the direct and specific efforts, manipulations or techniques of the psychotherapist. Within the framework of Ro9er's theory (1961), for example, the therapist approves and accepts the client and as a resuit the client comes to accept himself. The patient in psychoanalysis achieves personal satisfaction and comfort as the psychoanalyst interprets the patient's repressed psychosexual conflicts and the patient establishes a mature sexual adjustment (Fenichel, 1945). Shoben (1953) describes the goal of psychotherapy as the alleviation of symptoms in addition to the increase in a patient's affective comfort. Thus, in all psychotherapy, "specific factors" (therapist be:-vi~rs) aifiedly produce intended effects (personal comfort of the patient).

Eysenck (1952, 1961) states, and Rosenthal and Frank suggest (1956) that in addition to the intended efforts of the psychotherapist (specific factors), there are other, nonspecific factors in psychotherapy which significantly contribute to the outcome of treatment. "Nonspecific factors" are loosely defined as those "placebo effects," situational events and conditions in addition to the intended efforts of the therapist which reportedly function in all treatment situations (Shapiro, 1964). They serve to cloud Interpretation of "real"' effects of psychological

I







2


treatment (Rosenthal and Frank, 1956) for they produce manifest effects which are Indistinguishable from the effects produced by the Intentional treatment prescribed by the particular theory of psychotherapy.

Strupp (1962) indicates that because these nonspecific factors have been largely neglected in research in psychotherapy, the extent of their Influence is undetermined. Many significant theorists and researchers (e.g., Rogers, 1961) expound on the efficacy of their psychotherapeutic systems, although nonspecific factors, which may greatly contribute to the final outcome of psychological treatment are not effectively accounted for, controlled or measured. Understandably, successful outcome of psychotherapy Is viewed as proof of the effectiveness of psychotherapy. However, it has not been demonstrated that personality and behavioral modifications which occur as a result of psychotherapy are entirely (or even largely) due to the psychotherapist's specific techniques. Instead, the actual outcome of treatment is a function of both the specifc and nonspecific events occurring during psychotherapy.

Eysenck (1961), after reviewing and summarizing the literature, shows that actual psychotherapy does not achieve higher recovery rates then those reported from ordinary life experiences and the nonspecific effects of routine medical treatmert. After demonstrating that naaspecif c treatment produces effects c~9marable with actual psychotherapy, Eysenck implies that nonspecific factors in psychotherapy are essential factors responsible for bringing about successful psychotherapeutic change. Findings and statements such as these clarify the necessity for research in this area.








3

Effects of treatment with drug placebos are those nonspecific effects given the greatest attention In the literature. Based on his review of the literature, Shapiro (1964) states that placeao effects are subtle and common to all treatment situations and are the most important nonspecific factors in psychological treatment. He defines the placebo effect as the "psychologic, physiologic, or psychophysiologic effect of any medication or procedure given with therapeutic intent which is independent of or minimally related to the effects of the medication or to the specific effects of the procedure and which operates through a psychologic mechanism" (p. 298). In addition, the administration of this medication or procedure must ae recognized as legitimately therapeutic by the patient.

Rosenthal and Frank (1956) also refer to the placebo effect as a type of nonspecific event in psychological treatment and they report that since research in psychotherapy has not dealt directly with the placebo effect, the nature of its influence is undetermined. However, these authors speculate that it occurs with considerable regularity, and they stress the need to demonstrate through research that observed effects of psychotherapy are due to the therapist's techniques and exist separate from nonspecific effects.

Borgatta (1959) expands on the importance of the placebo effect by reporting that f there is no evidence that an agent is the cause of a given outcome, and there is also no evidence that a placebo is the cause of a given outcome, an appropriate statement is that there is no evidence that the agent is more efficient than the placebo in bringing








4

about the outtcom. Further, there is also no evidence that the placebo is not as effective as the agent in bringing about the outcome" (p. 331).

The literature is replete with references confirming the efficacy of nonspecific effects of drug placebos (see reviews by Shapiro, 1964 and Hoging field, 1963). However, despite the demonstrated effectiveness of drug placebos, and the alleged importance of the placebo effect in psychological treatment, adequately controlled studies evaluating the placebo effect in psychotherapy are conspicuously lacking (Rosenthal and Frank, 1956 and Strupp, 1962). One is currently forced, therefore, to rely on implication, inference, and analogy when constructing evidence of, and predictive hypotheses pertaining to, the influence of nonspecific placebo factors in psychotherapy.


The Present Study

"As Withering's foxglove was made more potent when digitalis was isolated from the concoction, and after it was recognized that only the dropsy of congestive heart failure was benefited, so will psychotherapy become more potent after the placebo effect is isolated and dissected free from the psychotherapeutic process" (Shapiro, 1964, p. 85).

As long as both specific and nonspecific (placebo) factors operate together during the treatment, one may never accurately determine the real contribution of the placebo effect to psychotherapy. Since nonspecific placebo factors reportedly function unbeknown to the psychotherapist, since their presence is reportedly a factor in the outcome of all psychotherapy, and since they remain a nebulous and poorly defined factor in psychotherapy, it is clear that the extraction of these placebo effects








5

from an investigation of actual psychotherapy is a complex, if not impossible task. Thus, a demonstration of the genuine effects of psychotherapy free of placebo effects seem never to have been made.

In this study an attempt is made to investigate a form of "psychotherapeutic encounter in which specific factors could not contribute to the outcome of treatment. That Is, any changes which accrue as a resuit of the treatment are not due to a psychotherapist's behavior, technique or manipulations during treatment. Change, instead, must ae attributed to the presence of the placebo factor and must ae referred to as a placebo effect.

A psychotherapeutic encounter" can take place when a patient reports to a tape recorder for one "therapy session" believing a psychotherapist will listen and respond to his talking. Measures of change in such "affective comfort" (Shoben, 1953) as moods and anxiety taken immediately after this treatment will indicate the effect of the treatment. 8orgatta (1959) suggests, satirically, that this technique be used to demonstrate the relative ineffectiveness of actual psychotherapy. Slack (1960), on the other hand, demonstrated that therapeutically inaccessible juvenile delinquents could be introduced to psychotherapy oy initially having them speak into a tape recorder. Eventually these juveniles were introduced to the therapist who had been giving them feedback from the tapes, and actual psychotherapy was Initiated. Martin, Lundy and Lewin (1960) evaluated the reinforcing effects of three degrees of therapist communication on the affectively toned verbalizations of their subjects. The group Intended to have virtually no communication from the therapist spoke into a tape recorder as if they were speaking to a psychotherapist.








6


In the present study, one half of the subjects will speak into the tape recorder while the other half will wait an identical length of time. Since both groups will be given the instructions that they are about to speak to a psychotherapi st through the tape recorder this will permit a comparison between a group given both the promise and the gesture of help and a group given only the promise of help. Changes in mood and anxiety in the group which does spea speak ma be vied as a demonstration of the nonspecific placebo effect in a single session of psychological treatment. The effects of this treatment should be utilized as a base line against which further research may be compared to demonstrate that an actual psychotherapeutic interview produces changes different from, or greater than the "placebo psychotherapy."

Shapiro (1964) reports that important features of a therapeutic relationship which are responsible for the placebo effect are basically elements of (1) the patient, (2) the doctor, and (3) the treatmet situation; in addition to a fourth factor: the doctor-patient relationship. That is, there are certain characteristics of each of these components of treatment which contribute to a nonspecific placebo effect.



Shapiro, in his review of drug research (1964), points out that many Individual differences of patients have been evaluated to assess possible relationships with the placebo effect. Such characteristics as sex, age, intelligence and diagnosis have been observed repeatedly in order to ascertain If a certain type of person responds more readily to placebo factors. However, as Shapiro reports, this body of research is equivocal and often contradictory. For example, Lasagna at al. (1954)








7


administered a placebo to a group of postoperative patients to relieve reported pain. Those patients reporting a decrease in pain after receiving the placebo had a mean age five years greater than the age of the nonreactors (p.< .05). On the other hand, Tibbetts and Hawkings (1956) found that patients who responded in two drug placebo situations were younger than the nonresponders. In studies by Kornetsky g~.g. (1957) and Abramson 1_4. (1955) no meaningful relationship existed between age and drug placebo responsiveness. Similarly, level of intelligence as a patient variable has produced contradictory results. In the Lasagna study reported above, intelligence did not significantly distinguish the placebo reactor from the nonreactor. On the other hand, Tibbetts and Hawkings (1956) found the placebo reactor to be less Intelligent than the nonreactor, while Abramson arlJ. (1955) found the placebo reactor to have a higher verbal intellectual ability and a lower performance intellectual ability.

Leiberman (19,4) reports that Eysenck has correlated neuroticism with the need for conformity in an individual's perceptions and judgments. Therefore, he contends that "In group situations, where an element of interpersonal pressure activates the need for conformity, the subjects with neurotic tendencies respond with placebo effects." In fact, Eysenck argues that a subject's degree of neuroticism is a consistent factor for picking out high placebo reactors. Eysenck (1901) further speculates that since introverts form conditioned responses more quickly than extroverts, and since he reports that response to a placebo is a learned or conditioned response, he expects introverts to respond more readily to placebo factors (p. 630).











However, despite the logical consistency of Eysenck's speculations, the literature concerning the placebo responsiveness of neurotics and introverts Is equivocal. For example, Fischer and Dlin (1956), using psychiatric patients, demonstrated that neurotics were the most responsive subjects in a drug placebo situation. On the other hand, Kurland (1958) disputes this conclusion on the basis of his finding that no difference exists between the placebo reactivity o psychotics and nonpsychotics. Joyce (1959) in a drug study with medical students, indicates that extroverts are the most responsive subjects, while Laverty (1958) demonstrates that introverted subjects are the most responsive in drug treatment situations.

In this study, age, Intelligence, introversion and neuroticism are the patient characteristics whose contribution to the placebo effect will be evaluated.

(2) T ct

The nature of the treater is reportedly an important factor for encouraging or discouraging placebo effects (Shapiro, 1964; Uhlenruth sltal., 1959). For example, Glidemen gj gi. (1957) demonstrated that a male physician gained a 70 per cent cure rate in a treatment for which a female nurse could achieve only a 25 per cent cure rate. He also has shown that if the administering physician is perceived as an expert healer, the response of the patient to treatment will be greater than if the physician is believed to be a medical quack. These effects were obtained, conceivably, because the patient maintains certain cognitive expectancies about the nature of treaters, and when these expectancies are











dissonant with the treatment situation, the effectiveness of treatment decreases (see Festinger and Bramel In Bachrach, 1962).

Simi arly, folk medicine in the Spanish Southwestern United States (Jaco, 958) and witchcraft healing in many parts of the world (Frank, 19i.) f ourish and prevent the establishment of modern medical program, at east in part, because scientifically trained physicians do not meet oca: expectancies about the nature of treaters. That is, to these people scientific physicians are cognitive y dissonant In the ro e of "doctor."' Conversely, the effectiveness of the native witch doctor is enhanced because he meets his ciientele's expectances concerning the nature of treaters.

In the present research, verbal instructions to the patients will estab ish two different types of doctors psychotherapistts. One i. the type of psychotherapist the patient expects and the other does not conform to his expectations (see Appendix A). This wi 1 produce two distinct groups of subjects regarding their personal percetion of their psychotherapists.

(3) The treatment situation.

Foulds (1958) and Shapiro (1904) refer to many studies which demonstrate that a treater's bias can significantly influence treatment outcome. For example, Hofiing (1955) and Frank (1958), in studies, demonstrate that positive staff attitudes and biases toward a particular piacebo or specific treatment increases the effectiveness of that treatment. Lyerly et a. (i962) in a drug study, have shown that subjects receiving instructions a)oropriate to the expected effects of ch oral hydrate, regardiess of the actua drug or placebo administered, reported subjective











changes appropriate to the choral hydrate. Abramson et ael. (1955) administered tap water to a group of subjects who believed the liquid to be LSD-25. A percentage of these subjects responded to the tap water with psychomimetic behavior as if the water actuaily were LSD-25.

Frank (i9,i) suggests that a patient's faith" in a set of treatment operations is often ail that is necessary to produce positive change, and this is so even when these treatment operations are scientifically absurd. Woif (ip5$) states that a placebo is effective when a patient has a "conviction" that a certain effect wiil follow.

In the present study, an attempt wi ;I be made to manipulate expectations concerning the va ue of the 'placebo psychotherapy" and the amount of profit to be anticipated from the treatment. One group will receive a positive set concerning that which can be gained from the treatment and the other wi I I receive a more neutral set (see Appendix A).

To summarize, the hypotheses under investigation in this study are:

I. Those subjects who receive the piacebo psychotherapy (speak into the tape recorder) wi I show
more marked change in a "therapeutic direction"
on measures of anxiety and mood than those subjacts who simply wait a comparable length of time.
However, neither group wils show a change in genera level of psychological adjustment in this
single treatment.

II. The subjects who believe that their therapist Is
an individual congruent with their expectations
wi i shokw more marked change in a "therapeutic
'direction"' on post measures of anxiety and mood
than those subjects whose expectations are dissonant with the therapist's description. Neither
group wili show a change in general levei of psychological adjustment after this sing treatment.











iIl. Those subjects receiving a positive set regarding the value and outcome of the treatment will
show more marked change in a "therapeutic direction" on post measures of anxiety and mood
than those subjects who receive a more neutral set.
However, neither group will show a change In general level of psychological adjustment in this
single treatment.

IV. The measured individual subject differences of
age, intelligence, level of introversion and of neuroticism will not identify the subjects who
respond best to this placebo situation. That is, there will be no differences between the placebo responsiveness of subjects in terms of their age, intelligence, level of Introversion-extroversion
and level of neuroticism.

May the "placebo psychotherapy, as introduced above, be legitimately referred to as a placebo treatment situation? There is a strong analogy between this treatment and that which exists when a physician administers a drug placebo. The physician who treats a patient with a placebo pill does so in the same treatment setting, giving the same directions, and offering the same encouragement as if the pill were pharmacologically sound. The only changed characteristic, and that which defines the situation as exclusively nonspecific, is the absence of the actual drug--that agent in the treatment situation which produces the specific effect. Reactions to this placebo are recorded as "placebo effects" although they may appear to be a result of a specific drug which is clearly absent from the treatment. Similarly, in a sham interview such as that mentioned above, the patient is given the same directions, the same encouragement and is placed in the same treatment setting as if it were an actual psychotherapeutic situation. The only changed characteristic, and that characteristic which defines the situation as nonspecific in this case, is t absence of the psychotherapist, that is, the only agent








12


through which specific effects may result. Reactions to this placebo situation can be viewed as "placebo effects." Without knowledge of the nature of the treatment, the effects may appear to result from the specific behavior or techniques of a psychotherapist. However, his absence from the treatment necessarily means that all effects are nonspecific, or placebo.

Since there is no psychotherapist present, can this situation be thought analogous to a therapeutic interview. The patient is talking to a therapist who exists only in the patient's fantasy and is instituted by the experimenter. Apfelbaum (1958) points out that Individuals tend to create, in part, interpersonal experience and that ongoing interpersonal process in therapy Is primarily a function of a patient's personal transferences. A patient, in utilizing past experience and partial cues, projects the image of his psychotherapist, and thereby unknowingly perceives him in part through his projective capacity. Consequently, the therapist is perceived largely as a function of the patient himself. The attitudes a patient holds toward his alleged therapist in the "shad' or "placebo" interview suggested above may be controlled and manipulated by experimental design. The therapist exists entirely in fantasy and as a function of the patient's projective capacity and description of him as reported _y the experimenter. The patient, however, in effect, establishes an intrapersonal relationship which he believes to be, and which appears to have the qualities of, an interpersonal psychotherapeutic relationship.

Thus, the sham interview will be regarded as both a placebo treatment and a therapeutic interview.














METHOD


The study involved 96 subjects ho vere placed into eight treatment conditions as diagra;med below in Table 1.


TABLE I

THE EXPERIMENTAL DESIGN

Exerimental GroControl Group
Congruent incongruent Congruent Incongruent
Therapist Therapist Therapist Therapist
Instructions Instructions Instructions Instructions

Pos ti ve
set
concerning treatment
outcome n 12 n = 12 n = 12 n = 12


Neutral
set
concern ng
treatment
outcome n = 12 n 12 n = 12 n 12 n 48 n = ib



The technique and procedure of establishing these treatment conditions, and the way in which the data were collected are fully explained in the Procedure section. The tests which were administered to the subjects and the use to which the test information was put in the design are described in the Means of Measurement section.





13











ts

The subjects for the study were 96 neuropsychiatric hospital patients in the Gulfport, Mbsilssippl, Veterans Administration Hospital. They are ail in-patient residents of the Rehabilitation and Placement Service Ward which houses only those patients judged capable of returning to the community. The ward offers a vocational testing service and is professionasly staffed by a psychiatrist, two counseling psycho ogists and a nwmber of nursing personnel.

The 96 subjects were drawn from the ward during 48 separate sessions extending from August 29, 1964, to October 19, i164. The sample consists of white adult males. There is a wide range of prehospita ization vocations. None were In psychotherapy when seen as subjects for the study. However, it was known that some of these patients had been in group or individual psychotherapy earlier in their hospitalization. The identification of former psychotherapy patients was not made, but there is no reason to suspect a nonrandom assignment of these patients to the different experimental treatment conditions. Similarly, although it was known that some of the sample were receiving drugs for their psychiatric conditions, these subjects were not identified and their random asJgnment to the different experiment treatment conditions was expected.

The mean age of the sample was 38.7&; range 18 to 58, standard deviation 7.41. Below Is a summary tabie of an analysis of variance (Lindquist, 1953) demonstrating that subject ages are randomly distributod across ai treatment t conditions.

The mean I.Q. of the sam le, as measured by the Beta Intelligence Test (Kellogget al., '4) was 98.54, which is within the normal range.







15


TABLE 2

SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANDOM DISTRIBUTION OF AGE

Source ss df V F

A (experimental vs. control) 25.01 1 25.01 .44 B (congruent vs incongruent) 3.76 1 3.76 .07 C (positive vs. neutral) 147.51 1 147.51 2.58 AB 25.01 1 25.01 .44 AC 11.35 1 11.35 .20 BC 25.01 1 25.01 .44 ABC 3.76 1 3.76 .07

w Calls 5.030.08 57.16

Total 5,271.49 95


(An F of 3.95 is significant at the .05 level.)

The range of I.Q.'s was from 73 to 129; standard deviation 12.14. Below is a summary table of an analysis of variance demonstrating that I.Q. Is randomly distributed across all treatment conditions.

The mean educational level of the sample is 10.9 years, and

ranged from 6 years to 20 years. The mean of the length of current hospitalization to the closest month is 27 months and ranged from three days to 14 years. The psychiatric diagnoses carried by these patients were made by a psychiatrist at the time of admission to the hospital. There is a wide range of psychiatric diagnoses, which, for the purposes of this study are divided into 56 psychotic and 40 nonpsychotic. Because all subjects were randomly assigned to the treatment conditions, there is








16


TABLE 3

SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANDOM DISTRIBUTION OF I.,Q,.

Source ss df V F

A (experimental vs.control) 96.00 1 96.00 .2 8 (congruent vs. Incongruent) .37 1 .37 .00 C (positive vs. neutral) 2.04 1 2.04 .01 AB 165.38 1 165.38 1.07 AC 3.37 1 3.37 .02 BC 60.67 I bb.7 .43 ABC 140.17 1 140.17 .90

w Cells 15565iZ3 155.20

Total 14,131.83 95


(An F of 3.95 Is significant at the .05 level.)

reason to expect that educational level, length of hospitalization and diagnosis are randomly distributed across treatment conditions. The psychotic patients all have shown good remission of their symptoms and all patients serving as subjects in the study were judged by a psychiatrist to be sufficiently psychologically sound to return to the community and be employed.


Means of Measurement

Two tests upon which change in feelings of personal comfort can be monitored (the Anxiety Adjective Check List and the Mood Adjective Check List), and one test evaluating general level of psychological








17


adjustment (The Adjustment Score) were administered twice to each subject

--once prior to and once immediately after the treatment. Changes in scores on these tests were uied to evaluate changes as described in the four hypotheses.

(I) The Active Check List (AACL (see Appendix 8).

The Anxiety Adjective Check List (AACL) (1960) has been shown by Zuckerman to be a "quick measure of anxiety level." It was chosen for this study because of its demonstrated validity and reliability and the fact that the instructions may be modified to permit sensitive eva uation of changes in the level of manifest anxiety over short periods of time.

The check list is an empirically developed pool of 61 adjectives with varying affective connotations. Twenty-one adjectives were identified, on the basis of an item analysis, as discriminating between psychiatric patients rated high on anxiety and normal control subjects. Of

these items, eleven are anxiety negative adjectives and ten are anxiety positive. Subjects may obtain a score from 0 to 21 on the check ist. Zuckerman has developed two forms. One instructs subjects to denote how they feel 'in general and the other asks how they fee "'today."' Aithouch standardized on co lege students, Zuckerman reports no differences in performance due to age, sex or level of education.

In two studies (960, 1962) ZucKerman reports internal and testretest reiabilities for the "in general form to be significant at greater than the .001 level. However, the "today' form, although slmiiarty internally reliable, had low test-retest reliability (r = .08, p <( .ut). These results were anticipated and suggest that the today form is sensitive to short-term fiuctuations in anxiety. Validity was











evaluated by administrations of the today form to a class in general psycho loy. The check iist was given on 10 nonexamination days and three examination days. The examination-day administrations resulted in a significantly higher reported level of anxiety than the nonexamination-day administrations (p <.00u5). The today form, with instructions modified to read "right now," wiil be used in this study.

(2) The Mood Adective Check List C (see Appendix C).

Nowlis (1956, 960, 196i) assumed mood to be a multidimensional

characterization of a person's feeling or behavior which Is accessibe to self report. The MACL resulted from a series of seven factor analytic studies by him (1960). In those seven studies Nos demonstrated that 40 adjectives, from his origina. pool of 2u0, have relatively consistent and high loadings on twelve separate factors. These 40 adjectives constitute the HACL which is scored separately for each of the twelve factors (which are the identified moods). Further, the tweive factors (moods) are divided into three groups: those "factors most consistently identified," those "factors identified fairly consistently" and "factors tentatively identified." Below are the twelve mood factors and the particular adjectives constituting the factors.

Certain specific moods on the MACL have particular relevance for the purpose of this study because they would be expected to change in a certain direction as a result of successful psychotherapy. Aggression and Deactivation wouid be expected to decrease, while e Social Affection would be expected to increase. Similarly, after successful psychotherapy the moods of Anxiety and Depression would be expected to decrease and Pleasantness and Activation wouid be expected to increase. At a lower











TABLE 4

THE TWELVE MOOD FACTORS OF THE MACL AND THE ADJECTIVES CONSTITUTING EACH MOOD FACTOR

Adjectives with high loadins on the factor

Factors Host Consistently identified:
AGGRESSION angry, bold, defiant, rebellious.
CONCENTRATION concentrating, earnest, engaged in tlhoght, serious.
DEACTIVATION drowsy, earnest, tired.
SOCIAL AFFECTION affectionate, forgiving, kindly, warmhearted.

Factors Identified
Fairly Consistently:
ANXIETY apprehensive, clutched up, fearful, insecure.
DEPRESSION blue, lonely, regretful.
EGOTISM boastful, cocky, egotistIc, self-centered.
PLEASANTNESS elated, lighthearted, overjoyed, pleased.

Factors Tentativel y
Identified:
ACTIVATION active, energetic, vigorous.
NONCHALANCE nonchalant, playful, witty.
SKEPTICISM skeptical, suspicious.
STARTLE startled, shocked.



level of certainty, because of difficulty in Interpreting its meaningfulness in a psychotherapeLtic context, Concentration and Nonchalance might

be expected to Increase, while Egotism, Skepticism and Startle might be

expected to decrease after treatment. In addition, changes in Skepticism

after the treatment may reflect the extent to which this treatment was beIleved to be real.

In his research developing the MACL, Nowlis used male .nd female

college students and U. S. Navy personnel in a variety of experimental

(drugs, films) and field situations. For example, one validation procedure

that Nowlis reports (1961) involved the administration of a drug which








20

has known effects on a person's feelings and moods. Continuous monitorIng of mood on the MICL for 16 hours demonstrated that fluctuations in the MACL were related to anticipated drug-induced changes. Without reporting actual correlational figures, Nowlis indicates that concurrent validity of the ACL is high, although the test-retest reliability is low. Nowlis describes temporality as one of the characteristics of a mood and therefore, b~a~use of the low reliability of the check list, he suggests that the scale is sensitive to short-term fluctuations in mood. Because of its short length and ease of administration, Nowlis points out, it is possible to monitor mood with this instrument over repeated intervals of any required length.

The subject is asked to circle one of four options for each word in the check list: a double plus if the adjective definitely describes a current strong feeling, a single plus if the word slightly applies to a current feeling, a question mark if the word is unclear or the subject is unable to decide if the adjective describes a current feeling and no if the subject is certain that the word does not describe a present feeling. Double plus is scored as 3, single plus is scored as 2, question mark is scored as i, and no is scored as 0. Twelve mood scores are obtained by summing the scores of the individual adjectives in each of the moods.

The scores may range from 0 to 12 on the moods of Aggression,

Concentration, Social Affection, Anxiety, Egotism and Pleasantness. The scores may range from 0 to 9 on the moods of Deactivation, Depression, Activation and Nonchalance. On the moods of Skepticism and Startle the scores may range from 0 to 6.







21


(Both the AACL and the Anxiety Mood on the MACL reportedly measure level of manifest anxiety. The hACL Anxiety mood contains four affectively toned adjectives, each of which the subject responds to by denoting the extent to which the adjective describes his current feelings. The AACL, on the other handu, asks a subject to check those adjectives that apply to his current feelings. Only one of the four adjectives on the MACL Anxiety mood is identical with the 21 adjectives which are scored on the AACL. A comparison between pretreatment performance on the two Anxiety measures resulted in a Pearson product-moment correlation of +.i54, and a Pearson product-moent correlation between the subject's changes in the two scales from pre- to posttreatment testing resulted in a correlation of +.57. Both these correlations are significant beyond the .001 level. This suggests that despite the basic uniqueness of the two measures, they are measuring essentially the same factor of "anxiety.'")

(3) The Adjustment Score (see Apoendix 0).

The Adjustment Score was chosen to evaluate changes In general level of psychological adjustment. It was originally constructed by Dymond (1953) as a Q-sort to evaluate the effects of nondirective counseling. The statements were compiled empirically by Butler and Haigh who noted that they were representative of the positive and negative comments made by people in psychotherapy.

Four "well-trained, practicing clinical psychologists" who were not client-centered by theoretical orientation "'agreed remarkably well" in choosing 37 negative indicators of adjustment and 37 statements describing positive adjustment (1953). These 74 statements rated on a "IIke me"' to 'unlike mer' dimension constitute the scale.








22


Dymond (1954) demonstrated that a significant difference existed between the mean adjustment scores on the Q.-sort of subjects in two months of client-centered psychotherapy as compared to control subjects who waited a comparable length of time (p <.01). Dymond reports that the test-retest reliability of this control group Is +.86. Validity was established in two ways: (1) rank order correlations between self-ideal correlations and the Q-sort adjustment score of clients before therapy began was +.83, and the rank order of these same subjects after therapy was +.92; (2) each therapist rated the success of his therapy for each client and these ratings correlated at setter than the 5 per cent level with the subject's own scores on the Q-sort.

For the purpose of the present study, the 74 items on the 0-sort were randomly organized into a questionnaire. The subject is asked to mark an item "true" if it pertains to him and mark "false"' if the item does not pertain to him. A total of the positive Items marked true and negative Items marked false constitute a subject's adjustment score.

The AACL, MACL and the Adjustment Score were administered twice to all subjects, once prior to and once immediately after the treatment. The change in scores frompre- to posttreatment testing constitute measured changes in manifest anxiety, 12 Independent moods and general level of psychological adjustment.

(4) The Maudslev Personality il.vetory (MPI).

In order to evaluate part of Hypothesis IV, the MPI (Eysenck, 1959, 1961) was administered to obtain subject differences in level of neuroticism and degree of introversion. "Extroversion, as opposed to introversion, refers to the out-going, uninhibited, social proclivities of








23


a person," and neuroticism represents "the general emtional liability of a person, his emotional overresponsiveness, and his liability to neurotic breakdown under stress' (1959, p. 3). The Inventory has 48 questions which have been factor analyzed to yield scores on an extroversionIntroversion dimension and a neuroticism dimension. Although the two dimensions are slightly negatively correlated (-.15), Eysenck submits that they are essentially orthogonal.

The original standardization took place on a sample of 200 normal English men and 200 normal English women. Eysenck has since supplemented this by adding groups of students, nurses, industrial apprentices and "a quota sample of the whole population." Data are also available on groups of hospital patients and prisoners. Split-half and Kuder-Richardson reliability coefficients for the Neuroticism scale fall between .85 and .90 and for the extroversion scale fall between .75 and .65. Using both concurrent and construct validation techniques, Eysenck reports that the MPI has yielded data supporting its validity.

(5) A Modification of the Expectation Qsort (Apfelbaum, 1958) (see Appendix E).

The Q-sort asks a subject to indicate his expectancies regarding the nature of a psychotherapist who may be assigned in the future. Thirty-six of the o0 items in Apfelbaum's item pool were chosen and randomly organized into a questionnaire. These 36 Items constitute those with heaviest factor loadings on three separate clusters: (1) those items which designate a guiding, giving, protective therapist; (2) those items which designate a tolerant, accepting and permissive therapist; and

(3) those items which designate a cold, rigid and condemnatory therapist.








24


in addition to the five tests, each subject was also asked to comp ete a questionnaire asking for personal Informatrbn (see: Appendix F). Each subject also had taken the Beta Intelligence Test (Kellogg .t al, 1946).


Procedure

Subjects were chosen from the ward by a secretary who was naive concerning the mechanics and goals of the study. She was informally questioned after the completion of the study and no bias In the subjects she chose, and her order of choosing them, was noted.

The subjects were seen two at a time. Once the subjects were

seated in the testing room they were given instructions which drew their interest to the tasks about to be presented but did not disclose the nature of the research nor the fact that It was research (see Appendix A). They were given the following tasks in the order noted:

I. Personal information sheet
2. Maudsley Personallty Inventory
3. Anxiety Adjctive Check List
4. Mood Adjective Check List 5. Expectation questionnaire
6. The Adjustment Score

These tasks were relatively short, took no longer than 40 minutes to complete and subjects did not appear to become fatigued. After conleting the six questionnaires the pair of subjects was told that this was an opportunity to speak with a psychiatrist, that he was unable to get to the hospi tel and that he requested they see into a taoe recorder, the tape of which would be sent to him by registered mail (see Appendix A for exact wording).







25


By a coin toss prior to the meeting between subjects and Exoerimenter, one subject of the pair was to be placed in the Experimental Group and the other of the pair was to be placed in the Control Group (treatment vs. wait). Two additional coin tosses by the Experimenter prior to the meeting determined whether this particular pair of subjects would receive (1) the positive or neutral set, and (2) the congruent or incongruent therapist instructions. Both subjects were then given the instructions appropriate to their designated treatment category (i.e.,

(1) positive set, congruent therapist; (2) positive set, incongruent therapist; (3) neutral set, congruent therapist; and (4) neutral set, incongruent therapist [ see Appendix A]). The Experimenter had sufficient time while the subjects were filling in the Adjustment Score, to choose the predetermined Items on the Expectation Questionnaire and manipulate the congruence and incongruence of the subject's alleged psychotherapy st (see Appendix A).

The Control Group subject of the pair was then asked to remain in the testing room. He was asked to relax, was offered the use of a stack of magazines and was told that the Experimenter would return in about 50 minutes (see Appendix A for exact wording). The Experrmental Group subject was then taken to the room which contained the tape recorder. This room had a one-way vision screen and was wired for sound. The subject was seated in an easy chair, instructed to speak to the "psychiatrist" and was told that the Experimenter would return in about 50 minutes to retrieve the tape and the tape recorder was turned on (see Appendix A for exact wording). The Experimenter then seated himself In an observation room where he was able to record the ExL erlmental Group subject's










verbalizations and to observe that the Control Group subject remained in his room. Record was made of the amount of the subject's speech, and in several cases the content of the subject's verbalizations. With a stop watch, amount of speech was rated as follows: constant speaier, no longer than 3-minute pauses; moderate speaker, no longer than 5-minute pauses; little speaker, at least some verbalization; no speaker, says nothing.

At the end of 45 minutes the Experimenter turned off the tape recorder and guided the Experimental Group subject back into the testing room. Both subjects were then asked to "retake several of the shorter tests.' The fol lwing were readministered:

7. Anxiety Adjective Check List
8. Mood Adjective Check List
9. The Adjustment Score

At this point, the collection of the data was completed. The Experimental Group subject was dismissed. If the subject asked about feedback, he was told that he would receive some type of feedback within several weeks. He was also encouraged not to tell anyone the nature of what had happened because the Experlmenter would be unable to see everyone for this project. The Control Group subject was then asked if he wished to "speak to the psychiatrist." If he agreed, he was given 45 minutes with the tape recorder, however, if he declined, he was dismissed. Twenty-one of the 48 Control Group subjects chose to speak.

If any subject discussed suicidal or homicidal content, mentioned intensely pressing issues or made special requests, the Experimenter promptly called this information to the attention of the ward psychiatrist. This occurred in twelve cases.








27


After all the data were collected, all subjects who spoke into the tape recorder were sent a letter (see Appendix G). The purpose of this letter was to thank the subject for his cooperation and inform him of the way in which he could contact the Experimenter for further clarification of the procedure. Twelve of the 69 subjects who spoke into the tape recorder contacted the Experimenter as a result of this letter.














RESULTS


Analysis of covarlance (Ray, 1960) was used to evaluate the tenability of hypotheses 1, II and iI (those hypotheses concerned with the effects of differential experimental treatment). In analyzing the three hypotheses, the eight independent treatment conditions, as diagramed in Table I were considered. Dependent variables in hypotheses I, II and lII were the changes In scores from pre-to posttreatment testing on the AACL, the 12 moods in the MCCL and Dymond's Adjustment Score. Thus, evaluation of the fi rst three hypotheses was made across the three independent vari-r ables utilizing 14 2 x 2 x 2 analyses of covariance--one for each dependent variable.

The technique of analysis of covarlance chosen to evaluate hypotheses 1, 11, and III utilizes the fact that differences between the treatment groups on a dependent variable posttreatment measure may be to some extent a reflection of differences between these groups on the initial pretreatment measure. It removes the variance due to these initial random differences from the final posttreatment variances, thus reducing the size of the error variance and Increasing the precision of the experiment (Gourlay, 1953). That is, by utilizing the regression of the posttreatment measures on the pretreatment measures, an adjustment Is made for the variability which is associated with Initial differences among the subjects. After this adjustment is made, the remaining variability may be analyzed with a more precise estimate of factors which may have produced an effect.

28








29


By inspection, the data appear to have significant homogeneity of variance. Norton (as cited Lindquist, 1953) demonstrates that unless the heterogeneity of variance is so extreme that it is readily apparent upon inspection, the effect upon the F distribution will be negligible. No formal tests of homogeneity of variance were conducted. Statistical significance was set at the .05 level prior to the analysis. Below are the summary tales of the analyses of covariance for each of the dependant variables (Tables 5 through 18). In addition, the mean change scores from pre-to posttesting in each treatment condition are presented for each of the dependent variables (Tables 5A through 18A).

Hypothesis I states that the Experimental Group (placebo treatment group) will show more 'positive therapeutic changes" in the AACL and the moods of the MACL than the Control Group (wait group). This prediction was supported in the analyses of the main effects of three of the fourteen dependent variables. On the AACL, the Experimental Group showed a greater decrease in reported anxiety after treatment than did the Control Group (p < .05) (see Table 5). On the I4ACL Social Affection mood, the difference was significant between the increase in Social Affection for the Experimental Group after treatment and the decrease after treatment in Social Affection for the Control Group (p< .05) (see Table 9). Similarly, the Anxiety mood on the MACL decreased after treatment for the Experimental Group and increased for the Control Group (p <.05) (see Table 10).

Two dependent variables indicated that an interaction exists between the experimental vs. control treatment (Hypothesis I) and the congruent vs. incongruent therapist treatment (see below, Hypotheses II).







30

TABLE 5

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE ZUCKERMN ADJECTIVE CHECK LIST

Source ss df V F A (experimental vs. control) 44.42 1 44.42 e.2J B (congruent vs incongruent) 38.55 1 98.55 13.96***

C (positive vs. neutral) .00 1 .00 .00 AB 69.98 1 69.98 9.91** AC 6.63 1 6.63 .94 BC 17.67 1 17.67 2.50 ABC 17.67 1 17.67 2.50 error [1g AZ 7.06 Total 899.9 94

*** x significant < .001.
a** significant < .01.
z* significant ( .05.


TABLE 5A
PRETREATMENT MEANS (x ), POSTTREATMENT MEANS (X2) AND MEAN
CHANGE SCORES (D) ~OR EACH TREATMENT CONDITION OF THE
DESIGN ON THE ZUCKERMAN ADJECTIVE CHECK LIST

Exoerimental Group Cqntrol Group
Congruent Incongruent Congruent incongruent Therapist Therapist Therapist Therap st Positive L : 11.08 K1 a 8.92 10.33 a 8.41
set x .0 1 x Ia x x L
D a 4.08 D I +1.67 : .58 0 .00 Neutral : 10.08 a4 a 9.75 U 11.33 3 9.41 set X2 = -21 X2 u x x2 *~ al D -2.75 D = -.67 D : -.17 D x +.42







31


TABLE 6

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL AGGRESSION MOOD

Source ss df V F A (experimental vs. control) 8.40 1 8.40 2.00 B (congruent vs. incongruent) 7.57 1 7.57 1.81 C (positive vs. neutral) .10 1 .10 .02 AB 31,84 1 31.64 7.60* AC 1.90 1 1.96 .47 Bc 4.G6 1 4.66 1.11 ABC 5.43 1 5.43 1.30 error 364.29 4.19 Total 424.25 94

r* = significant < .01.



TABLE 6A

PRETREATMENT MEANS (X ), POSTTREATMENT MEANS (X) AND MEAN
CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL AGGRESSION MOOD

Experimental Group Control Groqp
Congruent Incongruent Congruent Incongruen t The raea i .. Theraist_ Ter t TheraiLst Positive X, = J.33 X = 3.42 x 2 3.50 x1 2.33 Set X2 X X2= ~ X2 2= 200 0 -1.50 0 = + .25 D +1.00 -.33
- - - - - - - - - - - - - - - - - -

Neutral X 2.17 X 2.42 2.67 x 2.08 Set X2 12 X2 x -J X2 X2 = 2.
D = -.92 0 = +.75 D = .00 D = +.42










TABLE 7

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MAL CONCENTRATION MOOD

Source ss df V F A (exp.erlmental vs. control) 1.81 1 1.81 .49 B (congruent vs. incongruent) 6.549 6.59 1.7L C (positive vs neutral) 1.12 1 1.12 .30 AB 4.75 I 4.75 1.28 AC 6.14 I 6.14 1.65 BC 3.87 I 3.87 i.A4 ABC .45 1 .45 .12 error 123.02 3.71

Total 347.75 94




TABLE 7A

PRETREATMENT MEANS (Xi), POSTTREATMENT MEANS (X2) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL CONCENTRATION MOOD

Ex.ermental Group Control GroupCongruent Incongruent Congruent Incongruent Therapist Therapist Therapist Thera!ist Positive XI = 1.75 Xi = 10.08 '1 9.50 Xi = 10.33
Set X2 = 3 X2 = .67 8X2 8. 2 9.42
0 = +.17 0 = -.41 D -1.0 D -.91


Neutral xl I .u8 Xi 9.42 il 9.17 xI =1017
Set X2 11.25 X2 8_.2 X2- 2 X2 ".-L
= +.17 o = -1.17 D .00 -.42








33

TABLE 8

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL DEACTIVATION MOOD

Source ss df V F

A (experimental vs. control) 7.82 1 7. 2 1.67 b (congruent vs. Incongruent) 25.46 1 25.48 5.43* C (positive v.. neutral) .33 1 .33 .07 AB 1.24 i 1.24 .26 AC .96 1 .96 .20 BC .48 1 .4 .10 ABC .70 I .70 .15 error 407.95 4.69 Total 444.96 94


*Si gnlficant ( .05.



TALE 8A

PRETREATMENT MEANS (X1), POSTTREATMENT MEANS (X) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITIN OF THE
DESIGN ON THE MACL DEACTIVATION MOOD

Experimental Group Control Groy
Congruent Incongruent Congruent Incongruent Thera, st Therapist Therqgist Therapist Positive xA = 4.06 xI = 3.42 _I 3.67 l 4.08
Set X2 -. 11 x2 = 36 XX2 X2 .
o -1.17 0 = +.25 D = +.08 0 +.50 Neutral X, = 4.25 I = 4.17 _1 4.75 x1 = 2.25
Set X2 ll x2 -= X2, 4.00 X 23.
=I -1.o 0 = +.16 0 = -.75 D = +.83








34


TABLE 9

SUMMARY TABLE OF THE ANALYSt S OF COVARIANCE FOR THE MACL SOCIAL AFFECTION IOOD

Source ss df V F A (ex-erimntal vs. control) 27.15 1 27.15 4.35*

8 (congruent vs incongruent) 14.37 1 14.37 2.30 C (positive vs. neutral) .03 1 .03 .001 AB 19.b2 I 19.62 3.14 AC 10.84 1 10.84 1.74 1.36 1 1.36 .22 ABC 22.63 1 22.60 3.62 error 543.18 7 6.24 Total 639.15 94


*Significant U.05.


TABLE 9A

PRETREATMENT MEANS (X), POSTTREATMENT MEANS (x2) AND MEAN
CHANGE SCORES (D) OR EACH TREATENT CONDITION OF THE
DESIGN ON THE MACL SOCIAL AFFECTION MOOD

ExDerlmental Group Control Group
Congruent Incongruent Congruen Incongruent Therapist Therapist Therapist Therapist Positive xI = 6.17 XI = 6.25 xI = 6.08 x = 8.00
Set X2 Z.4 X2 = -50 X2 = X2 Zl
D = .67 o -1.25 D = -.58 D = -.17


Neutral X- = 7.58 X = 7.75 x 7.58 x 6.42
Set X2 = 42 X2 8.0& X2= X 2 =.25
0 = +.84 = +.33 0 = -.63 o -1.17







35

TABLE 10

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL ANXIETY MOOD

Source ss df V F

A (exerimentai vs. control) 35.35 1 35.35 6.29* B (congruent vs. Incongruent) 4.27 1 4.27 .76 C (positive vs. neutral) .00 1 .00 .00 AB 6.31 1 6.31 1.12 AC 4.00 1 4.00 .71 Bc .07 1 .07 .01 ABC 8.33 1 8.33 1.48 error 488. 5.62 Total 547.25 94


*Signlficant < .05.


TABLE 10A
PRETREATMENT MEANS ( ), POSTTREAThENT MEANS (2) AND MEAN
CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL ANXIETY MOOD Ex_ erimental Grow Control Groui
congruent Incongruent Congruent Incongruent Therapy st Therapjst Therapist Therapi st Positive XI 5.25 X1 4.92 Xj 4.50 X, 4.33
Set X2 3 75 x2 x 2 x 4.33
D = -1.50 0 +. D .00 Neutral 1 = 5.84 _l = 3.83 XI 5.67 x1 4.o00
Set X2 = 4.42 X2 -~ X2 X2 5s4
0 = -1.42 D -.6 0 m +.17 D = +1.00








36


TABLE I I

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL DEPRESSION O030

Source ss df V V

A (experimental vs. control) 3.04 1 3.04 .78 8 (congruent vs. incongruent) .00 I .00 .00 C (positive vs. neutral) .49 1 .49 .13 AB .86 .86 .22 AC 1.24 1 1.24 .32 BC .37 I .37 .10 ABC 5.55 1 5.55 1.43 error 17i 8Z 3.88 Total 394.16 94




TABLE IIA

PRETREATMENT MANS ( ) POSTTREATMENT MEANS () AND MEAN
CHANGE SCORES (0) 0R EACH TREATMENT CONDITI&4 OF THE
DESIGN ON THE MACL DEPRESSION MOOD

Experimental .Group Control Group
Congruent I conoruen t Congruent Incongruent S.Thera st r ai Therap I st Thera~ist Positive x = 3.92 x = 3.75 x = 3.58 x1 -3.75
Set X2 2 2. 2 X2 3.2 2 l
0 -1.25 0 -.83 D = +.b D -.92 Neutral x1 = 5.33 x 4.17 X- = 4.25 x- = 3.08
set xz 4.42 x2 -X2 .432 X2 =22
S= -.91 = -.-.3 -.16








37


TABLE 12

SUARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL EGOTISM MOOD

Source ss df V F A experimentall vs. control) 4.53 1 4.53 1.09 I (congruent vs. Incongruent) 2.I0 1 2.10 .51 C (positive vs. neutral) 3.19 I 3.19 .77 AB .57 1 .57 .14 AC 11.58 1 11.58 2.80 BC 1.33 1 1.33 .32 ABC 2.60 I 2.60 .63 error ) s (2 4.14 Tote l 385.76 94




TABLE 12A

PRETREATMENT MEANS (XI), POSTTREATMENT MEANS (X2) AND MEANS
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DE:G' ^2 T!" N'ACL EGOTISM MOOD Experi menta I Group Control Gro
Conngrent Incongruent Congruent Incongruent
Ther, i s t Th.raist Therapi TheraIst Positive 1 = 3.08 X = 3.25 _i = 3.00 1 = 2.67
Set X2 2j0 X2 = 2.8 X2 I X2 1
D -.5b D =-1.17 = +.08 D = +.58 Neutral 2.58 Xi = 2.50 2.58 i = 2.42
Set x2 ~ I X2 = X2 X2 2.1
O = +.50 D +.17 D +.42 D = -.25








38

TABLE 13

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL PLEASANTNESS MOOD

Source ss df V F A (experimental vs. control) 8.12 1 8.12 1.21 B (congruent vs. incongruent) 39.31 1 39.41 5.89* C (positive vs. neutral) 43.51 1 43.51 6.50* AB .01 1 .01 .001 AC .02 1 .02 .003 BC 3.57 i 3.57 .54 ABC 13.87 I 13.87 2.07 error 581.7 .69 Total 690.17 94


*Significant ( .05.



TABLE 13A

PRETREATMENT MEANS (X ), POSTTREATMENT MEANS (X2) AND MEAN
CHANGE SCORES (0) FOR EACH TREATMENT CONDITI& OF THE
DESIGN ON THE MACL PLEASANTNESS MOOD

xPerimental Grou Control Group
Congruent Incongruent Congruent Incongruent Therapist Therapist Therapist Theraist

Positive = 4.25 t 4.33 - 4.42 X 4.33
set x2m x2 x 2 =~ x2 4.0
+2.42 = .00 0 = +1.08 0 +.17 Neutral .17 4.75 x 4.33 xI 4.17
Set X2 2 4 X2 4.4 X2 2.7
0 .00 0 -.33 0 = +.09 0 -1.50










TABLE 14

SUMMARY TABLE OF THE ANALYSIS OF iOVARIANCE FOR THE MACL ACTIVATION MOOD

Source ss df V F

A (ex-erimintal vs. control) 2.90 1 2.90 .80 8 (congruent vs. incongruent) 1.80 1 1.80 .50 C (positive vs. neutral) .20 1 .20 .06 Ab 5.58 1 5.58 1.54 AC .47 1 .47 .13 BC .00 1 .00 .00 ABC .12 i .12 .03 error 314.94 3.62 Total 326.01 94




TABLE 14A

PRETREATMENT MEANS (x ), POSTTREATMENT MEANS (2) AND MEAN
CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL ACTIVATION MOOD Epcerimental Group... Control Group Congruent Incongruent Congruent Incongruent
Therapist Theracist Therapist Therapist

Positive xI = 4.25 X =- 4.00 x = 4.42 x = 4.25
Set x2 = 4.0oo0 x2 1- X2 = 3.-2 x2 4
D -.25 D = -.33 D = -.50 D = +.25


Neutral X = 3.75 xl 3.58 Xi 3.92 xi = 4.58
Set x w .0 -.3 x x2 =.
a -.25 D --.50 = -.34 +.34










TABLE 15

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL NONCHALANCE MOD

Source .,s df V F A (experimental vs. control) .87 1 .67 .23 B (congruent vs. Incongruent) 1.43 1 1.43 .38 C (positive vs. neutral) .17 I .17 .04 AB 3.52 1 3.52 .92 AC .24 1 .24 .06 C 8.46 1 8.46 2.22 ABC .37 1 .37 .10 error 331.90 3.81

Total 3436.9 94



TABLE 15A

PRETREATMENT MEANS (X ), POSTTREATMENT MEANS (X2) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL NONCHALANCE MOOD

Experiment Group n trol Group
Congruent Incongruent Congruent Incongruent
Th erap_ is t Therabist Theraist Theraoist Positive 3.08 Xi = 3.42 1 = 3.83 _i = 2.o8 set x2 Xz x = 2 x 32 x2 1 D = +.17 D = -.50 0 = -.50 D -.33
- - - m .- - - - - - - - - - - - - -

Neutral X 1 2.00 X = 3.08 58 i = 2.00 Set X2 22 X2 l 8 j 0 = -.17 D .00 D -1.00 0 +.







41


TABLE l

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL SKEPTI CISM HOOD

Source ss df V F

A (experimental vs. control) 1.98 1 1.98 .74 B (congruent vs. Incongruent) .49 1 .49 .18 C (positive vs. neutral) .10 1 .10 .04 As .36 1 .36 .13 AC 2.17 I 2.17 .81 BC .00 1 .00 .00 ABC .01 1 .01 .004 error 1I~1 81 2.67 Total 236.98 94




TABLE IGA

PRETREAThENT MEANS (i), POSTTREATMENT MEANS (X2) AND MEAN
CHANGE SCORES (D) 'OR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL SKEPTICISM MOOD

ri! menta rou Control Group
Congruent Incongruent Congruent Incongruent TherPist Therap Therapist Therap s Positive 1.92 2X = 2.75 x1 2.75 X = 1.07 Set X2 S XX2 m 22 2 I 5 D = -.17 0 -.25 D -.2 +.b08
---- -- -- -------------------- - --- --- -- -- -- -- -- -Neutral XI = 2.00 XI = 2.17 XI = 3.00 Xl = 1.75 Set x2 X2 L X2 X2 2.1 S= -.42 0 -.26 D -.17 D = +.42







42

TABLE 17

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL STARTLE MOO

Source ss df V F

A (experimental vs. control) .03 1 .03 .0

B (congruent vs. incongruent) 13.47 1 13.47 2. 4***

C (positive vs. neutral) 2.67 1 2.67 5.34* AB .04 I .04 .08 AC .37 1 .37 .74 SC 1.47 1 1.47 2.94 ABC .03 1 .03 .0O error tlao Az .50 Total 61.78 94


***Significant < .001.

**Significant < .05.

TABLE 17A

PRETREATMENT MEANS (XI), POSTTREAThENT MEANS (X2) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL STARTLE MOOD Ex.erimental Grou Control GroL
Congruent Incongruent Congruent Incongruent
Theraist Theapist TherapIst Therpist Positive = 1.42 X 1.17 XI 1.50 Xi = 1.25
Set X2 = 1.00 X2 1 .= 2 = .00 X2 -" 1
o -.42 = +.66 D = -.50 +.42 m - - - - - - - - - - - - - - - - -

Neutral = .92 1 = .92 I.0 i 1.25
Set X2 1.00 X2 i 1z 2 1. 92
) = +.08 +.58 ++.17 0 7







43

TABLE 18

SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE ADJUSTMENT SCORE

Source ss df V F

A (experimental vs. control) 40.12 1 40.12 3.04 B (congruent vs. Incongruent) .07 1 .07 .005 C (positive vs. neutral) 20.84 1 20.84 1.58 AB .16 1 .1 .01 AC 33.34 I 33.34 2.52 BC 3.77 1 3.77 .29 ABC 9.98 1 9.9 .76 error 1.4 70 13.21 Total 1,257.98 94



TABLE 18A

PRETREATMENT MEANS (XI), POSTTREATMENT MEANS (X2) AND MEAN
CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE ADJUSTMENT SCORE Exeri a Group Control Group
Congruent Incongruent Congruent Incongruent Therapy ist Therapist Therepist Therapist Positive Xm 45.42 xl 43.17 I 39.67 -X 44.83
Set X2 z- Z x2 = : x2 41.83 X2 I 725
D +.33 0 +.75 0 = +2.16 0 = +2.42 Neutral xi 42.00 XI 46.50 XI = 37.17 XI = 49.58
set X2 41.5 i x2 427.00 X2 3750 2 = 49. 58
0 -.42 D +.50 0D +.33 D = .00











Changes after treatment on the AACL indicated that the Congruent Therapist, Experimental Group and the Incongruent Therapist, Control Group significantly reduced anxiety when compared to the Incongruent Therapist, Experimental Group and the Congruent Therapist, Control Group (p 4 .01) (see Table 5). Similarly, the HACL Aggression mood indicated that the Congruent Therapist, Experimental Group and the Incongruent Therapist, Control Group significantly reduced aggression when compared to the Incongruent Therapist, Experimental Group and the Congruent Therapist, Control Group (p < .01) (see Table 6). These interactions suggest that as measured on the AACL and the MACL mood of Aggression, the superiority of the Experimental Group treatment pertains only for those subjects who are given the congruent therapist instructions, while Experimental Group subjects given the Ini cjgruent therapist instructions report "nontherapeutic" changes.

While only three dependent variables significantly distinguished between the Experimental Group and Control Group (in addition to two interaction effects) and the Adjustment Score indicated no difference as predicted, no dependent variables significantly changed in a direction opposite to that predicted. Thus Hypothesis I appears to be supported.

Hypothesis it states that subjects, when offered a "psychotherapist" who is coners;. with their expectancies, will show more "positive therapeutic changes" on the dependent variables than subjects assigned a "psychotherapist" who is incongruent with their expectations. This was found to be the case in the analysis of the main effects in four of the 14 dependent variables. The AACL showed that subjects given a therapist congruent with thai r expectations decreased anxiety after treatment










significantly oire than those subjects who were assigned a therapist incongruent with their expectations (who increased slightly in anxiety) (p e- .001) (see Table 5). Similarly, the Congruent Group showed a significantly higher increase in the i4CL mood of Pleasantness (p < .5) than the Incongruent Group which decreased slightly on this mood (see Table 13). Further, the Incongruent Therapist Group displayed a signifIcant increase on the MACL mood of Deactivation (p < .05) when compared to the Congruent Group, which decreased on this mood (see Table 8). Also, the Incongruent Therapist Group reported a significantly higher increase in the NEICL mood of Startle after treatment (p < .001) than the Congruent Therapist Group which decreased slightly in Startle (see Table 17). While only four dependent variables significantly distinguished between the Congruent and Incongruent Therapist Groups (in addition to two interaction effectss, and the Adjustment Score Indicated no difference as predicted, no dependent variables significantly changed in the unpredicted direction. Thus, Hypothesis II appears to be supported. (The factor analytic studies of Apfelbaum (1958) suggest that three basic types of psychotherapists may be described on the Expectation Q-sort. Inspection of the data suggest that the sample does not show a trend toward describing a particular type of psychotherapist. Similarly, an individual subject does not clearly choose one or another of Apfelbawu's three types of psychotherapists.)

Hypothesis II states that subjects who are given a more positive set concerning the value and outcome of treatment will display more "positive therapeutic changes on the dependent variables than subjects given a more neutral set. This was found to be the case upon analysis of two











dependent variables. The Positive Set Group increased significantly less on the MACL mood of Startle after treatment than did the Neutral Set Group which increased more (p 4 .05) (see Table 17). The Positive Set Group also reported a significantly higher increase on the M4CL mood of Pleasantness (p 4 .05) than the Neutral Set Group which decreased reported pleasantness after treatment (see Table 13). Only two of the 13 dependent variables which were expected to change significantly distinguished between the positive and neutral set groups. Of particular note is the fact that no significant changes were demonstrated on either the AACL or the MACL mood of Anxiety. However, as predicted, no change occurred on the Adjustment Score and no dependent variables changed significantly in the unpredicted direction. Thus, the hypothesis appears supported.

In summary, those dependent variables which demonstrated significant differences between one or more of the three independent variables were: the AACL and the MACL moods of Aggression, Social Affection, Deactivation, Anxiety, Pleasantness and Startle. Those dependent variables which did not significantlyy distinguish between any of the three independent variables were Dymond's Adjustment Score (as predicted) and the MACL moods of Concentration, Depression, Egotism, Skepticism, Nonchalance and Activation.

As noted earlier, after Nowlis factor analyzed the MACL he divided his 12 mood factors into three groups of four moods each: those moods most clearly identified, those less clearly identified and those least clarly identified (see Table 4). It is noteworthy that of the four moods in the group most clearly identified by Nowlis, three were found to











distinguish significantly between the Independent variables; of the four moods in the less clearly identified group, two were found to distinguish significantly between the independent variables; and of the four moods in the least clearly identified category, only one was found to distinguish significantly -etween the independent variables. The AACL, the scale with the greatest number of items, and probably, therefore, the scale with the highest reliability, showed signlfiant differences between two of three main effects (independent variables). Below is a summary table which on the vertical axis lists the three independent variables and their interactions. On the horizontal axis is listed the 14 dependent variables and entered into the body of the table are the levels of significance at which a null hypothesis might e rejected.

Although seven dependent variables did not reflect significant differences between any of the treatment conditions, there were changes worth noting in several of them. For example, the IACL mood of Depression was found to decrease after treatment in all treatment conditions except one: the Positive Set, Congruent Therapist, Control Group condition where a small increase in reported depression occurred (see Table hA). Evaluation by the Signs Test (Siegel, 1956) shows that the probability of seven out of eight cells decreasing by chance is at less than the .035 level. This indicates that a reported decrease in depression tends to occur after treatment, regardless of the specific nature of the treatment within the study. The MAL mood of Concentration, similarly, did not result in any statistically significant differences between the treatment conditions. However, the Congruent Therapist, Experimental Group condition was the only treatment which described an increase in













TABLE 19

SUMMARY OF THE DEPENDENT VARIABLES AND THE SIGNIFICANCE LEVELS AT WHICH THEY DISTINGUISHED BETWEEN THE INDEPENDENT VARIABLES Nowlis Clearly Nowlis Less Clearly Nowlis Least Clearly Identified Factors Identified Factors Identified Factors
So.
AACL Aggr. Conc. Deact. Aff. Anx. Depr. Egot. Pleas. Activ. Nonch. Skept. Start.

A (experimental
vs. control) .05 .05 .05

B (congruent vs.
incongruent) .001 .05 .05 .001

C (positive vs.
neutral) .05 .05 AB .01 .01

AC

BC

ABC











concentration (see Table 7A). The MACL mood of Skepticism also displayed no statistically significant differences between the independent variables. However, the incongruent Therapist, Control Group condition reported a slight increase in skepticism after treatment while all others reported a decrease in skepticism after treatment (see Table 16A).

Hypothesis IV states that no relationship exists between differences in certain subject characteristics and changes in the dependent variables. This was evaluated by multiple regression technique. The subject differences under consideration were age, level of intelligence (Beta Ilntelligence Test, 1941), level of neuroticism (HPI, 1953) and level of introversion (API, 1959). Analysis of variance (Lindquist, 1953) was used to demonstrate that each of these four subject variables was randomly distributed across the eight treatment conditions. (The summary tables for toese analyses on age and level of intelligence are presented in the Subjects section of the etiod chapter. Below are the summary tables demonstrating the random distribution of introversion (Table 20) and neuroticism (Table 21).

Pearson product-moment correlations were computed between each of these four variables and the change in scores fronpre- to posttreatment testing on each of the 14 dependent variables. This was done for each of the eight treatment conditions. This resulted in 448 correlations which are presented oelow in Tables 22, 23, 24, and 25. At the .05 level of significance, by chance alone w would expect 24 significant correlations out of the total 448. In the present data, by frequency count, there are 26 correlations wilch are significant at the .05 level. This suggests that, over-all, the null hypothesis acppe rs supported. It is difficult







50






TABLE 20
SUMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE
RANDOM DISTRIBUTION OF MPI INTROVERSION SCORES

Source ss df V F

A (experimental vs. control) 19.26 1 19.2u .34 8 (congruent vs incongruent) 31.51 1 31.51 .55 C (positive vs. neutral) 33.8 1 33.84 .59 AB 52.51 1 52.51 .92 AC .02 i .02 .00 BC 128.35 1 128.35 2.25 ABC 31.50 1 31.50 .55

w Cells U12.7;2 56.90 Total 5,309. 74 95

(An F of 3.95 is significant at the .05 level.)








51







TABLE 21
SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE
RANDOM DISTRIBUTION OF MPI NEUROTICISM SCORES
-~-I~-~ -- 'I- -i -- ............ ...
Source ss df V F

A (experimental vs. control) 2.06 1 2.0u .01 8 (congruent vs. Incongruent) 408.38 1 408.38 2.04 C (positive vs. neutral) 135.40 1 135.40 .08 AB .37 1 .37 .00 AC 57.02 1 57.02 .28 BC 376.03 1 376.03 1.88 ABC 63.37 I 63.37 .32

w CeIlls _,o23.3 200.27 Total 18,665.96 95

(An F of 3.95 is significant at the .05 level.)











to attribute a high degree of conclusiveness to results showing some consistency in the absence of significance. However, there are several observations which may be made and several trends worth identifying. It is important to note, also, that the omplexity of o-th the data and the independent variables tends to reduce the meaningfulness of interpretation.

It will be noted in Table 22 that there is a tendency for subjects, given a positive set and a congruent therapist and who score lower measured levels of intelligence, to decrease reported anxiety on the AACL after treatment. Conversely, this suggests that subjects with higher measured intelligence, when given a positive set and a congruent therapist, increase reported anxiety on the AACL after treatment. Similarly, subjects with lower measured intelligence, who are given a positive set and an incongruent therapist, seem to increase their scores on the AACL after treatment more than subjects with higher measured intelligence. It will be noted further, with the exception of the Experimental Group, Neutral Set, Congruent Therapist subjects (which showed an opposite trend), less bright subjects tend to increase the MACL mood of Activation after treatment. Fromthesedata, the conclusion may be tentatively advanced that people with lower intelligence seem to change in a therapeutic direction after certain types of placebo treatment. On the other hand, with the exception of the Experimental Group, Positive Set, Congruent Therapist, there appears to be a trend which suggests that subjects with higher measured Intelligence tend to increase more on the A4CL mood of Social Affection after treatment than subjects who score lower on intelligence. This finding suggests a conclusion which is not in line with that indicated above from the analysis of the AACL and the MACL mood of





53









TABLE 22

PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EIGHT OF THE DEPENDENT VARIABLES AND SCORES ON THE BETA INTELLIGENCE TEST. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS Moods of the MACL
So. Adjustment Treatment Conditions AACL Aqqr. Conc. Deact. Aff. Anx. Depr. Egot. Pleas. Activ. Nonch. Skept. Start. Score

Experimental Group
Positive Set, Congruent Therapist +.36 +.10 -.29 +.42 -.11 +.22 +.37 +.09 -.28 -.50 -.24 +.31 +.49 -.24

Positive Set, Incongruent Therapist -.51 +.05 +.11 +.20 +.43 -.44 -.23 -.15 +.39 -.23 +.52 -.18 -.61* +.48

Neutral Set, Congruent Therapist +.20 +.34 -.50 +.02 +.26 +.04 +.09 -.35 +.07 +.18 .00 +.01 +.52 -.50

Neutral Set, Incongruent Therapist -.13 -.39 +.19 -.89* +.54* -.44 -.11 +.60* -. -. -.32 -.18 -.10 -.73* -.27


Control Group
Positive Set, Congruent Therapist +.45 -.07 -.27 -.27 +.06 +.35 -.33 +.02 +.31 -.27 -.27 +.26 +.37 -.01

Positive Set, Incongruent Therapist -.33 +.02 -.36 +.13 +.10 -.53* -.18 -.10 +.08 -.25 -.10 -.37 -.12 +.08

Neutral Set, Congruent Therapist +.05 -.32 -.13 -.05 +.25 +.31 +.09 -.18 -.24 -.02 -.17 +.01 -.22 -.09

Neutral Set, Incongruent Therapist +.11 +.41 +.18 +.06 +.09 +.07 -.05 -.28 -.15 -.38 -.26 -.26 +.07 +.16


*Significant 4 .05.







54


Activation. That is, in this case, brighter subjects tend to increase Social Affection (personal comfort) as a result of the placebo situation, while in the former case, brighter subjects appear to increase anxiety and become less active. Of further interest Is another tentative finding suggesting that brighter subjects, when given a congruent therapist, tend to increase scores on the MACL Skepticism mood after treatment, but decrease skepticism after treatment when given an Incongruent therapist.

From Table 23 several Interesting, but also tentative relationships may be discussed. In the Experimental Group, the more introverted subjects who are 4iven the Positive Set, Congruent Therapist instructions tend to increase reported Social Affection after treatment. However, the more introverted subjects who are given the Neutral Set, Incongruent Therapist instructions tend to decrease the MACL mood of Social Affection. An additional trend (with the exception of the Control Group, Neutral Set, Incongruent Therapist subjects) appears to be that introverted subjects, more than extroverted subjects, tend to decrease scores on the MACL mood of Skepticism after treatment.

Several relationships presented in Table 24 may be tentatively

discussed. It appears that in the positive set cells of the Experimental Group and in the neutral set cells of the Control Group the older the subject, the larger the increase in the MACL mood of Social Affection. Similarly, in the positive set cells of both the Experimental and Control Groups, the younger the subject the greater the Increase in reported MACL Egotism mood after treatment. Another effect appears to exist in the relationships between age and change in anxiety after treatment. On both the AACL and the MACL Anxiety mood, it is noted that younger subjects





55







TABLE 23
PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND SCORES ON THE MPI INTROVERSION (LOW SCORES)-EXTROVERSION (HIGH SCORES). THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS Moods of the MACL
So. Adj ustment Treatment Conditions AACL Aggr. Conc. Deact. Aff. Anx. Depr. Eqot. Pleas. Activ. Nonch. Skept. Start. Score

Experimental Group
Positive Set, Congruent Therapist +.21 +.03 -.21 +.10 -.54* +.01 +.20 +.32 -.07 +.57* +.53* +.44 +.25 +.71*

Positive Set, Incongruent Therapist +.06 -.12 +.09 -.46 -.25 -.16 +.05 +-51 +.03 +.41 +.13 +.21 +.24 -.13

Neutral Set, Congruent Therapist +.07 +.02 -.11 -.33 -.14 +.03 -.06 +.14 -.31 -.11 +.29 +.40 -.13 +.26

Neutral Set, Incongruent Therapist -.02 -.31 +.09 -.11 +.60* +.41 -.28 -.36 +.02 +.01 +.01 +.01 -.51 +.01


Control Group
Positive Set, Congruent Therapist +.12 -.36 -.52 .00 -.59* +.30 -.08 -.06 -.16 +.06 +.10 +.30 +.57* -.03

Positive Set, Incongruent Therapist -.07 +.52 +.08 +.08 -.40 -.05 +.14 +.37 -.28 +.32 +.48 +.02 -.03 -.04

Neutral Set, Congruent Therapist +.31 +.04 -.05 +.11 +.35 -.37 +.40 -.24 +.37 +.29 -.04 +.61* -.15 -.36

Neutral Set, Incongruent Therapist -.27 -.18 -.14 -.25 -.17 +.01 +.11 +.29 -.17 -.65* +.15 -.33 -.19 -.30


*Significant 4 .05.





56









TABLE 24

PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND SUBJECTS' AGES. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS Moods of the MACL
So. Adjustment Treatment Conditions AACL A9qh. Conc. Deact. Aff. Anx. Depr. Egot. Pleas. Activ. Nonch. Skept. Start. Score

Experimental Group
Positive Set, Congruent Therapist +.07 +.07 +.19 +.06 +.49 +.11 -.19 -.36 -.35 -.41 -.56* +.14 +.14 -.51

Positive Set, Incongruent Therapist +.36 +.42 +.32 +.06 +.50 +.45 +.51 -.24 +.14 -.06 -.32 -.07 -.03 -.05

Neutral Set, Congruent Therapist +.51 +.01 -.34 +.13 -.22 +.32 -.20 +.30 +.03 -.05 +.06 +.52 +.50 +.50

Neutral Set, Incongruent Therapist +.03 +.23 -.10 .00 -.25 +.06 -.25 +.39 +.41 -.02 +.48 +.49 -.14 -.48


Control Group
Positive Set, Congruent Therapist +.37 -.22 +.16 -.33 -.45 +.18 -.07 -.28 -.20 -.01 -.36 -.66* -.08 -.9"

Positive Set, Incongruent Therapist +.25 +.15 +.20 -.43 -.21 +.26 -.19 -.18 +.21 +.24 -.22 -.21 -.29 -.15

Neutral Set, Congruent Therapist -.60* +.02 +.36 +.12 +.31 +.23 +.02 +.28 +.32 +.30 -.01 +.05 -.24 +.41

Neutral Set, Incongruent Therapist -.12 -.29 +.22 -.31 +.18 -.25 +.19 +.06 -.09 +.06 +.42 -.38 +.14 +.03


*Significant < .05.








57


decrease reported anxiety after Experimental Group treatment more readily than do older subjects. This is especially so for the Positive Set, Incongruent Therapist and Neutral Set, Congruent Therapist conditions.

Eysenck defined his Neuroticism Scale as a measure extending

along a highly neurotic to normal continuum, In the current study, the sample contains no normalss,' therefore, serious questions may be raised concerning the meaningfulness of the neuroticism score. For example, does a low score on neuroticism in this study mean that a patient is psychotic? A biserial correlation (McNemar, 1955) between diagnosis (psychotic vs. nonpsychotic) and scores on the Neuroticism Scale results in a correlation of .57 (p 4 .001). This suggests that subjects with nonpsychotic diagnoses tend to score high on the Neuroticism Scale while subjects with psychotic diagnoses tend to score lower. However, this correlation accounts for only 32.49 per cent of the variance, and it is with extreme caution that Interpretation of the neuroticism score data be made.

As noted in Table 25, an apparent relationship exists between

neuroticism scores and the MACL mood of Nonchalance. With the exception of the Experimental Group, Neutral Set, Congruent Therapist subjects, it appears that as neuroticism increases, nonchalance decreases after treatment. Another interesting relationship indicates that subjects scoring high on the Neuroticism Scale who are in the Positive Set, Congruent Therapist, Control Group decrease their MtCL mood of Startle after treatment; while subjects scoring high on the Neuroticism Scale who are in the Neutral Sot, incongruent Therapist, Control Group increase this Startle mood after treatment.





58









TABLE 25

PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND SCORES ON THE MPI NEUROTICISM SCALE. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS Moods of the MACL
So. Adjustment Treatment Conditions AACL Aqgr.. Conc. Deact. Aff. Anx. Depr. Egot. Pleas. Activ. Nonch. Skept. Start. Score

Experimental Group
Positive Set, Congruent Therapist -.08 +.19 +.51 -.12 +.52 +.17 -.43 -.46 -.36 -.39 -.43 -.29 -.24 -.56*

Positive Set, Incongruent Therapist +.18 -.23 -.37 -.03 -.06 +.07 +.41 -.22 +.01 -.38 -.59* -.38 +.03 -.43

Neutral Set, Congruent Therapist -.24 -.41 +.14 +.13 +.04 +.43 +.09 +.42 +.70* +.34 +.26 -.23 +.20 -.06

Neutral Set, Incongruent Therapist -.13 +.02 +.29 +.12 -.03 +.56* +.21 +.12 +.45 +.42 -.07 +.48 +.12 -.51


Control Group
Positive Set, Congruent Therapist -.01 +.03 +.52 -.17 +.68* -.14 +.08 +.03 -.16 +.19 -.18 -.17 -.69* -.10

Positive Set, Incongruent Therapist -.29 -.28 -.37 -.15 +.16 -.27 -.60* -.30 -.12 -.41 -.28 -.02 -.30 +.42

Neutral Set, Congruent Therapist -.38 +.26 +.37 +.10 -.28 +.39 -.36 +.35 -.36 +.13 -.14 -.30 -.11 +.43

Neutral Set, Incongruent Therapist +.17 +.20 -.34 +.40 -.19 +.47 +.15 -.16 +.23 -.22 -.35 +.20 +.56* +.04


*Significant (.05.











It is believed sufficiently important to stress once again the tentativeness of all relationships discussed under Hypothesis IV. It is this writer's opinion that due to the low number of significant correlations, the data basically support the null hypothesis that no conclusively meaningful relationships exist between the dependent variables and the measured Individual differences.

It was believed that a subject's reported level of desire for

psychotherapy would be related to his amount of verbalization when he was permitted to speak to the "psychotherapist." The tenability of this belief was evaluated by multiple regression technique utilizing only the Experimental Group and the correlations are presented below in Table 26.


TABLE 26

MULTIPLE REGRESSION COEFFICIENTS BETWEEN AMOUNT OF SPEECH AND
DESIRE FOR PSYCHOTHERAPY PRESENTED FOR EACH INDEPENDENT
VARIABLE (EXPERIMENTAL DATA ONLY)

Congruent Incongruent Positive Neutral Therapist Therapist Set Set Amount of
Speech .50 .02 .49 .08
VS.
Desire for
Psychotherapy n 24 n = 24 n a 24 n = 24



Amount of verbalization, which was originally rated on a fourpoint continuum (see p. 2b) was divided, for the purpose of this evaluatlon, into two groups: (1) constant speakers and moderate speakers, and

(2) little speakers and no speech. Desire for psychotherapy was rated on a three-point scale by each subject (see Appendix F).










The correlations presented in Table 26 suggest that if a subject's desire for psychotherapy is high, his amount of verbalization is high If he is given a congruent therapist and/or a positive set. If, however, he is given an Incongruent therapist and/or a neutral set there is no relationship between his desire for psychotherapy and his amount of speech.

Presented below in Table 27 ere co-relationships between (1)

amount of speech; (2) desire for psychotherapy; and (3) diagnosis on the one hand, and change scores from pre- to posttesting on the AACL and the moods most c'!arly identified by Nowlls on the other (data only for the Experimental Group).

One interesting conclusion which may be drawn rrom these data concerns the relationship between amount of speech and change scores on the AACL. It appears that subjects who speak the most increase anxiety from pre- to posttreatment testing. This is true in all treatment conditions and the correlation for all groups is +.30 (p < .05). Of further interest is the fact that level of desire for psychotherapy does not appear, by Inspection, to appreciably Influence the amount of change on these five dependent variables from pre- to posttreatment testing. Similarly, by Inspection, diagnosis (psychotic or nonpsychotic) does not appear to be related to the amount of reported change In personal comfort after treatment as measured by the five dependent variables.









6I






N 4 0 ON C C NOJ r d
*J m DO M NO 0 N- 0 -0
h ft + ++ +3 a ++ 1 +


-wi


I0 4.1 0 0 0 N a 0 %I -- O -0 IfO 4 n0




a

CL uj L) t b PVI \s OW 45 t rl oDQIh




S;




03 0 N- ~O EV30* C* * * * * C * C*
S1 + + ++ + ++ +g +
0 *2



A
LO



NN N 00 a l0 00
CF C . C *. C .C 3 +1 + 3+ + ++ a ++ + 4+
aalt




w -U'

~ W C0- N NO N 3. M 0< 0 + + + I + ++ + 1+ + ++ CA:I



j
Ac C4 0 C*4 00 :5

&i Sj 4 & #







W4 W 4 = 4C U
0 --3







1 -35 A 01 a a g a .. *.
rrcb4 J B p

o t a





C0 0 we

WZ 4UN C344N N N e


A S UC QL4IC 15 g gag Lucr









COW WM MMWOw



Cu UN e4am0 C 0 4 3
I 0 A 2V m I cl. 0 ~ 4 z 6 E 0V





Ws w y -c cn tn (Af V) 41 t4 th 4J N -4 1 *


"I " C A 1. lw c to:3
It% 'D 0L C 4) ii w 44 .0 .o 4 %.
0u C c 41 0,
u ,c p re E CI IO =CL













DISCUSSION


As mentioned in the introduction, the literature supporting the efficacy of psychotherapy has not demonstrated that personality and behavioral modifications which occur as a result of psychotherapy are entirely due to a psychotherapist's behavior and techniques (specific factors). The outcome of treatment, it was suggested, Is a function of moth the specific and the nonspecific (placebo) events occurring during psychotherapy, although, in the literature, the presence and the contribution of the placebo effect in psychotherapy appear to be entirely conjectural.

This study appears to demonstrate that the placebo effect in psychotherapy does exist. Subject-patients who spoke into a tape recorder for one session, believing that they were speaking to a psychotherapist, changed certain behaviors in therapeutic directions when compared to subjects who did not speak (see Hypothesis I). That is, it has been shown that verbally reported patient behaviors such as anxiety and certain moods which would be expected to change as a result of therapist behavior in actual psychotherapy, also change, in fact, as a result of the placebo psychotherapy (nonspecific treatment) in which a therapist is absent. The group who spoke to the "psychotherapist for one session, when compared to the group who waited, increased their Social Affection mood on the MACL and reduced their level of anxiety as measured on both the AACL and the IACL Anxiety mood.


62







63


As a result of these findings in the current study, it is proposed that the placebo effect does contribute to the outcome of psychotherapy, and that a measurement of the amount of contribution appears to have been made. However, It is important to question precisely what has been evaluated by a comparison between the Experimental Group (talkers) and the Control Group (waiters). If the subjects who spoke received placebo treatment and the subjects who waited received no treatment, the statistical comparison between these two groups was an accurate measurement of the placebo effect. However, it is believed that those in the Control Group, who waited, but were promised a therapist, and who changed on the dependent variables in a generally "nontherapeutic" direction, received some placebo benefit from the procedure (e.g., see Tables I1A and 13A). This finding does not Influence the conclusion that the placebo effect inpsychotherapy does exist, however, the precise therapeutic benefit of the placebo, as measured in this study, has not veen compared to the therapeutic benefit of no placebo treatment. Instead, the comparison between the Experimental and Control Groups appears to have been an analysis between a more effective (Experimental Group) and a questionably effective (Control Group) placebo treatment.

In addition to congruence of expectations regarding the characteristics of the assigned psychotherapist and positive set regarding treatment outcome (see below), may other "'psychologic mechanisms" through which the placebo effect takes place (Shapiro, 1964) be identified in order to account for the finding that this treatment did result in a placebo effect? Shapiro (19o4) reports that throughout the history of medicine, "methods of depletion (which are currently accepted as placebos)










were widely used (to effect cures): emitics, cathartics, enemas, purses, stomachics, sweating, bleeding, leeching, cupping, starvation and dehydration. Methods of depletion and expulsion . may relieve symptoms by symbolically expelling bad thoughts and conflictual ego-alien impulses." In our more sophisticated and verbally oriented culture, Shapiro continues, "relief of symptoms may occur when the patient is able to express verbally conflictual and guilt-ridden thoughts and feelings in the free, nonjudging and accepting atmosphere of the doctor's office" (p. 81). In other words, verbal catharsis, it appears, may theoretically lie at the vasis of the placebo effect in psychotherapy, Just as physical-mechanical catharsis lay at the basis of medical treatment which today is widely looked upon as placebo. The treatment in this experiment is conducted in an obviously free, accepting and nonjudgmental atmosphere and each patient-subject may establish his own personal optimal level of catharsis. It would seem logical then, that catharsis may be a major underlying factor for the placebo effect in psychotherapy, at least in so far as outcome of psychotherapy is measured in this study. It is interesting, in this regard, however, that the data indicate that a linear relationship does not exist between amount of speech (loosely, catharsis) during the "interview" and degree of change on the dependent variables. In fact, in this context, although one significant correlation among the 25 computed would be expected to be significant by chance alone, the only significant correlation (.05 level) which does exist, suggests that the more a subJect speaks, the greater the likelihood that his anxiety level (AACL) will increase. These correlation-. however, evaluate only the amount of speech in relation to the dependent variable changes after treatment. The







65

analysis does not assess content of speech or nature of affect which are certainly extremely significant aspects of catharsis. In addition, this analysis evaluates only the linear relationships between amount of speech and degree of change on the dependent variables. The analysis does not deal with nonlinear relationships, and it is conceivable that such relationships might exist.

Shapiro (1964) reports that expressed desire for treatment is another important concept which reportedly is basic in eliciting the placebo effect. That is, those subjects who are more highly motivated for treatment would be more responsive to placebo treatment. However, in this study, motivation of the subjects, as measured by their statements concerning their level of desire for psychotherapy, is not linearly related to the amount or nature of change after treatment as measured on the dependent variables. Interestingly, however, when subjects who report high motivation for psychotherapy are given a positive set concerning the outcome of the treatment and/or are offered a therapist who is congruent with their expectancies they tend to speak more. That is, greater reported desire for psychotherapy is related to a larger amount of verbal catharsis; however, it is unrelated to changes on the measures of mood and anxiety after placebo psychotherapy when a therapist is absent.

It appears, thus, that in this study, catharsis or degree of motivation, or both, have not been shown to underlie the placebo effect. Shapiro (1964), in his review of placebo literature, indicates that a great nurer o~ "psychologic mechanisms" have been proposed as being underlying characteristics of the placebo effect. Several are: catharsis,










motivation, faith, learning and coditioning, trust, confidence and previous experience with healers.

Shapiro (and Whitehorn, 1958) further indicate that no single characteristic of treatment has been consistently shown to underlie and to produce the placebo effect. The conflicting and inconclusive evidence in the literature Is most conceivably due to the unclear nature of the placebo effect, differences of measurement of the effect from research to research and the different theoretical frameworks within which different experiments are planned.

Thus far, the discussion has been concerned with the contrast between the Experimental Group (talkers) and the Control Group (walters) and several of the characteristics of psychological treatment which may determine the placebo effect. The conditions under which the placebo effect is most likely to take place or most likely to be lessened, are further clarified when attention is drawn to the way in which the treatment groups were divided within the present experiment. Analysis of Hypothesis 11, indicates that patients who are offered a "psychotherapist" congruent with their expectations, report more "positive therapeutic changes" on measures of anxiety and moods than do patients who are assigned a psychotherapist who is incongruent with their expectations concerning the nature of the psychotherapist to be assigned. That is, the subject's perception of the treater or doctor is a significant aspect in encouraging or discouraging the placebo effect.

Festinger (1957) and Festinger and Bramel (Bachrach, 1962) propose that dissonance results when two cognitions which a person holds are inconsistent with each other according to the expectations of the person.










Dissonance is said to be a motivating state which is comparable to other drive states, and just as hunger produces physical discomfort, dissonance results in psychological discomfort. When a patient-subject is instructed that the characteristics of his psychotherapist are incongruent with his personal expectations concerning this psychotherapist, the patientsubject experiences dissonance. This in turn, reduces the effectiveness of the placebo treatment, as is shown by the statistical analyses which indicate that the AACL and MACL moods of Startle and Deactivation decreased significantly more for the Congruent Therapist Group, while the #ACL mood of Pleasantness increased significantly more in the Congruent Therapist Group.

(it is noteworthy that although the Apfelbaum questionnaire [see Appendix E) instructs subjects to respond on the basis of their e tatgs., any individual subject may have completed the item in terms of his ideal psychotherapist or in terms of a desired psychotherapist. There is no apparent way in which to evaluate these possibilities. if, however, any individual subject did respond in terms of an ideal or a desired psychotherapist, he may have received additional, or reduced, positive set, in addition to having the nature of the treater confirmed or not confirmed.)

in addition to demonstrating that the patient's perception of the nature of the treater is one significant aspect of inducing a greater or lesser placebo effect, this research also suggests that the nature of the treatment situation itself is important. Analysis of Hypothesis III indicates that subjects who are told that the placebo psychotherapy will be extremely helpful to them (Positive Set Group) profit more from the










treatment than those subjects with whom positive treatment outcome is not discussed (Neutral Set Group). This conclusion is somewhat less certain than those drawn from hypotheses I and II. This is so because only two of the dependent variables (MACL moods of Startle and Pleasantness) indicated that the positive set was significantly more effective in encouraging the placebo effect than the neutral set. The literature which served as the impetus for introducing this independent variable (positive vs. neutral set) into the current study, apparently uniformly indicates that positive set or attitude toward a particular placebo treatment increases the placebo effect. (For example, see pp. 45-4( of the present study and also social psychological research on set and frames of reference: Newcomb, 1958, pp. 264-297; Maccoby t -Al., 1958, p. 95.)

Why hasn't the current study dminstrated more conclusively that a positive set concerning treacmnt outcome s even more important in securing a placebo effect, as might be anticipated on the basis of the literature? Two explanations are suggested. On the one hand, it is possible that the wording of the two instructional sets (constructed by the author) were not sufficiently different and thus, two entirely separate groups concerning expectation of treatment outcome may not have been established. On the other hand, it will be noted (see Appendix A) that the Positive Set instructions emphasized the reputation and demand for the alleged psychotherapist, in addition to his willingness to help the subject. It is possible that instructions stressing this psychotherapist's professional reputation and the demand for his services did not increase the positive nature of the set for this particular group of subjects as a whole. Thus, portions of the Positive Set instructions may not have










served the purpose for which they were Intended. (By stressing characteristics of the psychotherapist, this aspect of the instructions may have had little influence on the level of set but may have had some vicarious influence on the congruence or incongruence of the psychotherapist for any particular subject.)

In summarizing the results of the analysis of hypotheses I, II

and Iii, the following may be said: The placebo effect in psychotherapy does exist and through the technique of this study appears to be measurable. Further, the placebo effect is apparently influenced by both the perceived characteristics of the treater, and the projected outcome of the treatment as reported to the patient. The influence of the perceived characteristics of the treater has been approached by uil ysis of differences on the dependent variables between one group of subjects who were told that a therapist to be assigned to them will be congruent with their expectations, and another group who were told that the therapist will be incongruent with their expectations. The data apparently demonstrate that the group given the congruent therapist responds to the treatment with a greater positive placebo effect. The influence of the reported outcome of the treatment has been approached by analysis of differences on the dependent variables between one group of subjects who were given a positive set concerning treatment outcome and another group who were given a more neutral set. The data suggest that the group given the positive set tends to respond to treatment with a greater positive placebo effect. (However, congruent therapist instructions seem to be more instrumental in encourage ng the placebo effect than the "positive set" Instructions.)







70


Having found that offering a therapist who is congruent with a subject's expectations seems to encourage the placebo effect more than offering a positive set concerning treatment outcome may have implications beyond the present study. There is a considerable body of research concerned with the matching of patients and therapists and the resulting relationships in psychotherapy (e.g., see Leery and Gill [19591). In the present study it is seen that matching a therapist with a patient's expectations does, in fact, increase the effectiveness of the treatment to a greater extent than offering a positive set concerning treatment outcome. This is indicated in this placebo situation, where no therapist actually exists; however, it is suggested that in psychotherapy perception of the psychotherapist may also be more valuable than belief about treatment outcome. That is, the nature of the relationship (in this study, a relationship in which the patient perceives his therapist as being congruent with his expectations) may be more essential in successful psychotherapy than statements concerning the value of the treatment Itself.

An important finding of this study is that, as predicted, the Adjustment Score, which measures general level of psychological adjustment, does not change after the placebo Interview. This is so regardless of the particular treatment (talk vs. wait, congruent therapist vs. incongruent therapist and positive set vs. neutral set) or co~irnation of treatments in the study. This indicates that in a single treatment of placebo psychotherapy, no change in general level of psychological adJustment may take place. Had the placebo psychotherapy continued over a longer period of time would there have been changes in this Adustment







71


Score? That is, would extended placebo psychotherapy have produced changes in psychological adjustment similar to changes which occur after actual psychotherapy? (See Frank, 1961, and Rogers and Dymond, 1954, for evidence of changes in general level of psychological adjustment after psychotherapy.) It is difficult to speculate on precisely what may have happened to the subject's general level of psychological adjustment had the placebo treatment continued, although researchers such as Glasser and Whittlow (1953, 1954) submit that the effects of placebo treatment are not permanent and the gains from continued placebo treatment become less and less.

What might have happened to moods and anxiety had the placebo

psychotherapy continued? Martin t gI (1960), in a study concerned with levels of therapist communications (discussed more fully in the introduction), had subjects speak into a tape recorder for five 30-minute weekly sessions as if they were speaking to a psychotherapist. These authors demonstrate that although some initial positive therapeutic effect was indicated (lowering of anxiety), the procedure lost its positive therapeutic effectiveness as the number of sessions with the tape recorder Increased. Glasser and Whittlow (1953, 1954), in placebo drug studies, found that if subjects were given placebos over a period of time and required to answer questionnaires identifying the effects of these placebos, the positive placebo effect was demonstrated initially, but the effect was less upon each successive evaluation. Findings such as these suggest that the placebo procedure used In the current study, if extended for more than one session, might not have continued to have positive therapeutic effect on mood and anxiety.







72


Analysis of Hypothesis IV gives no clear picture of the contrl'ution to the placebo effect of the individual patient differenc s in age, Intelligence, introversion and neuroticism. This finding is not surprising in viewof the existing body of research which is equivocal and often contradictory in regard to this question (see Introduction). The evaluation of the data collected to explore this hypothesis resulted in borderline significance and lack of remarkable trends both within the four measured subject differences and within all of the dependent variables across Individual subject differences. Even when tentative or apparent trends were identified (as discussed on pp. 49-59), the complexity of the behaviors involved in the eight different combinations of the independent variables makes the meaningfulness of any interpretation of results doubtful. (In addition, the reader is reminded of the questionable use of the Neuroticism Scale or the MPI with this subject sample; see p. 57.) Pearson product-moment correlations were used to evaluate this hypothesis. These only identify linear relationships between the dependent variables and the measured subject differences, and the possibility of a nonlinear relationship cannot be dismissed on the basis of the analysis performed.

The problem of discussing the meaningfulness of the results and the inability to disclose significant trends in the data of this study apparently leads to the acceptance of the null hypotheses which state that, in the present saVle, age, intelligence, introversion or neuroticism will not identify the type of subject who responds favorably to placebo psychotherapy. This is in keeping with the findings of Wolf atalj. (1957) wo indicate that Individual subject reactions to placebos are








73


generally inconsistent, unpredictable and not uniform (see also Hagans ,atLL., 1957, and Kurland, 1957).

A pertinent issue may be raised concerning whether the subjects accepted the placebo treatment as a legitimate form of psychotherapy. There is no way in which to be certain of any conclusion in this regard. However, the nature of changes on the dependent variables certainly suggests that the subjects did perceive the procedure as actual psychotherapy. Moreover, it will Le noted that the Skepticism mood on the ACL remained rather constant from pre- to posttreatment testing, and the only groups to report a small and insignificant increase in Skepticism, after having completed the proceJre, were the Incongruent Therapist, Control Group subjects (see Table 16A). No subjects reported that they disbelived the legitimacy of the treatment and the comments several of the subjects made to the examiner and into the tape recorder appear related to the issue of subjects accepting the treatment as real. For example, one subject in the Experimental, Neutral Set, Congruent Therapist Group reported the following into the tape recorder during the placebo therapy: "I appreciate the privilege of being given a chance to express my feelings and views in this way. . This is the first chance I've had to honestly and openly express my feelings and opinions and I appreciate it very, very much."

It is hoped that the placebo effects demonstrated in this treatment will be kept in mind in future research on the effects of specific methods in psychotherapy. It is further hoped that this demonstration will lead to the practice of separating effects common to all psychotherapy (e.g., nonspecific placebo effects) from those effects due to the








74


specific aspects of the therapy being evaluated (e.g., therapist behaviors).


































APPENDICES

































APPENDIX A

INSTRUCTIONS TO SUBJECTS













INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS PRIOR TO
ESTABLISHING THE TREATMENT CONDITIONS IN THE STUDY

I'd like you men to do several things this afternoon. I think you will enjoy these tasks. You will learn something about yourself and may profit from what we do here for the next hour or so.

(Distribute the personal Information questionnaire.)

I'd like for you to answer these questions about yourselves.

(Distribute the Maudsley Personality Inventory.)

The next questionnaire is this one. (Read directions.) Begin.

(Distribute the Anxiety Adjective Check List.)

The next questionnaire is this one. (Read directions.) Please answer honestly the way you feel 19 n9. Begin.

(Distribute the Mood Adjective Check List.)

The next questionnaire is this one which asks you about your mood and the way you feel fight n. (Read directions.) Some of the items in this list are like items in the last list, but please answer them anyway. Be honest with yourself and just answer the way you feel right now.

(Distribute the Modification of the Expectation Q-Sort.)

This questionnaire is a little different. It says--(read directions). Now, don't think too long on any of these, but answer them as best you can.

(Distribute the Adjustment Score.)

This questionnaire asks you how you feel about yourself and also about things in general. Read the Instructions and work quickly


INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS TO DETERMINE
POSITIVE SET OR NEUTRAL SET

Positive Set Instructions

I can tell you now what this is all about. It's a real good opportunity for you men because you are being given a chance to talk with a great psychiatrist. Now, you can talk to him about anything you'd like 77










relating to problems you've had, or problems you are having, or problems you can see ahead of you in the future. This psychiatrist has such an excellent reputation that he is an extremely busy man, but he has offered to help you men to solve any problems you may have anyway. Because he is so busy he can't get here to the hospital so he suggested we do something a little differently because he is so interested in helping you. You will silly talk into a tape recorder, and the tape will be sent registered mail to him. I think this is really great because he has helped so many, many people and I 'm sure he can help you with any problems you may possibly have.


Neutral Set Instructions

I can tell you now what this is all about. It is an opportunity for you men to talk with a psychiatrist. Now, you can talk to him about anything you'd like relating to problems you've had, or problems you are having, or problems you can see ahead of you in the future. This psychiatrist is an extremely busy man, and he is unable to get here to the hospital so he suggested we do something a little differently. You will simply talk into a tape recorder, and the tape will be sent registered mail to him.


INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS TO DETERMINE
CONGRUENT THERAPIST R I NCONGRUENT THERAPIST

Conruen T a st Inst ructi ns

I've been looking over your questionnaire on the way you would
expect this psychotherapist or counselor to be, and I see that this psychiatrist is the way you thought he would be. For example, Mr. (subject seated on the Experimenter's right), he (read this subject's responses to items 7, 14 and 29; see Appendix E). And Mr. (subject seated on the Exoerimenter's left), he (read this subject's responses to items 1, 31 and 32; see Appendix E).


Inconr u snt Therit Ins tructions

I've been looking over your questionnaire on the way you would
expect this psychotherapist or counselor to be, and I see that this psychiatrist is not exactly the way you thought he would be. For example, Mr. (subject seated on the Experimenter's right), he (read a denial of the accuracy of this subject's responses to Items 7, 14 and 2); see Appendix E). And Mr. (subject seated on the Experimenter's left), he (read a denial of the accuracy of this subject's responses to items i, 31 and 32; see Appendlx E).








79


INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECT AFTER HAVING
ESTABLISHED THE FOUR TREATMENT CONDITIONS

Mr. (Control Group subject), I 'd li ie for you to wait here
for a while. Make yourself comfortable. You can look at some magazines, or rest, but please stay in this room. I'll be back in about forty-five minutes. Mr. (Exerimental Group subject), please come with me.


INSTRUCTIONS GIVEN TO THE EXPERIMENTAL GROUP SUBJECT AFTER
ING SEATED IN THE ROOM WITH THE TAPE RECORDER

Now, this is the tape recorder and the tape that wi, be sent to the psychiatrist. Just make yourself comfortable, and say whatever you would lixte int> this microphone. The controls on the machine are set, and please remain in your seat until I return. It's OK If there are pauses, and you don't have to talk all of the time. I'll be back in about fifty minutes to get the tape.


I NSTRUCTI ONS GIVEN TO EACH PAIR OF SUBJECTS AFTER THE EXPERIMENTAL
GROUP SUBJECT IS REUNITED WITH THE CONTROL GROUP SUBJECT

I'd like for you to retae several of these shorter tests. This will only take a few minutes nmre.

(Distribute the Anxiety Adjective Check List.)

This is the first one. You remember this one, it asks you to check the way In which you feel right na. Please answer honestly the way you feel iit now.

(Distribute the tod Adjective Check List.)

This is the text one. You remember, this one, it asks you to mark the way you feel right now. Please answer honestly the way you feel liaqlno.

(Distribute the Adjustment Score.)

This is the next one. Please work as quickly as you can.


INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS
AFTER THE PAIR HAS COMPLETED THE TESTS

Mr. (Exerimental Group subject), you may leave now. Mr.
(Control Group subject), do you care to speak to the psychiatrist? If the Control Group subject wishes to speak, he is taken to the tape recorder; if, however, he declines to speak, he is dismissed at thki time.)

































APPENDIX 8E

SAMPLE ANXIETY ADJECTIVE CHECK LIST













Name

Belw you will find words which describe different kinds of feelIngs. Check the words which describe how you feel RIGHT NOW. Some of the words may sound alike but we want you to check all of the words that describe your feelings.

i. AFRAID 21. GAY 41. PANICKY 2. AGITATED 22. ___GLOOMY 2. PEACEFUL 3. ANGRY 23. .GRIM 43. PLEASANT 4. BITTER 24. HAPPY 44. RATTLED

5. CALM 25. HELPLESS 45. SAD o. CHARMING 26. HOPELESS 46. SECURE

7. CHEERFUL 27. INSECURE 47.___SENTIMENTAL

8. COMPLAINING 2.. JEALOUS t. SERIOUS 9. CONTENTED 29. JOYFUL 49. SNAKEY 10. CONTRARY 30. KINDLY 50. SOLEMN 11. COOL 31. __LIGHT-HEARTED 51. STEADY 12. CROSS 32. LONELY 52. TENDER 13. DESPERATE 33. LOVING 53. TENSE 14._ EASY-GOING 34. MAD 54. TERRIFIED 15. FEARFUL 35. MEAN 55. THREATENED 16. FEARLESS 36. MERRY 56. THOUGHTFUL 17. FRETFUL 37. MISERABLE 57. UNCONCERNED 18. FRI ENGLY 38. NERVOUS 5S. UNEASY 19. FRIGHTENED 39. OVERCONCERNED 59. UPSET

2 ___FURIOUS 40. OVERWHELMED 60. WARM 61. WORRYING
































APPENDIX C

SAMlPLE MOOD ADJECTIVE CHECK LIST













Name

Each of the 40 words in the following list describes feelings or mood. Please use the list to describe your feelings at this moment. Mark each word according to these instructions: If the word definitely describes how you feel at the moment you read It, ircle the double plus
(++) to the right of the word. For example, If the word is calm and you are definitely -:eeling calm at the moment, circle the double plus as follows:

calm G + 7 no (This means that you definitely feel calm at this moment.)

If the word only slightly applies to your feelings at the moment, circle the single plus as follows:

calm ++ ) 7 no (This means that you feel slightly calm at this moment.)

if the word is not clear to you or if you cannot decide whether or not it describes your feelings at the moment, circle the question mark as follows:

calm ++ + ) no (This means you can't decide whether you are calm.) If you clearly decide that the word does not apply to your feelings at
the moment, circle the no as follows: calm ++ + ? ( (This means that you are sure that you are not calm at the moment.)

Work rapidly. Your first reaction is best. Work down the column before going to the next. Mark all the words. This should take only a few minutes.

angry .. .. + ? no boastful ...... ++ + ? no c oncentratlng ++ + ? no elated ...... ++ + ? no drowsy .. .. -+ + ? no active .... .. + ? no

affectionate + + ? no nonchalant .. ++ + no apPrehensive + ? no skeptical .. .. ++ + ? no blue .... + ? no shocked ...... ++ + ? no



83










bold .. .. .. ++ + ? no pleased .... ++ + ? no earnest .... ++ + ? no tired .... ++ + ? no sluggish .. ++ + ? no kindly .. .. ++ + ? no forgiving .. ++ + ? no fearful .... + + ? no clutched up .. ++ + ? no regretful .. ++ + ? no lonely .. .. ? no gotistic .. ++ + ? no cocky .... ++ + ? no overjoyed .. ++ + ? no lighthearted .. ++ + ? no vigorous .. ++ + ? no energetic .. 44 + 7 no witty .... ++ + 7 no playful .. .. + ? no rebellious .. ++ 4 ? no suspicious .. ++ + 7 no serious .... + + ? no startled .. ++ + ? no warnearted .. ++ + ? no defiant ? no Insecure .. ++ + ? no engaged In thought ++ + ? no self-centered 4+ + 7 no

































APPENDIX D

SAMPLE ADJUSTMENT SCORE













Name

If the statement is true of you, circle "True." If the statement is not true of you, circle 'False."'

1. I am no one. Nothing seems to t, me. True False 2. 1 am optimistic. True False 3. I am a hostile person True False 4. Self-control Is no problem to me. True False

5. I tend to be on my guard with people who are somewhat more friendly than I had expected. True False B. I try not to think about my problems. True False 7. I have an attractive personality. True False 8. I am a rational person. True False

9. I have a horror of falling In anything I want to accomplish. True False 10. I put on a false front. True False II. I feel helpless. True False 12. I am a good mixer. True False 13. It is difficult to control my aggression. True False 14. I feel insecure within myself. True False 15. I really am disturbed. True False

1. All you have to do is just insist with me, and
I give in. True False 17. I have to protect myself with excuses, with
rational zing. True False 18. I have few values and standards of my own. True False 19. I take a positive attitude toward myself True False







S7


20. I am ambitious. True False 21. I don't trust my emotions. True False 22. 1 have the feeling that I am just not facing facts. True False 23. 1 shrine from facing a crisis or difficulty. True False 24. 1 am assertive. True False 25. I am tolerant. True False 26. I have initiative. True False 27. I express my emotions freely. True False 2'. I can accept most social values and standards. True False 29. I feel hopetess. True False 30. It isoretty tough to be me. True False 31. 1 am contented. True False 32. My decisions are not my own. True False 33. 1 am a failure. True False 34. 1 make strong demands on myself. True False 35. 1 despise myself. True False 36. I doubt my sexual power. True False 37. I am afraid of a full-fledged disagreement with
a person. True False 38. I am self-reliant. True False 39. I am sexually attractive. True False 40. I am relaxed and nothing really bothers me. True False 41. 1 an likable. True False 42. I am worthless. True False 43. I often kick myself for the things I do. True False 44. 1 am different from others. True False










45. I am poised. True False 4. I am shy. True False 47. I usually feel driven. True False 48. I can't seem to make upc my mind one way or the other. True False 49. I feel emotionally mature. True False 5j. I can usually live comfortably with people. True False 51. 1 understand myself. True False 52. 1 am afraid of sex. True False 53. 1 can usually make up my mind and stick to it. True False 54. I have a warm emotional relationship with others. True False 55. I usually like people. True False 56. 1 am satisfied with myself. True False .7. 1 often feel humiliated. True False 58. I have a feeling of hopelessness. True False 59. I am Intelligent. True False 60. I dislike my own sexuality. True False 61. My hardest battles are with myself. True False 62. I am a hard worker. True False 63. I am disorganized. True False 64. I feel tired out. True False 65. I am responsible for my troubles. True False a6. I am unreliable. True False 67. I am a responsible person. True False 68. My persnalIty is attractive to the opposite sex. True False 69. I want to give up trying to cope with the world. True False 70. I an confused. True False










71. I feel adequate. True False 72. I am Iiked by imst p~le who ent me. True False 73. I just don't respect myself. True False 74. 1 am impulsive. True False
































APPENDIX E

SAMPLE OF TE MODI00FICATION OF THE EXPECTATION Q-SORT














Name

If you were to have a psychotherapist or counselor to talk with about your problems, what would you expect him to be like? If the statement is true concerning how you would expect the counselor or psychotherapist to be, put a circle around "true." If the statement is false concerning how you would expect this counselor or psychotherapist to be, put a circle around 'false.'

I. Calm, easygoing .. .. .. .. .. .. .. True False
2. Is careful not to let people waste his time.. .. True False 3. Looks for the good points In people .. .. .. True False
4. Likes to have a hand in managing other people's affairs .. .. .. .. .. .. .. .. .. True False
5. Is concerned with what's right .. .. True False o. Blunt, straightforward, calls a spade a spade .. True False 7. Hard to get to know. .. .. .. .. .. .. True False
8. Is likely to give advice and guidance .. .. .. True False 9. Cares what other people think of him .. .. .. True False lu. Is likely to overestimate a person's abilities .. True False 11. Is Indulgent, forgiving. .. .. .. .... True False
12. I able to sense other people's feelings .. .. True False 13. Is careful not to uoset others .. .. .. True False 14. Judges the behavior of others. .. .. .. .. True False 15. Expects the individual to shoulder his own responsibilities .. .. .. ........ True False
16. Is logical, sticks to the facts .. .. .. .. True False 17. Is likely to keep his irritations or resentments
to himself .. .. .. .. .. .. .. True False
18. Is gentle, tender .. .. .. .. .. .. .. True False
19. Self-satisfied .. .. .. .. .. .... True False
20. Never makes people fet uncomfortabe .. .. True False 21. Hard to deceive, does not accept things at face value True False 22. Businesslike. .. .. .. .. .. .. .. .. True False
23. Is conscientious about duties and responsibilities True False 24. LiKes to do a good job .. .. .. .. .. .. True False
25. Is not emotional .. .. .. .. .. .. .. True False
26. Reacts to most people in about the same way.. .. True False 27. Sympathetic.. .. .. .. .. .. True False
20. Tries to discover who's to blame for mistakes made True False 29. Is able to change his opinions easily .. .. .. True False 30. Diplomatic ... .. .. .. .. .... True False
31. Is troubled by the misfortunes of others .. .. True False 32. Persuasive .... ..T... .. .. .. .. True False
33. Well adjusted, gets along well in the world.. .. True False 34. Is quick to give encouragement and reassurance .. True False 35. Has no trouble getting along with people, makes
friends easily .. .. .. .. .. .. .. True False
3 Critical, not easily impressed .. .. .. .. True False
vji
































APPENDIX F

SAMPLE PERSONAL INFORMATION QUESTIONNAIRE




Full Text
3
Effects of treatment with drug placebos are those nonspecific ef
fects given the greatest attention in the literature. Based on his re
view of the literature, Shapiro (1964) states that placebo effects are
subtle and common to all treatment situations and are the most important
nonspecific factors in psychological treatment. He defines the placebo
effect as the "psychologic, physiologic, or psychophysiologic effect of
aw. medj.cat.i.on or procedure given with therapeutic intent which is Inde
pendent of or minimally related to the effects of the medication or to
the specific effects of the procedure and which operates through a psy
chologic mechanisnf' (p. 293). In addition, the administration of this
medication or procedure must be recognized as legitimately therapeutic by
the patient.
Rosenthal and Frank (1956) also refer to the placebo effect as a
type of nonspecific event in psychological treatment and they report that
since research in psychotherapy has not dealt directly with the placebo
effect, the nature of Its Influence is undetermined. However, these au
thors speculate that it occurs with considerable regularity, and they
stress the need to demonstrate through research that observed effects of
psychotherapy are due to the therapist's techniques and exist separate
from nonspecific effects.
Borgatta (1959) expands on the importance of the placebo effect
by reporting that "If there is no evidence that an agent is the cause of
a given outcome, and there is also no evidence that a placeoo is the
cause of a given outcome, an appropriate statement is that there is no
evidence that the agent is more efficient than the placebo in bringing


34
TABLE 9
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL SOCIAL AFFECTION MOOD
Source
ss
of
V
F
A (experimental vs. control)
27.15
1
27.15
4.35*
B (congruent vs incongruent)
14.37
1
14.37
2.30
C (positive vs. neutral)
03
1
.03
.001
AB
19.62
1
19.62
3.14
AC
10.84
1
10.84
1.74
SC
1.36
1
1.36
.22
ABC
22.60
l
22.60
3.62
error
543.18
8Z
6.24
Totai
639.15
94
^Significant ^.05.
TABLE 9A
PRETREATMENT MEANS (X.), POSTTREATMENT MEANS (X£) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL SOCIAL AFFECTION MOOD
Experimental Group
Control
Group
Congruent
Therapist
1ncongruent
Theraplst
Congruent
Therapist
Incongruent
Theraplst
Posl ti ve X | = 6.17
Set X2 = 7.84
D + I7
X, 6.25
X2 5.00
D * -1.25
X, 6.08
X2 0
D -.58
X, 8.00
x2 1
D -.17
Neutral Xt 7.58 Xj 7-75 X, 7-58 X, 6.42
Set X2 8.42 X2 8.08 X2 b.75 X2 5.25
o +.84 d +.33 0 -.83 0 -I.17


79
INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS AFTER HAVING
ESTABLISHED THE FOUR TREATMENT CONDITIONS
Mr. (Control Group subject), I'd like for you to wait here
for a while. Make yourself comfortable. You can look at some magazines,
or rest, but please stay in this room. I'll be back in about forty-five
minutes. Mr. (Experimental Group subject), please come with me.
INSTRUCTIONS GIVEN TO THE EXPERIMENTAL GROUP SUBJECT AFTER
BEING SEATED IN THE ROOM WITH THE TAPE RECORDER
Now, this Is the tape recorder and the tape that wi 11 be sent to
the psychiatrist. Just make yourself comfortable, and say whatever you
would like into this microphone. The controls on the machine are set,
and please remain In your seat until I return. It's OK if there are
pauses, and you don't have to talk all of the time. I'll be back In
about fifty minutes to get the tape.
INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS AFTER THE EXPERIMENTAL
GROUP SUBJECT IS REUNITED WITH THE CONTROL GROUP SUBJECT
I'd like for you to retake several of these shorter tests. This
will only take a few minutes more.
(Distribute the Anxiety Adjective Checx List.)
This is the first one. You remember this one, it asks you to
check the way in which you feel right now. Please answer honestly the
way you feel riqht now.
(Distribute the Mood Adjective Check List.)
This is the next one. You remember this one. It asks you to mark
the way you feel right now. Please answer honestly the way you feel
right now.
(Distribute the Adjustment Score.)
This is the next one. Please work as quickly as you can.
INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS
AFTER THE PAIR HAS COMPLETED THE TESTS
Mr. (Experimental Group subject), you may leave now. Mr.
(Control Group subject), do you care to speak to the psychiatrist?
(If the Control Group subject wishes to speak, he Is taken to the tape
recorder; If, however, he declines to speak, he is dismissed at this
time.)


APPENDIX H
SUBJECT RAW DATA


89
71.
1
feel adequate.
True
72.
1
am liked by must people who know me.
True
73.
1
just don't respect myself.
True
74.
1
am impulsive.
True
False
False
False
False


68
treatment than those subjects with whom positive treatment outcome is not
discussed (Neutral Set Group). This conclusion is somewhat less certain
than those drawn from hypotheses I and II. This Is so because only two
of the dependent variables (MACL moods of Startle and Pleasantness) indi*
cated that the positive set was significantly more effective in encourag
ing the placebo effect than the neutral set. The literature which served
as the impetus for introducing this Independent variable (positive vs.
neutral set) into the current study, apparently uniformly indicates that
positive set or attitude toward a particular placebo treatment increases
the placebo effect. (For example, see pp. 45-46 of the present study and
also social psychological research on set and frames of reference:
Newcomb, 1958, pp. 264-297* Maccoby et ai., 1958, p. 95.)
Why hasn't the current study demonstrated more conclusively that
a positive set concerning treacmont outcome Is even more important in se
curing a placebo effect, as might be anticipated on the basis of the lit
erature? Two explanations are suggested. On the one hand, it is possi
ble that the wording of the two instructional sets (constructed by the
author) were not sufficiently different and thus, two entirely separate
groups concerning expectation of treatment outcome may not have been es
tablished. On the other hand, it will be noted (see Appendix A) that the
Positive Set instructions emphasized the reputation and demand for the
alleged psychotherapist, in addition to his willingness to help the sub
ject. It is possible that Instructions stressing this psychotherapist's
professional reputation and the demand for his services did not increase
the positive nature of the set for this particular group of subjects as a
whole. Thus, portions of the Positive Set instructions may not have




APPENOIX E
SAMPLE OF THE MODIFICATION OF THE EXPECTATION Q-SORT


37
TABLE 12
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL EGOTISM MOOD
Source
ss
df
F
A (experimental vs. control)
4.53
1
4.53
1.09
B (congruent vs. incongruent)
2.10
1
2.10
.51
C (positive vs. neutral)
3.9
1
3-19
.77
AB
.57
1
57
.14
AC
11.58
1
11.58
2.80
BC
1.33
1
1.33
.32
ABC
2.60
1
2.60
.63
error
353.86
sz
4.14
Total
385-76
94
TABLE 12A
PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS (X2) AND MEANS
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN CM TPr MACL EGOTISM MOOD
Experimental Group
Control
Group
Congruent
Therapist
1ncongruent
Therapist
Congruent
Therapist
Incongruent
Therapist
Posit ve
Set
X, 3.08
x2 *2
D -.58
X, 3.25
x2 2.08
D -1.17
X, 3.00
X2 3.08
D +.08
X, 2.67
x2 USl
D +.58
X, = 2.58 X, 2.50
*2 -
0 +.50 D +.17
X, 2.42
*2 "
D -.25
Neutral
Set


In the present study, one half of the subjects will speak into
the tape recorder while the other half will wait an Identical length of
6
time. Since both groups will be given the instructions that they are
about to speak to a psychotherapist through the tape recorder this will
permit a comparison between a group given both the promise and the ges-
ture of help and a group given only the promise of help. Changes in mood
and anxiety in the group which does speak may be viewed as a demonstra
tion of the nonspecific placebo effect in a single session of psychologi
cal treatment. The effects of this treatment should be utilized as a
base line against which further research may be compared to demonstrate
that an actual psychotherapeutic interview produces changes different
from, or greater than the "placebo psychotherapy."
Shapiro (1964) reports that important features of a therapeutic
relationship which are responsible for the placebo effect are basically
elements of (I) the patient, (2) the doctor, and (3) the treatment situa
tion; in addition to a fourth factor: the doctor-patlent relationship.
That is, there are certain characteristics of each of these components of
treatment which contribute to a nonspecific placebo effect.
0) Ifaff. patient.
Shapiro, in his review of drug research (1964), points out that
many individual differences of patients have been evaluated to assess
possible relationships with the placebo effect. Such characteristics as
sex, age, intelligence and diagnosis have been observed repeatedly in
order to ascertain if a certain type of person responds more readily to
placebo factors. However, as Shapiro reports, this body of research is
equivocal and often contradictory. For example, Lasagna et al. (1954)


20
am ambitlous.
21.
22.
23.
24.
25.
26.
27.
28.
* 29.
30.
3K
32.
33.
34.
35.
36.
37.
38.
39-
40.
41.
42.
43.
44.
I don't trust my emotions.
I have the feeling that I am just not facing facts.
I shrink from facing a crisis or difficulty.
I am assertive.
I am tolerant.
I have initiative.
I express my emotions freely.
I can accept most social values and standards.
I feel hopeiess.
It isoretty tough to be me.
I am contented.
Hy decisions are not my own.
I am a failure.
I make strong demands on myself,
i despise myself.
I doubt my sexual power.
I am afraid of a full-fledged disagreement with
a person.
I am self-reliant.
I am sexually attractive.
I am relaxed and nothing realiy bothers me.
I an likable.
I am worthless.
I often kick myself for the things i do.
i am different from others.
87
True
False
True
Fal se
True
Fai se
True
Fal se
True
False
True
False
True
False
True
False
True
False
True
False
True
Fal se
True
False
True
False
True
Fal se
True
False
True
False
True
False
True
False
True
False
True
False
True
False
True
False
True
Fal se
True
False
True
False


50
TABLE 20
SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE
RANOOM DISTRIBUTION OF MPI INTROVERSION SCORES
Source
ss
df
V
F
A (experimental vs. control)
19.26
1
19.20
.34
B (congruent vs Incongruent)
31.51
1
31.51
.55
C (positive vs. neutral)
33.84
1
33.84
.59
AB
52.51
1
52.51
.92
AC
.02
1
.02
.00
BC
128.35
1
128.35
2.25
ABC
31.50
1
31.50
.55
w Cells
5.012.75
SSL
56.96
Total
5.309.74
95
(An F of 3.95 is significant at the .05 level.)


21
(Both the AACL and the Anxiety Mood on the MACL reportedly meas
ure level of manifest anxiety. The MACL Anxiety mood contains four af
fectively toned adjectives, each of which the subject responds to by de
noting the extant to which the adjective describes his current feelings.
The AACL, on the other hand, asks a subject to check those adjectives
that apply to his current feelings. Only one of the four adjectives on
the MACL Anxiety mood is identical with the 21 adjectives which are scored
on the AACL. A comparison between pretreatment performance on the two
Anxiety measures resulted In a Pearson product-moment correlation of +.4,
and a Pearson product-moment correlation between the subject's changes in
the two scales from pre- to posttreatment testing resulted In a correla
tion of +.57* Both these correlations are significant beyond the .001
level. This suggests that despite the basic uniqueness of the two meas
ures, they are measuring essentially the same factor of "anxiety.")
(3) The Adjustment Score (see Appendix D).
The Adjustment Score was chosen to evaluate changes in general
level of psychological adjustment, it was originally constructed by
Dymond (1953) as a Q-sort to evaluate the effects of nondirective coun
seling. The statements were compiled empirically by Butler and Haigh who
noted that they were representative of the positive and negative comments
made by people in psychotherapy.
Four 'Well-trained, practicing clinical psychologists" who were
not client-centered by theoretical orientation "agreed remarkably welI"
In choosing 37 negative indicators of adjustment and 37 statements de
scribing positive adjustment (1953). These 74 statements rated on a
"like me" to 'unlike me" dimension constitute the scale.


treatment (Rosenthal and Frank, 1956) for they produce manifest effects
which are indistinguishable from the effects produced by the intentional
treatment prescribed by the particular theory of psychotherapy.
2
Strupp (1962) indicates that because these nonspecific factors
have been largely neglected in research in psychotherapy, the extent of
their influence is undetermined. Many significant theorists and research
ers (e.g., Rogers, 1961) expound on the efficacy of their psychothera
peutic systems, although nonspecific factors, which may greatly contrib
ute to the final outcome of psychological treatment are not effectively
accounted for, controlled or measured. Understandably, successful out
come of psychotherapy Is viewed as proof of the effectiveness of psycho
therapy. However, It has not been demonstrated that personality and be
havioral modifications which occur as a result of psychotherapy are en
tirely (or even largely) due to the psychotherapist's specific techniques,
instead, the actual outcome of treatment is a function of both the spe
cific and nonspecific events occurring during psychotherapy.
Eysenck (1961), after reviewing and surnnarizing the literature,
shows that actual psychotherapy does not achieve higher recovery rates
than those reported from ordinary life experiences and the nonspecific
effects of routine medica) treatment. After demonstrating that nonspe
cific treatment produces effects comparable with actual psychotherapy,
Eysenck implies that nonspecific factors in psychotherapy are essential
factors responsible for bringing about successful psychotherapeutic
change. Findings and statements such as these clarify the necessity for
research In this area.


45.
1
am poised.
True
88
Fal se
46.
1
am shy.
True
Fal se
47.
1
usually feel driven.
True
False
48.
1
can't seem to make up my mind one way or the other.
True
False
49.
1
feel emotionally mature.
True
False
50.
1
can usually live comfortably with people.
True
False
5K
1
understand myself.
True
False
52.
1
am afraid of sex.
True
Fal se
53.
1
can usually make up my mind and stick to it.
True
False
54.
1
have a warm emotional relationship with others.
True
False
55.
1
usually like people.
True
Fal se
56.
1
am satisfied with myself.
True
Fal se
57-
1
often feel humiliated.
True
False
58.
1
have a feeling of hopelessness.
True
Fal se
vn
u>

1
am intel1igent.
True
Fal se
60.
1
dislike my own sexuality.
True
False
61.
My hardest battles are with myself.
True
False
62.
1
am a hard worker.
True
False
63.
1
am disorganized.
True
False
64.
1
feel tired out.
True
Fal se
65.
1
am responsible for my troubles.
True
False
66.
1
am unreliable.
True
False
67.
1
am a responsible person.
True
False
68.
My personality Is attractive to the opposite sex.
True
False
69.
1
want to give up trying to cope with the world.
True
False
70.
1
am confused.
True
False


39
TABLE 14
SUMMARY TABLE OF THE ANALYSIS OF iOVARIANCE FOR THE
MACL ACTIVATION MOOD
Source
ss
df
V
F
A (experimental vs. control)
2.90
1
2.90
.80
B (congruent vs. incongruent)
1.80
1
1.80
.50
C (positive vs. neutral)
.20
1
.20
.06
AB
5.58
1
5-58
1.54
AC
.47
1
47
.13
BC
.00
1
.00
.00
ABC
.12
1
.12
.03
error
i!4^4
SZ
3.62
Total
326.01
94
TABLE I4A
PRETREATMENT MEANS (X,). POSTTREATMENT MEANS (X.) AND MEAN
CHANGE SCORES (O) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL ACTIVATION MOOD
experimental Group Control Group
Congruent Incongruent Congruent Incongruent
Therapist
Therapist
Therapist
Therapist
Pos!tive
Set
X, 4.25
X2 4.00
D -.25
X, = 4.00
x2 .Liz
0 = -.33
X, 4.42
X2 3^92
D -.50
X, 4.25
X2 4.50
D +.25
Neutral
Set
X, 3.75
X2 -
0 -.25
X, 3.58
X2 1M
D -.50
X, 3-92
X2
D -.34
X, 4.58
x2 L2&
D +.34


4
about the outcome. Further, there is also no evidence that the placebo
is not as effective as the agent in bringing about the outcome" (p. 331)*
The literature is replete with references confirming the efficacy
of nonspecific effects of drug placebos (see reviews by Shapiro, 1964 and
Hongingfield, 1963) However, despite the demonstrated effectiveness of
drug placebos, and the alleged importance of the placebo effect in psy
chological treatment, adequately controlled studies evaluating the pla
cebo effect in psychotherapy are conspicuously lacking (Rosenthal and
Frank, 1956 and Strupp, 1962). One is currently forced, therefore, to
rely on implication, Inference, and analogy when constructing evidence of,
and predictive hypotheses pertaining to, the influence of nonspecific
placebo factors In psychotherapy.
Ib.e Preeq.t.,
"As Withering's foxglove was made more potent when digitalis was
Isolated from the concoction, and after it was recognized that only the
dropsy of congestive heart failure was benefited, so will psychotherapy
become more potent after the placebo effect is isolated and dissected
free from the psychotherapeutic process" (Shapiro, 1964, p. 85).
As long as both specific and nonspecific (placebo) factors oper
ate together during the treatment, one may never accurately determine the
real contribution of the placebo effect to psychotherapy. Since nonspe
cific placebo factors reportedly function unbeknown to the psychothera
pist, since their presence is reportedly a factor in the outcome of all
psychotherapy, and since they remain a nebulous and poorly defined factor
in psychotherapy, it is clear that the extraction of these placebo effects


12
through which specific effects may result. Reactions to this placebo
situation can be viewed as "placebo effects." Without knowledge of the
nature of the treatment, the effects may appear to result from the spe
cific behavior or techniques of a psychotherapist. However, his absence
from the treatment necessarily means that all effects are nonspecific, or
placebo.
Since there is no psychotherapist present, can this situation be
thought analogous to a therapeutic interviews The patient is talking to
a therapist who exists only in the patient's fantasy and is instituted by
the experimenter. Apfelbaum (1958) points out that individuals tend to
create, in part, interpersonal experience and that ongoing interpersonal
process in therapy Is primarily a function of a patient's personal trans
ferences. A patient, in utilizing past experience and partial cues, pro
jects the image of his psychotherapist, and thereby unknowingly perceives
him in part through his projective capacity. Consequently, the therapist
is perceived largely as a function of the patient himself. The attitudes
a patient holds toward his alleged therapist In the "shanf* or "placebo"
interview suggested above may be controlled and manipulated by experi
mental design. The therapist exists entirely in fantasy and as a func
tion of the patient's projective capacity and description of him as re
ported by the experimenter. The patient, however, in effect, establishes
an intrapersonal relationship which he believes to be, and which appears
to have the qualities of, an interpersonal psychotherapeutic relationship.
Thus, the sham interview will be regarded as both a placebo
treatment and a therapeutic interview.


PLACEBO PSYCHOTHERAPY AND CHANGE IN
ANXIETY, MOOD AND ADJUSTMENT
By
HERBERT GOLDSTEIN
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
April, 1965

ACKNOWLEDGMENTS
This study would not have been completed had It not been for the
foresight, wisdom and encouragement of many people to whom the writer is
indebted and grateful.
A special measure of gratitude is due Dr. Audrey S. Schumacher,
chairman of the supervisory committee, for her genuine interest, her un
tiring patience and her constant support and guidance. The assistance
and encouragement of the remainder of the supervisory committee, Dr. R.
J. Anderson, Dr. 8. Barger, Dr. J. J. Wright and Or. V. A. Hines, is
warmly accepted and greatly appreciated.
The writer wishes to acknowledge the staff and office workers of
the Gulfport, Mississippi Veterans' Administration Hospital who gave un
sparingly of their time, their cooperation and their resources. A par
ticular note of gratitude is extended to three people who specifically
asked to remain unmentioned, but without whom the progress of this study
would certainly have been impeded.
The writer wishes to thank his parents for their faith, for their
Investment and for wanting someday to say: "there's my son, the Doctor."
Most of all, the writer thanks his wife, Alee, for her under
standing, her support, her guidance and her remarkable patience which
sustained the writer in his efforts to complete this study.
ii

TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS II
LIST OF TABLES Iv
INTRODUCTION . I
The Present Study 4
METHOD. 13
Subjects 14
Means of Measurement 16
Procedure 24
RESULTS 28
DISCUSSION 62
APPENDICES 75
A INSTRUCTIONS TO SUBJECTS 76
B SAMPLE ANXIETY ADJECTIVE CHECK LIST 80
C SAMPLE MOOD ADJECTIVE CHECK LIST 82
D SAMPLE ADJUSTMENT SCORE 85
E SAMPLE OF THE MODIFICATION OF THE EXPECTATION
Q-SORT 90
F SAMPLE PERSONAL INFORMATION QUESTIONNAIRE 92
G SAMPLE LETTER SENT TO ALL SUBJECTS WHO SPOKE
INTO THE TAPE RECORDER 94
H SUBJECT RAW DATA 96
BIBLIOGRAPHY 103
Hi

LIST OF TABLES
Table page
1 THE EXPERIMENTAL DESIGN 13
2 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE
DEMONSTRATING THE RANDOM DISTRIBUTION OF
AGE 15
3 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE
DEMONSTRATING THE RANDOM DISTRIBUTION OF
i .a 16
4 THE TWELVE MOOD FACTORS OF THE MACL AND
THE ADJECTIVES CONSTITUTING EACH MOOD
FACTOR >9
5 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE ZUCKERMAN ADJECTIVE CHECK LIST 30
5A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
ZUCKERMAN ADJECTIVE CHECK LIST 30
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL AGGRESSION MOOD 31
6A PRETREATMENT MEANS (X.), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE CORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL AGGRESSION MOOD 31
7 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL CONCENTRATION MOOD . 32
7A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (0) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL CONCENTRATION MOOD 32
8 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL DEACTIVATION MOOD 33
Iv

LIST OF TABLESContinued
Tabie Page
8A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS
(XJ AND MEAN CHANCE SCORES (0) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL DEACTIVATION MOOD 33
9 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL SOCIAL AFFECTION MOOD 34
9A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL SOCIAL AFFECTION MOOD. 34
10 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL ANXIETY MOOD 35
10A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL ANXIETY MOOD 35
11 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL DEPRESSION MOOD 36
11A PRETREATMENT MEANS (X|), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL DEPRESSION MOOD 36
12 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL EGOTISM MOOD 37
12A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEANS CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL EGOTISM MOOD 37
13 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL PLEASANTNESS MOOD 38
13A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL PLEASANTNESS MOOD 38
v

LIST OF TABLESContinued
Table Page
14 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL ACTIVATION MOOD 39
14A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
TREATMENT CONDITION OF THE DESIGN ON THE
MACL ACTIVATION MOOD 39
13 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE,
FOR THE MACL NONCHALANCE MOOD 40
15A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL NONCHALANCE MOOO 40
16 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL SKEPTICISM MOOD 41
ioA PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL SKEPTICISM MOOD 41
17 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL STARTLE MOOO 42
I7A PRETREATMENT MEANS (X.), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL STARTLE MOOD 42
18 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE ADJUSTMENT SCORE 43
I8A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
ADJUSTMENT SCORE 43
19 SUMMARY OF THE DEPENDENT VARIABLES AND THE
SIGNIFICANCE LEVELS AT WHICH THEY
DISTINGUISHED 8ETWEEN THE INDEPENDENT
VARIABLES 48
vl

LIST OF TABLESContinued
Table Page
20 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE
DEMONSTRATING THE RANDOM DISTRIBUTION OF
MPI INTROVERSION SCORES 50
21 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE
DEMONSTRATING THE RANDOM DISTRIBUTION OF
MPI NEUROTICISM SCORES 51
22 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN
SUBJECTS' CHANGE SCORES FROM PRE- TO POST
TREATMENT TESTING ON EIGHT OF THE DEPENDENT
VARIABLES AND SCORES ON THE BETA INTELLIGENCE
TEST. THE DATA ARE PRESENTED FOR EACH OF THE
EIGHT TREATMENT CONDITIONS 53
23 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN
SUBJECTS' CHANGE SCORES FROM PRE- TO POST
TREATMENT TESTING ON EACH OF THE DEPENDENT
VARIABLES AND SCORES ON THE MPI INTROVERSION
(LOW SCORES)-EXTROVERSION (HIGH SCORES). THE
DATA ARE PRESENTED FOR EACH OF THE EIGHT
TREATMENT CONDITIONS 55
24 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN
SUBJECTS' CHANGE SCORES FROM PRE- TO POST
TREATMENT TESTING ON EACH OF THE DEPENDENT
VARIABLES AND SUBJECTS' AGES. THE DATA ARE
PRESENTED FOR EACH OF THE EIGHT TREATMENT
CONDITIONS 56
25 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN
SUBJECTS' CHANGE SCORES FROM PRE- TO POST
TREATMENT TESTING ON EACH OF THE DEPENDENT
VARIABLES AND SCORES ON THE MPI NEUROTICISM
SCALE. THE DATA ARE PRESENTED FOR EACH OF
THE EIGHT TREATMENT CONDITIONS 58
26 MULTIPLE REGRESSION COEFFICIENTS BETWEEN AMOUNT
OF SPEECH AND DESIRE FOR PSYCHOTHERAPY PRESENTED
FOR EACH INDEPENDENT VARIABLE (EXPERIMENTAL
DATA ONLY) 59
vli

LIST OF TABLESContinued
Table Page
27 MULTIPLE REGRESSION COEFFICIENTS BETWEEN THE
VARIABLES LISTED ON THE VERTICAL AXIS AND FIVE
DEPENDENT VARIABLE CHANGE SCORES FOR THE EX
PERIMENTAL GROUP. THE DATA ARE PRESENTED
SEPARATELY FOR EACH OF THE INDEPENDENT VARIA
BLES IN ADDITION TO THE TOTAL EXPERIMENTAL
GROUP 61
vili


INTRODUCTION
The extensive literature dealing with processes and systems of
psychotherapy generally makes the explicit assumption that behavioral
changes following psychotherapeutic encounter are a function of the in*
tended efforts of the psychotherapist. That is, behavioral modification
results from the direct and specific efforts, manipulations or techniques
of the psychotherapist. Within the framework of Roger's theory (1961),
for example, the therapist approves and accepts the client and as a re*
suit the client comes to accept himself. The patient in psychoanalysis
achieves personal satisfaction and comfort as the psychoanalyst Inter*
prets the patient's repressed psychosexual conflicts and the patient es*
tablishes a mature sexual adjustment (Fenichel, 1945) Shoben (1953) de
scribes the goal of psychotherapy as the alleviation of symptoms in addi
tion to the increase in a patient's affective comfort. Thus, In all psy
chotherapy, "specific factors" (therapist bohsviors) aJl^geily produce
intended effects (personal comfort of the patient).
Eysenck (1952, 1961) states, and Rosenthal and Frank suggest
(1956) that in addition to the intended efforts of the psychotherapist
(specific factors), there are other, nonspecific factors in psychotherapy
which significantly contribute to the outcome of treatment. "Nonspecific
factors" are loosely defined as those "placebo effects," situational
events and conditions in addition to the intended efforts of the therapist
which reportedly function in all treatment situations (Shapiro, 1964).
They serve to cloud interpretation of "real" effects of psychological
I

treatment (Rosenthal and Frank, 1956) for they produce manifest effects
which are indistinguishable from the effects produced by the intentional
treatment prescribed by the particular theory of psychotherapy.
2
Strupp (1962) indicates that because these nonspecific factors
have been largely neglected in research in psychotherapy, the extent of
their influence is undetermined. Many significant theorists and research
ers (e.g., Rogers, 1961) expound on the efficacy of their psychothera
peutic systems, although nonspecific factors, which may greatly contrib
ute to the final outcome of psychological treatment are not effectively
accounted for, controlled or measured. Understandably, successful out
come of psychotherapy Is viewed as proof of the effectiveness of psycho
therapy. However, It has not been demonstrated that personality and be
havioral modifications which occur as a result of psychotherapy are en
tirely (or even largely) due to the psychotherapist's specific techniques,
instead, the actual outcome of treatment is a function of both the spe
cific and nonspecific events occurring during psychotherapy.
Eysenck (1961), after reviewing and surnnarizing the literature,
shows that actual psychotherapy does not achieve higher recovery rates
than those reported from ordinary life experiences and the nonspecific
effects of routine medica) treatment. After demonstrating that nonspe
cific treatment produces effects comparable with actual psychotherapy,
Eysenck implies that nonspecific factors in psychotherapy are essential
factors responsible for bringing about successful psychotherapeutic
change. Findings and statements such as these clarify the necessity for
research In this area.

3
Effects of treatment with drug placebos are those nonspecific ef
fects given the greatest attention in the literature. Based on his re
view of the literature, Shapiro (1964) states that placebo effects are
subtle and common to all treatment situations and are the most important
nonspecific factors in psychological treatment. He defines the placebo
effect as the "psychologic, physiologic, or psychophysiologic effect of
aw. medj.cat.i.on or procedure given with therapeutic intent which is Inde
pendent of or minimally related to the effects of the medication or to
the specific effects of the procedure and which operates through a psy
chologic mechanisnf' (p. 293). In addition, the administration of this
medication or procedure must be recognized as legitimately therapeutic by
the patient.
Rosenthal and Frank (1956) also refer to the placebo effect as a
type of nonspecific event in psychological treatment and they report that
since research in psychotherapy has not dealt directly with the placebo
effect, the nature of Its Influence is undetermined. However, these au
thors speculate that it occurs with considerable regularity, and they
stress the need to demonstrate through research that observed effects of
psychotherapy are due to the therapist's techniques and exist separate
from nonspecific effects.
Borgatta (1959) expands on the importance of the placebo effect
by reporting that "If there is no evidence that an agent is the cause of
a given outcome, and there is also no evidence that a placeoo is the
cause of a given outcome, an appropriate statement is that there is no
evidence that the agent is more efficient than the placebo in bringing

4
about the outcome. Further, there is also no evidence that the placebo
is not as effective as the agent in bringing about the outcome" (p. 331)*
The literature is replete with references confirming the efficacy
of nonspecific effects of drug placebos (see reviews by Shapiro, 1964 and
Hongingfield, 1963) However, despite the demonstrated effectiveness of
drug placebos, and the alleged importance of the placebo effect in psy
chological treatment, adequately controlled studies evaluating the pla
cebo effect in psychotherapy are conspicuously lacking (Rosenthal and
Frank, 1956 and Strupp, 1962). One is currently forced, therefore, to
rely on implication, Inference, and analogy when constructing evidence of,
and predictive hypotheses pertaining to, the influence of nonspecific
placebo factors In psychotherapy.
Ib.e Preeq.t.,
"As Withering's foxglove was made more potent when digitalis was
Isolated from the concoction, and after it was recognized that only the
dropsy of congestive heart failure was benefited, so will psychotherapy
become more potent after the placebo effect is isolated and dissected
free from the psychotherapeutic process" (Shapiro, 1964, p. 85).
As long as both specific and nonspecific (placebo) factors oper
ate together during the treatment, one may never accurately determine the
real contribution of the placebo effect to psychotherapy. Since nonspe
cific placebo factors reportedly function unbeknown to the psychothera
pist, since their presence is reportedly a factor in the outcome of all
psychotherapy, and since they remain a nebulous and poorly defined factor
in psychotherapy, it is clear that the extraction of these placebo effects

5
from an investigation of actual psychotherapy is a complex, if not impos
sible task. Thus, a demonstration of the genuine effects of psychother
apy free of placebo effects seems never to have been made.
in this study an attempt is made to Investigate a form of "psy
chotherapeutic encounter11 in which specific factors could not contribute
to the outcome of treatment. That is, any changes which accrue as a re
sult of the treatment are not due to a psychotherapist's behavior, tech
nique or manipulations during treatment. Change, tnstead, must oe at
tributed to the presence of the placebo factor and must be referred to as
a placebo effect.
A psychotherapeutic encounter" can take place when a patient re
ports to a tape recorder for one "therapy session" believing a psycho
therapist will listen and respond to his talking. Measures of change in
such "affective comfort" (Shoben, 1953) as moods and anxiety taken imme
diately after this treatment will indicate the effect of the treatment.
Borgatta (1959) suggests, satirically, that this technique be used to
demonstrate the relative ineffectiveness of actual psychotherapy. Slack
(I960), on the other hand, demonstrated that therapeutically inaccessible
juvenile delinquents could be introduced to psychotherapy by initially
having them speak into a tape recorder. Eventually these juveniles were
introduced to the therapist who had been giving them feedback from the
tapes, and actual psychotherapy was initiated. Martin, Lundy and Lewin
(I960) evaluated the reinforcing effects of three degrees of therapist
communication on the affectively toned verbalizations of their subjects.
The group intended to have virtually no communication from the therapist
spoke into a tape recorder as If they were speaking to a psychotherapist.

In the present study, one half of the subjects will speak into
the tape recorder while the other half will wait an Identical length of
6
time. Since both groups will be given the instructions that they are
about to speak to a psychotherapist through the tape recorder this will
permit a comparison between a group given both the promise and the ges-
ture of help and a group given only the promise of help. Changes in mood
and anxiety in the group which does speak may be viewed as a demonstra
tion of the nonspecific placebo effect in a single session of psychologi
cal treatment. The effects of this treatment should be utilized as a
base line against which further research may be compared to demonstrate
that an actual psychotherapeutic interview produces changes different
from, or greater than the "placebo psychotherapy."
Shapiro (1964) reports that important features of a therapeutic
relationship which are responsible for the placebo effect are basically
elements of (I) the patient, (2) the doctor, and (3) the treatment situa
tion; in addition to a fourth factor: the doctor-patlent relationship.
That is, there are certain characteristics of each of these components of
treatment which contribute to a nonspecific placebo effect.
0) Ifaff. patient.
Shapiro, in his review of drug research (1964), points out that
many individual differences of patients have been evaluated to assess
possible relationships with the placebo effect. Such characteristics as
sex, age, intelligence and diagnosis have been observed repeatedly in
order to ascertain if a certain type of person responds more readily to
placebo factors. However, as Shapiro reports, this body of research is
equivocal and often contradictory. For example, Lasagna et al. (1954)

7
administered a placebo to a group of postoperative patients to relieve
reported pain. Those patients reporting a decrease in pain after receiv
ing the placebo had a mean age five years greater than the age of the
nonreactors (p ^ .05). On the other hand, Tibbetts and Hawkings (1956)
found that patients who responded in two drug placebo situations were
younger than the nonresponders. In studies by Kornetsky et al. (1957)
and Abramson et ai. (1955) no meaningful relationship existed between age
and drug placebo responsiveness. Similarly, level of intelligence as a
patient variable has produced contradictory results. In the Lasagna
study reported above, intelligence did not significantly distinguish the
placebo reactor from the nonreactor. On the other hand, Tibbetts and
Hawkings (1956) found the placebo reactor to be less intelligent than the
nonreactor, while Abramson et al. (1955) found the placebo reactor to
have a higher verbal intellectual ability and a lower performance intel
lectual ability.
Leiberman (1964) reports that Eysenck has correlated neurotic!sm
with the need for conformity in an individual's perceptions and judgments.
Therefore, he contends that "in group situations, where an element of in
terpersonal pressure activates the need for conformity, the subjects with
neurotic tendencies respond with placebo effects." In fact, Eysenck ar
gues that a subject's degree of neurotic!sm is a consistent factor for
picking out high placeoo reactors. Eysenck (1961) further speculates
that since introverts form conditioned responses more quickly than extro
verts, and since he reports that response to a placebo is a learned or
conditioned response, he expects Introverts to respond more readily to
placebo factors (p. 636).

8
However, despite the logical consistency of Eysenck's specula
tions, the literature concerning the placebo responsiveness of neurotics
and introverts is equivocal. For example, Fischer and Oiin (1956), using
psychiatric patients, demonstrated that neurotics were the most respon
sive subjects in a drug placebo situation. On the other hand, Kurland
(1958) disputes this conclusion on the basts of his finding that no dif
ference exists between the placebo reactivity of psychotics and nonpsy-
chotics. Joyce (1959) in a drug study with medical students, indicates
that extroverts are the most responsive subjects, while Laverty (1958)
demonstrates that introverted subjects are the most responsive in drug
treatment situations.
In this study, age, intelligence, introversion and neurotic!sm
are the patient characteristics whose contribution to the placebo effect
will be evaluated.
(2) The doctor-
The nature of the treater is reportedly an important factor for
encouraging or discouraging placebo effects (Shapiro, 1964; Uhlenruth
et al.. 1959). For example, Glldeman et al. (1957) demonstrated that a
male physician gained a 70 per cent cure rate in a treatment for which a
female nurse could achieve only a 25 per cent cure rate. He also has
shown that if the administering physician Is perceived as an expert
healer, the response of the patient to treatment will be greater than if
the physician Is believed to be a medical quack. These effects were ob
tained, conceivably, because the patient maintains certain cognitive ex
pectancies about the nature of treaters, and when these expectancies are

9
dissonant with the treatment situation, the effectiveness of treatment
decreases (see Fastinger and Brame I in Bachrach, 1962).
Similarly, folk medicine In the Spanish Southwestern United
States (Jaco, l958) and witchcraft healing in many parts of the world
(Frank, I96i) flourish and prevent the establishment of modern medical
program, at least in part, because scientifically trained physicians do
not meet loca; expectancies about the nature of treaters. That Is, to
these people scientific physicians are cognitive y dissonant in the roe
of "doctor." Conversely, the effectiveness of the native witch doctor is
enhanced because he meets his clientele's expectancies concerning the na
ture of treaters.
in the present research, verbal instructions to the patients will
estab Ish two different types of doctors (psychotherapists). One is the
type of psychotherapist the patient expects and the other does not con
form to his expectations (see Appendix A). This will produce two dis
tinct groups of subjects regarding their personal perception of their
psychotherapists.
(3) The treatment situation.
Foulds (1958) and Shapiro (1964) refer to many studies which dem
onstrate that a treater's bias can significantly Influence treatment out
come. For example, Hof ling (1955) and Frank (1958), In studies, demon
strate that positive staff attitudes and biases toward a particular pla
cebo or specific treatment increases the effectiveness of that treatment.
Lyerly et al. (¡962) In a drug study, have shown that subjects receiving
instructions appropriate to the expected effects of chloral hydrate, re
gardless of the actual drug or placebo administered, reported subjective

10
changes appropriate to the chloral hydrate. Abramson et al. (¡955) admin
istered tap water to a group of subjects who believed the liquid to be
LSD-25. A percentage of these subjects responded to the tap water with
psychomlmetlc behavior as If the water actually were LSD-25-
Frank (I9l) suggests that a patient's faith In a set of treat
ment operations is often ail that is necessary to produce positive change
and this is so even when these treatment operations are scientifically
absurd. Wolf (1959) states that a placebo is effective when a patient
has a "conviction that a certain effect will follow.
In the present study, an attempt will be made to manipulate ex
pectations concerning the vaue of the "placebo psychotherapy and the
amount of profit to be anticipated from the treatment. One group will
receive a positive set concerning that which can be gained from the
treatment and the other will receive a more neutral set (see Appendix A).
To summarize, the hypotheses under investigation in this study
are:
i. Those subjects who receive the placebo psycho
therapy (speak into the tape recorder) will show
more marked change in a "therapeutic direction
on measures of anxiety and mood than those sub
jects who simply wait a comparable length of time.
However, neither group will show a change in gen
eral level of psychological adjustment in this
single treatment.
II. The subjects who believe that their therapist is
an individual congruent with their expectations
will show more marked change in a "therapeutic
"direction on post measures of anxiety and mood
than those subjects whose expectations are disso
nant with the therapist's description. Neither
group will show a change in general level of psy
chological adjustment after this single treatment.

11
III. Those subjects receiving a positive set regard
ing the value and outcome of the treatment will
show more marked change in a "therapeutic di
rect ionf* on post measures of anxiety and mood
than those subjects who receive a more neutral set.
However, neither group will show a change In gen
eral level of psychological adjustment in this
single treatment.
IV. The measured individual subject differences of
age, intelligence, level of introversion and of
neurotlclsm will not identify the subjects who
respond best to this placebo situation. That Is,
there will be no differences between the placebo
responsiveness of subjects In terms of their age,
intelligence, level of introversion-extroversion
and level of neurotlclsm.
Hay the "placebo psychotherapy,1 as introduced above, be legiti
mately referred to as a placebo treatment situation? There is a strong
analogy between this treatment and that which exists when a physician ad
ministers a drug placebo. The physician who treats a patient with a pla
cebo pill does so in the same treatment setting, giving the same direc
tions, and offering the same encouragement as If the pill were pharmaco
logically sound. The only changed character!stlc, and that which defines
the situation as exclusively nonspecific, is the absence of the actual
drugthat agent in the treatment situation which produces the specific
effect. Reactions to this placebo are recorded as "placebo effects" al
though they may appear to be a result of a specific drug which is clearly
absent from the treatment. Similarly, in a sham interview such as that
mentioned above, the patient is given the same directions, the same en
couragement and is placed in the same treatment setting as if it were an
actual psychotherapeutic situation. The only changed characteristic, and
that characteristic which defines the situation as nonspecific in this
case, is the absence of the psychotherapist, that is, the only agent

12
through which specific effects may result. Reactions to this placebo
situation can be viewed as "placebo effects." Without knowledge of the
nature of the treatment, the effects may appear to result from the spe
cific behavior or techniques of a psychotherapist. However, his absence
from the treatment necessarily means that all effects are nonspecific, or
placebo.
Since there is no psychotherapist present, can this situation be
thought analogous to a therapeutic interviews The patient is talking to
a therapist who exists only in the patient's fantasy and is instituted by
the experimenter. Apfelbaum (1958) points out that individuals tend to
create, in part, interpersonal experience and that ongoing interpersonal
process in therapy Is primarily a function of a patient's personal trans
ferences. A patient, in utilizing past experience and partial cues, pro
jects the image of his psychotherapist, and thereby unknowingly perceives
him in part through his projective capacity. Consequently, the therapist
is perceived largely as a function of the patient himself. The attitudes
a patient holds toward his alleged therapist In the "shanf* or "placebo"
interview suggested above may be controlled and manipulated by experi
mental design. The therapist exists entirely in fantasy and as a func
tion of the patient's projective capacity and description of him as re
ported by the experimenter. The patient, however, in effect, establishes
an intrapersonal relationship which he believes to be, and which appears
to have the qualities of, an interpersonal psychotherapeutic relationship.
Thus, the sham interview will be regarded as both a placebo
treatment and a therapeutic interview.

METHOD
The study involved 96 subjects Mho Mere placed into eight treat*
roent conditions as diagrammed below in Tabie i.
TABLE i
THE EXPERIMENTAL DESIGN
Experimental Group
Control
Group
Congruent
Therapist
1nstructions
1ncongruent
Therapist
Instructions
Congruent
Therapist
Instructions
1ncongruent
Therapi st
1nstructions
Posi ti ve
set
concerning
treatment
outcome
n a 12
n 12
n 12
n 12
Neutral
set
concern!ng
treatment
outcome
n 12
n 12
n 12
n 12
n 48
n 48
The technique and procedure of establishing these treatment conditions,
and the way in which the data were collected are fully explained in the
Procedure section. The tests which were administered to the subjects and
the use to which the test information was put in the design are described
In the Means of Measurement section.
13

Subjects
The subjects for the study were 96 neuropsychiatric hospital pa
tients in the Gulfport, Mississippi, Veterans Administration Hospital.
They are all in-patient residents of the Rehabilitation and Placement
Service Ward which houses oniy those patients judged capable of returning
to the coomunlty. The ward offers a vocational testing service and is
professionally staffed by a psychiatrist, two counseling psychologists
and a number of nursing personnel.
The 96 subjects were drawn from the ward during 48 separate ses
sions extending from August 29, 1964, to October 19, 1964. The sample
consists of white adult males. There is a wide range of prehospitaliza
tion vocations. None were In psychotherapy when seen as subjects for the
study. However, it was known that some of these patients had been in
group or individual psychotherapy earlier in their hospitalization. The
identification of former psychotherapy oatients was not made, but there
is no reason to suspect a nonrandom assignment of these patients to the
different experimental treatment conditions. Similarly, although it was
known that some of the sanp.e were receiving drugs for their psychiatric
conditions, these subjects were not identified and their random assign
ment to the different experimental treatment conditions was expected.
The mean age of the sample was 38.76; range 18 to 58, standard
deviation 7*41. Below is a summary table of an analysis of variance
(Lindquist, 1953) demonstrating that subject ages are randomly distrib
uted across all treatment conditions.
The mean I .-(£. of the sample, as measured by the Beta Intelligence
Test (Kel logg et al.. 1946) was 98.54, which is within the normal range.

15
TABLE 2
SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE
RANDOM DISTRIBUTION OF AGE
Source
ss
df
V
F
A (experimental vs. control)
25.01
1
25.01
.44
B (congruent vs. incongruent)
3.76
1
3.76
.07
C (positive vs. neutral)
147.51
1
147.51
2.58
AB
25.01
1
25.01
.44
AC
11.35
1
11.35
.20
BC
25.01
1
25.01
.44
ABC
3.76
3.76
.07
w Cells
ixQlQ'M
88
57.16
Total
5,271.49
95
(An F of 3*95 is significant at the .05 level.)
The range of I.Q.'s was from 73 to 129; standard deviation 12.14. Below
is a summary table of an analysis of variance demonstrating that I .Q. is
randomly distributed across all treatment conditions.
The mean educational level of the sample is 10.9 years and
ranged from 6 years to 20 years. The mean of the length of current hos
pitalization to the closest month is 27 months and ranged from three days
to 14 years. The psychiatric diagnoses carried by these patients were
made by a psychiatrist at the time of admission to the hospital. There
is a wide range of psychiatric diagnoses, which for the purposes of this
study are divided into 54 psychotic and 40 nonpsychotlc. Because all
subjects were randomly assigned to the treatment conditions, there is

16
TABLE 3
SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE
RANDOM DISTRIBUTION OF I.Q.
Source
ss
df
V
F
A (experimental vs. control)
96.00
1
96.00
.62
B (congruent vs. incongruent)
.37
1
.37
.00
C (positive vs. neutral)
2.04
1
2.04
.01
AB
165.38
1
165.38
1.07
AC
3.37
1
3.37
.02
BC
66.67
1
66.67
.43
ABC
140.17
1
140.17
.90
w Cells
H-657-83
SSL
155.20
Total
14,131.83
95
(An F of 3*95 Is significant at the .OS level.)
reason to expect that educational level, length of hospitalization and
diagnosis are randomly distributed across treatment conditions. The psy-
chotic patients all have shown good remission of their symptoms and all
patients serving as subjects in the study were judged by a psychiatrist
to be sufficiently psychologically sound to return to the conmunity and
be employed.
Means of Measurement
Two tests upon which change in feelings of personal comfort can
be monitored (the Anxiety Adjective Check List and the Mood Adjective
Check List), and one test evaluating general level of psychological

17
adjustment (The Adjustment Score) were administered twice to each subject
--once prior to and once immediately after the treatment. Changes In
scores on these tests were used to evaluate changes as described in the
four hypotheses.
(I) The Anxiety Adjective Check List (AACL) (see Appendix B).
The Anxiety Adjective Check List (AACL) (i960) has been shown by
Zuckerman to be a "quick measure of anxiety level." It was chosen for
this study because of its demonstrated validity and reliability and the
fact that the instructions may be modified to permit sensitive evaluation
of changes in the level of manifest anxiety over short periods of time.
The check list is an empirically developed pool of 61 adjectives
with varying affective connotations. Twenty-one adjectives were identi
fied, on the basis of an item analysis, as discriminating between psychi
atric patients rated high on anxiety and normal control subjects. Of
these items, eleven are anxiety negative adjectives and ten are anxiety
positive. Subjects may obtain a score from 0 to 21 on the check list.
Zuckerman has developed two forms. One instructs subjects to denote how
they feel "In general' and the other asks how they fee "today." Al
though standardized on college students, Zuckerman reports no differences
in performance due to age, sex or level of education.
in two studies (i960, 1962) Zuckerman reports internal and test-
retest reiiabiIities for the "In general" form to be significant at
greater than the .001 level. However, the "today" form, although simi-
arly internally reliable, had law test-retest reliability (r .68,
o ^ ,00i). These results were anticipated and suggest that the today
form is sensitive to short-term fluctuations in anxiety. Validity was

18
evaluated by administrations of the today form to a class in general psy
chology. The check list was given on 10 nonexamination days and three
examination days. The examination-day administrations resulted In a sig
nificantly higher reported level of anxiety than the nonexamination-day
administrations (p<.0005). The today form, with instructions modified
to read "right now," will be used in this study.
(2) The Hood Adjective Check List (MACL) (see Appendix C).
Nowlis (1956, 19, 19I) assumed mood to be a multidimensional
characterization of a person's feeling or behavior which is accessible to
seif report. The MACL resulted from a series of seven factor analytic
studies by him (i960). In those seven studies Nowlis demonstrated that
40 adjectives, from his original pool of 200, have relatively consistent
and high loadings on twelve separate factors. These 40 adjectives con
stitute the MACL which Is scored separately for each of the twelve fac
tors (which are the identified moods). Further, the twelve factors
(moods) are divided into three groups: those "factors most consistently
identified," those "factors identified fairly consistently" and "factors
tentatively identified." Below are the twelve mood factors and the par
ticular adjectives constituting the factors.
Certain specific moods on the MACL have particular relevance for
the purpose of this study because they would be expected to change in a
certain direction as a result of successful psychotherapy. Aggression
and Deactivation would be expected to decrease, while Social Affection
would be expected to increase. Similarly, after successful psychotherapy
the moods of Anxiety and Depression would be expected to decrease and
Pleasantness and Activation would be expected to increase. At a lower

19
TABLE 4
THE TWELVE MOOD FACTORS OF THE MCL AND THE ADJECTIVES
CONSTITUTING EACH MOOD FACTOR
?-raraarajir. 1 irmr.i r:.mg-i'-aita a. u-j.-.u. xiinrti.r.7r n,u;anjgg.z.w^snx,x,a n, a rmi amsszsBsam
Adjectives with high loadings on the factor
Factors Most Con-
si stently Identified:
AGGRESSION
CONCENTRATION
DEACTIVATION
SOCIAL AFFECTION
angry, bold, defiant, rebellious,
concentrating, earnest, engaged in thoight,
serious.
drowsy, earnest, tired.
affectionate, forgiving, kindly, warmhearted.
Factors Identified
Fairly Consistently:
ANXIETY
DEPRESSION
EGOTISM
PLEASANTNESS
apprehensive, clutched up, fearful, Insecure,
blue, lonely, regretful,
boastful, cocky, egotistic, self-centered,
elated, lighthearted, overjoyed, pleased.
Factors Tentatively
I dent!fIed:
ACTIVATION
NONCHALANCE
SKEPTICISM
STARTLE
active, energetic, vigorous,
nonchalant, playful, witty,
skeptical, suspicious,
startled, shocked.
level of certainty, because of difficulty in Interpreting its meanlngful-
ness in a psychotherapeutic context, Concentration and Nonchalance might
be expected to increase, while Egotism, Skepticism and Startle might be
expected to decrease after treatment. In addition, changes in Skepticism
after the treatment may reflect the extent to which this treatment was be-
Iieved to be real.
In his research developing the MACL, Now!is used male and female
college students and U. S. Navy personnel In a variety of experimental
(drugs, films) and field situations. For example, one validation procedure
that Nowlis reports (1961) involved the administration of a drug which

20
has known effects on a person's feelings and moods. Continuous monitor"
ing of mood on the MACL for 16 hours demonstrated that fluctuations in
the MACL were related to anticipated drug-induced changes. Without re*
porting actual correlational figures, Now!Is indicates that concurrent
validity of the MACL is high, although the test-retest reliability is low.
Nowlis describes temporality as one of the characteristics of a mood and
therefore, b*c&use of the low reliability of the check list, he suggests
that the scale is sensitive to short-term fluctuations in mood. Because
of its short length and ease of administration, Nowlis points out, it is
possible to monitor mood with this instrument over repeated intervals of
any required length."
The subject is asked to circle one of four options for each word
in the check list: a double plus if the adjective definitely describes a
current strong feeling, a single plus if the word slightly applies to a
current feeling, a question mark if the word is unclear or the subject is
unable to decide if the adjective describes a current feeling and no if
the subject is certain that the word does not describe a present feeling.
Oouble plus is scored as 3 single plus is scored as 2, question mark is
scored as I, and no is scored as 0. Twelve mood scores are obtained by
summing the scores of the individual adjectives in each of the moods.
The scores may range from 0 to 12 on the moods of Aggression,
Concentration, Social Affection, Anxiety, Egotism and Pleasantness. The
scores may range from 0 to 9 on the moods of Deactivation, Depression,
Activation and Nonchalance. On the moods of Skepticism and Startle the
scores may range from 0 to 6.

21
(Both the AACL and the Anxiety Mood on the MACL reportedly meas
ure level of manifest anxiety. The MACL Anxiety mood contains four af
fectively toned adjectives, each of which the subject responds to by de
noting the extant to which the adjective describes his current feelings.
The AACL, on the other hand, asks a subject to check those adjectives
that apply to his current feelings. Only one of the four adjectives on
the MACL Anxiety mood is identical with the 21 adjectives which are scored
on the AACL. A comparison between pretreatment performance on the two
Anxiety measures resulted In a Pearson product-moment correlation of +.4,
and a Pearson product-moment correlation between the subject's changes in
the two scales from pre- to posttreatment testing resulted In a correla
tion of +.57* Both these correlations are significant beyond the .001
level. This suggests that despite the basic uniqueness of the two meas
ures, they are measuring essentially the same factor of "anxiety.")
(3) The Adjustment Score (see Appendix D).
The Adjustment Score was chosen to evaluate changes in general
level of psychological adjustment, it was originally constructed by
Dymond (1953) as a Q-sort to evaluate the effects of nondirective coun
seling. The statements were compiled empirically by Butler and Haigh who
noted that they were representative of the positive and negative comments
made by people in psychotherapy.
Four 'Well-trained, practicing clinical psychologists" who were
not client-centered by theoretical orientation "agreed remarkably welI"
In choosing 37 negative indicators of adjustment and 37 statements de
scribing positive adjustment (1953). These 74 statements rated on a
"like me" to 'unlike me" dimension constitute the scale.

22
Dymond (1954) demonstrated that a significant difference existed
between the mean adjustment scores on the Q-sort of subjects in two months
of client-centered psychotherapy as compared to control subjects who
waited a comparable length of time (p ^.01). Dymond reports that the
test-retest reliability of this control group is +.86. Validity was es
tablished in two ways: (I) rank order correlations between self-ideal
correlations and the Q-sort adjustment score of clients before therapy
began was +.83 and the rank order of these same subjects after therapy
was +.92; (2) each therapist rated the success of his therapy for each
client and these ratings correlated at better than the 5 per cent level
with the subject's own scores on the 0-sort.
For the purpose of the present study, the 74 items on the Q-sort
were randomly organized into a questionnaire. The subject is asked to
mark an item "true" if it pertains to him and mark "false" if the item
does not pertain to him. A total of the positive items marked true and
negative items marked false constitute a subject's adjustment score.
The AACL, MACL and the Adjustment Score were administered twice
to all subjects, once prior to and once immediately after the treatment.
The change In scores frompre- to posttreatment testing constitute meas
ured changes in manifest anxiety, 12 independent moods and general level
of psychological adjustment.
(4) Ite .Mauds 1 ffiL.Pa.riffiHfll Uy I nyent;ory (HP.Q .
In order to evaluate part of Hypothesis IV, the HPI (Eysenck,
1959, 1961) was administered to obtain subject differences in level of
neurotic!sm and degree of introversion. "Extroversion, as opposed to In
troversion, refers to the out-going, uninhibited, social proclivities of

23
a person," and neurotlcism represents "the general emotional liability of
a person, his emotional overresponsiveness, and his liability to neurotic
breakdown under stress" (1959, p. 3)- The Inventory has 48 questions
which have been factor analyzed to yield scores on an extroversion-
introversion dimension and a neurotlcism dimension. Although the two di
mensions are slightly negatively correlated (-.15), Eysenck submits that
they are essentially orthogonal.
The original standardization took place on a sample of 200 normal
English men and 200 normal English women. Eysenck has since supplemented
this by adding groups of students, nurses, industrial apprentices and "a
quota sample of the whole population." Data are also available on groups
of hospital patients and prisoners. Split-half and Kuder-Richardson re
liability coefficients for the Neurotlcism scale fall between .85 and .90
and for the extroversion scale fall between .75 and .85. Using both con
current arid construct validation techniques, Eysenck reports that the MPI
has yielded data supporting its validity.
(5) A Modification of the Expectation Q-sort (Apfelbaum, I95&)
(see Appendix £).
The Q-sort asks a subject to indicate his expectancies regarding
the nature of a psychotherapist who may be assigned in the future.
Thirty-six of the 60 items in Apfelbaum's item pool were chosen and ran
domly organized into a questionnaire. These 3 items constitute those
with heaviest factor loadings on three separate clusters: (1) those
items which designate a guiding, giving, protective therapist; (2) those
items which designate a tolerant, accepting and permissive therapist; and
(3) those items which designate a cold, rigid and condemnatory therapist.

24
In addition to the five tests, each subject was also asked to
complete a questionnaire asking for personal Information (see Appendix
F). Each subject also had taken the Beta Intelligence Test (Kellogg
et ai.. 1946).
Procedure
Subjects were chosen from the ward by a secretary who was naive
concerning the mechanics and goals of the study. She was informally
questioned after the completion of the study and no bias In the subjects
she chose, and her order of choosing them, was noted.
The subjects were seen two at a time. Once the subjects were
seated in the testing room they were given instructions which drew their
Interest to the tasks about to be presented but did not disclose the na
ture of the research nor the fact that it was research (see Appendix A).
They were given the following tasks in the order noted:
1. Personal information sheet
2. Mauds ley Personality Inventory
3. Anxiety Adjective Check List
4. Mood Adjective Check List
5. Expectation questionnaire
6. The Adjustment Score
These tasks were relatively short, took no longer than 40 minutes to com
plete and subjects did not appear to become fatigued. After completing
the six questionnaires the pair of subjects was told that this was an op-
portunlty to speak with a psychiatrist, that he was unable to get to the
hospital and that he requested they soeek Into a taoe recorder, the tape
of which would be sent to him by registered mall (see Appendix A for ex
act wording).

25
By a coin toss prior to the meeting between subjects and Expert-
mentar, one subject of the pair was to be placed in the Experimental
Group and the other of the pair was to be placed in the Control Group
(treatment vs. wait). Two additional coin tosses by the Experimenter
prior to the meeting determined whether this particular pair of subjects
would receive (1) the positive or neutral set, and (2) the congruent or
incongruent therapist instructions. Both subjects were then given the
instructions appropriate to their designated treatment category (i.e.,
(1) positive set, congruent therapist; (2) positive set, incongruent
therapist; (3) neutral set, congruent therapist; and (4) neutral set, in
congruent therapist [see Appendix A]). The Experimenter had sufficient
time while the subjects were filling in the Adjustment Score, to choose
the predetermined items on the Expectation Questionnaire and manipulate
the congruence and incongruence of the subject's alleged psychotherapist
(see Appendix A).
The Control Group subject of the pair was then asked to remain in
the testing room. He was asked to relax, was offered the use of a stack
of magazines and was told that the Experimenter would return in about 50
minutes (see Appendix A for exact wording). The Experimental Group sub
ject was then taken to the room which contained the tape recorder. This
room had a one-way vision screen and was wired for sound. The subject
was seated in an easy chair, instructed to speak to the "psychiatrist"
and was told that the Experimenter would return In about 50 minutes to
retrieve the tape and the tape recorder was turned on (see Appendix A for
exact wording). The Experimenter then seated himself in an observation
room where he was able to record the Experimental Group subject's

26
verbalizations and to observe that the Control Group subject remained in
his room. Record was made of the amount of the subject's speech, and in
several cases the content of the subject's verbalizations. With a stop
watch, amount of speech was rated as follows: constant speaker, no longer
than 3~m¡nute pauses; moderate speaker, no longer than 5~minute pauses;
little speaker, at least some verbalization; no speaker, says nothing.
At the end of 45 minutes the Experimenter turned off the tape re
corder and guided the Experimental Group subject back into the testing
room. Both subjects were then asked to "retake several of the shorter
tests." The following were readministered:
7. Anxiety Adjective Check List
8. Mood Adjective Check List
9. The Adjustment Score
At this point, the collection of the data was completed. The Ex
perimental Group subject was dismissed. If the subject asked about feed
back, he was told that he would receive some type of feedback within sev
eral weeks. He was also encouraged not to tell anyone the nature of what
had happened because the Experimenter would be unable to see everyone for
this project. The Control Group subject was then asked If he wished to
"speak to the psychiatrist." If he agreed, he was given 45 minutes with
the tape recorder, however, If he declined, he was dismissed. Twenty-one
of the 48 Control Group subjects chose to speak.
If any subject discussed suicidal or homicidal content, mentioned
intensely pressing issues or made special requests, the Experimenter
promptly called this Information to the attention of the ward psychiatrist.
This occurred in twelve cases.

27
After all the data were collected, all subjects who spoke into
the tape recorder were sent a letter (see Appendix G). The purpose of
this tetter was to thank the subject for his cooperation and Inform him
of the way in which he could contact the Experimenter for further clari
fication of the procedure. Twelve of the 69 subjects who spoke into the
tape recorder contacted the Experimenter as a result of this letter.

RESULTS
Analysis of covariance (Ray, I960) was used to evaluate the tena-
billty of hypotheses I, II and III (those hypotheses concerned with the
effects of differential experimental treatment). In analyzing the three
hypotheses, the eight independent treatment conditions, as diagramed in
Table 1 were considered. Dependent variables in hypotheses I, II and III
were the changes in scores frompre-to posttreatment testing on the AACL,
the 12 moods in the MACL and Dymond's Adjustment Score. Thus, evaluation
of the first three hypotheses was made across the three independent vari
ables utilizing 14 2 x 2 x 2 analyses of covarianceone for each de
pendent variable.
The technique of analysis of covariance chosen to evaluate hy
potheses I, II, and III utilizes the fact that differences between the
treatment groups on a dependent variable posttreatment measure may be to
some extent a reflection of differences between these groups on the ini
tial pretreatment measure. It removes the variance due to these initial
random differences from the final posttreatment variances, thus reducing
the size of the error variance and increasing the precision of the expe
riment (Gourlay, 1953)* That is, by utilizing the regression of the
posttreatment measures on the pretreatment measures, an adjustment Is
made for the variability which is associated with Initial differences
among the subjects. After this adjustment is made, the remaining varia
bility may be analyzed with a more precise estimate of factors which may
have produced an effect.
28

29
f
By inspection, the data appear to have significant homogeneity of
variance. Norton (as cited by Lindquist, 1953) demonstrates that unless
the heterogeneity of variance is so extreme that it is readily apparent
upon inspection, the effect upon the F distribution will be negligible.
No formal tests of homogeneity of variance were conducted. Statistical
significance was set at the .05 level prior to the analysis. Below are
the summary tables of the analyses of covariance for each of the depend
ent variables (Tables 5 through 18). In addition, the mean change scores
frompre-to posttesting in each treatment condition are presented for
each of the dependent variables (Tables 5A through 18A).
Hypothesis I states that the Experimental Group (placebo treat
ment group) will show more "positive therapeutic changes" in the AACL and
the moods of the MACL than the Control Group (wait group). This predic
tion was supported in the analyses of the main effects of three of the
fourteen dependent variables. On the MCL, the Experimental Group showed
a greater decrease in reported anxiety after treatment than did the Con
trol Group (p ^ .05) (see Table 5)* On the MACL Social Affection mood,
the difference was significant between the increase in Social Affection
for the Experimental Group after treatment and the decrease after treat
ment in Social Affection for the Control Group (p^ .05) (see Table 9).
Similarly, the Anxiety mood on the MACL decreased after treatment for the
Experimental Group and increased for the Control Group (p^.05) (see
Table 10).
Two dependent variables Indicated that an interaction exists be
tween the experimental vs. control treatment (Hypothesis l) and the con
gruent vs. incongruent therapist treatment (see below. Hypotheses II).

30
TABLE 5
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE ZUCKERMAN
ADJECTIVE CHECK LIST
Source
ss
df
V
F
A (experimental vs. control)
44.42
1
44.42
6.29*
B (congruent vs incongruent)
98.55
1
98.55
13.96***
C (positive vs. neutral)
.00
l
.00
.00
AB
69.98
1
69.98
9.91**
AC
6.63
l
6.63
94
BC
17.67
1
17.67
2.50
ABC
17.67
1
17.67
2.50
error
615.07
SI
7.06
Total
869.96
94
*** s significant K .001.
** a significant < .01.
* significant <, .05.
TABLE 5A
PRETREATMENT MEANS (X,). POSTTREATMENT MEANS (X,) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE
ZUCKERMAN ADJECTIVE CHECK LIST
Experimental Group
Control Group

Congruent
Therapist
Incongruent
Therapist
Congruent
Therapist
1ncong ruent
Therapist
Positive
Set
X, a 11.08
X2 5 7.00
D s -4.08
X| a 8.92
X2 10.59
D a +1.67
X, a 10.33
x2 JU1
D s -.58
X. a 8.41
X, 8.41
D a .00
Neutral
Set
X, s 10.08
*2 s
D -2.75
X, 9.75
X2 2^08
D a -.67
L 11.33
X2 11.16
D = -.17
X, a 9.41
x2 9.83
D a +.42

31
TABLE 6
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL AGGRESSION MOOD
Source
ss
df
V
F
A (experimental vs. control)
8.40
l
8.40
2.00
B (congruent vs. incongruent)
7.57
1
7-57
1.81
C (positive vs. neutral)
.10
1
.10
.02
AB
31,84
1
31.84
7.60**
AC
1.95
1
1.96
.47
BC
4.Go
1
4.66
1.11
ABC
5.43
1
5-43
1.30
error
8Z
4.19
Total
424.25
94
** significant < .Oi.
TABLE 6A
PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X,) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL AGGRESSION MOOD
Experimental Group
Control
Group
Congruent
Therapist
Incongruent
Therapist
Congruent
Therapi st
1ncongruent
Therapist
Positive
Set
X, 3.33
X, 1.83
D -1.50
X, 3.42
X2 diz
D + .25
X, 3-50
x2 4.50
D = +1.00
X, 2.33
X2 2.00
0 -.33
Neutral
Set
X, 2.17
x2 hli
D = -.92
X, 2.42
x2 UI
0 = +.75
X, 2.67
X2 2^
D =* .00
X, 2.08
x2 2^50
D +.42

32
TABLE 7
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL CONCENTRATION MOOD
Source
ss
df
V
F
A (experimental vs. control)
1.81
1
1.81
.49
B (congruent vs. incongruent)
6.59
1
6.59
1.78
C (positive vs neutral)
1.12
1
1.12
.30
AB
4.75
1
4.75
1.28
AC
6.14
1
6.14
1.65
BC
3.87
1
3.87
1.04
ABC
.45
1
.45
.12
error
3.21,02
§1
3.71
Total
347.75
94
TABLE 7A
PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS (X,) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL CONCENTRATION MOOD
Experimental Group
Control
Group
Congruent
Therapist
Incongruent
Therapist
Congruent
Therapist
1ncongruent
Therapist
Pos!tlve
Set
x, 8.75
X2 8.92
D * +.17
X| 10.08
X2 9.67
0 -.41
x, 9.50
X2 8.50
0 -1.00
x, 10.33
X2 9-42
D -.91
Neutral
Set
X, 11.08
x2 11.25
D + .17
X, = 9.42
*2 8,2*
D =* -I.I7
x, 9.17
x2 Idl
D .00
x, 10.17
x2
D -.42

33
TABLE 8
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL DEACTIVATION MOOD
Source
ss
df
V
f
A (experimental vs. control)
7.82
1
7.82
1.67
B (congruent vs. Incongruent)
25-48
1
25.48
5.43*
C (positive vs. neutral)
33
1
.33
.07
AB
1.24
1
1.24
.26
AC
.96
1
.96
.20
BC
.48
1
.48
.10
ABC
70
1
70
.15
error
40,7-25
8Z
4.69
Total
444.56
94
Significant <£. .05.
TABLE 8A
PRETREATMENT MEANS (Xi), POSTTREATMENT MEANS (j) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL DEACTIVATION MOOD
Experimental Group
Control
Group
Congruent
Therapist
1ncongruent
Therapist
Congruent
Therapist
Incongruent
Therapist
Pos!tive
Set
X, 4.08
x2JL2L
D -1.17
X, 3.42
X2 iiiZ
D +.25
X, 3.67
x2 "US
0 +.08
X, 4.08
X2 ibM
D +.50
Neutral
Set
x, 4.25
X2 UL2
D -1.08
X, 4.17
x2 4.33
0 +.16
X, 4.75
X2 4.00
0 -.75
X| 2.25
X2 3.08
D +.83

34
TABLE 9
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL SOCIAL AFFECTION MOOD
Source
ss
of
V
F
A (experimental vs. control)
27.15
1
27.15
4.35*
B (congruent vs incongruent)
14.37
1
14.37
2.30
C (positive vs. neutral)
03
1
.03
.001
AB
19.62
1
19.62
3.14
AC
10.84
1
10.84
1.74
SC
1.36
1
1.36
.22
ABC
22.60
l
22.60
3.62
error
543.18
8Z
6.24
Totai
639.15
94
^Significant ^.05.
TABLE 9A
PRETREATMENT MEANS (X.), POSTTREATMENT MEANS (X£) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL SOCIAL AFFECTION MOOD
Experimental Group
Control
Group
Congruent
Therapist
1ncongruent
Theraplst
Congruent
Therapist
Incongruent
Theraplst
Posl ti ve X | = 6.17
Set X2 = 7.84
D + I7
X, 6.25
X2 5.00
D * -1.25
X, 6.08
X2 0
D -.58
X, 8.00
x2 1
D -.17
Neutral Xt 7.58 Xj 7-75 X, 7-58 X, 6.42
Set X2 8.42 X2 8.08 X2 b.75 X2 5.25
o +.84 d +.33 0 -.83 0 -I.17

35
TABLE 10
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL ANXIETY MOOD
Source
ss
df V
F
A (experimental vs. control) 35*35
1 35.35
6.29*
B (congruent vs. Incongruent) 4.27
1 4.27
.76
C (positive vs. neutral)
.00
1 .00
.00
A6
6.31
1 6.31
1.12
AC
4.00
1 4.00
.71
BC
.07
1 .07
.01
ABC
8.33
1 8.33
1.48
error
488.92
82 5-62
Total
547.25
94
Significant ^ .05.
TABLE I0A
PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X2) AND MEAN
CHANCE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL ANXIETY MOOD
Experimental Grouo
Control Group
Congruent
Therapist
1ncongruent
Therapist
Congruent Incongruent
Therapist Therapist
Positive X, 5-25
Set X2 3J5
d = -1.50
X, 4.92
X2 5.08
d +7K
Xj 4.50 X, -
X- 5.08 X, -
D +3H D -
4.33
lii
.00
Neutral Xj 5.84
Set X2 4.42
0 -1.42
X, 3.83
*2 hdl
D -.66
X, 5.67 X, 4.00
X2 5.84 X2 5.00
0 17 D =* -H.00

36
TABLE 11
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL DEPRESSION MOOD
Source
ss
df
V
V
A (experimental vs. control)
3.04
1
3.04
.78
B (congruent vs. incongruent)
.00
1
.00
.00
C (positive vs. neutral)
.49
1
.49
.13
AB
.86
1
.86
.22
AC
1.24
1
1.24
.32
BC
.37
1
.37
.10
ABC
5-55
1
5-55
1.43
error
317,51
8Z
3.88
Total
394.06
94
TABLE 11A
PRETREATMENT MEANS (X.), POSTTREATMENT MEANS (X,) ANO MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL DEPRESSION MOOD
Experimental Group
Control
Group
Congruent
Therap1st
1nconnruent
Therapist
Congruent
Therapist
1ncongruent
Therapist
Posltlve
Set
X, 3.92
D -1.25
X, 3.75
*2-1^22
D -.83
X, 3.58
X2 -i,66
D 4.08
X, 3.75
X2 2^81
D -.92
Neutral
Set
X, 5.33
X*2 4.42
0 -.91
X, 4.17
*2 L31
D -.84
X, 4.25
*2 IM
0 -.83
X, 3.08
X2 2^22
D -.16

37
TABLE 12
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL EGOTISM MOOD
Source
ss
df
F
A (experimental vs. control)
4.53
1
4.53
1.09
B (congruent vs. incongruent)
2.10
1
2.10
.51
C (positive vs. neutral)
3.9
1
3-19
.77
AB
.57
1
57
.14
AC
11.58
1
11.58
2.80
BC
1.33
1
1.33
.32
ABC
2.60
1
2.60
.63
error
353.86
sz
4.14
Total
385-76
94
TABLE 12A
PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS (X2) AND MEANS
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN CM TPr MACL EGOTISM MOOD
Experimental Group
Control
Group
Congruent
Therapist
1ncongruent
Therapist
Congruent
Therapist
Incongruent
Therapist
Posit ve
Set
X, 3.08
x2 *2
D -.58
X, 3.25
x2 2.08
D -1.17
X, 3.00
X2 3.08
D +.08
X, 2.67
x2 USl
D +.58
X, = 2.58 X, 2.50
*2 -
0 +.50 D +.17
X, 2.42
*2 "
D -.25
Neutral
Set

38
TABLE 13
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL PLEASANTNESS MOOD
Source
ss
4 f
V
F
A (experimental vs. control)
8.12
1
8.12
1.21
B (congruent vs. incongruent)
39.31
1
39.41
5.89*
C (positive vs. neutral)
43-51
1
43.51
6.50*
AB
.01
1
.01
.001
AC
.02
i
.02
.003
BC
3.57
1
3-57
54
ABC
3.87
1
13.87
2.07
error
581.76
£Z
6.69
Total
690.17
94
Significant ^ .05.
TABLE 13A
PRETREATMENT MEANS (X.), POSTTREATMENT MEANS (X,) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL PLEASANTNESS MOOD
Experimental Group
Control
Group
Congruent
Therapist
Incongruent
Therapist
Congruent
Therapist
Incongruent
Therapist
Pos!tive
Set
X, = 4.25
X2 6.67
0 +2.42
x, 4.33
x2 itJl
0 .00
X, 4.42
X2 5.50
D +1.08
X, 4.33
X2 4^
0 +.17
Neutral
Set
X, 4.17
X2 4J2
D .00
X, 4.75
x2 4.42
D -.33
X, 4.33
x2 4.42
D +.09
X, 4.17
X2 IM
0 -1.50

39
TABLE 14
SUMMARY TABLE OF THE ANALYSIS OF iOVARIANCE FOR THE
MACL ACTIVATION MOOD
Source
ss
df
V
F
A (experimental vs. control)
2.90
1
2.90
.80
B (congruent vs. incongruent)
1.80
1
1.80
.50
C (positive vs. neutral)
.20
1
.20
.06
AB
5.58
1
5-58
1.54
AC
.47
1
47
.13
BC
.00
1
.00
.00
ABC
.12
1
.12
.03
error
i!4^4
SZ
3.62
Total
326.01
94
TABLE I4A
PRETREATMENT MEANS (X,). POSTTREATMENT MEANS (X.) AND MEAN
CHANGE SCORES (O) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL ACTIVATION MOOD
experimental Group Control Group
Congruent Incongruent Congruent Incongruent
Therapist
Therapist
Therapist
Therapist
Pos!tive
Set
X, 4.25
X2 4.00
D -.25
X, = 4.00
x2 .Liz
0 = -.33
X, 4.42
X2 3^92
D -.50
X, 4.25
X2 4.50
D +.25
Neutral
Set
X, 3.75
X2 -
0 -.25
X, 3.58
X2 1M
D -.50
X, 3-92
X2
D -.34
X, 4.58
x2 L2&
D +.34

4
TABLE 15
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL NONCHALANCE MOOD
Source
df
V
F
A (experimental vs. control)
.87
1
.87
.23
B (congruent vs. incongruent)
1.43
1
1.43
.38
C (positive vs. neutral)
.17
1
.17
.04
AB
3.52
1
3.52
.92
AC
.24
1
.24
.06
BC
8.46
1
8.46
2.22
ABC
.37
1
.37
.10
error
331. SO
IZ
3.81
Total
346.96
94
TABLE I5A
PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X,) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL NONCHALANCE MOOD
Experiment Group Control Group
Congruent Incongruent Congruent Incongruent
Therapist
Therapist
Therapist
Therapist
Post tive
Set
X, 3.08
X2 3.2S
D +.17
X, 3.42
*2 -
D -.50
X, 3.83
x2-Ui
D -.50
X] 2.08
x2 US.
D -.33
Neutral
Set
X, 2.00
x2 1>83
0 -.17
Xj a 3.08
x2 1*08
D .00
X, 3.58
*2 1^8
D -1.00
X, 2.00
*2 U&
D +.58

4!
TABLE 16
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL SKEPTICISM MOOD
Source
$s
4#
V
r
A (experimental vs. control)
1.98
l
1.98
.74
B (congruent vs. Incongruent)
.49
l
.49
.18
C (positive vs. neutral)
.10
l
.10
.04
AB
.36
l
.36
.13
AC
2.17
l
2.17
.81
BC
.00
l
.00
.00
ABC
.01
1
.01
.004
error
ULM
liZ
2.67
Total
236.98
94
TABLE I6A
PRETREATMENT MEANS (X.), POSTTREATMENT MEANS (X,) AN0 MEAN
CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL SKEPTICISM MOOD
Experimental Group
Control
Group
Congruent
Therapist
Incongruent
Theraplst
Congruent
Theraplst
1ncongruent
Therapist
Posi tive
X, 1.92
X, 2.75
X, 2.75
X, 1.67
Set
x2 105
0 -.17
x2 2*50
D -.25
X2 -
D -.42
X2 1.75
D +.08
Neutral
X, 2.00
X, 2.17
X, 3.00
x2 iOl
D -.17
X, = 1.75
Set
X2 1.55
D = -.42
X2 1.91
0 -.26
X2 2J2
0 +.42

42
TABLE 17
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL STARTLE MOOD
Source
ss
df
V
F
A (experimental vs. control)
.03
1
.03
.06
B (congruent vs. incongruent)
13.47
1
13.47
26.3k***
C (positive vs. neutral)
2.67
1
2.67
5.34*
AB
.04
1
.04
.08
AC
37
1
.37
.74
BC
1.47
1
1.47
2.94
ABC
.03
1
.03
.06
error
43.70
81
.50
Total
61.78
94
***Slgnlflcant ^ .001.
**$Ignificant < .05*
TABLE I7A
PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X2) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL STARTLE MOOD
Exjerimental Group
Control
Group
Congruent
Therapist
1ncongruent
Therapist
Congruent
Therapist
1ncongruent
Therapist
Positive
Set
X, 1.42
x2 1.00
0 -te
x, 1.17
x2 1-83
D +.66
x, 1.50
x2 1.00
D -.50
X, 1.25
X? 1.67
D
Neutral
Set
x, = .92
x2 = 1.00
D +.08
x, .92
X2 L20
D +.58
X, 1.00
x2 1.17
D +.17
X, 1.25
x2 L21
0 +.07

43
TABLE 18
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
ADJUSTMENT SCORE
Source
ss
df
V
F
A (experimental vs. control)
40.12
1
40.12
3.04
B (congruent vs. Incongruent)
.07
1
.07
.005
C (positive vs. neutral)
20.84
1
20.84
1.58
AB
.16
1
.16
.01
AC
33.34
l
33.34
2.52
BC
3.77
l
3-77
.29
ABC
9.98
1
9.98
.76
error
uisa^o
8Z
13.21
Total
1,257.98
94
TABLE I8A
PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X2) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE ADJUSTMENT SCORE
Experimental Group
Control
Group
Congruent
Therapist
Incongruent
Theraplst
Congruent
Therapist
Incongruent
Therapist
PosItiva
Set
X, 45.42
*2 &--25
D +.33
X, 43.17
h !&&
D +.75
X, 39.67
x2 4U8i
D +2.16
X, 44.83
*2 "
D +2.42
Neutral
Set
X, o 42.00
X, 41,58
D -.42
X, 46.50
x2 47.00
0 = +.50
X, 37.17
x2 37.50
0 +.33
X, 49.58
x2 *
D .00

44
Changes after treatment on the AACL indicated that the Congruent Thera
pist, Experimental Group and the Incongruent Therapist, Control Group
significantly reduced anxiety when compared to the Incongruent Therapist,
Experimental Group and the Congruent Therapist, Control Group (p ^ .01)
(see Table 5). Similarly, the MACL Aggression mood indicated that the
Congruent Therapist, Experimental Group and the Incongruent Therapist,
Control Group significantly reduced aggression when compared to the In
congruent Therapist, Experimental Group and the Congruent Therapist, Con
trol Group (p 4 *01) (see Table 6). These interactions suggest that as
measured on the AACL and the MACL mood of Aggression, the superiority of
the Experimental Group treatment pertains only for those subjects who are
given the congruent therapist Instructions, while Experimental Group sub
jects given the incongruent therapist instructions report "nontherapeutlc"
changes.
While only three dependent variables significantly distinguished
between the Experimental Group and Control Group (in addition to two in
teraction effects) and the Adjustment Score Indicated no difference as
predicted, no dependent variables significantly changed In a direction
opposite to that predicted. Thus Hypothesis I appears to be supported.
Hypothesis it states that subjects, when offered a "psychothera
pist" who is congrccoc with their expectancies, will show more "positive
therapeutic changes" on the dependent variables than subjects assigned a
"psychotherapist" who is incongruent with their expectations. This was
found to be the case in the analysis of the main effects in four of the
14 dependent variables. The AACL showed that subjects given a therapist
congruent with their expectations decreased anxiety after treatment

45
significantly more than those subjects Mho were assigned a therapist in-
congruent with their expectations (who increased slightly in anxiety)
(p ^ .001) (see Table 5) Similarly, the Congruent Group showed a sig
nificantly higher increase In the MACL mood of Pleasantness (p < .05)
than the Incongruent Group which decreased slightly on this mood (see
Table 13) Further, the Incongruent Therapist Group displayed a signif
icant increase on the MACL mood of Deactivation (p .05) when compared
to the Congruent Group, which decreased on this mood (see Table 8).
Also, the Incongruent Therapist Group reported a significantly higher in
crease in the MACL mood of Startle after treatment (p .001) than the
Congruent Therapist Group which decreased slightly in Startle (see Table
17). While only four dependent variables significantly distinguished be
tween the Congruent and Incongruent Therapist Groups (in addition to two
interaction effects), and the Adjustment Score indicated no difference as
predicted, no dependent variables significantly changed in the unpre-
dlcted direction. Thus, Hypothesis II appears to be supported. (The
factor analytic studies of Apfelbaum (1958) suggest that three basic
types of psychotherapists may be described on the Expectation Q-sort.
Inspection of the data suggest that the sample does not show a trend to
ward describing a particular type of psychotherapist. Similarly, an in
dividual subject does not clearly choose one or another of Apfelbaum1s three
types of psychotherapists.)
Hypothesis III states that subjects who are given a more positive
set concerning the value and outcome of treatment will display more "pos
itive therapeutic changes on the dependent variables than subjects given
a more neutral set. This was found to be the case upon analysis of two

dependent variables. The Positive Set Group increased significantly less
on the HACL mood of Startle after treatment than did the Neutral Set
Group which increased more (p < .05) (see Table 17). The Positive Set
Group also reported a significantly higher increase on the HACL mood of
Pleasantness (p ^ .05) than the Neutral Set Group which decreased re*
ported pleasantness after treatment (see Table 13). Only two of the 13
dependent variables which were expected to change significantly distin
guished between the positive and neutral set groups. Of particular note
is the fact that no significant changes were demonstrated on either the
AACL or the HACL mood of Anxiety. However, as predicted, no change oc
curred on the Adjustment Score and no dependent variables changed sig
nificantly in the unpredicted direction. Thus, the hypothesis appears
supported.
In summary, those dependent variables which demonstrated signifi
cant differences between one or more of the three independent variables
were: the AACL and the HACL moods of Aggression, Social Affection, De
activation, Anxiety, Pleasantness and Startle. Those dependent variables
which did not significantly distinguish between any of the three inde
pendent variables were Dymond's Adjustment Score (as predicted) and the
HACL moods of Concentration, Depression, Egotism, Skepticism, Nonchalance
and Activation.
As noted earlier, after Nowlis factor analyzed the HACL he di
vided his 12 mood factors into three groups of four moods each: those
moods most clearly identified, those less clearly identified and those
least clarly identified (see Table 4). It is noteworthy that of the four
moods in the group most clearly identified by Nowlis, three were found to

47
distinguish significantly between the Independent variables; of the four
moods in the less clearly identified group, two were found to distinguish
significantly between the independent variables; and of the four moods in
the least clearly identified category, only one was found to distinguish
significantly etween the independent variables. The AACL, the scale
with the greatest number of items, and probably, therefore, the scale
with the highest reliability, showed significant differences between two
of three main effects (independent variables). Below is a summary table
which on the vertical axis lists the three independent variables and
their interactions. On the horizontal axis is listed the 14 dependent
variables and entered into the body of the table are the levels of sig
nificance at which a null hypothesis might be rejected.
Although seven dependent variables did not reflect significant
differences between any of the treatment conditions, there were changes
worth noting in severa) of them. For example, the MACL mood of Depres
sion was found to decrease after treatment in all treatment conditions
except one: the Positive Set, Congruent Therapist, Control Group condi
tion where a small increase in reported depression occurred (see Table
HA). Evaluation by the Signs Test (Siegel, 1956) shows that the proba
bility of seven out of eight cel 1 s decreasing by chance is at less than
the .035 level. This indicates that a reported decrease in depression
tends to occur after treatment, regardless of the specific nature of the
treatment within the study. The MACL mood of Concentration, similarly,
did not result in any statistically significant differences between the
treatment conditions. However, the Congruent Therapist, Experimental
Group condition was the only treatment which described an Increase In

48
TABLE 19
SUMMARY OF THE DEPENDENT VARIABLES AND THE SIGNIFICANCE LEVELS AT WHICH THEY DISTINGUISHED
BETWEEN THE INDEPENDENT VARIABLES
Now 1 Is C1ea r1y
1dentified Factors
Nowlis Less Clearly
1 dent!f1ed Factors
Nowlis Least Clearly
1dentified Factors
AACL
Aggr. Cone. Deact.
So.
Aff.
Anx. Depr. Eqot.
Pleas.
Activ. Nonch. Skept. Start.
A (experimental
vs. control)
05
05
05
B (congruent vs.
incongruent)
.001
.05
05
.001
C (positive vs.
neutral)
05
.05
AB
.01
.01
AC
BC
ABC

49
concentration (see Table 7A). The HACL mood of Skepticism also displayed
no statistically significant differences between the independent varia
bles. However, the Incongruent Therapist, Control Group condition re
ported a slight increase in skepticism after treatment while all others
reported a decrease in skepticism after treatment (see Table 16A).
Hypothesis IV states that no relationship exists between differ
ences in certain subject characteristics and changes in the dependent
variables. This was evaluated by multiple regression technique. The
subject differences under consideration were age, level of intelligence
(Beta Intelligence Test, 1946), level of neuroticism (HPI, 1959) and
level of introversion (HPI, 1959). Analysis of variance (Lindquist,
1953) was used to demonstrate that each of these four subject variables
was randomly distributed across the eight treatment conditions. (The
summary tables for these analyses on age and level of intelligence are
presented in the Subjects section of the Method chapter. Below are the
summary tables demonstrating the random distribution of introversion
(Table 20) and neuroticism (Table 21).
Pearson product-moment correlations were computed between each of
these four variables and the change in scores fromp re- to posttreatment
testing on each of the 14 dependent variables. This was done for each of
the eight treatment conditions. This resulted in 448 correlations which
are presented below in Tables 22, 23, 24, and 25. At the .05 level of
significance, by chance alone wo would expect 24 significant correlations
out of the total 448. In the present data, by frequency count, there are
26 correlations which are significant at the .05 level. This suggests
that, over-all, the null hypothesis appears supported. It is difficult

50
TABLE 20
SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE
RANOOM DISTRIBUTION OF MPI INTROVERSION SCORES
Source
ss
df
V
F
A (experimental vs. control)
19.26
1
19.20
.34
B (congruent vs Incongruent)
31.51
1
31.51
.55
C (positive vs. neutral)
33.84
1
33.84
.59
AB
52.51
1
52.51
.92
AC
.02
1
.02
.00
BC
128.35
1
128.35
2.25
ABC
31.50
1
31.50
.55
w Cells
5.012.75
SSL
56.96
Total
5.309.74
95
(An F of 3.95 is significant at the .05 level.)

51
TABLE 21
SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE
RANDOM DISTRIBUTION OF MPI NEUROTICISM SCORES
Source
ss
m
V
F
A (experimental vs. control)
2.06
i
2.06
.01
B (congruent vs. incongruent)
408.38
i
408.38
2.04
C (positive vs. neutral)
135.40
i
135.40
.68
AB
.37
i
.37
.00
AC
57.02
i
57.02
.28
BC
376.03
i
376.03
1.88
ABC
63.37
i
63.37
.32
w Cells
IZi&ldi
&
200.27
Total
18,665.96
95
(An F of 3*95 is significant at the .05 level.)

52
to attribute a high degree of conclusiveness to results showing some con
sistency in the absence of significance. However there are several ob
servations which may be made and several trends worth Identifying. It is
important to note, also, that the complexity of both the data and the in
dependent variables tends to reduce the meaningfulness of Interpretation.
It will be noted In Table 22 that there Is a tendency for subjects,
given a positive set and a congruent therapist and who score lower meas
ured levels of intelligence, to decrease reported anxiety on the AACL
after treatment. Conversely, this suggests that subjects with higher
measured intelligence, when given a positive set and a congruent thera
pist, increase reported anxiety on the AACL after treatment. Similarly,
subjects with lower measured intelligence, who are given a positive set
and an incongruent therapist, seem to increase their scores on the AACL
after treatment more than subjects with higher measured intelligence. It
will be noted further, with the exception of the Experimental Group,
Neutral Set, Congruent Therapist subjects (which showed an opposite
trend), iess bright subjects tend to increase the MACL mood of Activation
after treatment. From these data, the conclusion may be tentatively ad
vanced that people with lower intelligence seem to change in a thera
peutic direction after certain types of placebo treatment. On the other
hand, with the exception of the Experimental Group, Positive Set, Congru
ent Therapist, there appears to be a trend which suggests that subjects
with higher measured intelligence tend to Increase more on the MACL mood
of Social Affection after treatment than subjects who score lower on In
telligence. This finding suggests a conclusion which is not in line with
that indicated above from the analysis of the AACL and the MACL mood of

53
TABLE 22
PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EIGHT OF THE DEPENDENT VARIABLES AND
SCORES ON THE BETA INTELLIGENCE TEST. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS
Moods of the MACL
So. Adjustment
Treatment Conditions
AACL
Aqar.
Cone.
Deact.
Aff.
Anx.
Depr.
Eqot.
PIeas.
Activ.
Nonch.
Skept.
Start.
Score
Experimental Group
Positive Set, Congruent Therapist
+ .36
+ .10
-.29
+ .42
-.11
+ .22
+ .37
+ .09
O
Csl
l
i
VJ1
O
I
N>
-P-
+ .31
+. 49
-.24
Pos i ti ve Set,
Incongruent Therapist
-.51
+.05
+ 1 1
+ .20
+. 43
-.44
-.23
-.15
+.39
cr\
Csl
I
+.52
-.18
-.61*
+ .48
Neutral Set,
Congruent Therapist
+.20
+. 34
-.50
+ .02
+ .26
+. 04
+ .09
-.35
+.07
+.18
.00
+ .01
+ .52
-.50
Neutral Set,
Incongruent Therapist
-.13
-.39
+ .19
-. 89*
+.54* -.44
-.1 1
+ .60*
-.04
-.32
-.18
-.10
-.73*
-.21
Control Group
Pos i tive Set,
Congruent Therapist
+. 1+5
-.07
-.27
-.21
+.06 +.35
-.33
+.02
+.31
-.27
-.21
+ .26
+ .37
-.01
Positive Set,
Incongruent Therapist
-.33
+ .02
-.36
+. 13
+.10 -.53*
-.18
-.10
+ .08
-.25
-.10
-.37
-.12
+ .08
Neutral Set,
Congruent Therapist
+ .05
-.32
-.13
-.05
+ .25
+ .31
+ .09
-.18
-.24
-.02
-.17
+ .01
-.22
-.09
Neutral Set,
Incongruent Therapist
+ 11
+.41
+.18
+. 06
+. 09
+ .07
-.05
-.28
-.15
OO
-.26
26
+ .07
+ .16
*Significant ^ .05-

54
Activation. That is, in this case, brighter subjects tend to increase
Social Affection (personal comfort) as a result of the placebo situation,
while In the former case, brighter subjects appear to increase anxiety
and become less active. Of further interest Is another tentative finding
suggesting that brighter subjects, when given a congruent therapist, tend
to Increase scores on the MACL Skepticism mood after treatment, but de
crease skepticism after treatment when given an incongruent therapist.
From Table 23 several interesting, but also tentative relation
ships may be discussed. In the Experimental Group, the more Introverted
subjects who are *lven the Positive Set, Congruent Therapist Instructions
tend to increase reported Social Affection after treatment. However, the
more introverted subjects who are given the Neutral Set, Incongruent
Therapist instructions tend to decrease the MACL mood of Social Affection.
An additional trend (with the exception of the Control Group, Neutral Set,
Incongruent Therapist subjects) appears to be that introverted subjects,
more than extroverted subjects, tend to decrease scores on the MACL mood
of Skepticism after treatment.
Several relationships presented in Table 24 may be tentatively
discussed. It appears that in the positive set cells of the Experimental
Group and in the neutral set cells of the Control Group the older the
subject, the larger the increase in the MACL mood of Social Affection.
Similarly, In the positive set cells of both the Experimental and Control
Groups, the younger the subject the greater the increase in reported MACL
Egotism mood after treatment. Another effect appears to exist in the re
lationships between age and change in anxiety after treatment. On both
the AACL and the MACL Anxiety mood, it Is noted that younger subjects

55
TABLE 23
PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND
SCORES ON THE MPI INTROVERSION (LOW SCORES)-EXTROVERSION (HIGH SCORES). THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS
Moods of the MACL
So. Adjustment
Treatment Conditions
AA.CL
Aggr.
Cone.
Deact.
Aff. Anx.
Depr.
Eqot-
Pleas.
Activ.
Nonch.
Skept.
Start.
Score
Experimental Group
Positive Set, Congruent Therapist
+.21
+.03
-.21
+. 1 0
-.54* +.01
+ .20
+.32
-.07
+ .57*
+ .53*
+ .44
+ .25
+ .71*
Positive Set, Incongruent Therapist
+. 06
-. 12
+ 09
-.46
vO
1
LTV
CM
+.05
+.51
+ .03
+ .41
+.13
+.21
+ .24
-.13
Neutral Set, Congruent Therapist
+.07
+ .02
-.1 1
-.33
-.14 +.03
-.06
+ .14
-.31
-.11
+ .29
+ .40
-.13
+ .26
Neutral Set, Incongruent Therapist
-.02
-.31
+.09
-.11
+.60* +.41
-.28
-.36
+.02
+ .01
+ .01
+ .01
-.51
+ .01
Control Group
Positive Set, Congruent Therapist
+. 12
-.36
-.52
.00
*
-.59* +.30
00
0
1
-.06
-.16
+ 06
+ .10
+ .30
+.57*
-.03
Positive Set, Incongruent Therapist
-.07
+.52
+ .08
+. 08
LA
O
O
-3r
1
+. 14
+.37
-.28
+ .32
+ 48
+ .02
-.03
-.04
Neutral Set, Congruent Therapist
+ .31
+.04
-.05
+. 11
+.35 -.37
+.40
-.24
+.37
+ .29
-.04
+ .61*
-.15
-.36
Neutral Set, Incongruent Therapist
-.27
-.18
-.14
-.25
-.17 +.01
+. 11
+.29
-.17
-. 65*
+. 15
-33
-.19
-.30
*Significant ^ .05.

56
TABLE 24
PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND
SUBJECTS1 AGES. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS
Moods
of the
MACL
Treatment Conditions AACL
Aggn.
Cone.
Deact.
So.
Aff.
Anx.
Depr.
Eqot.
Pleas.
Acti v.
Nonch.
Skept.
Start.
Adj ustment
Score
Experimental Group
Positive Set, Congruent Therapist +.07
+ 07
+ 19
+ .06
+.49
+. 1 1
-.19
-.36
-.35
-.41
-.56*
+ .14
+ 14
-.51
Positive Set, Incongruent Therapist +.36
+.42
+ .32
+. 06
+. 50
+.45
+ .51
-.24
+. 14
-.06
-.32
-.07
-.03
-.05
Neutral Set, Congruent Therapist +.51
+.01
-.34
+ .13
-.22
+.32
-.20
+.30
+. 03
-.05
+ 06
+.52
+ .50
+ .50
Neutral Set, Incongruent Therapist +.03
+.23
-.10
.00
-.25
+. 06
-.25
+.39
+. 41
-.02
+.48
+.49
-. 14
-.48
Control Group
Positive Set, Congruent Therapist +.37
-.22
+. 16
-.33
-.45
+ .18
-.07
-.28
-.20
-.01
-.36
-.66*
-.08
-.59*
Positive Set, Incongruent Therapist +.25
+.15
+ .20
-.43
-.21
+ 26
-.19
-.18
+. 21
+.24
-.22
-.21
-.29
-.15
Neutral Set, Congruent Therapist -.60*
+.02
+. 36
+. 12
+ .31
+.23
+.02
+. 28
+.32
+.30
-.01
+ .05
-.24
+ .41
Neutral Set, Incongruent Therapist -.12
-.29
+ .22
-.31
+. 18
-.25
+.19
+. 06
-.09
+ .06
+ .42
-.38
+ .14
+ .03
'^Significant .05.

57
decrease reported anxiety after Experimental Group treatment more readily
than do older subjects. This Is especially so for the Positive Set, in-
congruent Therapist and Neutra) Set, Congruent Therapist conditions.
Eysenck defined his Neurotic)sm Scale as a measure extending
along a highly neurotic to normal continuum, in the current study, the
sample contains no "normals," therefore, serious questions may be raised
concerning the meaningfulness of the neurotic!sm score. For example,
does a low score on neuroticlsm in this study mean that a patient is psy
chotic? A biserial correlation (McNemar, 1955) between diagnosis (psy
chotic vs. nonpsychotic) and scores on the Neuroticlsm Scale results In a
correlation of .57 (p ^ .001). This suggests that subjects with nonpsy
chotic diagnoses tend to score high on the Neuroticlsm Scale whtie sub
jects with psychotic diagnoses tend to score lower. However, this corre
lation accounts for only 32.49 per cent of the variance, and it is with
extreme caution that interpretation of the neuroticism score data be made.
As noted in Table 25, an apparent relationship exists between
neuroticism scores and the MACL mood of Nonchalance. With the exception
of the Experimental Group, Neutral Set, Congruent Therapist subjects, it
appears that as neuroticism increases, nonchalance decreases after treat
ment. Another interesting relationship indicates that subjects scoring
high on the Neuroticism Scale who are In the Positive Set, Congruent
Therapist, Control Group decrease their HACL mood of Startle after treat
ment; while subjects scoring high on the Neuroticlsm Scale who are in the
Neutral Set, Incongruent Therapist, Control Group increase this Startle
mood after treatment.

58
TABLE 25
PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND
SCORES ON THE MPI NEUROTICISM SCALE. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS
Moods
of the
MACL
Treatment Conditions
AACL
Jtefl.Lv
Cone.
Deact.
So.
Aff.
Anx.
Depr.
Eqot.
Pleas.
Acti v.
Nonch.
Skept.
Start.
Adj ustment
Score
Experimental Group
Positive Set, Congruent Therapist
-.08
+. 19
+. 51
-.12
+.52
+.17
-.43
-.46
-.36
-.39
-.43
-.29
-.24
-.56*
Positive Set, Incongruent Therapist
+.18
-.23
-.37
-.03
-. 06
+.07
+. 41
-.22
+. 01
-.38
-.59*
-.38
+ .03
-.43
Neutral Set, Congruent Therapist
-.24
-.41
+. 14
+. 13
+. 04
+. 43
+.09
+.42
+ .70*
+. 34
+ .26
-.23
+ .20
-.06
Neutral Set, Incongruent Therapist
-.13
+.02
+.29
+ .12
-.03
+.56*
+.21
+. 12
+ .45
+.42
-.07
+ .48
+ 12
-.51
Control Group
Positive Set, Congruent Therapist
-.01
+.03
+ .52
-.17
Hr.
-.14
+. 08
+.03
-.16
+ 19
-.18
-.17
-. 69*
-.10
Positive Set, Incongruent Therapist
-.29
-.28
-.37
-.15
+. 16
-.27
-. 60*
-.30
-.12
-.41
-.28
-.02
-.30
+ .42
Neutral Set, Congruent Therapist
-.38
+ 26
+.37
+.10
-.28
+.39
1
LO
+.35
-.36
+. 13
-.14
-.30
-.1 I
+ .43
Neutral Set, Incongruent Therapist
+.17
+. 20
-.34
+. 40
-.19
+.47
+. 15
-.16
+.23
-.22
-.35
+.20
+. 56*
+. 04
''Significant ^.05.

59
It is believed sufficiently Important to stress once again the
tentativeness of all relationships discussed under Hypothesis IV. It is
this writer's opinion that due to the low number of significant correla
tions, the data basically support the null hypothesis that no conclu
sively meaningful relationships exist between the dependent variables and
the measured individual differences.
It was believed that a subject's reported level of desire for
psychotherapy would be related to his amount of verbalization when he was
permitted to speak to the "psychotherapist." The tenability of this be
lief was evaluated by multiple regression technique utilizing only the
Experimental Group and the correlations are presented below in Table 26.
TABLE 26
HULTIPLE REGRESSION COEFFICIENTS BETWEEN AMOUNT OF SPEECH ANO
DESIRE FOR PSYCHOTHERAPY PRESENTED FOR EACH INDEPENDENT
VARIABLE (EXPERIMENTAL DATA ONLY)
Congruent
Therapist
Incongruent
Therapist
Positive
Set
Neutral
Set
Amount of
Speech
.50
.02
.49
.08
vs.
Desire for
Psychotherapy
n s 24
o
c
s

e
n s 24
Amount of verbalization, which was originally rated on a four-
point continuum (see p. 26) was divided, for the purpose of this evalua
tion, Into two groups: (1) constant speakers and moderate speakers, and
(2) little speakers and no speech. Desire for psychotherapy was rated on
a three-point scale by each subject (see Appendix F).

6o
The correlations presented in Table 26 suggest that if a subject's
desire for psychotherapy is high, his amount of verbalization is high if
he is given a congruent therapist and/or a positive set. If, however, he
is given an incongruent therapist and/or a neutral set there Is no rela
tionship between his desire for psychotherapy and his amount of speech.
Presented below in Table 27 are co-relationships between (I)
amount of speech; (2) desire for psychotherapy; and (3) diagnosis on the
one hand, and change scores from pre- to posttesting on the AACL and the
moods most clearly identified by Now!is on the other (data only for the
ExperI mental Group).
One interesting conclusion which may be drawn rrom these data
concerns the relationship between amount of speech and change scores on
the AACL. It appears that subjects who speak the most Increase anxiety
from pre- to posttreatment testing. This Is true in all treatment con
ditions and the correlation for ail groups is +.30 (p < .05). Of
further interest Is the fact that level of desire for psychotherapy does
not appear, by inspection, to appreciably Influence the amount of change
on these five dependent variables from pre- to posttreatment testing.
Similarly, by inspection, diagnosis (psychotic or nonpsychotic) does not
appear to be related to the amount of reported change In personal comfort
after treatment as measured by the five dependent variables.

TABLE 27
MULTIPLE REGRESSION COEFFICIENTS BETWEEN THE VARIABLES LISTED ON THE VERTICAL AXIS AND FIVE
DEPENDENT VARIABLE CHANGE SCORES FOR THE EXPERIMENTAL GROUP. THE DATA ARE PRESENTED
SEPARATELY FOR EACH OF THE INDEPENDENT VARIABLES IN ADDITION TO THE
TOTAL EXPERIMENTAL GROUP
AACL
4r
Moods of
Cone.
the MACL
Deact
So.
Aff.
Congruent
Therapist

Amount of speech
DesI re for
+.11
-.29
-.13
-.06
-.32
Instructions
Psychotherapy
+.06
+.22
-.02
+.17
+.02
(n = 24)
Diagnosis
+.14
-.09
+.09
-.15
-.09
1ncongruent
Therapist
Amount of speech
Desire for
+.29
+.21
+.09
-.06
+.34
Instructions
Psychotherapy
-.20
-.07
+.20
+ .10
+ .27
(n 24)
Diagnosis
+.03
+.13
-16
+.03
+.04
Pos!tive
Set
Amount of speech
Desire for
+.26
+ .23
+.13
+.16
-.00
(n 24)
Psychotherapy
+.08
+.07
+ .04
+.59**
+.03
Diagnosis
+.04
+.05
+.08
+ .02
-.10
Neutral
Set
Amount of speech
Desire for
+.39
-.01
-.32
-.13
-.08
(n 24)
Psychotherapy
-.08
+.08
+.07
-.38
+ .22
Diagnosis
+.06
+.01
-.18
-.19
+.13
All
Experimental
Amount of speech
Desi re for
+.30*
+.11
-.06
+.04
-.01
Group Subjects
Psychotherapy
+.01
+.09
+.06
+.17
+.11
(n 48)
Diagnosis
+.05
+.02
-.04
-.07
-.01
^Significant ^ .05.
**Significant .01.

DISCUSSION
As mentioned in the Introduction, the literature supporting the
efficacy of psychotherapy has not demonstrated that personality and be*
havioral modifications which occur as a result of psychotherapy are en-
tirely due to a psychotherapist's behavior and techniques (specific fac
tors). The outcome of treatment, it was suggested, is a function of both
the specific and the nonspecific (placebo) events occurring during psy
chotherapy, although, in the literature, the presence and the contribu
tion of the placebo effect in psychotherapy appear to be entirely
conjectural.
This study appears to demonstrate that the placebo effect in psy
chotherapy does exist. Subject-patients who spoke into a tape recorder
for one session, believing that they were speaking to a psychotherapist,
changed certain behaviors In therapeutic directions when compared to sub
jects who did not speak (see Hypothesis l). That Is, it has been shown
that verbally reported patient behaviors such as anxiety and certain
moods which would be expected to change as a result of therapist behav
ior in actual psychotherapy, also change, in fact, as a result of the
placebo psychotherapy (nonspecific treatment) in which a therapist is ab
sent. The group who spoke to the "psychotherapist for one session, when
compared to the group who waited, increased their Social Affection mood on
the MACL and reduced their level of anxiety as measured on both the AACL
and the HA CL Anxiety mood.
62

63
As a result of these findings in the current study, it is pro
posed that the placebo effect does contribute to the outcome of psycho
therapy, and that a measurement of the amount of contribution appears to
have been made. However, it is important to question precisely what has
been evaluated by a comparison between the Experimental Group (talkers)
and the Control Group (waiters). If the subjects who spoke received pla
cebo treatment and the subjects who waited received no treatment, the
statistical comparison between these two groups was an accurate measure
ment of the placebo effect. However, it is believed that those in the
Control Group, who waited, but were promised a therapist, and who changed
on the dependent variables in a generally "nontherapeutic" direction, re
ceived some placebo benefit from the procedure (e.g., see Tables I1A and
I3A). This finding does not Tnftuence the conclusion that the placebo
effect in psychotherapy does exist, however, the precise therapeutic bene
fit of the placebo, as measured in this study, has not oeen compared to
the therapeutic benefit of no placebo treatment. Instead, the comparison
between the Experimental and Control Groups appears to have been an anal
ysis between a more effective (Experimental Group) and a questionably ef
fective (Control Group) placebo treatment.
In addition to congruence of expectations regarding the charac
teristics of the assigned psychotherapist and positive set regarding
treatment outcome (see below), may other "psychologic mechanisms" through
which the placebo effect takes place (Shapiro, 1964) be identified in
order to account for the finding that this treatment did result In a pla
cebo effect? Shapiro (1964) reports that throughout the history of medi
cine, "methods of depletion (which are currently accepted as placebos)

were widely used (to effect cures): entities, cathartics, enemas, purges,
stomachics, sweating, bleeding, leeching, cupping, starvation and dehy
dration, Methods of depletion and expulsion . may relieve symptoms
by symbolically expelling bad thoughts and conflictual ego-alien impulses."
In our more sophisticated and verbally oriented culture, Shapiro contin
ues, "relief of symptoms may occur when the patient is able to express
verbally conflictual and guilt-ridden thoughts and feelings in the free,
nonjudging and accepting atmosphere of the doctor's office" (p. 81). in
other words, verbal catharsis, it appears, may theoretically lie at the
basis of the placebo effect in psychotherapy, just as physical-mechanical
catharsis lay at the basis of medical treatment which today is widely
looked upon as placebo. The treatment in this experiment is conducted in
an obviously free, accepting and nonjudgmental atmosphere and each
patient-subject may establish his own personal optimal level of catharsis.
It would seem logical then, that catharsis may be a major underlying fac
tor for the placebo effect in psychotherapy, at least in so far as out
come of psychotherapy is measured in this study. It is interesting, in
this regard, however, that the data indicate that a linear relationship
does not exist between amount of speech (loosely, catharsis) during the
"interview" and degree of change on the dependent variables. In fact, in
this context, although one significant correlation among the 25 computed
would be expected to be significant by chance alone, the only significant
correlation (.05 level) which does exist, suggests that the more a sub
ject speaks, the greater the likelihood that his anxiety level (AACL)
will increase. These correlation*- however, evaluate only the amount of
speech in relation to the dependent variable changes after treatment. The

65
analysis does not assess content of speech or nature of affect which are
certainly extremely significant aspects of catharsis. In addition, this
analysis evaluates only the linear relationships between amount of speech
and degree of change on the dependent variables. The analysis does not
deal with nonlinear relationships, and It is conceivable that such rela
tionships might exist.
Shapiro (1964) reports that expressed desire for treatment is an
other important concept which reportedly is basic in eliciting the pla
cebo effect. That is, those subjects who are more highly motivated for
treatment would be more responsive to placebo treatment. However, In
this study, motivation of the subjects, as measured by their statements
concerning their level of desire for psychotherapy, Is not linearly re
lated to the amount or nature of change after treatment as measured on
the dependent variables. Interestingly, however, when subjects who re
port high motivation for psychotherapy are given a positive set concern
ing the outcome of the treatment and/or are offered a therapist who Is
congruent with their expectancies they tend to speak more. That is,
greater reported desire for psychotherapy is related to a larger amount
of verbal catharsis; however, It is unrelated to changes on the measures
of mood and anxiety after placebo psychotherapy when a therapist is
absent,
It appears, thus, that in this study, catharsis or degree of mo
tivation, or both, have not been shown to underlie the placebo effect.
Shapiro (1964), in his review of placebo literature, indicates that a
great number uf "psychologic mechanisms" have been proposed as being un
derlying characteristics of the placebo effect. Several are: catharsis,

66
motivation, faith, learning and cerxiitioning, trust, confidence and
previous experience with healers.
Shapiro (and Whitehorn, 1953) further indicate that no single
characteristic of treatment has been consistently shown to underlie and
to produce the placebo effect. The conflicting and inconclusive evidence
in the literature Is most conceivably due to the unclear nature of the
placebo effect, differences of measurement of the effect from research to
research and the different theoretical frameworks within which different
experiments are planned.
Thus far, the discussion has been concerned with the contrast be*
tween the Experimental Group (talkers) and the Control Group (waiters)
and several of the characteristics of psychological treatment which may
determine the placebo effect. The conditions under which the placebo ef
fect is most likely to take place or most likely to be lessened, are
further clarified when attention is drawn to the way in which the treat
ment groups were divided within the present experiment. Analysis of
Hypothesis II, indicates that patients who are offered a "psychothera
pist" congruent with their expectations, report more "positive thera
peutic changes" on measures of anxiety and moods than do patients who are
assigned a psychotherapist who is incongruent with their expectations
concerning the nature of the psychotherapist to be assigned. That is,
the subject's perception of the treater or doctor is a significant aspect
in encouraging or discouraging the placebo effect.
Festinger (1957) and Festlnger and Bramel (Bachrach, 1962) pro
pose that dissonance results when two cognitions which a person holds are
inconsistent with each other according to the expectations of the person.

67
Dissonance is said to be a motivating state which is comparable to other
drive states, and just as hunger produces physical discomfort, dissonance
results in psychological discomfort. When a patient-subject is instructed
that the characteristics of his psychotherapist are incongruent with his
personal expectations concerning this psychotherapist, the patient-
subject experiences dissonance. This in turn, reduces the effectiveness
of the placebo treatment, as is shown by the statistical analyses which
indicate that the AACL and MACL moods of Startle and Deactivation de
creased significantly more for the Congruent Therapist Group, while the
MACL mood of Pleasantness increased significantly more in the Congruent
Therapist Group.
(It is noteworthy that although the Apfelbaum questionnaire [see
Appendix E) instructs subjects to respond on the basis of their expecta
tions. any individual subject may have completed the items in terms of
his ideal psychotherapist or in terms of a desired psychotherapist.
There is no apparent way in which to evaluate these possibilities. If,
however, any individual subject did respond In terms of an ideal or a de
sired psychotherapist, he may have received additional, or reduced, posi
tive set, in addition to having the nature of the treater confirmed or
not confIrmed.)
In addition to demonstrating that the patient's perception of the
nature of the treater Is one significant aspect of inducing a greater or
lesser placebo effect, this research also suggests that the nature of the
treatment situation itself is important. Analysis of Hypothesis III in
dicates that subjects who are told that the placebo psychotherapy will be
extremely helpful to them (Positive Set Group) profit more from the

68
treatment than those subjects with whom positive treatment outcome is not
discussed (Neutral Set Group). This conclusion is somewhat less certain
than those drawn from hypotheses I and II. This Is so because only two
of the dependent variables (MACL moods of Startle and Pleasantness) indi*
cated that the positive set was significantly more effective in encourag
ing the placebo effect than the neutral set. The literature which served
as the impetus for introducing this Independent variable (positive vs.
neutral set) into the current study, apparently uniformly indicates that
positive set or attitude toward a particular placebo treatment increases
the placebo effect. (For example, see pp. 45-46 of the present study and
also social psychological research on set and frames of reference:
Newcomb, 1958, pp. 264-297* Maccoby et ai., 1958, p. 95.)
Why hasn't the current study demonstrated more conclusively that
a positive set concerning treacmont outcome Is even more important in se
curing a placebo effect, as might be anticipated on the basis of the lit
erature? Two explanations are suggested. On the one hand, it is possi
ble that the wording of the two instructional sets (constructed by the
author) were not sufficiently different and thus, two entirely separate
groups concerning expectation of treatment outcome may not have been es
tablished. On the other hand, it will be noted (see Appendix A) that the
Positive Set instructions emphasized the reputation and demand for the
alleged psychotherapist, in addition to his willingness to help the sub
ject. It is possible that Instructions stressing this psychotherapist's
professional reputation and the demand for his services did not increase
the positive nature of the set for this particular group of subjects as a
whole. Thus, portions of the Positive Set instructions may not have

69
served the purpose for which they were intended. (By stressing charac
teristics of the psychotherapist, this aspect of the instructions may
have had little influence on the level of set but may have had some vi
carious influence on the congruence or incongruence of the psychothera
pist for any particular subject.) ^
In summarizing the results of the analysis of hypotheses 1, 11
and III, the following may be said: The placebo effect in psychotherapy
does exist and through the technique of this study appears to be measura
ble. Further, the placebo effect is apparently influenced by both the
perceived characteristics of the treater, and the projected outcome of
the treatment as reported to the patient. The influence of the perceived
characteristics of the treater has been approached by rulysls of differ
ences on the dependent variables between one group of subjects who were
told that a therapist to be assigned to them will be congruent with their
expectations, and another group who were told that the therapist will be
incongruent with their expectations. The data apparently demonstrate
that the group given the congruent therapist responds to the treatment
with a greater positive placebo effect. The influence of the reported
outcome of the treatment has been approached by analysis of differences
on the dependent variables between one group of subjects who were given a
positive set concerning treatment outcome and another group who were
given a more neutral set. The data suggest that the group given the
positive set tends to respond to treatment with a greater positive pla
cebo effect. (However, congruent therapist instructions seem to be more
instrumental in encouraging the placebo effect than the "positive set"
instructions.)

70
Having found that offering a therapist who is congruent with a
subject's expectations seems to encourage the placebo effect more than
offering a positive set concerning treatment outcome may have Implica
tions beyond the present study. There is a considerable body of research
concerned with the matching of patients and therapists and the resulting
relationships in psychotherapy (e.g., see Leary and Gill [1959])* In the
present study it is seen that matching a therapist with a patient's ex
pectations does, in fact, increase the effectiveness of the treatment to
a greater extent than offering a positive set concerning treatment out
come. This is indicated in this placebo situation, where no therapist
actually exists; however, it is suggested that in psychotherapy percep
tion of the psychotherapist may also be more valuable than belief about
treatment outcome. That is, the nature of the relationship (in this
study, a relationship In which the patient perceives his therapist as
being congruent with his expectations) may be more essential In success
ful psychotherapy than statements concerning the value of the treatment
1tself.
An important finding of this study is that, as predicted, the
Adjustment Score, which measures general level of psychological adjust
ment, does not change after the placebo Interview. This is so regardless
of the particular treatment (talk vs. wait, congruent therapist vs. in-
congruent therapist and positive set vs. neutral set) or combination of
treatments in the study. This indicates that in a single treatment of
placebo psychotherapy, no change in general level of psychological ad
justment may take place. Had the placebo psychotherapy continued over a
longer period of time would there have been changes In this Adjustment

71
Score? That is, would extended placebo psychotherapy have produced
changes In psychological adjustment similar to changes which occur after
actual psychotherapy? (See Frank, 1961, and Rogers and Dymond, 195**, for
evidence of changes in general level of psychological adjustment after
psychotherapy.) it is difficult to speculate on precisely what may have
happened to the subject's general level of psychological adjustment had
the placebo treatment continued, although researchers such as Glasser and
Whitt low (1953> 1954) submit that the effects of placebo treatment are
not permanent and the gains from continued placebo treatment become less
and less.
What might have happened to moods and anxiety had the placebo
psychotherapy continued? Martin et ai (I960), In a study concerned with
levels of therapist comaunications (discussed more fully in the Introduc
tion), had subjects speak into a tape recorder for five 30-minute weekly
sessions as if they were speaking to a psychotherapist. These authors
demonstrate that although some initial positive therapeutic effect was
indicated (lowering of anxiety), the procedure lost its positive thera
peutic effectiveness as the number of sessions with the tape recorder In
creased. Glasser and Whlttlow (1953, 1954), in placebo drug studies,
found that if subjects were given placebos over a period of time and re
quired to answer questionnaires identifying the effects of these placebos,
the positive placebo effect was demonstrated Initially, but the effect
was less upon each successive evaluation. Findings such as these suggest
that the placebo procedure used in the current study, if extended for
more than one session, might not have continued to have positive thera
peutic effect on mood and anxiety.

72
Analysis of Hypothesis IV gives no clear picture of the contribu
tion to the placebo effect of the individual patient differences in age,
Intelligence, introversion and neurotic!sm. This finding is not surpris
ing in viewof the existing body of research which is equivocal and often
contradictory in regard to this question (see Introduction). The evalua
tion of the data collected to explore this hypothesis resulted in border
line significance and lack of remarkable trends both within the four
measured subject differences and within all of the dependent variables
across individual subject differences. Even when tentative or apparent
trends were identified (as discussed on pp. 49-59), the complexity of the
behaviors involved In the eight different combinations of the independent
variables makes the meaningfulness of any interpretation of results
doubtful. (In addition, the reader is reminded of the questionable use
of the Neurotic!sm Scale or the MPi with this subject sample; see p. 57*)
/
Pearson product-moment correlations were used to evaluate this hypothesis.
These only identify linear relationships between the dependent variables
and the measured subject differences, and the possibility of a nonlinear
relationship cannot be dismissed on the basis of the analysis performed.
The problem of discussing the meaningfulness of the results and
the inability to disclose significant trends in the data of this study
apparently leads to the acceptance of the null hypotheses which state
that, in the present sample, age, intelligence, Introversion or neuroti-
v
cism will not identify the type of subject who responds favorably to pla
cebo psychotherapy. This Is in keeping with the findings of Wolf et al.
(1957) who indicate that individual subject reactions to placebos are

73
generally inconsistent, unpredictable and not uniform (see also Hagans
et al.. 1957, and Kurland, 1957)*
A pertinent issue may be raised concerning whether the subjects
accepted the placebo treatment as a legitimate form of psychotherapy.
There is no way In which to be certain of any conclusion in this regard.
However, the nature of changes on the dependent variables certainly sug
gests that the subjects did perceive the procedure as actual psychother
apy. Moreover, it will be noted that the Skepticism mood on the MACL
remained rather constant from pre- to posttreatment testing, and the only
groups to report a small and insignificant Increase in Skepticism, after
having completed the procedure, were the Incongruent Therapist, Control
Group subjects (see Table I6A). No subjects reported that they disbe
lieved the legitimacy of the treatment and the comments several of the
subjects made to the examiner and into the tape recorder appear related
to the issue of subjects accepting the treatment as real. For example,
one subject in the Experimental, Neutral Set, Congruent Therapist Group
reported the following Into the tape recorder during the placebo therapy:
"I appreciate the privilege of being given a chance to express my feel
ings and views in this way. . This is the first chance I've had to
honestly and openly express my feelings and opinions and I appreciate it
very, very much."
It is hoped that the placebo effects demonstrated In this treat
ment will be kept in mind in future research on the effects of specific
methods In psychotherapy. It is further hoped that this demonstration
will lead to the practice of separating effects common to all psycho
therapy (e.g., nonspecific placebo effects) from those effects due to the

specific aspects of the therapy being evaluated (e.g., therapist
behaviors).

APPENDICES

APPENDIX A
INSTRUCTIONS TO SUBJECTS

INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS PRIOR TO
ESTABLISHING THE TREATMENT CONDITIONS IN THE STUDY
I'd like you men to do several things this afternoon. I think
you will enjoy these tasks. You will learn something about yourself and
may profIt from what we do here for the next hour or so.
(Distribute the personal Information questionnaire.)
I'd like for you to answer these questions about yourselves.
(Distribute the Maudsley Personality Inventory.)
The next questionnaire Is this one. (Read directions.) Begin.
(Distribute the Anxiety Adjective Check List.)
The next questionnaire is this one. (Read directions.) Please
answer honestly the way you feel right now. Begin.
(Distribute the Mood Adjective Check List.)
The next questionnaire is this one which asks you about your mood
and the way you feel right now. (Read directions.) Some of the items in
this list are like items in the last list, but please answer them anyway.
Be honest with yourself and just answer the way you feel right now.
(Distribute the Modification of the Expectation Q-$ort.)
This questionnaire is a little different. It says(read direc
tions). Now, don't think too long on any of these, but answer them as
best you can.
(Distribute the Adjustment Score.)
This questionnaire asks you ho* you feel about yourself and also
about things in general. Read the instructions and work quickly
INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS TO DETERMINE
POSITIVE SET OR NEUTRAL SET
Positlve Set Instructions
I can tell you now what this is all about. It's a real good op
portunity for you men because you are being given a chance to talk with a
great psychiatrist. Now, you can talk to him about anything you'd like
77

78
relating to problems you've had, or problems you are having, or problems
you can see ahead of you In the future* This psychiatrist has such an
excellent reputation that he is an extremely busy man, but he has offered
to help you men to solve any problems you may have anyway. Because he is
so busy he can't get here to the hospital so he suggested we do something
a little differently because he is so interested in helping you. You
will siftply talk Into a tape recorder, and the tape will be sent regis
tered mail to him. I think this Is really great because he has helped so
many, many people and I'm sure he can help you with any problems you may
possibly have.
Neutral Set instructions
i can tell you now what this is all about. It is an opportunity
for you men to talk with a psychiatrist. Now, you can talk to him about
anything you'd like relating to problems you've had, or problems you are
having, or problems you can see ahead of you In the future. This psy
chiatrist is an extremely busy man, and he Is unable to get here to the
hospital so he suggested we do something a little differently. You will
simply talk into a tape recorder, and the tape will be sent registered
mai1 to him.
INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS TO DETERMINE
CONGRUENT THERAPIST OR INCONGRUENT THERAPIST
Congruent Therapist Instructions
I've been looking over your questionnaire on the way you would
expect this psychotherapist or counselor to be, and I see that this psy
chiatrist Is the way you thought he would be. For example, Mr.
(subject seated on the Experimenter's right), he (read this subject's re
sponses to I terns 7. 14 and 29; see Appendix E). And Mr. (sub
ject seated on the Experimenter's left), he (read this subject's responses
to items 1, 31 and 32; see Appendix E).
Inconqruent Therapist instructions
I've been looking over your questionnaire on the way you would
expect this psychotherapist or counselor to be, and I see that this psy
chiatrist is not exactly the way you thought he would be. For example,
Mr. (subject seated on the Experimenter's right), he (read a denial
of the accuracy of this subject's responses to items 7, 14 and 29; see
Appendix E). And Mr. (subject seated on the Experimenter's left),
he (read a denial of the accuracy of this subject's responses to items I,
31 and 32; see Appendix E).

79
INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS AFTER HAVING
ESTABLISHED THE FOUR TREATMENT CONDITIONS
Mr. (Control Group subject), I'd like for you to wait here
for a while. Make yourself comfortable. You can look at some magazines,
or rest, but please stay in this room. I'll be back in about forty-five
minutes. Mr. (Experimental Group subject), please come with me.
INSTRUCTIONS GIVEN TO THE EXPERIMENTAL GROUP SUBJECT AFTER
BEING SEATED IN THE ROOM WITH THE TAPE RECORDER
Now, this Is the tape recorder and the tape that wi 11 be sent to
the psychiatrist. Just make yourself comfortable, and say whatever you
would like into this microphone. The controls on the machine are set,
and please remain In your seat until I return. It's OK if there are
pauses, and you don't have to talk all of the time. I'll be back In
about fifty minutes to get the tape.
INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS AFTER THE EXPERIMENTAL
GROUP SUBJECT IS REUNITED WITH THE CONTROL GROUP SUBJECT
I'd like for you to retake several of these shorter tests. This
will only take a few minutes more.
(Distribute the Anxiety Adjective Checx List.)
This is the first one. You remember this one, it asks you to
check the way in which you feel right now. Please answer honestly the
way you feel riqht now.
(Distribute the Mood Adjective Check List.)
This is the next one. You remember this one. It asks you to mark
the way you feel right now. Please answer honestly the way you feel
right now.
(Distribute the Adjustment Score.)
This is the next one. Please work as quickly as you can.
INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS
AFTER THE PAIR HAS COMPLETED THE TESTS
Mr. (Experimental Group subject), you may leave now. Mr.
(Control Group subject), do you care to speak to the psychiatrist?
(If the Control Group subject wishes to speak, he Is taken to the tape
recorder; If, however, he declines to speak, he is dismissed at this
time.)

APPENDIX t
SAMPLE ANXIETY ADJECTIVE CHECK LIST

Name
Below you will find words which describe different kinds of feel
ings. Check the words which describe how you feel RIGHT NOW. Some of
the words may sound alike but we want you to check al 1 of the words that
describe your feelings.
l._
AFRAI0
21.
GAY
41.
PANICKY
2.
AGITATED
22.
GLOOMY
42.
PEACEFUL
3-
ANGRY
23.
GRIM
1

PLEASANT
4.__
BITTER
24.
HAPPY
44.
RATTLED
5-_
CALM
25-
HELPLESS
45._
SAD
6.
CHARMING
26.
HOPELESS
46.
SECURE
7.
CHEERFUL
27-
1NSECURE
47-
SENTIMENTAL
8.
COMPLAINING
28.
JEALOUS
48.
SERIOUS
9-
CONTENTED
29-
JOYFUL
49. _
SHAKEY
IQ.__
CONTRARY
SjJ
o

1
KlNDLY
50.
SOLEMN
l-
COOL
31.
LIGHT-HEARTED
51.
STEADY
I2._
CR0SS
32.
LONELY
52.
TENDER
13.
..DESPERATE
33.
LOVING
53-
TENSE
!4.__
EASY-GOING
34.
MAD
54.
TERRIFI£D
*5-
FEARFUL
35.
MEAN
55.
THREATENED
16.
..FEARLESS
36.
MERRY
56.
THOUGHTFUL
17.
^FRETFUL
37.
MlSERABLE
57.
UNCONCERNED
18.
FRIENOLY
38.
NERVOUS
58.
UNEASY
I9.__
_FRIGHTENED
39.
OVERCONCERNEO
59.
UPSET
20.
^FURIOUS
40.
OVERWHELMED
60.
WARM
61.
WORRYING
81

APPENDIX C
SAMPLE MOOD ADJECTIVE CHECK LIST

Name
Each of the 40 words In the following list describes feelings or
mood. Please use the list to describe your feelings at this moment.
Hark each word according to these Instructions: If the word definitely
describes how you feel at the moment you read it, circle the double plus
(++) to the right of the word. For example, If the word Is calm and you
are definitely reeling calm at the moment, circle the double plus as
follows:
calm 0- ? no (This means that you definitely feel calm at this
moment.)
If the word only slightly applies to your feelings at the moment, circle
the single plus as follows:
calm ++ 0 ? no (This means that you feel slightly calm at this
moment.)
If the word is not clear to you or If you cannot decide whether or not
It describes your feelings at the moment, circle the question mark as
follows:
calm ++ + (j) no (This means you can't decide whether you are calm.)
If you clearly decide that the word does not apply to your feelings at
the moment, circle the no as follows:
calm ++ + ? (no) (This means that you are sure that you are not calm
at the moment.)
Work rapidly. Your first reaction Is best. Work down the column before
going to the
mi utes.
next.
Mark
all the words.
This should
take only a
few
angry ..
++
?
no
boastful ..
+
?
no
coceentrat I ng
++
+
?
no
elated
+
?
no
drowsy ..
++
+
?
no
active
?
no
affectionate
++
+
?
no
nonchalant
.. .. +
+
?
no
aporehens ive
++
+
?
no
skeptical
. .. ++
f
?
no
blue
++
+
?
no
shocked ..
+
?
no
83

84
\
bold ..
++
+
7
no
pleased ..
++
+
7
no
earnest

++
7
no
tired
++
+
7
no
sluggish

++
+
7
no
kindly .. ..
++
+
7
no
forglvlng
t
-H-
+
7
no
fearful
++
+
7
no
clutched up

+
+
7
no
regretful
++

7
no
lonely

+
+
7
no
egotistic
++
+
7
no
cocky

++

7
no
overjoyed
++
+
7
no
i Ighthearted

++
+
7
no
vigorous .*
+
+
7
no
energetic

-H-

?
no
witty ..
++

7
no
playful ..
a
++
?
no
rebellious
++
+
7
no
suspicious

++
+
7
no
serious .. ..
++

*
1
no
startled
*
++

?
no
warmhearted ..
++
7
no
defiant ..
*
++
+
?
no
Insecure
++

7
no
engaged In thought
++
+
?
no
self-centered
++
+
7
no

APPENDIX D
SAMPLE ADJUSTMENT SCORE

Name
i s
If the statement is true of you, circle "True."
not true of you, circle "False."
1 f the
statement
1.
1 am no one. Nothing seems to h<* me.
True
False
2.
1 am optimlStic.
True
False
3-
1 am a hostile person
True
False
4.
Self-control Is no problem to me.
True
False
5.
1 tend to be on my guard with people who are
somewhat more friendly than 1 had expected.
True
False
6.
1 try not to think about my problems.
True
False
7.
1 have an attractive personality.
True
Fal se
6.
1 am a rational person.
True
False
9.
1 have a horror of failing in anything 1 want
to accomplish.
True
False
10.
1 put on a false front.
True
False
11.
1 feel helpless.
True
Fal se
12.
1 am a good mixer.
True
False
13.
It is difficult to control my aggression.
True
False
14.
1 feel insecure within myself.
True
Fal se
15.
1 really am disturbed.
True
Fal se
16.
All you have to do is just insist with me, and
i give in.
True
Fa) se
17.
1 have to protect myself with excuses, with
rational izlng.
True
False
18.
1 have few values and standards of my own.
True
Fal se
19.
1 take a positive attitude toward myself
True
False
86

20
am ambitlous.
21.
22.
23.
24.
25.
26.
27.
28.
* 29.
30.
3K
32.
33.
34.
35.
36.
37.
38.
39-
40.
41.
42.
43.
44.
I don't trust my emotions.
I have the feeling that I am just not facing facts.
I shrink from facing a crisis or difficulty.
I am assertive.
I am tolerant.
I have initiative.
I express my emotions freely.
I can accept most social values and standards.
I feel hopeiess.
It isoretty tough to be me.
I am contented.
Hy decisions are not my own.
I am a failure.
I make strong demands on myself,
i despise myself.
I doubt my sexual power.
I am afraid of a full-fledged disagreement with
a person.
I am self-reliant.
I am sexually attractive.
I am relaxed and nothing realiy bothers me.
I an likable.
I am worthless.
I often kick myself for the things i do.
i am different from others.
87
True
False
True
Fal se
True
Fai se
True
Fal se
True
False
True
False
True
False
True
False
True
False
True
False
True
Fal se
True
False
True
False
True
Fal se
True
False
True
False
True
False
True
False
True
False
True
False
True
False
True
False
True
Fal se
True
False
True
False

45.
1
am poised.
True
88
Fal se
46.
1
am shy.
True
Fal se
47.
1
usually feel driven.
True
False
48.
1
can't seem to make up my mind one way or the other.
True
False
49.
1
feel emotionally mature.
True
False
50.
1
can usually live comfortably with people.
True
False
5K
1
understand myself.
True
False
52.
1
am afraid of sex.
True
Fal se
53.
1
can usually make up my mind and stick to it.
True
False
54.
1
have a warm emotional relationship with others.
True
False
55.
1
usually like people.
True
Fal se
56.
1
am satisfied with myself.
True
Fal se
57-
1
often feel humiliated.
True
False
58.
1
have a feeling of hopelessness.
True
Fal se
vn
u>

1
am intel1igent.
True
Fal se
60.
1
dislike my own sexuality.
True
False
61.
My hardest battles are with myself.
True
False
62.
1
am a hard worker.
True
False
63.
1
am disorganized.
True
False
64.
1
feel tired out.
True
Fal se
65.
1
am responsible for my troubles.
True
False
66.
1
am unreliable.
True
False
67.
1
am a responsible person.
True
False
68.
My personality Is attractive to the opposite sex.
True
False
69.
1
want to give up trying to cope with the world.
True
False
70.
1
am confused.
True
False

89
71.
1
feel adequate.
True
72.
1
am liked by must people who know me.
True
73.
1
just don't respect myself.
True
74.
1
am impulsive.
True
False
False
False
False

APPENOIX E
SAMPLE OF THE MODIFICATION OF THE EXPECTATION Q-SORT

If you were to have a psychotherapist or counselor to talk with
about your problems, what would you expect him to be like? If the state
ment is true concerning how you would expect the counselor or psycho
therapist to be, put a circle around "true." If the statement is false
concerning how you would expect this counselor or psychotherapist to be,
put a circle around "false."
1.
Calm, easygoing
True
False
2.
Is careful not to let people waste his time..
True
False
3.
Looks for the good points In people
True
False
4.
Likes to have a hand in managing other people's
affairs *
True
Fal se
5.
Is concerned with what's right
True
False
6.
Blunt, straightforward, calls a spade a spade
True
False
7.
Hard to get to know
True
Fal se
8.
is likely to give advice and guidance
True
False
9.
Cares what other people think of him
True
False
10.
Is likely to overestimate a person's abilities ..
True
False
11.
Is Indulgent, forgiving
True
Fal se
12.
I* able to sense other people's feelings
True
Fal se
13.
Is careful not to upset others
True
False
14.
Judges the behavior of others. .. ..
True
Fal se
15.
Expects the individual to shoulder his own re-
sponsibl11 ties
True
False
16.
Is logical, sticks to the facts .. .. ..
True
False
17-
Is likely to keep his irritations or resentments
to himself
True
False
18.
Is gentle, tender
True
Fal se
19.
Self-satisfied
True
False
20.
Never makes people feet uncomfortable ..
True
False
21.
Hard to deceive, does not accept things at face value
True
False
22.
Businesslike
True
False
23.
Is conscientious about duties and responsibilities
True
False
24.
Likes to do a good job ..
True
Fal se
25.
Is not emotional
True
Fal se
26.
Reacts to most people in about the same way..
True
False
27.
Sympathetic
True
False
28.
Tries to discover who's to blame for mistakes made
True
False
29.
Is able to change his opinions easily
True
False
30.
Diplomatic
True
Fal se
31.
Is troubled by the misfortunes of others
True
False
32.
Persuasive
True
False
33.
Well adjusted, gets along well in the world..
True
False
34.
Is quick to qive encouragement and reassurance ..
True
False
35.
Has no trouble getting along with people, makes
friends easily
True
Fal se
36.
Critical, not easily impressed
True
False

APPENDIX F
SAMPLE PERSONAL INFORMATION QUESTIONNAIRE

Name:
Ward:
Age:
Number of years of school:
Type of work before hospital:
Harital status: Number of children:
Are you service connected: %
Number of times in a mental hospital:
How long have you been In the hospital this time?
If you had a chance to get psychotherapy or counseling would you want it:
very much?
a little?
not at all?
93

APPENDIX 6
SAMPLE LETTER SENT TO ALL SUBJECTS WHO SPOKE
INTO THE TAPE RECORDER

November 13* 1964
NAME
Veterans Administration Hospital
Ward D
Gulfport, Mississippi
Dear Mr. :
Several weeks ago you were asked to take a series of tests and
then talk about yourself into a tape recorder. We were studying a new
way of having patients discuss their personal problems.
What you said into the tape recorder and your test performance
has been reviewed. If you have any questions or wish to discuss any
aspect of what you did, please contact me at the Biloxi Veterans Admin-
1strat ion Hospital, ID-21541, extension 283.
In case I do not see you personally, I would like to express my
sincere appreciation for your assistance in this exploratory study. With
out your cooperation, and the cooperation of the other patients who par
ticipated, we would have been unable to evaluate this new approach.
Sincerely yours,
Herbert Goldstein, MA.
Clinical Psychology Trainee
95

APPENDIX H
SUBJECT RAW DATA

(Note: in Column 2. I represents Experimental Group, Congruent
Therapist, Positive Set; 2 represents Experimental Group, Incongruent
Therapist, Positive Set; 3 represents Experimental Group, Congruent Ther
apist, Neutral Set; 4 represents Experimental Group, Incongruent Thera
pist, Neutral Set; 5 represents Control Group, Congruent Therapist, Posi
tive Set; 6 represents Control Group, Incongruent Therapist, Positive
Set; 7 represents Control Group, Congruent Therapist, Neutral Set; 8 rep
resents Control Group, Incongruent Therapist, Neutral Set. Columns 7
through 20 represent changes In the dependent variables from pre- to
posttreatment testing. A change score of 0 is represented by 09. Nega
tive change scores are represented by increasing numbers and positive
change scores are represented by decreasing numbers. That is, a change
score of -1 is represented by 10, a change score of -2 is represented by
II, etc. A change score of +1 is represented by 08, a change score of +2
is represented by 07. etc. In Column 21. 01 represents Tittle, or no
speech during treatment, and 02 represents a moderate amount or constant
speech during treatment. In Column 22. 01 represents the greatest stated
desire for psychotherapy, 02 represents a moderate stated desire for psy
chotherapy and 03 represents a stated lack of desire for psychotherapy.
In Column 23. 01 represents a nonpsychotic diagnosis and 02 represents a
psychotic diagnosis.)
97

98
Subj Treat Adj So Amt of Des
No
Group
i .a.
1 ntro
Neuro
Aqe
AACL
Score
Agg.r
Cone
Deact
Aff
Anx
Dep r
Eqot
Pleas
ActI v
Noneh
Skept
Start
Speech
Psycho
Di aq
1
1
107
28
04
33
15
09
09
09
09
07
12
09
08
01
10
07
08
09
02
03
02
2
1
107
24
02
27
10
07
12
1 1
09
09
09
09
09
08
09
09
09
09
02
03
02
3
1
076
28
30
30
13
06
08
08
13
07
07
10
09
09
08
05
09
14
01
01
02
4
1
099
13
46
43
09
14
11
08
12
08
07
13
11
08
13
12
13
1 1
01
01
02
5
1
113
28
34
48
12
07
09
12
09
09
09
09
10
09
09
06
09
09
t
01
02
01
6
1
088
20
48
41
16
10
11
06
1 1
04
10
12
15
06
10
1 1
11
09
01
01
01
7
1
099
36
32
26
13
09
10
08
10
06
13
1 1
08
07
07
08
10
09
1
01
02
02
8
1
121
29
36
38
1 1
09
10
09
09
09
08
08
09
09
1 1
10
07
09
!
01
01
01
9
1
093
12
40
48
20
1 1
10
09
10
02
17
09
08
03
08
10
10
10
01
01
01
10
1
102
29
25
51
12
09
12
09
12
07
10
11
11
10
08
12
07
08
01
01
02
11
1
088
28
30
37
1 1
07
12
05
06
08
10
14
08
07
08
10
06
09
02
03
02
12
1
078
34
08
18
15
06
12
12
12
12
14
09
09
02
06
06
11
09
02
01
02
13
2
089
24
12
37
04
05
05
06
10
05
07
08
06
09
06
09
06
06
02
02
02
14
2
099
14
40
43
12
13
1 1
09
09
07
10
11
09
03
10
11
14
09
01
01
01
15
2
088
36
26
29
06
18
09
08
12
13
08
09
06
09
09
11
09
09
02
01
01
16
2
119
18
44
40
12
08
12
09
1 1
07
12
09
09
08
09
09
09
09
02
02
01
17
2
102
24
22
22
12
06
09
12
05
11
16
16
10
10
12
07
14
09
01
02
02
18
2
099
14
06
32
09
09
09
08
07
1 1
07
12
11
12
09
09
07
09
01
02
02
19
2
106
10
44
45
05
13
06
09
06
09
09
09
12
09
15
12
09
09
02
03
01
20
2
109
22
30
43
05
04
05
07
07
06
08
08
06
06
07
07
08
08
02
02
02
21
2
103
12
38
43
06
09
08
12
09
12
05
10
11
11
09
09
11
09
02
01
02

99
Subj
No
T reat
Group
i .a.
1 ntro
Neuro
Aqe
AACL
Adj
Score
Aqq.r
Cone
Deact
So
Aff
Anx
Depr
Eqot
Pleas
Acti v
Nonch
Skept
Start
Amt of
Speech
Des
Psycho
Di aq
22
2
095
26
24
22
08
10
14
13
1 1
18
09
09
10
09
08
09
06
07
01
02
02
23
2
112
18
18
40
09
07
07
08
10
09
10
1 1
16
07
09
06
09
10
01
01
02
24
2
089
18
48
44
00
15
10
12
09
15
05
06
16
15
09
15
09
06
01
01
01
25
3
098
30
42
46
13
09
12
09
13
06
12
12
07
07
07
07
1 1
09
02
02
02
26
3
104
40
10
45
10
10
08
09
09
10
09
09
08
16
09
09
09
09
01
01
02
27
3
077
18
36
45
07
07
09
09
09
13
09
09
07
08
12
09
10
09
02
02
02
28
3
090
21
20
45
09
09
09
09
09
09
09
09
09
12
09
12
09
09
01
01
02
29
3
104
15
28
35
13
1 1
06
09
09
03
09
13
10
07
09
09
10
09
01
03
02
30
3
109
13
38
41
07
10
09
09
09
06
09
09
09
11
09
09
09
09
01
01
01
31
3
124
19
42
58
10
1 1
08
09
08
08
1 1
10
08
06
07
09
09
08
01
01
01
32
3
093
24
12
41
12
09
09
09
14
08
1 1
1 1
11
11
12
09
09
09
02
02
02
33
3
111
30
38
48
13
07
09
09
12
09
1 1
09
09
05
12
09
06
09
01
03
01
34
3
077
18
38
31
19
08
10
07
09
06
13
10
07
06
07
08
08
10
01
01
02
35
3
093
14
42
41
14
06
13
09
09
11
10
11
07
10
09
10
10
08
01
01
01
36
3
108
11
44
20
14
16
09
09
11
09
12
07
10
09
09
10
13
09
01
02
01
37
4
102
26
11
40
12
10
06
12
09
09
12
09
09
09
12
06
09
09
01
02
02
38
4
098
19
34
45
11
09
11
09
10
07
09
12
09
06
09
10
10
09
01
02
02
39
4
093
14
08
37
05
03
04
15
08
14
12
1 1
09
12
09
10
12
07
02
02
02
40
4
097
12
28
35
12
05
11
06
09
09
06
12
09
06
09
09
09
06
01
03
01
41
4
108
23
13
40
11
07
11
10
09
07
09
09
09
09
10
09
09
09
02
02
01
42
4
101
38
34
40
08
11
09
10
09
08
1 1
08
10
10
09
10
09
09
02
01
02
43
4
106
40
00
30
10
05
09
09
09
06
1 1
09
09
11
09
08
09
09
02
03
02
44
4
087
10
43
42
10
10
03
11
07
12
06
07
09
09
09
07
07
07
01
01
01

100
Subj Treat Adj So Amt of Des
No
Group
1 .4.
1 ntro
Neuro
Aqe
AACL
Score
Aggr
Conc
Deact
Aff
Anx
Depr
Eqot
Pleas
Act i v
Nonch
Skept
Start
Speech
Psycho
D! aq
45
4
088
24
26
43
09
10
09
10
07
09
09
1 1
10
10
09
08
08
08
02
01
01
46
4
102
26
30
24
10
08
10
10
09
06
10
07
09
11
09
12
10
09
02
02
02
47
4
116
17
32
49
09
12
07
10
10
08
1 1
10
05
09
09
08
07
10
02
02
01
48
4
111
14
16
36
09
12
09
10
10
09
10
13
09
10
1 1
11
12
09
01
03
02
49
5
108
26
42
30
06
03
11
11
09
08
09
07
07
10
09
10
09
09
50
5
092
10
44
44
09
09
09
09
09
08
09
08
09
10
09
12
12
11
51
5
107
29
25
55
06
10
10
10
17
10
06
09
1 1
11
07
13
13
09
52
5
079
28
36
42
12
08
09
08
1 1
09
07
11
09
10
08
09
09
11
53
5
086
36
10
44
09
09
08
11
05
14
11
09
08
06
12
09
10
09
54
5
083
23
35
37
10
10
08
07
07
11
09
06
16
09
08
07
10
10
55
5
118
22
35
31
12
06
07
09
10
11
09
12
11
08
09
09
09
09
56
5
119
30
24
31
08
10
11
17
09
09
02
09
09
10
09
1 1
04
09
57
5
101
25
17
29
10
01
07
10
1 1
09
10
09
09
03
09
09
09
09
58
5
129
13
48
39
07
09
06
07
09
09
09
10
09
02
13
09
10
10
59
5
098
28
11
43
11
04
06
12
04
14
08
09
07
07
12
07
09
08
60
5
092
18
42
28
15
03
04
09
06
03
12
08
02
09
09
09
09
10
61
6
093
22
46
43
10
00
10
09
09
13
07
11
10
09
09
09
09
09
62
6
091
26
32
44
05
09
07
1 1
09
13
07
13
09
14
10
09
10
10
63
6
098
20
35
34
07
00
09
09
08
07
12
09
10
10
11
09
09
09
64
6
102
26
38
41
13
00
09
1 1
14
07
14
11
06
08
07
10
09
08
65
6
073
26
18
40
09
09
06
08
09
09
09
09
09
08
09
07
09
09
66
6
081
12
42
41
08
05
14
10
08
05
06
09
08
09
09
14
10
08
67
6
087
28
10
32
10
04
11
11
09
11
09
09
07
11
10
08
09
06

101
Subj Treat Adj So Amt of Des
No
Group
i .a.
1 ntro
Neuro
Aqe
AACL
Score
Aqqr.
Cone
Deact
Aff
Anx
Dep r
Eqot
Pleas
Acti v
Nonch
Skept
Start Speech Psycho Diaq
68
6
100
22
39
35
12
16
10
13
08
08
10
13
08
06
09
06
07
09
69
6
098
18
18
47
08
11
09
09
09
09
10
09
09
05
08
11
09
09
70
6
087
32
06
37
07
12
09
08
06
09
05
06
07
08
03
09
09
09
71
6
103
18
22
44
09
1 1
09
10
08
10
08
09
09
08
09
09
08
08
72
6
117
28
40
30
10
05
09
10
05
09
11
09
09
10
1 1
11
09
09
73
7
097
10
44
45
12
06
12
08
09
09
08
10
08
06
11
12
10
09
74
7
100
13
22
38
10
07
10
1 1
15
10
07
11
09
11
09
06
10
09
75
7
109
29
15
35
04
09
08
09
09
10
09
07
09
11
11
13
08
09
76
7
081
21
16
38
09
10
09
10
10
10
12
11
09
10
11
13
11
06
77
7
108
36
02
38
09
10
12
09
09
09
09
09
09
06
09
09
09
09
78
7
093
24
48
44
10
04
08
08
10
10
08
13
09
10
06
11
10
09
79
7
116
19
42
43
10
10
09
09
08
09
07
06
09
12
09
11
09
09
80
7
090
32
26
43
10
12
07
08
12
10
10
06
07
09
09
11
06
09
81
7
103
17
42
40
1 1
08
09
11
1 1
09
11
17
09
09
09
09
10
09
82
7
076
34
08
42
09
10
09
10
08
09
10
08
10
07
08
07
08
10
83
7
103
1 1
43
35
09
12
09
09
09
11
09
10
09
11
09
10
09
09
84
7
086
14
42
41
07
06
06
06
07
12
06
10
06
11
11
08
10
09
85
8
086
16
08
32
08
01
08
09
05
09
09
09
10
09
04
09
07
09
86
8
110
24
06
37
05
13
09
09
09
09
07
09
09
12
12
09
09
09
87
8
080
27
06
44
11
12
12
1 1
11
12
12
09
06
13
10
04
13
11
88
8
095
27
22
41
09
06
09
10
09
1 1
09
09
09
1 1
1 1
09
09
09
89
8
105
24
48
40
08
13
09
09
09
15
09
10
12
13
12
09
09
06
90
8
089
23
42
41
04
08
09
14
03
06
00
06
06
03
06
09
04
06

102
Subj
No
T reat
Group
i .a.
1 ntro
Neuro
Aqe
AACL
Adj
Score
Aggr
Cone
Deact
So
Aff
Anx
Depr
Eqot
Pleas
Acti v
Nonch
Skept
Amt of
Start Speech
Des
Psycho Diaq
91
8
089
23
02
44
10
10
07
07
09
07
09
09
09
09
07
09
09
09
92
8
124
28
44
41
10
04
07
09
06
09
07
08
09
10
10
09
11
09
93
8
115
18
19
53
08
07
09
09
09
09
09
09
09
11
07
07
09
09
94
8
095
23
22
46
11
10
09
06
1 1
11
11
1 1
13
13
07
09
10
05
95
8
101
23
28
31
10
09
07
11
08
09
08
12
10
10
09
09
10
09
96
8
089
24
24
32
09
15
09
09
09
15
06
09
09
12
09
09
03
09

BIBLIOGRAPHY
Abramson, H. A., Jarvlk, M. I,, Levine, A., Kaufman, H. R. and Hirsch,
M. U. Lysergic acid diethylamide (LSD-25)* XV. The effects
produced by substitution of a tap water placebo. J. psychoi..
1955, 4&, 367-383*
Apfelbaum, B. Dimensions of transference in psychotherapy. Berkeley,
Calif.: Univ. of California Press, 1958.
Boryatta, E. F. The new principle of psychotherapy. J. ciin. psvchoi..
1959, , 330-339*
Dymond, R. F. An adjustment score for Q-sorts. J. of consult. Dsvchol..
1953, U. 339-342.
Dymond, R. Adjustment changes over therapy from self-sorts, In Rogers,
C. R. and Dymond, R. (eds.) Psychotherapy and personality
change. Chicago, Illinois: Unlv. of Chicago Press, 1954, 76-85*
Eysenck, H. J. The effects of psychotherapy: an evaluation. J. con
sult. psvchoi.. 1952, 319-324.
Eysenck, H. J. Manual of the Haudsiey personality inventory. London,
England: Univ. of London Press, 1959*
Eysenck, H. J. Handbook of abnormal psychology. New York, New York:
Basic Books, 1961.
Fenichel, 0. The psychoanalytic theory of neurosis* New York, New York:
Norton and Company, 1945*
Festinger, L. A theory of punitive dissonance. Evanston, III.: Row,
Peterson, 1957-
Festinger, L. and Brame!, D. The reactions of humans to cognitive dis
sonance, in Bachrach, A. J. (ed.) Experimental foundations of
clinI cal psychology. New York, New York: Basic Books, 1962.
f'ischer, H. K. and Dlin, B. M. The dynamics of placebo therapy: a
clinical study. Am, J. med. sci.. 1956, 232. 504-512.
Foulds, G. Clinical research In psychiatry. J. ment. _scj_., 1958, Jj&,
259-265.
103

104
Frank, J. D. Son affects of expectancy and influence in psychotherapy,
in Hasserraan, J. H. and Horeno, J. I. (eds.) Progress in psycho
therapy. New York, New York: Grue and Stratton, 1958.
Frank, J. D. Persuasion and healing. Baltimore, Md.: Johns Hopkins
Press, 1961.
Glasser, E. H. and Whittlow, G. C. Evidence for a nonspecific mechanism
of habituation. J. physiol.. 1953. 122. 43-44.
Glasser, E. H. and Whittlow, G. C. Experimental errors in clinical
trials. Clin, scl.. 1954, H, Number 2, 199-210.
Glideman, L. H., Gantt, W. H. and Teitelbaum, H. A. Some implications of
conditional reflex studies for placebo research. Am. J. psychiat..
1957. JLL. 1103-1107.
i : 1 )
Gourlay, N. Covariance analysis and its applications in psychological
research. Brit- J, 1953. &. 25-33.
Hagans. J. A., Doering, C. R., Ashley, F. W. and Wolf, S. The thera
peutic experiment. J. lab, clin, med.. 1957. itii, 282-285.
Hofling, C. K. The place of placebos in medical practice. Gen, pract..
1955, iL 103-118.
Hongingfield, G. The placebo effecta review of nonspecific factors in
treatment. Cooperative studies in psychiatry. Nov., 1963.
Jaco, E. G. Patients, physicians and illnesses. New York, New York:
Free Press, 1958.
Joyce, C. R. Consistent differences in Individual reactions to drugs and
dummies. Brit. J. of Pharmacol.. 1959. 14 512-521.
Kellogg, C. £., Horton, N. W., Lindner, R. H. and Gurvitz, H. Revi sed
beta examination: manual. New York, New York: Psychological
Corporation, 1946.
Kornetsky, C. and Humphries, 0. Relationship between effects of a number
of centrally acting drugs and personality. A.H.A. arch, neural.
psychiat.. 1957, 22. 325-327.
Kurland, A. A. The drug placeboIts psychodynamic and conditional re
flex action. Behav. sci.. 1957. 2, 101-110.
Kurland, A. A. The placebo, in Masserman, J. H. and Horeno, J. i. (eds.)
Progress in psychotherapy. New York, New York: Grue and
Stratton, 1958.

105
Lasagna, L., Hosteler, F., Von Felsinger, J. H. and Beecher, H. K. A
study of the placebo response. Am. J. of med.. 1954, jj, 770-779
Laverty, S. G. Sodium amytal and extroversion. J. of neurol. neurosurg.
and aaifitLtt.* 1958, L 50-54.
Leary, T. and Gill, H, The dimensions and a measure of the process of
psychotherapy: a system for the analysis of the content of
clinical evaluations and patient-therapist verbalizations, in
Rubinstein, E. A. and Parloff, H. B. (eds.) Research In psycho
therapy. Washington, 0. C.: Am Psychol. Assn., 1959. 62-95.
Leiberman, R. An analysis of the placebo phenomenon. J. chron disea..
1964, JLi, 761-785.
Lindquist, E. F. Design and analysis of experiments in psychology apA
education. Bcston, Mass.: Houghton Mifflin, 1953.
Locke, J. Locke's works. London, England: Davidson, 1823, 275.
Lyerly, S. B., Ross, S., Krugman, A. D. and Clyde, D. J. Drugs and pla
cebos: the effects of instructions upon performance and mood
under amphetamine sulphate aid chloral hydrate. Psvchol. rep..
1962, 383-392.
McNemar, Q. Psychological statistics. New York, New York: Wiley and
Sons, 1955.
Maccoby, E. E., Newcomb, T. H. and Hartley, E. L. (eds.). Readings in
social psychology. New York, New York: Holt and Company, 1958.
Martin, B., Lundy, R. M. and Lewin, M. H. Verbal and G.S.R. responses in
experimental Interviews as a function of three degrees of "ther
apist" communication. J. of ab. and soc. psycho!.. I960, bO.
234-240.
Newcomb, T. M. Social psychology. New York, New York: Holt and Company,
1958.
Nowlls.V. Description and analysis of mood. N. Y. acad. of sci. annals.
1956, 6*, 345-355.
Nowlis, V. Some studies of the influence of films on mood and attitude.
Technical report number 7. Research project number NR 171 34Z,
Jan., i960.
Nowlis, V. Methods for studying mood changes produced by drugs. Rev.
asvcho 1, app_l.. 1961, jJL, 373-386.
Ray, W. S. An introduction to experimental design. New York, New York:
Macro!11an Company, 1960.

106
Rogers, C. R. Client centered therapy. Boston, Hass.: Houghton Hlfflin,
1954.
Rogers, C. R.
cago,
and Dymond, R. Psychotherapy and personality change.
Illinois: Univ. of Chicago Press, 1954a.
Chi -
Rogers, C. R. On becoming a person. Boston, Hass.: Houghton Hifflin,
1961.
Rosenthal, 0. and Frank, J. D. Psychopathology and the placebo effect.
Psychol, bull.. 1956, I, 294-302.
Shapiro, A. K. Factors contributing to the placebo effect. Aro. J. of
psvchoth.. 1964, JjJ, supplement I, 73-88.
Shoben, E. H. Some observations on psychotherapy and the learning proc
ess, In Howrer, 0. H. (ed.) Psychotherapy, theory and research.
New York, New York: Ronald Press, 1953.
Siegel, S. Nonparametrlc statistics. New York, New York: HcGraw-Hill,
1956.
Slack, C. W. Experimenter-subject psychotherapy: a new method of in
troducing intensive office treatment for unreachable cases.
Kent, hyg.. i960, 238-256.
Strupp, H. H. Psychotherapy. Ann, rev, psvchol.. 1962, 445-478.
Tibbetts, R. W. and Hawkings, J. R. The piacebo response. J. of ment.
- sci.. 1956, 60-66.
Uhlenruth, E. H., Canter, A., Neustadt, J. 0. and Payson, H. E. The
symptomatic relief of anxiety with meprobanate, phenobarbital
and placebo. Am, J. psvchiat.. 1959, 115. 905-910.
Whltehorn, J. C. Psychiatric implications of the "placebo effect." Am.
J. Psychiat.. 1958, 114, 662-664.
Wolf, S., Doering, C. R., Clark, H. L. and Hagans, J. A. Chance distri
bution and the piacebo "reactor." J. lab, clin, reed.. 1957 49.
837-841.
Wolf, S. The pharmacology of placebos. Pharmacol, rev.. 1959, i 1
698-702.
luckerman, H. The development of an affect adjective check list for the
measurement of anxiety. J. consult, psvchol.. I960, 4, 457-462.
Zuckerman, H. and Blase, D. V. Replication and further data on the affect
adjective check list measure of anxiety. J. consult, psvchol..
1962, 6, 291.

BIOGRAPHICAL NOTE
Herbert Goldstein was born on July 20, 1936 in New York City, New
York. He graduated from Allentown High School, Allentown, Pennsylvania,
in June, 1954. He received the Bachelor of Arts degree from the Univer
sity of Florida In June, 1958
He served In the United States Army from September, 1958 to Sep
tember, I960.
He entered the Graduate School of the University of Florida in
September, I960. In December, 1962 he received the Master of Arts de
gree with a major in psychology and a minor in educational foundations.
While in graduate school in the University of Florida he held graduate
assistantships in the Psychology Department, In the Reading Laboratory
and Clinic and a cl Ini clanship at the University Counseling Center.
He served as Psychology Trainee In the Gulfport-BIloxl, Missis
sippi Veterans Administration Hospitals from January, 1964 until Decem
ber, 1964. He is currently a USVRA fellow completing his predoctoral In
ternship at the J. HI 11 is Miller Health Center.

This dissertation was prepared under the direction of the chair
man of the candidate's supervisory committee and has been approved by all
members of that committee. It was submitted to the Dean of the College
of Arts and Sciences and to the Graduate Council and was approved as
partial fulfillment of the requirements for the degree of Doctor of
Philosophy.
April 24, 1965
Dean, Graduate School
Supervisory Committee:
d. fiJcL
Chai
,/A



57
decrease reported anxiety after Experimental Group treatment more readily
than do older subjects. This Is especially so for the Positive Set, in-
congruent Therapist and Neutra) Set, Congruent Therapist conditions.
Eysenck defined his Neurotic)sm Scale as a measure extending
along a highly neurotic to normal continuum, in the current study, the
sample contains no "normals," therefore, serious questions may be raised
concerning the meaningfulness of the neurotic!sm score. For example,
does a low score on neuroticlsm in this study mean that a patient is psy
chotic? A biserial correlation (McNemar, 1955) between diagnosis (psy
chotic vs. nonpsychotic) and scores on the Neuroticlsm Scale results In a
correlation of .57 (p ^ .001). This suggests that subjects with nonpsy
chotic diagnoses tend to score high on the Neuroticlsm Scale whtie sub
jects with psychotic diagnoses tend to score lower. However, this corre
lation accounts for only 32.49 per cent of the variance, and it is with
extreme caution that interpretation of the neuroticism score data be made.
As noted in Table 25, an apparent relationship exists between
neuroticism scores and the MACL mood of Nonchalance. With the exception
of the Experimental Group, Neutral Set, Congruent Therapist subjects, it
appears that as neuroticism increases, nonchalance decreases after treat
ment. Another interesting relationship indicates that subjects scoring
high on the Neuroticism Scale who are In the Positive Set, Congruent
Therapist, Control Group decrease their HACL mood of Startle after treat
ment; while subjects scoring high on the Neuroticlsm Scale who are in the
Neutral Set, Incongruent Therapist, Control Group increase this Startle
mood after treatment.


24
In addition to the five tests, each subject was also asked to
complete a questionnaire asking for personal Information (see Appendix
F). Each subject also had taken the Beta Intelligence Test (Kellogg
et ai.. 1946).
Procedure
Subjects were chosen from the ward by a secretary who was naive
concerning the mechanics and goals of the study. She was informally
questioned after the completion of the study and no bias In the subjects
she chose, and her order of choosing them, was noted.
The subjects were seen two at a time. Once the subjects were
seated in the testing room they were given instructions which drew their
Interest to the tasks about to be presented but did not disclose the na
ture of the research nor the fact that it was research (see Appendix A).
They were given the following tasks in the order noted:
1. Personal information sheet
2. Mauds ley Personality Inventory
3. Anxiety Adjective Check List
4. Mood Adjective Check List
5. Expectation questionnaire
6. The Adjustment Score
These tasks were relatively short, took no longer than 40 minutes to com
plete and subjects did not appear to become fatigued. After completing
the six questionnaires the pair of subjects was told that this was an op-
portunlty to speak with a psychiatrist, that he was unable to get to the
hospital and that he requested they soeek Into a taoe recorder, the tape
of which would be sent to him by registered mall (see Appendix A for ex
act wording).


dependent variables. The Positive Set Group increased significantly less
on the HACL mood of Startle after treatment than did the Neutral Set
Group which increased more (p < .05) (see Table 17). The Positive Set
Group also reported a significantly higher increase on the HACL mood of
Pleasantness (p ^ .05) than the Neutral Set Group which decreased re*
ported pleasantness after treatment (see Table 13). Only two of the 13
dependent variables which were expected to change significantly distin
guished between the positive and neutral set groups. Of particular note
is the fact that no significant changes were demonstrated on either the
AACL or the HACL mood of Anxiety. However, as predicted, no change oc
curred on the Adjustment Score and no dependent variables changed sig
nificantly in the unpredicted direction. Thus, the hypothesis appears
supported.
In summary, those dependent variables which demonstrated signifi
cant differences between one or more of the three independent variables
were: the AACL and the HACL moods of Aggression, Social Affection, De
activation, Anxiety, Pleasantness and Startle. Those dependent variables
which did not significantly distinguish between any of the three inde
pendent variables were Dymond's Adjustment Score (as predicted) and the
HACL moods of Concentration, Depression, Egotism, Skepticism, Nonchalance
and Activation.
As noted earlier, after Nowlis factor analyzed the HACL he di
vided his 12 mood factors into three groups of four moods each: those
moods most clearly identified, those less clearly identified and those
least clarly identified (see Table 4). It is noteworthy that of the four
moods in the group most clearly identified by Nowlis, three were found to


Subjects
The subjects for the study were 96 neuropsychiatric hospital pa
tients in the Gulfport, Mississippi, Veterans Administration Hospital.
They are all in-patient residents of the Rehabilitation and Placement
Service Ward which houses oniy those patients judged capable of returning
to the coomunlty. The ward offers a vocational testing service and is
professionally staffed by a psychiatrist, two counseling psychologists
and a number of nursing personnel.
The 96 subjects were drawn from the ward during 48 separate ses
sions extending from August 29, 1964, to October 19, 1964. The sample
consists of white adult males. There is a wide range of prehospitaliza
tion vocations. None were In psychotherapy when seen as subjects for the
study. However, it was known that some of these patients had been in
group or individual psychotherapy earlier in their hospitalization. The
identification of former psychotherapy oatients was not made, but there
is no reason to suspect a nonrandom assignment of these patients to the
different experimental treatment conditions. Similarly, although it was
known that some of the sanp.e were receiving drugs for their psychiatric
conditions, these subjects were not identified and their random assign
ment to the different experimental treatment conditions was expected.
The mean age of the sample was 38.76; range 18 to 58, standard
deviation 7*41. Below is a summary table of an analysis of variance
(Lindquist, 1953) demonstrating that subject ages are randomly distrib
uted across all treatment conditions.
The mean I .-(£. of the sample, as measured by the Beta Intelligence
Test (Kel logg et al.. 1946) was 98.54, which is within the normal range.


71
Score? That is, would extended placebo psychotherapy have produced
changes In psychological adjustment similar to changes which occur after
actual psychotherapy? (See Frank, 1961, and Rogers and Dymond, 195**, for
evidence of changes in general level of psychological adjustment after
psychotherapy.) it is difficult to speculate on precisely what may have
happened to the subject's general level of psychological adjustment had
the placebo treatment continued, although researchers such as Glasser and
Whitt low (1953> 1954) submit that the effects of placebo treatment are
not permanent and the gains from continued placebo treatment become less
and less.
What might have happened to moods and anxiety had the placebo
psychotherapy continued? Martin et ai (I960), In a study concerned with
levels of therapist comaunications (discussed more fully in the Introduc
tion), had subjects speak into a tape recorder for five 30-minute weekly
sessions as if they were speaking to a psychotherapist. These authors
demonstrate that although some initial positive therapeutic effect was
indicated (lowering of anxiety), the procedure lost its positive thera
peutic effectiveness as the number of sessions with the tape recorder In
creased. Glasser and Whlttlow (1953, 1954), in placebo drug studies,
found that if subjects were given placebos over a period of time and re
quired to answer questionnaires identifying the effects of these placebos,
the positive placebo effect was demonstrated Initially, but the effect
was less upon each successive evaluation. Findings such as these suggest
that the placebo procedure used in the current study, if extended for
more than one session, might not have continued to have positive thera
peutic effect on mood and anxiety.


30
TABLE 5
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE ZUCKERMAN
ADJECTIVE CHECK LIST
Source
ss
df
V
F
A (experimental vs. control)
44.42
1
44.42
6.29*
B (congruent vs incongruent)
98.55
1
98.55
13.96***
C (positive vs. neutral)
.00
l
.00
.00
AB
69.98
1
69.98
9.91**
AC
6.63
l
6.63
94
BC
17.67
1
17.67
2.50
ABC
17.67
1
17.67
2.50
error
615.07
SI
7.06
Total
869.96
94
*** s significant K .001.
** a significant < .01.
* significant <, .05.
TABLE 5A
PRETREATMENT MEANS (X,). POSTTREATMENT MEANS (X,) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE
ZUCKERMAN ADJECTIVE CHECK LIST
Experimental Group
Control Group

Congruent
Therapist
Incongruent
Therapist
Congruent
Therapist
1ncong ruent
Therapist
Positive
Set
X, a 11.08
X2 5 7.00
D s -4.08
X| a 8.92
X2 10.59
D a +1.67
X, a 10.33
x2 JU1
D s -.58
X. a 8.41
X, 8.41
D a .00
Neutral
Set
X, s 10.08
*2 s
D -2.75
X, 9.75
X2 2^08
D a -.67
L 11.33
X2 11.16
D = -.17
X, a 9.41
x2 9.83
D a +.42


BIBLIOGRAPHY
Abramson, H. A., Jarvlk, M. I,, Levine, A., Kaufman, H. R. and Hirsch,
M. U. Lysergic acid diethylamide (LSD-25)* XV. The effects
produced by substitution of a tap water placebo. J. psychoi..
1955, 4&, 367-383*
Apfelbaum, B. Dimensions of transference in psychotherapy. Berkeley,
Calif.: Univ. of California Press, 1958.
Boryatta, E. F. The new principle of psychotherapy. J. ciin. psvchoi..
1959, , 330-339*
Dymond, R. F. An adjustment score for Q-sorts. J. of consult. Dsvchol..
1953, U. 339-342.
Dymond, R. Adjustment changes over therapy from self-sorts, In Rogers,
C. R. and Dymond, R. (eds.) Psychotherapy and personality
change. Chicago, Illinois: Unlv. of Chicago Press, 1954, 76-85*
Eysenck, H. J. The effects of psychotherapy: an evaluation. J. con
sult. psvchoi.. 1952, 319-324.
Eysenck, H. J. Manual of the Haudsiey personality inventory. London,
England: Univ. of London Press, 1959*
Eysenck, H. J. Handbook of abnormal psychology. New York, New York:
Basic Books, 1961.
Fenichel, 0. The psychoanalytic theory of neurosis* New York, New York:
Norton and Company, 1945*
Festinger, L. A theory of punitive dissonance. Evanston, III.: Row,
Peterson, 1957-
Festinger, L. and Brame!, D. The reactions of humans to cognitive dis
sonance, in Bachrach, A. J. (ed.) Experimental foundations of
clinI cal psychology. New York, New York: Basic Books, 1962.
f'ischer, H. K. and Dlin, B. M. The dynamics of placebo therapy: a
clinical study. Am, J. med. sci.. 1956, 232. 504-512.
Foulds, G. Clinical research In psychiatry. J. ment. _scj_., 1958, Jj&,
259-265.
103


42
TABLE 17
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL STARTLE MOOD
Source
ss
df
V
F
A (experimental vs. control)
.03
1
.03
.06
B (congruent vs. incongruent)
13.47
1
13.47
26.3k***
C (positive vs. neutral)
2.67
1
2.67
5.34*
AB
.04
1
.04
.08
AC
37
1
.37
.74
BC
1.47
1
1.47
2.94
ABC
.03
1
.03
.06
error
43.70
81
.50
Total
61.78
94
***Slgnlflcant ^ .001.
**$Ignificant < .05*
TABLE I7A
PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X2) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL STARTLE MOOD
Exjerimental Group
Control
Group
Congruent
Therapist
1ncongruent
Therapist
Congruent
Therapist
1ncongruent
Therapist
Positive
Set
X, 1.42
x2 1.00
0 -te
x, 1.17
x2 1-83
D +.66
x, 1.50
x2 1.00
D -.50
X, 1.25
X? 1.67
D
Neutral
Set
x, = .92
x2 = 1.00
D +.08
x, .92
X2 L20
D +.58
X, 1.00
x2 1.17
D +.17
X, 1.25
x2 L21
0 +.07


BIOGRAPHICAL NOTE
Herbert Goldstein was born on July 20, 1936 in New York City, New
York. He graduated from Allentown High School, Allentown, Pennsylvania,
in June, 1954. He received the Bachelor of Arts degree from the Univer
sity of Florida In June, 1958
He served In the United States Army from September, 1958 to Sep
tember, I960.
He entered the Graduate School of the University of Florida in
September, I960. In December, 1962 he received the Master of Arts de
gree with a major in psychology and a minor in educational foundations.
While in graduate school in the University of Florida he held graduate
assistantships in the Psychology Department, In the Reading Laboratory
and Clinic and a cl Ini clanship at the University Counseling Center.
He served as Psychology Trainee In the Gulfport-BIloxl, Missis
sippi Veterans Administration Hospitals from January, 1964 until Decem
ber, 1964. He is currently a USVRA fellow completing his predoctoral In
ternship at the J. HI 11 is Miller Health Center.


10
changes appropriate to the chloral hydrate. Abramson et al. (¡955) admin
istered tap water to a group of subjects who believed the liquid to be
LSD-25. A percentage of these subjects responded to the tap water with
psychomlmetlc behavior as If the water actually were LSD-25-
Frank (I9l) suggests that a patient's faith In a set of treat
ment operations is often ail that is necessary to produce positive change
and this is so even when these treatment operations are scientifically
absurd. Wolf (1959) states that a placebo is effective when a patient
has a "conviction that a certain effect will follow.
In the present study, an attempt will be made to manipulate ex
pectations concerning the vaue of the "placebo psychotherapy and the
amount of profit to be anticipated from the treatment. One group will
receive a positive set concerning that which can be gained from the
treatment and the other will receive a more neutral set (see Appendix A).
To summarize, the hypotheses under investigation in this study
are:
i. Those subjects who receive the placebo psycho
therapy (speak into the tape recorder) will show
more marked change in a "therapeutic direction
on measures of anxiety and mood than those sub
jects who simply wait a comparable length of time.
However, neither group will show a change in gen
eral level of psychological adjustment in this
single treatment.
II. The subjects who believe that their therapist is
an individual congruent with their expectations
will show more marked change in a "therapeutic
"direction on post measures of anxiety and mood
than those subjects whose expectations are disso
nant with the therapist's description. Neither
group will show a change in general level of psy
chological adjustment after this single treatment.


52
to attribute a high degree of conclusiveness to results showing some con
sistency in the absence of significance. However there are several ob
servations which may be made and several trends worth Identifying. It is
important to note, also, that the complexity of both the data and the in
dependent variables tends to reduce the meaningfulness of Interpretation.
It will be noted In Table 22 that there Is a tendency for subjects,
given a positive set and a congruent therapist and who score lower meas
ured levels of intelligence, to decrease reported anxiety on the AACL
after treatment. Conversely, this suggests that subjects with higher
measured intelligence, when given a positive set and a congruent thera
pist, increase reported anxiety on the AACL after treatment. Similarly,
subjects with lower measured intelligence, who are given a positive set
and an incongruent therapist, seem to increase their scores on the AACL
after treatment more than subjects with higher measured intelligence. It
will be noted further, with the exception of the Experimental Group,
Neutral Set, Congruent Therapist subjects (which showed an opposite
trend), iess bright subjects tend to increase the MACL mood of Activation
after treatment. From these data, the conclusion may be tentatively ad
vanced that people with lower intelligence seem to change in a thera
peutic direction after certain types of placebo treatment. On the other
hand, with the exception of the Experimental Group, Positive Set, Congru
ent Therapist, there appears to be a trend which suggests that subjects
with higher measured intelligence tend to Increase more on the MACL mood
of Social Affection after treatment than subjects who score lower on In
telligence. This finding suggests a conclusion which is not in line with
that indicated above from the analysis of the AACL and the MACL mood of


65
analysis does not assess content of speech or nature of affect which are
certainly extremely significant aspects of catharsis. In addition, this
analysis evaluates only the linear relationships between amount of speech
and degree of change on the dependent variables. The analysis does not
deal with nonlinear relationships, and It is conceivable that such rela
tionships might exist.
Shapiro (1964) reports that expressed desire for treatment is an
other important concept which reportedly is basic in eliciting the pla
cebo effect. That is, those subjects who are more highly motivated for
treatment would be more responsive to placebo treatment. However, In
this study, motivation of the subjects, as measured by their statements
concerning their level of desire for psychotherapy, Is not linearly re
lated to the amount or nature of change after treatment as measured on
the dependent variables. Interestingly, however, when subjects who re
port high motivation for psychotherapy are given a positive set concern
ing the outcome of the treatment and/or are offered a therapist who Is
congruent with their expectancies they tend to speak more. That is,
greater reported desire for psychotherapy is related to a larger amount
of verbal catharsis; however, It is unrelated to changes on the measures
of mood and anxiety after placebo psychotherapy when a therapist is
absent,
It appears, thus, that in this study, catharsis or degree of mo
tivation, or both, have not been shown to underlie the placebo effect.
Shapiro (1964), in his review of placebo literature, indicates that a
great number uf "psychologic mechanisms" have been proposed as being un
derlying characteristics of the placebo effect. Several are: catharsis,


47
distinguish significantly between the Independent variables; of the four
moods in the less clearly identified group, two were found to distinguish
significantly between the independent variables; and of the four moods in
the least clearly identified category, only one was found to distinguish
significantly etween the independent variables. The AACL, the scale
with the greatest number of items, and probably, therefore, the scale
with the highest reliability, showed significant differences between two
of three main effects (independent variables). Below is a summary table
which on the vertical axis lists the three independent variables and
their interactions. On the horizontal axis is listed the 14 dependent
variables and entered into the body of the table are the levels of sig
nificance at which a null hypothesis might be rejected.
Although seven dependent variables did not reflect significant
differences between any of the treatment conditions, there were changes
worth noting in severa) of them. For example, the MACL mood of Depres
sion was found to decrease after treatment in all treatment conditions
except one: the Positive Set, Congruent Therapist, Control Group condi
tion where a small increase in reported depression occurred (see Table
HA). Evaluation by the Signs Test (Siegel, 1956) shows that the proba
bility of seven out of eight cel 1 s decreasing by chance is at less than
the .035 level. This indicates that a reported decrease in depression
tends to occur after treatment, regardless of the specific nature of the
treatment within the study. The MACL mood of Concentration, similarly,
did not result in any statistically significant differences between the
treatment conditions. However, the Congruent Therapist, Experimental
Group condition was the only treatment which described an Increase In


TABLE 27
MULTIPLE REGRESSION COEFFICIENTS BETWEEN THE VARIABLES LISTED ON THE VERTICAL AXIS AND FIVE
DEPENDENT VARIABLE CHANGE SCORES FOR THE EXPERIMENTAL GROUP. THE DATA ARE PRESENTED
SEPARATELY FOR EACH OF THE INDEPENDENT VARIABLES IN ADDITION TO THE
TOTAL EXPERIMENTAL GROUP
AACL
4r
Moods of
Cone.
the MACL
Deact
So.
Aff.
Congruent
Therapist

Amount of speech
DesI re for
+.11
-.29
-.13
-.06
-.32
Instructions
Psychotherapy
+.06
+.22
-.02
+.17
+.02
(n = 24)
Diagnosis
+.14
-.09
+.09
-.15
-.09
1ncongruent
Therapist
Amount of speech
Desire for
+.29
+.21
+.09
-.06
+.34
Instructions
Psychotherapy
-.20
-.07
+.20
+ .10
+ .27
(n 24)
Diagnosis
+.03
+.13
-16
+.03
+.04
Pos!tive
Set
Amount of speech
Desire for
+.26
+ .23
+.13
+.16
-.00
(n 24)
Psychotherapy
+.08
+.07
+ .04
+.59**
+.03
Diagnosis
+.04
+.05
+.08
+ .02
-.10
Neutral
Set
Amount of speech
Desire for
+.39
-.01
-.32
-.13
-.08
(n 24)
Psychotherapy
-.08
+.08
+.07
-.38
+ .22
Diagnosis
+.06
+.01
-.18
-.19
+.13
All
Experimental
Amount of speech
Desi re for
+.30*
+.11
-.06
+.04
-.01
Group Subjects
Psychotherapy
+.01
+.09
+.06
+.17
+.11
(n 48)
Diagnosis
+.05
+.02
-.04
-.07
-.01
^Significant ^ .05.
**Significant .01.


32
TABLE 7
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL CONCENTRATION MOOD
Source
ss
df
V
F
A (experimental vs. control)
1.81
1
1.81
.49
B (congruent vs. incongruent)
6.59
1
6.59
1.78
C (positive vs neutral)
1.12
1
1.12
.30
AB
4.75
1
4.75
1.28
AC
6.14
1
6.14
1.65
BC
3.87
1
3.87
1.04
ABC
.45
1
.45
.12
error
3.21,02
§1
3.71
Total
347.75
94
TABLE 7A
PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS (X,) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL CONCENTRATION MOOD
Experimental Group
Control
Group
Congruent
Therapist
Incongruent
Therapist
Congruent
Therapist
1ncongruent
Therapist
Pos!tlve
Set
x, 8.75
X2 8.92
D * +.17
X| 10.08
X2 9.67
0 -.41
x, 9.50
X2 8.50
0 -1.00
x, 10.33
X2 9-42
D -.91
Neutral
Set
X, 11.08
x2 11.25
D + .17
X, = 9.42
*2 8,2*
D =* -I.I7
x, 9.17
x2 Idl
D .00
x, 10.17
x2
D -.42


84
\
bold ..
++
+
7
no
pleased ..
++
+
7
no
earnest

++
7
no
tired
++
+
7
no
sluggish

++
+
7
no
kindly .. ..
++
+
7
no
forglvlng
t
-H-
+
7
no
fearful
++
+
7
no
clutched up

+
+
7
no
regretful
++

7
no
lonely

+
+
7
no
egotistic
++
+
7
no
cocky

++

7
no
overjoyed
++
+
7
no
i Ighthearted

++
+
7
no
vigorous .*
+
+
7
no
energetic

-H-

?
no
witty ..
++

7
no
playful ..
a
++
?
no
rebellious
++
+
7
no
suspicious

++
+
7
no
serious .. ..
++

*
1
no
startled
*
++

?
no
warmhearted ..
++
7
no
defiant ..
*
++
+
?
no
Insecure
++

7
no
engaged In thought
++
+
?
no
self-centered
++
+
7
no


APPENDIX F
SAMPLE PERSONAL INFORMATION QUESTIONNAIRE


69
served the purpose for which they were intended. (By stressing charac
teristics of the psychotherapist, this aspect of the instructions may
have had little influence on the level of set but may have had some vi
carious influence on the congruence or incongruence of the psychothera
pist for any particular subject.) ^
In summarizing the results of the analysis of hypotheses 1, 11
and III, the following may be said: The placebo effect in psychotherapy
does exist and through the technique of this study appears to be measura
ble. Further, the placebo effect is apparently influenced by both the
perceived characteristics of the treater, and the projected outcome of
the treatment as reported to the patient. The influence of the perceived
characteristics of the treater has been approached by rulysls of differ
ences on the dependent variables between one group of subjects who were
told that a therapist to be assigned to them will be congruent with their
expectations, and another group who were told that the therapist will be
incongruent with their expectations. The data apparently demonstrate
that the group given the congruent therapist responds to the treatment
with a greater positive placebo effect. The influence of the reported
outcome of the treatment has been approached by analysis of differences
on the dependent variables between one group of subjects who were given a
positive set concerning treatment outcome and another group who were
given a more neutral set. The data suggest that the group given the
positive set tends to respond to treatment with a greater positive pla
cebo effect. (However, congruent therapist instructions seem to be more
instrumental in encouraging the placebo effect than the "positive set"
instructions.)


LIST OF TABLESContinued
Tabie Page
8A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS
(XJ AND MEAN CHANCE SCORES (0) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL DEACTIVATION MOOD 33
9 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL SOCIAL AFFECTION MOOD 34
9A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL SOCIAL AFFECTION MOOD. 34
10 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL ANXIETY MOOD 35
10A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL ANXIETY MOOD 35
11 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL DEPRESSION MOOD 36
11A PRETREATMENT MEANS (X|), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL DEPRESSION MOOD 36
12 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL EGOTISM MOOD 37
12A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEANS CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL EGOTISM MOOD 37
13 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL PLEASANTNESS MOOD 38
13A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL PLEASANTNESS MOOD 38
v


15
TABLE 2
SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE
RANDOM DISTRIBUTION OF AGE
Source
ss
df
V
F
A (experimental vs. control)
25.01
1
25.01
.44
B (congruent vs. incongruent)
3.76
1
3.76
.07
C (positive vs. neutral)
147.51
1
147.51
2.58
AB
25.01
1
25.01
.44
AC
11.35
1
11.35
.20
BC
25.01
1
25.01
.44
ABC
3.76
3.76
.07
w Cells
ixQlQ'M
88
57.16
Total
5,271.49
95
(An F of 3*95 is significant at the .05 level.)
The range of I.Q.'s was from 73 to 129; standard deviation 12.14. Below
is a summary table of an analysis of variance demonstrating that I .Q. is
randomly distributed across all treatment conditions.
The mean educational level of the sample is 10.9 years and
ranged from 6 years to 20 years. The mean of the length of current hos
pitalization to the closest month is 27 months and ranged from three days
to 14 years. The psychiatric diagnoses carried by these patients were
made by a psychiatrist at the time of admission to the hospital. There
is a wide range of psychiatric diagnoses, which for the purposes of this
study are divided into 54 psychotic and 40 nonpsychotlc. Because all
subjects were randomly assigned to the treatment conditions, there is


19
TABLE 4
THE TWELVE MOOD FACTORS OF THE MCL AND THE ADJECTIVES
CONSTITUTING EACH MOOD FACTOR
?-raraarajir. 1 irmr.i r:.mg-i'-aita a. u-j.-.u. xiinrti.r.7r n,u;anjgg.z.w^snx,x,a n, a rmi amsszsBsam
Adjectives with high loadings on the factor
Factors Most Con-
si stently Identified:
AGGRESSION
CONCENTRATION
DEACTIVATION
SOCIAL AFFECTION
angry, bold, defiant, rebellious,
concentrating, earnest, engaged in thoight,
serious.
drowsy, earnest, tired.
affectionate, forgiving, kindly, warmhearted.
Factors Identified
Fairly Consistently:
ANXIETY
DEPRESSION
EGOTISM
PLEASANTNESS
apprehensive, clutched up, fearful, Insecure,
blue, lonely, regretful,
boastful, cocky, egotistic, self-centered,
elated, lighthearted, overjoyed, pleased.
Factors Tentatively
I dent!fIed:
ACTIVATION
NONCHALANCE
SKEPTICISM
STARTLE
active, energetic, vigorous,
nonchalant, playful, witty,
skeptical, suspicious,
startled, shocked.
level of certainty, because of difficulty in Interpreting its meanlngful-
ness in a psychotherapeutic context, Concentration and Nonchalance might
be expected to increase, while Egotism, Skepticism and Startle might be
expected to decrease after treatment. In addition, changes in Skepticism
after the treatment may reflect the extent to which this treatment was be-
Iieved to be real.
In his research developing the MACL, Now!is used male and female
college students and U. S. Navy personnel In a variety of experimental
(drugs, films) and field situations. For example, one validation procedure
that Nowlis reports (1961) involved the administration of a drug which


were widely used (to effect cures): entities, cathartics, enemas, purges,
stomachics, sweating, bleeding, leeching, cupping, starvation and dehy
dration, Methods of depletion and expulsion . may relieve symptoms
by symbolically expelling bad thoughts and conflictual ego-alien impulses."
In our more sophisticated and verbally oriented culture, Shapiro contin
ues, "relief of symptoms may occur when the patient is able to express
verbally conflictual and guilt-ridden thoughts and feelings in the free,
nonjudging and accepting atmosphere of the doctor's office" (p. 81). in
other words, verbal catharsis, it appears, may theoretically lie at the
basis of the placebo effect in psychotherapy, just as physical-mechanical
catharsis lay at the basis of medical treatment which today is widely
looked upon as placebo. The treatment in this experiment is conducted in
an obviously free, accepting and nonjudgmental atmosphere and each
patient-subject may establish his own personal optimal level of catharsis.
It would seem logical then, that catharsis may be a major underlying fac
tor for the placebo effect in psychotherapy, at least in so far as out
come of psychotherapy is measured in this study. It is interesting, in
this regard, however, that the data indicate that a linear relationship
does not exist between amount of speech (loosely, catharsis) during the
"interview" and degree of change on the dependent variables. In fact, in
this context, although one significant correlation among the 25 computed
would be expected to be significant by chance alone, the only significant
correlation (.05 level) which does exist, suggests that the more a sub
ject speaks, the greater the likelihood that his anxiety level (AACL)
will increase. These correlation*- however, evaluate only the amount of
speech in relation to the dependent variable changes after treatment. The


DISCUSSION
As mentioned in the Introduction, the literature supporting the
efficacy of psychotherapy has not demonstrated that personality and be*
havioral modifications which occur as a result of psychotherapy are en-
tirely due to a psychotherapist's behavior and techniques (specific fac
tors). The outcome of treatment, it was suggested, is a function of both
the specific and the nonspecific (placebo) events occurring during psy
chotherapy, although, in the literature, the presence and the contribu
tion of the placebo effect in psychotherapy appear to be entirely
conjectural.
This study appears to demonstrate that the placebo effect in psy
chotherapy does exist. Subject-patients who spoke into a tape recorder
for one session, believing that they were speaking to a psychotherapist,
changed certain behaviors In therapeutic directions when compared to sub
jects who did not speak (see Hypothesis l). That Is, it has been shown
that verbally reported patient behaviors such as anxiety and certain
moods which would be expected to change as a result of therapist behav
ior in actual psychotherapy, also change, in fact, as a result of the
placebo psychotherapy (nonspecific treatment) in which a therapist is ab
sent. The group who spoke to the "psychotherapist for one session, when
compared to the group who waited, increased their Social Affection mood on
the MACL and reduced their level of anxiety as measured on both the AACL
and the HA CL Anxiety mood.
62


This dissertation was prepared under the direction of the chair
man of the candidate's supervisory committee and has been approved by all
members of that committee. It was submitted to the Dean of the College
of Arts and Sciences and to the Graduate Council and was approved as
partial fulfillment of the requirements for the degree of Doctor of
Philosophy.
April 24, 1965
Dean, Graduate School
Supervisory Committee:
d. fiJcL
Chai
,/A


100
Subj Treat Adj So Amt of Des
No
Group
1 .4.
1 ntro
Neuro
Aqe
AACL
Score
Aggr
Conc
Deact
Aff
Anx
Depr
Eqot
Pleas
Act i v
Nonch
Skept
Start
Speech
Psycho
D! aq
45
4
088
24
26
43
09
10
09
10
07
09
09
1 1
10
10
09
08
08
08
02
01
01
46
4
102
26
30
24
10
08
10
10
09
06
10
07
09
11
09
12
10
09
02
02
02
47
4
116
17
32
49
09
12
07
10
10
08
1 1
10
05
09
09
08
07
10
02
02
01
48
4
111
14
16
36
09
12
09
10
10
09
10
13
09
10
1 1
11
12
09
01
03
02
49
5
108
26
42
30
06
03
11
11
09
08
09
07
07
10
09
10
09
09
50
5
092
10
44
44
09
09
09
09
09
08
09
08
09
10
09
12
12
11
51
5
107
29
25
55
06
10
10
10
17
10
06
09
1 1
11
07
13
13
09
52
5
079
28
36
42
12
08
09
08
1 1
09
07
11
09
10
08
09
09
11
53
5
086
36
10
44
09
09
08
11
05
14
11
09
08
06
12
09
10
09
54
5
083
23
35
37
10
10
08
07
07
11
09
06
16
09
08
07
10
10
55
5
118
22
35
31
12
06
07
09
10
11
09
12
11
08
09
09
09
09
56
5
119
30
24
31
08
10
11
17
09
09
02
09
09
10
09
1 1
04
09
57
5
101
25
17
29
10
01
07
10
1 1
09
10
09
09
03
09
09
09
09
58
5
129
13
48
39
07
09
06
07
09
09
09
10
09
02
13
09
10
10
59
5
098
28
11
43
11
04
06
12
04
14
08
09
07
07
12
07
09
08
60
5
092
18
42
28
15
03
04
09
06
03
12
08
02
09
09
09
09
10
61
6
093
22
46
43
10
00
10
09
09
13
07
11
10
09
09
09
09
09
62
6
091
26
32
44
05
09
07
1 1
09
13
07
13
09
14
10
09
10
10
63
6
098
20
35
34
07
00
09
09
08
07
12
09
10
10
11
09
09
09
64
6
102
26
38
41
13
00
09
1 1
14
07
14
11
06
08
07
10
09
08
65
6
073
26
18
40
09
09
06
08
09
09
09
09
09
08
09
07
09
09
66
6
081
12
42
41
08
05
14
10
08
05
06
09
08
09
09
14
10
08
67
6
087
28
10
32
10
04
11
11
09
11
09
09
07
11
10
08
09
06


38
TABLE 13
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL PLEASANTNESS MOOD
Source
ss
4 f
V
F
A (experimental vs. control)
8.12
1
8.12
1.21
B (congruent vs. incongruent)
39.31
1
39.41
5.89*
C (positive vs. neutral)
43-51
1
43.51
6.50*
AB
.01
1
.01
.001
AC
.02
i
.02
.003
BC
3.57
1
3-57
54
ABC
3.87
1
13.87
2.07
error
581.76
£Z
6.69
Total
690.17
94
Significant ^ .05.
TABLE 13A
PRETREATMENT MEANS (X.), POSTTREATMENT MEANS (X,) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL PLEASANTNESS MOOD
Experimental Group
Control
Group
Congruent
Therapist
Incongruent
Therapist
Congruent
Therapist
Incongruent
Therapist
Pos!tive
Set
X, = 4.25
X2 6.67
0 +2.42
x, 4.33
x2 itJl
0 .00
X, 4.42
X2 5.50
D +1.08
X, 4.33
X2 4^
0 +.17
Neutral
Set
X, 4.17
X2 4J2
D .00
X, 4.75
x2 4.42
D -.33
X, 4.33
x2 4.42
D +.09
X, 4.17
X2 IM
0 -1.50


72
Analysis of Hypothesis IV gives no clear picture of the contribu
tion to the placebo effect of the individual patient differences in age,
Intelligence, introversion and neurotic!sm. This finding is not surpris
ing in viewof the existing body of research which is equivocal and often
contradictory in regard to this question (see Introduction). The evalua
tion of the data collected to explore this hypothesis resulted in border
line significance and lack of remarkable trends both within the four
measured subject differences and within all of the dependent variables
across individual subject differences. Even when tentative or apparent
trends were identified (as discussed on pp. 49-59), the complexity of the
behaviors involved In the eight different combinations of the independent
variables makes the meaningfulness of any interpretation of results
doubtful. (In addition, the reader is reminded of the questionable use
of the Neurotic!sm Scale or the MPi with this subject sample; see p. 57*)
/
Pearson product-moment correlations were used to evaluate this hypothesis.
These only identify linear relationships between the dependent variables
and the measured subject differences, and the possibility of a nonlinear
relationship cannot be dismissed on the basis of the analysis performed.
The problem of discussing the meaningfulness of the results and
the inability to disclose significant trends in the data of this study
apparently leads to the acceptance of the null hypotheses which state
that, in the present sample, age, intelligence, Introversion or neuroti-
v
cism will not identify the type of subject who responds favorably to pla
cebo psychotherapy. This Is in keeping with the findings of Wolf et al.
(1957) who indicate that individual subject reactions to placebos are


20
has known effects on a person's feelings and moods. Continuous monitor"
ing of mood on the MACL for 16 hours demonstrated that fluctuations in
the MACL were related to anticipated drug-induced changes. Without re*
porting actual correlational figures, Now!Is indicates that concurrent
validity of the MACL is high, although the test-retest reliability is low.
Nowlis describes temporality as one of the characteristics of a mood and
therefore, b*c&use of the low reliability of the check list, he suggests
that the scale is sensitive to short-term fluctuations in mood. Because
of its short length and ease of administration, Nowlis points out, it is
possible to monitor mood with this instrument over repeated intervals of
any required length."
The subject is asked to circle one of four options for each word
in the check list: a double plus if the adjective definitely describes a
current strong feeling, a single plus if the word slightly applies to a
current feeling, a question mark if the word is unclear or the subject is
unable to decide if the adjective describes a current feeling and no if
the subject is certain that the word does not describe a present feeling.
Oouble plus is scored as 3 single plus is scored as 2, question mark is
scored as I, and no is scored as 0. Twelve mood scores are obtained by
summing the scores of the individual adjectives in each of the moods.
The scores may range from 0 to 12 on the moods of Aggression,
Concentration, Social Affection, Anxiety, Egotism and Pleasantness. The
scores may range from 0 to 9 on the moods of Deactivation, Depression,
Activation and Nonchalance. On the moods of Skepticism and Startle the
scores may range from 0 to 6.


Name
i s
If the statement is true of you, circle "True."
not true of you, circle "False."
1 f the
statement
1.
1 am no one. Nothing seems to h<* me.
True
False
2.
1 am optimlStic.
True
False
3-
1 am a hostile person
True
False
4.
Self-control Is no problem to me.
True
False
5.
1 tend to be on my guard with people who are
somewhat more friendly than 1 had expected.
True
False
6.
1 try not to think about my problems.
True
False
7.
1 have an attractive personality.
True
Fal se
6.
1 am a rational person.
True
False
9.
1 have a horror of failing in anything 1 want
to accomplish.
True
False
10.
1 put on a false front.
True
False
11.
1 feel helpless.
True
Fal se
12.
1 am a good mixer.
True
False
13.
It is difficult to control my aggression.
True
False
14.
1 feel insecure within myself.
True
Fal se
15.
1 really am disturbed.
True
Fal se
16.
All you have to do is just insist with me, and
i give in.
True
Fa) se
17.
1 have to protect myself with excuses, with
rational izlng.
True
False
18.
1 have few values and standards of my own.
True
Fal se
19.
1 take a positive attitude toward myself
True
False
86


35
TABLE 10
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL ANXIETY MOOD
Source
ss
df V
F
A (experimental vs. control) 35*35
1 35.35
6.29*
B (congruent vs. Incongruent) 4.27
1 4.27
.76
C (positive vs. neutral)
.00
1 .00
.00
A6
6.31
1 6.31
1.12
AC
4.00
1 4.00
.71
BC
.07
1 .07
.01
ABC
8.33
1 8.33
1.48
error
488.92
82 5-62
Total
547.25
94
Significant ^ .05.
TABLE I0A
PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X2) AND MEAN
CHANCE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL ANXIETY MOOD
Experimental Grouo
Control Group
Congruent
Therapist
1ncongruent
Therapist
Congruent Incongruent
Therapist Therapist
Positive X, 5-25
Set X2 3J5
d = -1.50
X, 4.92
X2 5.08
d +7K
Xj 4.50 X, -
X- 5.08 X, -
D +3H D -
4.33
lii
.00
Neutral Xj 5.84
Set X2 4.42
0 -1.42
X, 3.83
*2 hdl
D -.66
X, 5.67 X, 4.00
X2 5.84 X2 5.00
0 17 D =* -H.00


9
dissonant with the treatment situation, the effectiveness of treatment
decreases (see Fastinger and Brame I in Bachrach, 1962).
Similarly, folk medicine In the Spanish Southwestern United
States (Jaco, l958) and witchcraft healing in many parts of the world
(Frank, I96i) flourish and prevent the establishment of modern medical
program, at least in part, because scientifically trained physicians do
not meet loca; expectancies about the nature of treaters. That Is, to
these people scientific physicians are cognitive y dissonant in the roe
of "doctor." Conversely, the effectiveness of the native witch doctor is
enhanced because he meets his clientele's expectancies concerning the na
ture of treaters.
in the present research, verbal instructions to the patients will
estab Ish two different types of doctors (psychotherapists). One is the
type of psychotherapist the patient expects and the other does not con
form to his expectations (see Appendix A). This will produce two dis
tinct groups of subjects regarding their personal perception of their
psychotherapists.
(3) The treatment situation.
Foulds (1958) and Shapiro (1964) refer to many studies which dem
onstrate that a treater's bias can significantly Influence treatment out
come. For example, Hof ling (1955) and Frank (1958), In studies, demon
strate that positive staff attitudes and biases toward a particular pla
cebo or specific treatment increases the effectiveness of that treatment.
Lyerly et al. (¡962) In a drug study, have shown that subjects receiving
instructions appropriate to the expected effects of chloral hydrate, re
gardless of the actual drug or placebo administered, reported subjective


25
By a coin toss prior to the meeting between subjects and Expert-
mentar, one subject of the pair was to be placed in the Experimental
Group and the other of the pair was to be placed in the Control Group
(treatment vs. wait). Two additional coin tosses by the Experimenter
prior to the meeting determined whether this particular pair of subjects
would receive (1) the positive or neutral set, and (2) the congruent or
incongruent therapist instructions. Both subjects were then given the
instructions appropriate to their designated treatment category (i.e.,
(1) positive set, congruent therapist; (2) positive set, incongruent
therapist; (3) neutral set, congruent therapist; and (4) neutral set, in
congruent therapist [see Appendix A]). The Experimenter had sufficient
time while the subjects were filling in the Adjustment Score, to choose
the predetermined items on the Expectation Questionnaire and manipulate
the congruence and incongruence of the subject's alleged psychotherapist
(see Appendix A).
The Control Group subject of the pair was then asked to remain in
the testing room. He was asked to relax, was offered the use of a stack
of magazines and was told that the Experimenter would return in about 50
minutes (see Appendix A for exact wording). The Experimental Group sub
ject was then taken to the room which contained the tape recorder. This
room had a one-way vision screen and was wired for sound. The subject
was seated in an easy chair, instructed to speak to the "psychiatrist"
and was told that the Experimenter would return In about 50 minutes to
retrieve the tape and the tape recorder was turned on (see Appendix A for
exact wording). The Experimenter then seated himself in an observation
room where he was able to record the Experimental Group subject's


APPENDIX 6
SAMPLE LETTER SENT TO ALL SUBJECTS WHO SPOKE
INTO THE TAPE RECORDER


33
TABLE 8
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL DEACTIVATION MOOD
Source
ss
df
V
f
A (experimental vs. control)
7.82
1
7.82
1.67
B (congruent vs. Incongruent)
25-48
1
25.48
5.43*
C (positive vs. neutral)
33
1
.33
.07
AB
1.24
1
1.24
.26
AC
.96
1
.96
.20
BC
.48
1
.48
.10
ABC
70
1
70
.15
error
40,7-25
8Z
4.69
Total
444.56
94
Significant <£. .05.
TABLE 8A
PRETREATMENT MEANS (Xi), POSTTREATMENT MEANS (j) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL DEACTIVATION MOOD
Experimental Group
Control
Group
Congruent
Therapist
1ncongruent
Therapist
Congruent
Therapist
Incongruent
Therapist
Pos!tive
Set
X, 4.08
x2JL2L
D -1.17
X, 3.42
X2 iiiZ
D +.25
X, 3.67
x2 "US
0 +.08
X, 4.08
X2 ibM
D +.50
Neutral
Set
x, 4.25
X2 UL2
D -1.08
X, 4.17
x2 4.33
0 +.16
X, 4.75
X2 4.00
0 -.75
X| 2.25
X2 3.08
D +.83


106
Rogers, C. R. Client centered therapy. Boston, Hass.: Houghton Hlfflin,
1954.
Rogers, C. R.
cago,
and Dymond, R. Psychotherapy and personality change.
Illinois: Univ. of Chicago Press, 1954a.
Chi -
Rogers, C. R. On becoming a person. Boston, Hass.: Houghton Hifflin,
1961.
Rosenthal, 0. and Frank, J. D. Psychopathology and the placebo effect.
Psychol, bull.. 1956, I, 294-302.
Shapiro, A. K. Factors contributing to the placebo effect. Aro. J. of
psvchoth.. 1964, JjJ, supplement I, 73-88.
Shoben, E. H. Some observations on psychotherapy and the learning proc
ess, In Howrer, 0. H. (ed.) Psychotherapy, theory and research.
New York, New York: Ronald Press, 1953.
Siegel, S. Nonparametrlc statistics. New York, New York: HcGraw-Hill,
1956.
Slack, C. W. Experimenter-subject psychotherapy: a new method of in
troducing intensive office treatment for unreachable cases.
Kent, hyg.. i960, 238-256.
Strupp, H. H. Psychotherapy. Ann, rev, psvchol.. 1962, 445-478.
Tibbetts, R. W. and Hawkings, J. R. The piacebo response. J. of ment.
- sci.. 1956, 60-66.
Uhlenruth, E. H., Canter, A., Neustadt, J. 0. and Payson, H. E. The
symptomatic relief of anxiety with meprobanate, phenobarbital
and placebo. Am, J. psvchiat.. 1959, 115. 905-910.
Whltehorn, J. C. Psychiatric implications of the "placebo effect." Am.
J. Psychiat.. 1958, 114, 662-664.
Wolf, S., Doering, C. R., Clark, H. L. and Hagans, J. A. Chance distri
bution and the piacebo "reactor." J. lab, clin, reed.. 1957 49.
837-841.
Wolf, S. The pharmacology of placebos. Pharmacol, rev.. 1959, i 1
698-702.
luckerman, H. The development of an affect adjective check list for the
measurement of anxiety. J. consult, psvchol.. I960, 4, 457-462.
Zuckerman, H. and Blase, D. V. Replication and further data on the affect
adjective check list measure of anxiety. J. consult, psvchol..
1962, 6, 291.


70
Having found that offering a therapist who is congruent with a
subject's expectations seems to encourage the placebo effect more than
offering a positive set concerning treatment outcome may have Implica
tions beyond the present study. There is a considerable body of research
concerned with the matching of patients and therapists and the resulting
relationships in psychotherapy (e.g., see Leary and Gill [1959])* In the
present study it is seen that matching a therapist with a patient's ex
pectations does, in fact, increase the effectiveness of the treatment to
a greater extent than offering a positive set concerning treatment out
come. This is indicated in this placebo situation, where no therapist
actually exists; however, it is suggested that in psychotherapy percep
tion of the psychotherapist may also be more valuable than belief about
treatment outcome. That is, the nature of the relationship (in this
study, a relationship In which the patient perceives his therapist as
being congruent with his expectations) may be more essential In success
ful psychotherapy than statements concerning the value of the treatment
1tself.
An important finding of this study is that, as predicted, the
Adjustment Score, which measures general level of psychological adjust
ment, does not change after the placebo Interview. This is so regardless
of the particular treatment (talk vs. wait, congruent therapist vs. in-
congruent therapist and positive set vs. neutral set) or combination of
treatments in the study. This indicates that in a single treatment of
placebo psychotherapy, no change in general level of psychological ad
justment may take place. Had the placebo psychotherapy continued over a
longer period of time would there have been changes In this Adjustment


11
III. Those subjects receiving a positive set regard
ing the value and outcome of the treatment will
show more marked change in a "therapeutic di
rect ionf* on post measures of anxiety and mood
than those subjects who receive a more neutral set.
However, neither group will show a change In gen
eral level of psychological adjustment in this
single treatment.
IV. The measured individual subject differences of
age, intelligence, level of introversion and of
neurotlclsm will not identify the subjects who
respond best to this placebo situation. That Is,
there will be no differences between the placebo
responsiveness of subjects In terms of their age,
intelligence, level of introversion-extroversion
and level of neurotlclsm.
Hay the "placebo psychotherapy,1 as introduced above, be legiti
mately referred to as a placebo treatment situation? There is a strong
analogy between this treatment and that which exists when a physician ad
ministers a drug placebo. The physician who treats a patient with a pla
cebo pill does so in the same treatment setting, giving the same direc
tions, and offering the same encouragement as If the pill were pharmaco
logically sound. The only changed character!stlc, and that which defines
the situation as exclusively nonspecific, is the absence of the actual
drugthat agent in the treatment situation which produces the specific
effect. Reactions to this placebo are recorded as "placebo effects" al
though they may appear to be a result of a specific drug which is clearly
absent from the treatment. Similarly, in a sham interview such as that
mentioned above, the patient is given the same directions, the same en
couragement and is placed in the same treatment setting as if it were an
actual psychotherapeutic situation. The only changed characteristic, and
that characteristic which defines the situation as nonspecific in this
case, is the absence of the psychotherapist, that is, the only agent


APPENDIX D
SAMPLE ADJUSTMENT SCORE


Name
Each of the 40 words In the following list describes feelings or
mood. Please use the list to describe your feelings at this moment.
Hark each word according to these Instructions: If the word definitely
describes how you feel at the moment you read it, circle the double plus
(++) to the right of the word. For example, If the word Is calm and you
are definitely reeling calm at the moment, circle the double plus as
follows:
calm 0- ? no (This means that you definitely feel calm at this
moment.)
If the word only slightly applies to your feelings at the moment, circle
the single plus as follows:
calm ++ 0 ? no (This means that you feel slightly calm at this
moment.)
If the word is not clear to you or If you cannot decide whether or not
It describes your feelings at the moment, circle the question mark as
follows:
calm ++ + (j) no (This means you can't decide whether you are calm.)
If you clearly decide that the word does not apply to your feelings at
the moment, circle the no as follows:
calm ++ + ? (no) (This means that you are sure that you are not calm
at the moment.)
Work rapidly. Your first reaction Is best. Work down the column before
going to the
mi utes.
next.
Mark
all the words.
This should
take only a
few
angry ..
++
?
no
boastful ..
+
?
no
coceentrat I ng
++
+
?
no
elated
+
?
no
drowsy ..
++
+
?
no
active
?
no
affectionate
++
+
?
no
nonchalant
.. .. +
+
?
no
aporehens ive
++
+
?
no
skeptical
. .. ++
f
?
no
blue
++
+
?
no
shocked ..
+
?
no
83


LIST OF TABLES
Table page
1 THE EXPERIMENTAL DESIGN 13
2 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE
DEMONSTRATING THE RANDOM DISTRIBUTION OF
AGE 15
3 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE
DEMONSTRATING THE RANDOM DISTRIBUTION OF
i .a 16
4 THE TWELVE MOOD FACTORS OF THE MACL AND
THE ADJECTIVES CONSTITUTING EACH MOOD
FACTOR >9
5 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE ZUCKERMAN ADJECTIVE CHECK LIST 30
5A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
ZUCKERMAN ADJECTIVE CHECK LIST 30
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL AGGRESSION MOOD 31
6A PRETREATMENT MEANS (X.), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE CORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL AGGRESSION MOOD 31
7 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL CONCENTRATION MOOD . 32
7A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (0) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL CONCENTRATION MOOD 32
8 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL DEACTIVATION MOOD 33
Iv


27
After all the data were collected, all subjects who spoke into
the tape recorder were sent a letter (see Appendix G). The purpose of
this tetter was to thank the subject for his cooperation and Inform him
of the way in which he could contact the Experimenter for further clari
fication of the procedure. Twelve of the 69 subjects who spoke into the
tape recorder contacted the Experimenter as a result of this letter.


LIST OF TABLESContinued
Table Page
20 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE
DEMONSTRATING THE RANDOM DISTRIBUTION OF
MPI INTROVERSION SCORES 50
21 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE
DEMONSTRATING THE RANDOM DISTRIBUTION OF
MPI NEUROTICISM SCORES 51
22 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN
SUBJECTS' CHANGE SCORES FROM PRE- TO POST
TREATMENT TESTING ON EIGHT OF THE DEPENDENT
VARIABLES AND SCORES ON THE BETA INTELLIGENCE
TEST. THE DATA ARE PRESENTED FOR EACH OF THE
EIGHT TREATMENT CONDITIONS 53
23 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN
SUBJECTS' CHANGE SCORES FROM PRE- TO POST
TREATMENT TESTING ON EACH OF THE DEPENDENT
VARIABLES AND SCORES ON THE MPI INTROVERSION
(LOW SCORES)-EXTROVERSION (HIGH SCORES). THE
DATA ARE PRESENTED FOR EACH OF THE EIGHT
TREATMENT CONDITIONS 55
24 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN
SUBJECTS' CHANGE SCORES FROM PRE- TO POST
TREATMENT TESTING ON EACH OF THE DEPENDENT
VARIABLES AND SUBJECTS' AGES. THE DATA ARE
PRESENTED FOR EACH OF THE EIGHT TREATMENT
CONDITIONS 56
25 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN
SUBJECTS' CHANGE SCORES FROM PRE- TO POST
TREATMENT TESTING ON EACH OF THE DEPENDENT
VARIABLES AND SCORES ON THE MPI NEUROTICISM
SCALE. THE DATA ARE PRESENTED FOR EACH OF
THE EIGHT TREATMENT CONDITIONS 58
26 MULTIPLE REGRESSION COEFFICIENTS BETWEEN AMOUNT
OF SPEECH AND DESIRE FOR PSYCHOTHERAPY PRESENTED
FOR EACH INDEPENDENT VARIABLE (EXPERIMENTAL
DATA ONLY) 59
vli


73
generally inconsistent, unpredictable and not uniform (see also Hagans
et al.. 1957, and Kurland, 1957)*
A pertinent issue may be raised concerning whether the subjects
accepted the placebo treatment as a legitimate form of psychotherapy.
There is no way In which to be certain of any conclusion in this regard.
However, the nature of changes on the dependent variables certainly sug
gests that the subjects did perceive the procedure as actual psychother
apy. Moreover, it will be noted that the Skepticism mood on the MACL
remained rather constant from pre- to posttreatment testing, and the only
groups to report a small and insignificant Increase in Skepticism, after
having completed the procedure, were the Incongruent Therapist, Control
Group subjects (see Table I6A). No subjects reported that they disbe
lieved the legitimacy of the treatment and the comments several of the
subjects made to the examiner and into the tape recorder appear related
to the issue of subjects accepting the treatment as real. For example,
one subject in the Experimental, Neutral Set, Congruent Therapist Group
reported the following Into the tape recorder during the placebo therapy:
"I appreciate the privilege of being given a chance to express my feel
ings and views in this way. . This is the first chance I've had to
honestly and openly express my feelings and opinions and I appreciate it
very, very much."
It is hoped that the placebo effects demonstrated In this treat
ment will be kept in mind in future research on the effects of specific
methods In psychotherapy. It is further hoped that this demonstration
will lead to the practice of separating effects common to all psycho
therapy (e.g., nonspecific placebo effects) from those effects due to the


23
a person," and neurotlcism represents "the general emotional liability of
a person, his emotional overresponsiveness, and his liability to neurotic
breakdown under stress" (1959, p. 3)- The Inventory has 48 questions
which have been factor analyzed to yield scores on an extroversion-
introversion dimension and a neurotlcism dimension. Although the two di
mensions are slightly negatively correlated (-.15), Eysenck submits that
they are essentially orthogonal.
The original standardization took place on a sample of 200 normal
English men and 200 normal English women. Eysenck has since supplemented
this by adding groups of students, nurses, industrial apprentices and "a
quota sample of the whole population." Data are also available on groups
of hospital patients and prisoners. Split-half and Kuder-Richardson re
liability coefficients for the Neurotlcism scale fall between .85 and .90
and for the extroversion scale fall between .75 and .85. Using both con
current arid construct validation techniques, Eysenck reports that the MPI
has yielded data supporting its validity.
(5) A Modification of the Expectation Q-sort (Apfelbaum, I95&)
(see Appendix £).
The Q-sort asks a subject to indicate his expectancies regarding
the nature of a psychotherapist who may be assigned in the future.
Thirty-six of the 60 items in Apfelbaum's item pool were chosen and ran
domly organized into a questionnaire. These 3 items constitute those
with heaviest factor loadings on three separate clusters: (1) those
items which designate a guiding, giving, protective therapist; (2) those
items which designate a tolerant, accepting and permissive therapist; and
(3) those items which designate a cold, rigid and condemnatory therapist.


RESULTS
Analysis of covariance (Ray, I960) was used to evaluate the tena-
billty of hypotheses I, II and III (those hypotheses concerned with the
effects of differential experimental treatment). In analyzing the three
hypotheses, the eight independent treatment conditions, as diagramed in
Table 1 were considered. Dependent variables in hypotheses I, II and III
were the changes in scores frompre-to posttreatment testing on the AACL,
the 12 moods in the MACL and Dymond's Adjustment Score. Thus, evaluation
of the first three hypotheses was made across the three independent vari
ables utilizing 14 2 x 2 x 2 analyses of covarianceone for each de
pendent variable.
The technique of analysis of covariance chosen to evaluate hy
potheses I, II, and III utilizes the fact that differences between the
treatment groups on a dependent variable posttreatment measure may be to
some extent a reflection of differences between these groups on the ini
tial pretreatment measure. It removes the variance due to these initial
random differences from the final posttreatment variances, thus reducing
the size of the error variance and increasing the precision of the expe
riment (Gourlay, 1953)* That is, by utilizing the regression of the
posttreatment measures on the pretreatment measures, an adjustment Is
made for the variability which is associated with Initial differences
among the subjects. After this adjustment is made, the remaining varia
bility may be analyzed with a more precise estimate of factors which may
have produced an effect.
28


31
TABLE 6
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL AGGRESSION MOOD
Source
ss
df
V
F
A (experimental vs. control)
8.40
l
8.40
2.00
B (congruent vs. incongruent)
7.57
1
7-57
1.81
C (positive vs. neutral)
.10
1
.10
.02
AB
31,84
1
31.84
7.60**
AC
1.95
1
1.96
.47
BC
4.Go
1
4.66
1.11
ABC
5.43
1
5-43
1.30
error
8Z
4.19
Total
424.25
94
** significant < .Oi.
TABLE 6A
PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X,) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL AGGRESSION MOOD
Experimental Group
Control
Group
Congruent
Therapist
Incongruent
Therapist
Congruent
Therapi st
1ncongruent
Therapist
Positive
Set
X, 3.33
X, 1.83
D -1.50
X, 3.42
X2 diz
D + .25
X, 3-50
x2 4.50
D = +1.00
X, 2.33
X2 2.00
0 -.33
Neutral
Set
X, 2.17
x2 hli
D = -.92
X, 2.42
x2 UI
0 = +.75
X, 2.67
X2 2^
D =* .00
X, 2.08
x2 2^50
D +.42


APPENDIX A
INSTRUCTIONS TO SUBJECTS


specific aspects of the therapy being evaluated (e.g., therapist
behaviors).


43
TABLE 18
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
ADJUSTMENT SCORE
Source
ss
df
V
F
A (experimental vs. control)
40.12
1
40.12
3.04
B (congruent vs. Incongruent)
.07
1
.07
.005
C (positive vs. neutral)
20.84
1
20.84
1.58
AB
.16
1
.16
.01
AC
33.34
l
33.34
2.52
BC
3.77
l
3-77
.29
ABC
9.98
1
9.98
.76
error
uisa^o
8Z
13.21
Total
1,257.98
94
TABLE I8A
PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X2) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE ADJUSTMENT SCORE
Experimental Group
Control
Group
Congruent
Therapist
Incongruent
Theraplst
Congruent
Therapist
Incongruent
Therapist
PosItiva
Set
X, 45.42
*2 &--25
D +.33
X, 43.17
h !&&
D +.75
X, 39.67
x2 4U8i
D +2.16
X, 44.83
*2 "
D +2.42
Neutral
Set
X, o 42.00
X, 41,58
D -.42
X, 46.50
x2 47.00
0 = +.50
X, 37.17
x2 37.50
0 +.33
X, 49.58
x2 *
D .00


APPENDIX t
SAMPLE ANXIETY ADJECTIVE CHECK LIST


99
Subj
No
T reat
Group
i .a.
1 ntro
Neuro
Aqe
AACL
Adj
Score
Aqq.r
Cone
Deact
So
Aff
Anx
Depr
Eqot
Pleas
Acti v
Nonch
Skept
Start
Amt of
Speech
Des
Psycho
Di aq
22
2
095
26
24
22
08
10
14
13
1 1
18
09
09
10
09
08
09
06
07
01
02
02
23
2
112
18
18
40
09
07
07
08
10
09
10
1 1
16
07
09
06
09
10
01
01
02
24
2
089
18
48
44
00
15
10
12
09
15
05
06
16
15
09
15
09
06
01
01
01
25
3
098
30
42
46
13
09
12
09
13
06
12
12
07
07
07
07
1 1
09
02
02
02
26
3
104
40
10
45
10
10
08
09
09
10
09
09
08
16
09
09
09
09
01
01
02
27
3
077
18
36
45
07
07
09
09
09
13
09
09
07
08
12
09
10
09
02
02
02
28
3
090
21
20
45
09
09
09
09
09
09
09
09
09
12
09
12
09
09
01
01
02
29
3
104
15
28
35
13
1 1
06
09
09
03
09
13
10
07
09
09
10
09
01
03
02
30
3
109
13
38
41
07
10
09
09
09
06
09
09
09
11
09
09
09
09
01
01
01
31
3
124
19
42
58
10
1 1
08
09
08
08
1 1
10
08
06
07
09
09
08
01
01
01
32
3
093
24
12
41
12
09
09
09
14
08
1 1
1 1
11
11
12
09
09
09
02
02
02
33
3
111
30
38
48
13
07
09
09
12
09
1 1
09
09
05
12
09
06
09
01
03
01
34
3
077
18
38
31
19
08
10
07
09
06
13
10
07
06
07
08
08
10
01
01
02
35
3
093
14
42
41
14
06
13
09
09
11
10
11
07
10
09
10
10
08
01
01
01
36
3
108
11
44
20
14
16
09
09
11
09
12
07
10
09
09
10
13
09
01
02
01
37
4
102
26
11
40
12
10
06
12
09
09
12
09
09
09
12
06
09
09
01
02
02
38
4
098
19
34
45
11
09
11
09
10
07
09
12
09
06
09
10
10
09
01
02
02
39
4
093
14
08
37
05
03
04
15
08
14
12
1 1
09
12
09
10
12
07
02
02
02
40
4
097
12
28
35
12
05
11
06
09
09
06
12
09
06
09
09
09
06
01
03
01
41
4
108
23
13
40
11
07
11
10
09
07
09
09
09
09
10
09
09
09
02
02
01
42
4
101
38
34
40
08
11
09
10
09
08
1 1
08
10
10
09
10
09
09
02
01
02
43
4
106
40
00
30
10
05
09
09
09
06
1 1
09
09
11
09
08
09
09
02
03
02
44
4
087
10
43
42
10
10
03
11
07
12
06
07
09
09
09
07
07
07
01
01
01


LIST OF TABLESContinued
Table Page
14 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL ACTIVATION MOOD 39
14A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
TREATMENT CONDITION OF THE DESIGN ON THE
MACL ACTIVATION MOOD 39
13 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE,
FOR THE MACL NONCHALANCE MOOD 40
15A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL NONCHALANCE MOOO 40
16 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL SKEPTICISM MOOD 41
ioA PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL SKEPTICISM MOOD 41
17 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE MACL STARTLE MOOO 42
I7A PRETREATMENT MEANS (X.), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
MACL STARTLE MOOD 42
18 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE
FOR THE ADJUSTMENT SCORE 43
I8A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS
(X2) AND MEAN CHANGE SCORES (D) FOR EACH
TREATMENT CONDITION OF THE DESIGN ON THE
ADJUSTMENT SCORE 43
19 SUMMARY OF THE DEPENDENT VARIABLES AND THE
SIGNIFICANCE LEVELS AT WHICH THEY
DISTINGUISHED 8ETWEEN THE INDEPENDENT
VARIABLES 48
vl


45
significantly more than those subjects Mho were assigned a therapist in-
congruent with their expectations (who increased slightly in anxiety)
(p ^ .001) (see Table 5) Similarly, the Congruent Group showed a sig
nificantly higher increase In the MACL mood of Pleasantness (p < .05)
than the Incongruent Group which decreased slightly on this mood (see
Table 13) Further, the Incongruent Therapist Group displayed a signif
icant increase on the MACL mood of Deactivation (p .05) when compared
to the Congruent Group, which decreased on this mood (see Table 8).
Also, the Incongruent Therapist Group reported a significantly higher in
crease in the MACL mood of Startle after treatment (p .001) than the
Congruent Therapist Group which decreased slightly in Startle (see Table
17). While only four dependent variables significantly distinguished be
tween the Congruent and Incongruent Therapist Groups (in addition to two
interaction effects), and the Adjustment Score indicated no difference as
predicted, no dependent variables significantly changed in the unpre-
dlcted direction. Thus, Hypothesis II appears to be supported. (The
factor analytic studies of Apfelbaum (1958) suggest that three basic
types of psychotherapists may be described on the Expectation Q-sort.
Inspection of the data suggest that the sample does not show a trend to
ward describing a particular type of psychotherapist. Similarly, an in
dividual subject does not clearly choose one or another of Apfelbaum1s three
types of psychotherapists.)
Hypothesis III states that subjects who are given a more positive
set concerning the value and outcome of treatment will display more "pos
itive therapeutic changes on the dependent variables than subjects given
a more neutral set. This was found to be the case upon analysis of two


7
administered a placebo to a group of postoperative patients to relieve
reported pain. Those patients reporting a decrease in pain after receiv
ing the placebo had a mean age five years greater than the age of the
nonreactors (p ^ .05). On the other hand, Tibbetts and Hawkings (1956)
found that patients who responded in two drug placebo situations were
younger than the nonresponders. In studies by Kornetsky et al. (1957)
and Abramson et ai. (1955) no meaningful relationship existed between age
and drug placebo responsiveness. Similarly, level of intelligence as a
patient variable has produced contradictory results. In the Lasagna
study reported above, intelligence did not significantly distinguish the
placebo reactor from the nonreactor. On the other hand, Tibbetts and
Hawkings (1956) found the placebo reactor to be less intelligent than the
nonreactor, while Abramson et al. (1955) found the placebo reactor to
have a higher verbal intellectual ability and a lower performance intel
lectual ability.
Leiberman (1964) reports that Eysenck has correlated neurotic!sm
with the need for conformity in an individual's perceptions and judgments.
Therefore, he contends that "in group situations, where an element of in
terpersonal pressure activates the need for conformity, the subjects with
neurotic tendencies respond with placebo effects." In fact, Eysenck ar
gues that a subject's degree of neurotic!sm is a consistent factor for
picking out high placeoo reactors. Eysenck (1961) further speculates
that since introverts form conditioned responses more quickly than extro
verts, and since he reports that response to a placebo is a learned or
conditioned response, he expects Introverts to respond more readily to
placebo factors (p. 636).



PAGE 1

PLACEBO PSYCHOTHERAPY AND CHANGE IN ANXIETY, MOOD AND ADJUSTMENT By HERBERT GOLDSTEIN A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA April, 1965

PAGE 2

ACKNOWLEDGMENTS This study would not have been completed had it not been for the foresight, wisdom and encouragement of many people to whom the writer is indebted and grateful. A special measure of gratitude is due Or. Audrey S. Schumacher, chairman of the supervisory committee, for her genuine interest, her untiring patience and her constant support and guidance. The assistance and encouragement of the remainder of the supervisory committee, Or. R. J. Anderson, Dr. B. Barger, Dr. J. J. Wright and Dr. V. A. Hines, Is warmly accepted and greatly appreciated. The writer wishes to acknowledge the staff and office workers of the Gulfport, Mississippi Veterans' Administration Hospital who gave unsparingly of their time, their cooperation and their resources. A particular note of gratitude is extended to three people who specifically asked to remain unmentioned, but without whom the progress of this study would certainly have been impeded. The writer wishes to thank his parents for their faith, for their investment and for wanting someday to say: "there's my son, the Doctor." Most of all, the writer thanks his wife, Alene, for her understanding, her support, her guidance and her remarkable patience which sustained the writer in his efforts to complete this study. i i

PAGE 3

TABLE OF CONTENTS ACKNOWLEDGMENTS 11 LIST OF TABLES Iv INTRODUCTION 1 The Present Study METHOD 13 Subject* M Means of Measurement 16 Procedure 2k RESULTS 2 DISCUSSION 62 APPENDICES 75 A INSTRUCTIONS TO SUBJECTS 76 B SAMPLE ANXIETY ADJECTIVE CHECK LIST 80 C SAMPLE MOOO ADJECTIVE CHECK LIST 82 D SAMPLE ADJUSTMENT SCORE 85 E SAMPLE OF THE MODIFICATION OF THE EXPECTATION Q-SORT 90 F SAMPLE PERSONAL INFORMATION QUESTIONNAIRE 92 G SAMPLE LETTER SENT TO ALL SUBJECTS WHO SPOKE INTO THE TAPE RECORDER 3k H SUBJECT RAW DATA 96 BIBLIOGRAPHY 103 Hi

PAGE 4

LIST OF TABLES Table Pa 5 e 1 THE experimental design 13 2 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANDOM DISTRIBUTION OF AGE U 3 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANDOM DISTRIBUTION OF I Q ••• • ••••• •••• io k THE TWELVE MOOD FACTORS OF THE MACL AND THE ADJECTIVES CONSTITUTING EACH MOOD FACTOR • • 5 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE ZUCKERMAN ADJECTIVE CHECK LIST 30 5A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE ZUCKERMAN ADJECTIVE CHECK LIST 30 6 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL AGGRESSION MOOD 31 6A PRETREATMENT MEANS (Xi), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL AGGRESSION MOOD 31 7 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL CONCENTRATION MOOD 32 7A PRETREATMENT MEANS (X, ) POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL CONCENTRATION MOOD 32 8 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL DEACTIVATION MOOD 33 i V

PAGE 5

LIST OF TABLES— Continued Table J ag, M PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X,) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL DEACTIVATION MOOD 33 9 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL SOCIAL AFFECTION MOOD }k 9A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL SOCIAL AFFECTION MOOD yk 10 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL ANXIETY MOOD 35 I OA PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL ANXIETY MOOD 35 11 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE r'OR THE MACL DEPRESSION MOOD 36 I IA PRETREATMENT MEANS (X j ) POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL DEPRESSION MOOD 36 12 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL EGOTISM MOOD 37 I2A PRETREATMENT MEANS (X|), POSTTREATMENT MEANS Gf 2 ) AND MEANS CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL EGOTISM MOOD 37 13 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL PLEASANTNESS MOOD 38 13A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL PLEASANTNESS MOOD 38 v

PAGE 6

LIST OF TABLES— Continued Table Page \k SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL ACTIVATION MOOD 39 li*A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL ACTIVATION MOOD 39 15 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE. FOR THE MACL NONCHALANCE MOOD kQ ISA PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL NONCHALANCE MOOD kO 16 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL SKEPTICISM MOOD Ul ISA PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL SKEPTICISM MOOD 41 17 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL STARTLE MOOD k2 I7A PRETREATMENT MEANS (X.), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL STARTLE MOOD k2 18 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE ADJUSTMENT SCORE 1*3 \oA PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE ADJUSTMENT SCORE k} 19 SUMMARY OF THE DEPENDENT VARIABLES AND THE SIGNIFICANCE LEVELS AT WHICH THEY DISTINGUISHED BETWEEN THE INDEPENDENT VARIABLES 48 vi

PAGE 7

LIST OF TABLES— Continued Table 20 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANOOM DISTRIBUTION OF MPI INTROVERSION SCORES $0 21 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANDOM DISTRIBUTION OF MPI NEUROTICISM SCORES 51 22 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRETO POSTTREATMENT TESTING ON EIGHT OF THE DEPENDENT VARIABLES AND SCORES ON THE BETA INTELLIGENCE TEST. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS 53 23 PEARSON PRODUCT-MOMENT CORRELATI ONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRETO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND SCORES ON THE MPI INTROVERSION (LOW SCORES) -EXTROVERSION (HIGH SCORES). THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS 55 Ik PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRETO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND SUBJECTS' AGES. THE OATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS 5 25 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRETO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND SCORES ON THE MPI NEUROTICISM SCALE. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS 58 26 MULTIPLE REGRESSION COEFFICIENTS BETWEEN AMOUNT OF SPEECH AND DESIRE FOR PSYCHOTHERAPY PRESENTED FOR EACH INDEPENDENT VARIABLE (EXPERIMENTAL DATA ONLY) 59 vll

PAGE 8

LIST OF TABLES — Continued MULTIPLE REGRESS I OH COEFFICIENTS BETWEEN THE VARIABLES LISTED ON THE VERTICAL AXIS AND FIVE DEPENDENT VARIABLE CHANGE SCORES FOR THE EXPERIMENTAL GROUP. THE DATA ARE PRESENTED SEPARATELY FOR EACH OF THE INDEPENDENT VARIABLES IN ADDITION TO THE TOTAL EXPERIMENTAL GROUP vili

PAGE 10

I NTRQDUCTION The extensive literature dealing with processes and systems of psychotherapy generally makes the explicit assumption that behavioral changes following psychotherapeutic encounter are a function of the intended efforts of the psychotherapist. That is, behavioral modification results from the direct and specific efforts, manipulations or techniques of the psychotherapist. Within the framework of Roger's theory (1961), for example, the therapist approves and accepts the client and as a result the client comes to accept himself. The patient in psychoanalysis achieves persona) satisfaction and comfort as the psychoanalyst interprets the patient's repressed psychosexual conflicts and the patient establishes a mature sexual adjustment (Fenichel, 1945). Shoben (1953) describes the goal of psychotherapy as the alleviation of symptoms in addition to the increase in a patient's affective comfort. Thus, in all psychotherapy, "specific factors" (therapist boh&vi^rs) aSUge^ly produce intended effects (personal comfort of the patient). Eysenck (1952, 1961) states, and Rosenthal and Frank suggest (1956) that in addition to the intended efforts of the psychotherapist (specific factors), there are other, nonspecific factors in psychotherapy which significantly contribute to the outcome of treatment. "Nonspecific factors" are loosely defined as those "placebo effects," situational events and conditions in addition to the intended efforts of the therapist which reportedly function in all treatment situations (Shapiro, 1964). They serve to cloud interpretation of "real" effects of psychological 1

PAGE 11

treatment (Rosenthal and Frank, 1956) for they produce manifest effects which are Indistinguishable from the effects produced by the intentional treatment prescribed by the particular theory of psychotherapy. Strupp (1962) indicates that because these nonspecific factors have been largely neglected in research in psychotherapy, the extent of their influence is undetermined. Many significant theorists and research ers (e.g., Rogers, 1961) expound on the efficacy of their psychotherapeutic systems, although nonspecific factors, which may greatly contribute to the final outcome of psychological treatment are not effectively accounted for, controlled or measured. Understandably, successful outcome of psychotherapy Is viewed as proof of the effectiveness of psychotherapy. However, it has not bean demonstrated that personality and behavioral modifications which occur as a result of psychotherapy are entirely (or even largely) due to the psychotherapist's specific techniques Instead, the actual outcome of treatment is a function of both the specific and nonspecific events occurring during psychotherapy. Eysenck (1961), after reviewing and summarizing the literature, shows that actual psychotherapy does not achieve higher recovery rates than those reported from ordinary life experiences and the nonspecific effects of routine medical treatmer.t. After demon st rating that iw.speclfic treatment produces effects comparable with actual psychotherapy, Eysenck implies that nonspecific factors in psychotherapy are essential factors responsible for bringing about successful psychotherapeutic change. Findings and statements such as these clarify the necessity for research in this area.

PAGE 12

3 Effects of treatment with drug placebos are those nonspecific affects given the greatest attention in the literature. Based on his review of the literature, Shapiro (1964) states that placebo effects are subtle and common to all treatment situations and are the most important nonspecific factors in psychological treatment. He defines the placebo effect as the "psychologic, physiologic, or psychophysiologic effect of any medication or procedure given with therapeutic intent which is independent of or minimally related to the effects of the medication or to the specific effects of the procedure and which operates through a psychologic mechanism" (p. 298). In addition, the administration of this medication or procedure must be recognized as legitimately therapeutic by the patient. Rosenthal and Frank (195b) also refer to the placebo effect as a type of nonspecific event in psychological treatment and they report that since research in psychotherapy has not dealt directly with the placebo effect, the nature of its influence is undetermined. However, these authors speculate that it occurs with considerable regularity, and they stress the need to demonstrate through research that observed effects of psychotherapy are due to the therapist's techniques and exist separate from nonspecific effects. Borgatta (1959) expands on the importance of the placebo effect by reporting that "If there is no evidence that an agent is the cause of a given outcome, and there Is also no evidence that a placebo is the cause of a given outcome, an appropriate statement is that there Is no evidence that the agent is more efficient than the placebo in bringing

PAGE 13

k about the outcome. Further, there is also no evidence that the placebo is not as effective as the agent In bringing about the outcome" (p. 331). The literature is replete with references confirming the efficacy of nonspecific effects of drug placebos (see reviews by Shapiro, )36h and Hongingf ield, 1963). However, despite the demonstrated effectiveness of drug placebos, and the alleged importance of the placebo effect in psychological treatment, adequately controlled studies evaluating the placebo effect in psychotherapy are conspicuously lacking (Rosenthal and Frank, 1956 and Strupp, 1962) One is currently forced, therefore, to rely on implication, inference, and analogy when constructing evidence of, and predictive hypotheses pertaining to, the influence of nonspecific placebo factors in psychotherapy. The Present Study "As Withering 1 s foxglove was made more potent when digitalis was isolated from the concoction, and after it was recognized that only the dropsy of congestive heart failure was benefited, so will psychotherapy become more potent after the placebo effect is Isolated and dissected free from the psychotherapeutic process" (Shapiro, 1964, p. 85). As long as both specific and nonspecific (placebo) factors operate together during the treatment, one may never accurately determine the real contribution of the placebo effect to psychotherapy. Since nonspecific placebo factors reportedly function unbeknown to the psychotherapist, since their presence is reportedly a factor in the outcome of all psychotherapy, and since they remain a nebulous and poorly defined factor in psychotherapy, it is clear that the extraction of these placebo effects

PAGE 14

from an investigation of actual psychotherapy is a complex, if not impossible task. Thus, a demonstration of the genuine effects of psychotherapy free of placebo effects seems never to have been made. In this study an attempt is made to Investigate a form of "psychotherapeutic encounter" in which specific factors could not contribute to the outcome of treatment. That is, any changes which accrue as a result of the treatment are not due to a psychotherapist's behavior, technique or manipulations during treatment. Change, Instead, must oe attributed to the presence of the placebo factor and must .jb referred to as a placebo effect. A psychotherapeutic encounter" can take place when a patient reports to a tape recorder for one "therapy session' believing a psychotherapist will listen and respond to his talking. Measures of change In such "affective comfort" (Shoben, 1953) as moods and anxiety taken immediately after this treatment will Indicate the effect of the treatment. Borgatta (1959) suggests, satirically, that this technique be used to demonstrate the relative Ineffectiveness of actual psychotherapy. Slack (I960), on the other hand, demonstrated that therapeutically inaccessible juvenile delinquents could be introduced to psychotherapy by initially having them speak into a tape recorder. Eventually these juveniles were introduced to the therapist who had been giving them feedback from the tapes, and actual psychotherapy was initiated. Hart in, Lundy and Lewin (1960) evaluated the reinforcing effects of three degrees of therapist communication on the affectively toned verbalizations of their subjects. The group Intended to have virtually no communication from the therapist spoke into a tape recorder as if they were speaking to a psychotherapist.

PAGE 15

In the present study, one half of the subjects will speak into the tape recorder while the other half will wait an I dent I cat length of time. Since both groups will be given the Instructions that they are about to speak to a psychotherapist through the tape recorder this will permit a comparison between a group given both the promise and the gesture of help and a group given only the promise of help. Changes In mood and anxiety in the group which does speak nay be viewed as a demonstration of the nonspecific placebo effect in a single session of psychological treatment. The effects of this treatment should be utilized as a base line against which further research may be compared to demonstrate that an actual psychotherapeutic Interview produces changes different from, or greater than the "placebo psychotherapy." Shapiro (1964) reports that Important features of a therapeutic relationship which are responsible for the placebo effect are basically •laments of (I) the patient, (2) the doctor, and (3) the treatment situation; in addition to a fourth factor: the doctor-patient relationship. That is, there are certain characteristics of each of these components of treatment which contribute to a nonspecific placebo effect. (1) The patient Shapiro, in his review of drug research (1964), points out that many Individual differences of patients have been evaluated to assess possible relationships with the placebo effect. Such characteristics as sex, age, intelligence and diagnosis have been observed repeatedly In order to ascertain if a certain type of person responds more readily to placebo factors. However, as Shapiro reports, this body of research is equivocal and often contradictory. For example, Lasagna et al (195 1 *)

PAGE 16

7 administered a placebo to a group of postoperative patients to relieve reported pain. Those patients reporting a decrease In pain after receiving the placebo had a mean age five years greater than the age of the nonreactors (p ^ .05). On the other hand, Tibbetts and Hawkings (1956) found that patients who responded in two drug placebo situations were younger than the non res ponders. In studies by Kornetsky et al (1957) and Abramson et al (1955) no meaningful relationship existed between age and drug placebo responsiveness. Similarly, level of intelligence as a patient variable has produced contradictory results. In the Lasagna study reported above, intelligence did not significantly distinguish the placebo reactor from the non reactor. On the other hand, Tibbetts and Hawkings (1956) found the placebo reactor to be less intelligent than the nonreactor, while Abramson et al (1955) found the placebo reactor to have a higher verbal intellectual ability and a lower performance intellectual ability. Lelberman (196<0 reports that Eysenck has correlated neurotic! sm with the need for conformity in an Individual's perceptions and judgments. Therefore, he contends that "in group situations, where an element of interpersonal pressure activates the need for conformity, the subjects with neurotic tendencies respond with placebo effects." In fact, Eysenck argues that a subject's degree of neurotic! sm is a consistent factor for picking out high placebo reactors. Eysenck (1961) further speculates that since introverts form conditioned responses more quickly than extroverts, and since he reports that response to a placebo Is a learned or conditioned response, he expects introverts to respond more readily to placebo factors (p. 636).

PAGE 17

I However, despite the logical consistency of Eysenck's speculations, the literature concerning the placebo responsiveness of neurotics and introverts Is equivocal. For ex ample, Fischer and 01 in (1956), using psychiatric patients, demonstrated that neurotics were the roost responsive subjects in a drug placebo situation. On the other hand, Kurland (1953) disputes this conclusion on the basis of his finding that no difference exists between the placebo reactivity cT psychotics and nonpsychotics. Joyce (1959) in a drug study with medical students, indicates that extroverts are the most responsive subjects, while Laverty (1958) demonstrates that introverted subjects are the most responsive in drug treatment situations. In this study, age, intelligence, introversion and neuroticism are the patient characteristics whose contribution to the placebo effect wl 1 1 be evaluated. (2) The doctor. The nature of the treater is reportedly an important factor for encouraging or discouraging placebo effects (Shapiro, 1964; Uhlenruth at al .. 1959). For example, 611 daman et al (1957) demonstrated that a mala physician gained a 70 per cent cure rate in a treatment for which a female nurse could achieve only a 25 per cent cure rate. He also hat shown that if the administering physician is percoived as an expert healer, the response of the patient to treatment will be greater than if the physician is believed to be a medical quack. These effects were obtained, conceivably, because the patient maintains certain cognitive expectancies about the nature of treaters, and when these expectancies are

PAGE 18

9 dissonant with the treatment situation, the effectiveness of treatment decreases (>ee Festinger and Bramei In Bachrach, 1962). Similarly, folk medicine in the Spanish Southwestern United States (Jaco, 1958) and witchcraft healing in many parts of the world (Frank, I96I) flourish and prevent the establishment of modern medical program*, at ieast in part, because scientifically trained physicians do not meet ioca* expectancies about the nature of treaters. That Is, to these people scientific physicians are cognitive y dissonant In the ro a of "doctor." Conversely, the effectiveness of the native witch doctor is enhanced because he meets his clientele's expectancies concerning the nature of treaters. In the present research, verbal instructions to the patients will estab Ish two different types of doctors (psychotherapists). One I s the type of psychotherapist the patient expects and the other does not conform to his expectations (see Appendix A). This wi II produce two distinct groups of subjects regarding their personal perception of their psycho therapi sts. (3) The tr eatment situation Foulds (1958) and Shapiro (\36k) refer to many studies which demonstrate that a treater's bias can significantly Influence treatment outcome. For example, Hofling (1955) and Frank (1958), In studies, demonstrate that positive staff attitudes and biases toward a particular placebo or specific treatment increases the effectiveness of that treatment. Lyeriy et al. (i962) in a drug study, have shown that subjects receiving instructions aporopriate to the expected effects of chloral hydrate, regardless of the actual drug or placebo administered, reported subjective

PAGE 19

10 changes appropriate to the chiorai hydrate. Abramson et al (1955) administered tap water to a group of subjects who believed the liquid to be LSD-25. A percentage of these subjects responded to the tap water with psychomimetl c behavior as If the water actualy were LS0-25. Frank (l9ol) suggests that a patient's faith'' in a set of treatment operations is often ail that is necessary to produce positive change, and this Is so even when these treatment operations are scientifically absurd. Wolf (1959) states that a placebo is effective when a patient has a "conviction" that a certain effect will follow. In the present study, an attempt wl 1 1 be made to manipulate expectations concerning the va ue of the "piacebo psychotherapy" and the amount of profit to be anticipated from the treatment. One group will receive a positive set concerning that which can be gained from the treatment and the other will race! ve a mora neutral sat (sea Appendix A). To summarize, the hypotheses under investigation in this study are: I. Those subjects who receive the placebo psychotherapy (speak into the tape recorder) wi II show more marked change in a "therapeutic direction" on measures of anxiety and mood than those subjects who simply wait a comparable length of time. However, neither group will show a change in general level of psychological adjustment in this single treatment. II. The subjects who believe that their therapist is an individual congruent wi th their expectations will show more marked change in a "therapeutic "direction" on post measures of anxiety and mood than those subjects whose expectations are dissonant wl th the therapist's description. Neither group will show a change In general level of psychological adjustment after this single treatment.

PAGE 20

11 III. Those subjects receiving a positive set regarding the value and outcome of the treatment will show more marked change in a "therapeutic direction" on post measures of anxiety and mood than those subjects who receive a more neutral set. However, neither group will show a change in general level of psychological adjustment in this single treatment. IV. The measured Individual subject differences of age, intelligence, level of introversion and of neurotic! sm will not identify the subjects who respond best to this placebo situation. That is, there will be no differences between the placebo responsiveness of subjects In terms of their age, intelligence, level of introversion-extroversion and level of neurotic! sm. Hay the "placebo psychotherapy, as introduced above, be legitimately referred to as a placebo treatment situation? There is a strong analogy between this treatment and that which exists when a physician administers a drug placebo. The physician who treats a patient with a placebo pill does so in the same treatment setting, giving the same directions, and offering the same encouragement as If the pill were pharmacologically sound. The only changed characteristic, and that which defines the situation as exclusively nonspecific, Is the absence of the actual drug— that agent in the treatment situation which produces the specific effect. Reactions to this placebo are recorded as "placebo effects" although they may appear to be a result of a specific drug which is clearly absent from the treatment. Similarly, In a sham interview such as that mentioned above, the patient is given the same directions, the same encouragement and is placed in the same treatment setting as if it were an actual psychotherapeutic situation. The only changed characteristic, and that characteristic which defines the situation as nonspecific in this case, Is the absence of the psychotherapist, that is, the only agent

PAGE 21

II through which specific effects may result. Reactions to this placebo situation can be viewed as "placebo effects." Without knowledge of the nature of the treatment, the effects may appear to result from the specific behavior or techniques of a psychotherapist. However, his absence from the treatment necessarily means that all effects are nonspecific, or placebo. Since there is no psychotherapist present, can this situation be thought analogous to a therapeutic interview: The patient is talking to a therapist who exists only in the patient's fantasy and is instituted by the experimenter. Apfelbaum (1958) points out that individuals tend to create, in part, interpersonal experience and that ongoing interpersonal process in therapy is primarily a function of a patient's personal transferences. A patient, in utilizing past experience and partial cues, projects the Image of his psychotherapist, and thereby unknowingly perceives him in part through his projective capacity. Consequently, the therapist is perceived largely as a function of the patient himself. The attitudes a patient holds toward his alleged therapist In the "sham" or "placebo" interview suggested above may be controlled and manipulated by experimental design. The therapist exists entirely in fantasy and as a function of the patient's projective capacity and description of him as reported by the experimenter. The patient, however, in effect, establishes an intrapersona) relationship which he believes to be, and which appears to have the qualities of, an interpersonal psychotherapeutic relationship. Thus, the sham interview wi 1 1 be regarded as both a placebo treatment and a therapeutic interview.

PAGE 22

KCTNOB The study Involved 96 subjects who were placed into eight treatment conditions as diagrammed below In Table 1. TABLE 1 THE EXPERIMENTAL DESIGN Experimental Group C ontrol Group Congruent Incongruent Congruent Incongruent Therapist Therapist Therapist Therapist Instructions Instructions Instruct! ons Instructions PosI ti ve set concerni ng treatment outcome n 12 n 12 n 12 n 12 Neutral concerni ng treatment outcome n 12 n 12 n 12 n 12 n 48 n i8 The technique and procedure of establishing these treatment conditions, and the way in which the data were collected are fully explained in the Procedure section. The tests which were administered to the subjects and the use to which the test information was put in the design are described in the Means of Measurement section.

PAGE 23

Subjects The subjects for the study were 96 neuropsychiatri c hospital patients in the Gulfport, Mississippi, Veterans Administration Hospital. They are all in-patient residents of the Rehabi titation and Placement Service Ward which houses only those patients judged capable of returning to the community. The ward offers a vocational testing service and is professional ly staffed by a psychiatrist, two counseling psycho ogi sts and a number of nursing personnel. The 96 subjects were drawn from the ward during 48 separate sessions extending from August 29, 1964, to October 19, 1964. The sample consists of white adult males. There is a wide range of prehospi ta I zation vocations. None were in psychotherapy when seen as subjects for the study. However, it was known that some of these patients had bean in group or individual psychotherapy earlier In their hospitalization. The identification of former psychotherapy patients was not made, but there is no reason to suspect a nonrandom assignment of these patients to the different experimental treatment conditions. Similarly, although It was known that some of the samp.e were receiving drugs for their psychiatric conditions, these subjects were not identified and their random assignment to the different experimenta* treatment conditions was expected. The mean age of the sample was 38.76; range 18 to 58, standard deviation 7-4l. Below is a summary table of an analysis of variance (Lindqulst, 1953) demonstrating that subject ages are randomly distributed across all treatment conditions. The mean I .~0_. of the sample, as measured by the Beta Intelligence Test (Kellogg et al .. 1946) was 98.54, which is within the normal range.

PAGE 24

IS TABLE 2 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANDOM DISTRIBUTION OF AGE Source a df v V f m vexpen raont a i vs • cone ro i / Zp ul i O.UI hit. o rufsnv vs i nuony ruent/ 7f l J. /o • uy C (positive vs. neutral) 1^7.51 1 147.51 2.56 AB 25.01 1 25.01 .44 AC 11.35 1 11.35 .20 BC 25.01 1 25.01 .44 ABC 3.76 1 3.76 •07 w Cel Is 5.030,0$ 57.16 Total 5,271.49 95 (An F of 3.95 is significant at the .05 level.) The range of I.Q.'s Mas from 73 to 129; standard deviation 12.14. Below is a summary table of an analysis of variance demonstrating that I .Q. is randomly distributed across all treatment conditions. The mean educational level of the sample is 10.9 years* and ranged from 6 years to 20 years. The mean of the length of current hospitalization to the closest month is 27 months and ranged from three days to 14 years. The psychiatric diagnoses carried by these patients were made by a psychiatrist at the time of admission to the hospital. There is a wide range of psychiatric diagnoses, which, for the purposes of this study are divided into 5b psychotic and kO nonpsychotlc. Because all subjects were randomly assigned to the treatment conditions, there Is

PAGE 25

TABLE 3 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANDOM DISTRIBUTION OF I.Q. Source ss df V F A (experimental vs. control) 96.00 1 96.00 .62 B (congruent vs. > Incongruent) .37 1 .37 .00 C (positive vs. neutral) 2.04 2.04 .01 AB 165.38 165.38 1.07 AC 3.37 3.37 .02 BC 66.67 66.67 .43 ABC 140.17 140.17 .90 w Cells 3,657, m 155.20 Total 14,131.83 95 (An F of 3.95 is significant at the .05 level.) reason to expect that educational level, length of hospitalization and diagnosis are randomly distributed across treatment conditions. The psychotic patients all have shown good remission of their symptoms and all patients serving as subjects in the study were judged by a psychiatrist to be sufficiently psychologically sound to return to the community and be employed. Weans of Measurement Two tests upon which change in feelings of personal comfort can be monitored (the Anxiety Adjective Check List and the Mood Adjective Check List), and one test evaluating general level of psychological

PAGE 26

17 adjustment (The Adjustment Score) were administered twice to each subject — once prior to and once immediately after the treatment. Changes in scores on these tests were u*ed to evaluate changes as described in the four hypotheses. CO The An xiety Adjective Check List (AACL) (see Appendix B) The Anxiety Adjective Check List (AACL) (i960) has bean shown by Zuckerman to be a "quick measure of anxiety I aval." It was chosen for this study because of its demonstrated validity and reliability and the fact that the instructions may be modified to permit sensitive evaluation of changes in the lave) of manifest anxiety over short periods of time. The check list is an empirically developed pool of 61 adjectives with varying affective connotations. Twenty-one adjectives ware identified, on the basis of an I tarn analysis, as discriminating between psychiatric patients rated high on anxiety and normal control subjects. Of these items, eleven are anxiety negative adjectives and ten are anxiety positive. Subjects may obtain a score from 0 to 21 on the check list. Zuckerman has developed two forms. One instructs subjects to denote how they feei "In general and the other asks how they fee "today." Although standardized on co lege students, Zuckerman reports no differences In performance due to age, sex or leva! of education. In two studies ( 960, 1962) ZucKerman reports interna! and testretest reliabi Hties for the "In general form to be significant at greater than the .001 level. However, the "today" form, although similarly Internally reliable, had low test-retest reliability (r .68, p < .oOi). These results were anticipated and suggest that the today form is sensitive to short-term fluctuations in anxiety. Validity was

PAGE 27

evaluated by administrations of the today form to a class in general psychology. The check list was given on 10 nonexami nati on days and three examination days. The examination-day administrations resulted in a significantly higher reported level of anxiety than the nonexami nati on-day administrations (p<.0005). The today form, with instructions modified to read "right now," will be used in this study. (2) The Hood Ad jective Check List ( HACL) (see Appendix C) Nowlis (1956, 960, I96I) assumed mood to be a multidimensional characterization of a person's feeling or behavior which is accessible to self report. The HACL resulted from a series of seven factor analytic studies by him (i960). In those seven studies Now! Is demonstrated that 4u adjectives, from his original pool of 200, have relatively consistent and high loadings on twelve separate factors. These kO adjectives constitute the HACL which is scored separately for each of the twelve factors (which are the identified moods). Further, the twelve factors (moods) are divided into three groups: those "factors most consistently identified," those "factors identified fairly consistently" and "factors tentatively identified." Below are the twelve mood factors and the particular adjectives constituting the factors. Certain specific moods on the HACL have particular relevance for the purpose of this study because they would be expected to change in a certain direction as a result of successful psychotherapy. Aggression and Deactivation would be expected to decrease, whi le Social Affection would be expected to increase. Similarly, after successful psychotherapy the moods of Anxiety and Depression would be expected to decrease and Pleasantness and Activation would be expected to increase. At a iower

PAGE 28

IS TABLE k THE TWELVE HOOD FACTORS OF THE MACL AND THE ADJECTIVES CONSTITUTING EACH HOOD FACTOR Adjectives with high loadings on the factor Factors Most Consistently Identified: AGGRESSION CONCENTRATION DEACTIVATION SOCIAL AFFECTION Factors Identified Fairly Consistently: ANXIETY DEPRESSION EGOTISM PLEASANTNESS Factors Tentatively Identified: ACTIVATION NONCHALANCE SKEPTICISM STARTLE angry, bold, defiant, rebellious, concentrating, earnest, engaged in thought, serious, drowsy, earnest, tired. affectionate, forgiving, kindly, warmhearted. apprehensive, clutched up, fearful, Insecure, blue, lonely, regretful, boastful, cocky, egotistic, self -centered, elated, I ighthearted, overjoyed, pleased. active, energetic, vigorous, nonchalant, playful, witty, skeptical, suspicious, startled, shocked. level of certainty, because of difficulty in Interpreting its meaningfulness in a psychotherapeutic context, Concentration and Nonchalance might be expected to increase, while Egotism, Skepticism and Startle might be expected to decrease after treatment. In addition, changes in Skepticism after the treatment may reflect the extent to which this treatment was beI i eved to be real In his research developing the MACL, Now! is used male and female college students and U. S. Navy personnel In a variety of experimental (drugs, films) and field situations. For example, one validation procedure that Nowlis reports (1961) involved the administration of a drug which

PAGE 29

20 has known effects on a person's feelings and moods. Continuous monitoring of mood on the MACL for 16 hours demonstrated that fluctuations in the MACL were related to anticipated druginduced changes. Without reporting actual correlational figures, Nowlis indicates that concurrent validity of the MACL is high, although the test-retest reliability Is low. Nowlis describes temporality as one of the characteristics of a mood and therefore, b*csuse of the low reliability of the check list, he suggests that the scale is sensitive to short-terra fluctuations in mood. Because of its short length and ease of administration, Nowlis points out, It Is possible to monitor mood with this instrument over repeated intervals of any required length. The subject is asked to circle one of four options for each word in the check list: a double plus if the adjective definitely describes a current strong feeling, a single plus if the word slightly applies to a current feeling, a question mark if the word Is unclear or the subject Is unable to decide if the adjective describes a current feeling and no if the subject Is certain that the word does not describe a present feeling. Double plus is scored as 3. single plus Is scored as 2, question mark is scored as 1, and no is scored as 0. Twelve mood scores are obtained by summing the scores of the individual adjectives in each of the moods. The scores may range from 0 to 12 on the moods of Aggression, Concentration, Social Affection, Anxiety, Egotism and Pleasantness. The scores may range from 0 to 9 on the moods of Deactivation, Depression, Activation and Nonchalance. On the moods of Skepticism and Startle the scores may range from 0 to 6.

PAGE 30

21 (Both the AACL and the Anxiety Hood on the MACL reportedly measure level of manifest anxiety. The MACL Anxiety mood contains four affectively toned adjectives, each of which the subject responds to by denoting the extent to which the adjective describes his current feelings. The AACL, on the other hand, asks a subject to check those adjectives that apply to his current feelings. Only one of the four adjectives on the MACL Anxiety mood is identical with the 21 adjectives which are scored on the AACL. A comparison between pretreatment performance on the two Anxiety measures resulted In a Pearson product-moment correlation of +.64, and a Pearson product-moment correlation between the subject's changes in the two scales from preto post treatment testing resulted in a correlation of +.57Both these correlations are significant beyond the .001 level. This suggests that despite the basic uniqueness of the two measures, they are measuring essentially the same factor of "anxiety.") (3) The Adju stment Score, (see Apsendix D) The Adjustment Score was chosen to evaluate changes in general Uvl of psychological adjustment. It was originally constructed by Dymond (1953) as a Q-sort to evaluate the effects of nondi recti ve counseling. The statements were compiled empirically by Butler and Haigh who noted that they were representative of the positive and negative comments made by people in psychotherapy. Four Wei I -trained, practicing clinical psycho I ogi sts" who were not client-centered by theoretical orientation "agreed remarkably well" in choosing 37 negative indicators of adjustment and 37 statements describing positive adjustment (1953). These 7^ statements rated on a "like ma" to unlike me" dimension constitute the scale.

PAGE 31

Dymond 095*0 demonstrated that a significant difference existed between the mean adjustment scores on the Q-sort of subjects in two months of cl lent -centered psychotherapy as compared to control subjects who waited a comparable length of time (p ^.01). Dymond reports that the test-retest reliability of this control group Is +.86. Validity was established in two ways: (I) rank order correlations between self -ideal correlations and the Q-sort adjustment score of clients before therapy began was +.83, and the rank order of these same subjects after therapy was +.92; (2) each therapist rated the success of his therapy for each client and these ratings correlated at oetter than the 5 per cent level with the subject's own scores on the Q-sort For the purpose of the present study, the 7k items on the Q-sort were randomly organized into a questionnaire. The subject is asked to mark an item "true" if It pertains to him and mark "false" If the Item does not pertain to him. A total of the positive items marked true and negative items marked false constitute a subject's adjustment score. The AACL, HACL and the Adjustment Score were administered twice to all subjects, once prior to and once immediately after the treatment. The change in scores frompre-to posttreatment testing constitute measured changes in manifest anxiety, 12 independent moods and general level of psychological adjustment. (*0 The Maudsl ev Personality Inventory (MPl) In order to evaluate part of Hypothesis IV, the MPI (Eysenck, 1959, 1961) was administered to obtain subject differences In level of neuroticism and degree of introversion. "Extroversion, as opposed to Introversion, refers to the out-going, uninhibited, social proclivities of

PAGE 32

23 a person," and neuroticism represents "the general emotional liability of a person, his emotional overresponsi veness, and his liability to neurotic breakdown under stress" (1959, p. 3). The Inventory has 48 questions which have been factor analyzed to yield scores on an extroversionintroversion dimension and a neuroticism dimension. Although the two dimensions are slightly negatively correlated (-.15), Eysenck submits that they are essentially orthogonal. The original standardization took place on a sample of 200 normal English men and 200 normal English women. Eysenck has since supplemented this by adding groups of students, nurses, industrial apprentices and "a quota sample of the whole population." Data are also available on groups of hospital patients and prisoners. Split-half and Kuder-Rlchardson reliability coefficients for the Neuroticism scale fall between .85 and .90 and for the extroversion scale fall between .75 and .85. Using both concurrent and construct validation techniques, Eysenck reports that the MPI has yielded data supporting its validity. (5) A Modification of the Ex pectation Q-sort (Apfelbaum, I 958) (see Appendix E) The Q-sort asks a subject to indicate his expectancies regarding the nature of a psychotherapist who may be assigned in the future. Thirty-six of the 60 items in Apfelbaum 1 s item pool were chosen and randomly organized into a questionnaire. These 36 items constitute those with heaviest factor loadings on three separate clusters: (1) those items which designate a guiding, giving, protective therapist; (2) those items which designate a tolerant, accepting and permissive therapist; and (3) those items which designate a cold, rigid and condemnatory therapist.

PAGE 33

In addition to the five tests, each subject was also asked to complete a questionnaire: asking for personal information (see' Appendix F). Each subject also had taken the Beta Intelligence Test (Kellogg et al .. 1946). Procedure Subjects were chosen from the ward by a secretary who was naive concerning the mechanics and goals of the study. She was informally questioned after the completion of the study and no bias In the subjects she chose, and her order of choosing them, was noted. The subjects were seen two at a time. Once the subjects were seated in the testing room they were given instructions which drew their interest to the tasks about to be presented but did not disclose the nature of the research nor the fact that it was research (sea Appendix A). They were given the following tasks in the order noted: 1. Persona) information sheet 2. Mauds ley Personality Inventory 3. Anxiety Adjective Check List k. Hood Adjective Check List 5. Expectation questionnaire 6. The Adjustment Score These tasks were relatively short, took no longer than kO minutes to complete and subjects did not appear to become fatigued. After completing the six questionnaires the pair of subjects was told that this was an opportunity to speak with a psychiatrist, that he was unable to get to the hospital and that ha requested they speai^ Into a taoe recorder, the tape of which would be sent to him by registered mall (see Appendix A for exact wording)

PAGE 34

25 By a coin toss prior to the meeting between subjects and Experimenter, one subject of the pair was to be placed in the Experimental Group and the other of the pair was to be placed in the Control Group (treatment vs. wait). Two additional coin tosses by the Experimenter prior to the meeting determined whether this particular pair of subjects would receive (1) the positive or neutral set, and (2) the congruent or incongruent therapist instructions. Both subjects were then given the instructions appropriate to their designated treatment category (i.e., (1) positive set, congruent therapist; (2) positive set, incongruent therapist; (3) neutral set, congruent therapist; and (4) neutral set, Incongruent therapist [see Appendix A]). The Experimenter had sufficient time while the subjects were filling in the Adjustment Score, to choose the predetermined Items on the Expectation Questionnaire and manipulate the congruence and incongruence of the subject's alleged psychotherapist (see Appendix A) The Control Group subject of the pair was then asked to remain in the testing room. He was asked to relax, was offered the use of a stack of magazines and was told that the Experimenter would return in about 50 minutes (see Appendix A for exact wording). The Experimental Group subject was than taken to the room which contained the tape recorder. This room had a one-way vision screen and was wired for sound. The subject was seated in an easy chair, instructed to speak to the "psychiatrist" and was told that the Experimenter would return In about 50 minutes to retrieve the tape and the tape recorder was turned on (see Appendix A for exact wording). The Experimenter then seated himself in an observation room where he was able to record the Experimental Group subject's

PAGE 35

2o verbalizations and to observe that the Control Group subject remained in his room. Record was made of the amount of the subject's speech, and in several cases the content of the subject's verbalizations. With a stop watch, amount of speech was rated as follows: constant speaker, no longer than 3-minute pauses; moderate speaker, no longer than 5-minute pauses; little speaker, at least some verbalization; no speaker, says nothing. At the end of 45 minutes the Experimenter turned off the tape recorder and guided the Experimental Group subject back into the testing room. Both subjects were then asked to "retake several of the shorter tests. The fol lowing were readminl stered: 7. Anxiety Adjective Check List 8. Mood Adjective Check List 9. The Adjustment Score At this point, the collection of the data was completed. The Experimental Group subject was dismissed. If the subject asked about feedback, he was told that he would receive some type of feedback within several weeks. He was also encouraged not to tell anyone the nature of what had happened because the Experimenter would be unable to see everyone for this project. The Control Group subject was then asked If he wished to "speak to the psychiatrist." If he agreed, he was given 45 minutes with the tape recorder, however, if he declined, he was dismissed. Twenty-one of the 48 Control Group subjects chose to speak. If any subject discussed suicidal or homicidal content, mentioned intensely pressing issues or made special requests, the Experimenter promptly called this Information to the attention of the ward psychiatrist. This occurred in twelve cases.

PAGE 36

After all the data Mere collected, all subjects who spoke into the tape recorder were sent a letter (see Appendix 6). The purpose of this letter was to thank the subject for his cooperation and Inform him of the way In which he could contact the Experimenter for further clarification of the procedure. Twelve of the 69 subjects who spoke Into the tape recorder contacted the Experimenter as a result of this letter.

PAGE 37

RESULTS Analysis of covariance (Ray, I960) was used to evaluate the testability of hypotheses I, II and III (those hypotheses concerned with the effects of differential experimental treatment). In analyzing the three hypotheses, the eight independent treatment conditions, as diagramed in Table I were considered. Dependent variables in hypotheses I, II and III were the changes in scores from preto post treatment testing on the AACL, the 12 moods In the MACL and Oymond's Adjustment Score. Thus, evaluation of the first three hypotheses was made across the three independent variables utilizing \k 2 x 2 x 2 analyses of covariance— one for each dependent variable. The technique of analysis of covariance chosen to evaluate hypotheses I, II, and III utilizes the fact that differences between the treatment groups on a dependent variable post treatment measure may be to some extent a reflection of differences between these groups on the initial pretreatment measure. It removes the variance due to these initial random differences from the final posttreatment variances, thus reducing the size of the error variance and increasing the precision of the experiment (Gourlay, 1953). That is, by utilizing the regression of the posttreatment measures on the pretreatment measures, an adjustment Is made for the variability which is associated with initial differences among the subjects. After this adjustment is made, the remaining variability may be analyzed with a more precise estimate of factors which My have produced an effect. 28

PAGE 38

29 By inspection, the data appear to have significant homogeneity of variance. Norton (as cited iy Lindquist, 1953) demonst rates that unless the heterogeneity of variance is so extreme that it is readily apparent upon inspection, the effect upon the F distribution will be negligible. No formal tests of homogeneity of variance were conducted. Statistical significance was set at the .05 level prior to the analysis. Below are the summary tables of the analyses of covariance for each of the dependent variables (Tables 5 through 18). In addition, the mean change scores frompre-to posttesting In each treatment condition are presented for each of the dependent variables (Tables 5A through 18A). Hypothesis I states that the Experimental Group (placebo treatment group) will show more positive therapeutic changes" In the AACL and the moods of the MACL than the Control Group (wait group). This prediction was supported in the analyses of the main effects of three of the fourteen dependent variables. On the AACL, the Experimental Group showed a greater decrease in reported anxiety after treatment than did the Control Group (p< .05) (see Table 5). On the MACL Social Affection mood, the difference was significant between the increase in Social Affection for the Experimental Group after treatment and the decrease after treatment in Social Affection for the Control Group (p
PAGE 39

3i TABLE 5 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE ZUCKERMAN ADJECTIVE CHECK LIST Source ss df v F A (experimental vs. control; I.I. 1. A 44.42 III. 1.1 44.42 6.29* B (congruent vs i ncongruent) 98.55 98.55 C (positive vs. neutral) .00 .00 .00 AB 69.98 69.98 9.91** AC 6.63 6.63 BC 17.67 17.67 2.50 ABC 17.67 17.67 2.50 error 611*22 SZ 7.06 Total 869.96 9^ *** s significant 4. .001. ** a significant < .01. a significant < .05. TABLE 5A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X,) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE ZUCKERMAN ADJECTIVE CHECK LIST Exoeri mental Grouo Control Grouo Congruent Therapj s $ 1 ncongruent Therapist Cong ruent Thefapi s{ 1 ncong ruent Therapi st Positive Set X, a 11.08 X 2 = 7.00 D a -4.08 X, a 8.92 X 2 10.59 0 a +1.67 X, m 10.33 X. a 8.41 x 2 iJil 0 a .00 Neutral Set J, s 10.08 H 7r33 D -2.75 L 9.75 X 2 2*08. 0 a -.67 l\ 11.33 X 2 a 11.16 D a -.17 Zl 9.^1 x 2 3.83 D a +.42

PAGE 40

51 TABLE 6 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL AGGRESSION MOOD Source ss df V F A (experimental vs. control) 8.40 8.40 2.00 B (congruent vs. incongruent) 7-57 7-57 1.81 C (positive vs. neutral) .10 .10 .02 AB 31,84 31.84 7.60** AC 1.94 1.96 .47 BC 4.66 4.66 1.11 ABC 5.43 5-43 1.30 error 364.2* 4. 19 Total 424.25 94 ** significant < .01. TABLE 6A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X,) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL AGGRESSION MOOD Experimental Grouo Control Jr-ju) Congruent The rap i st Incongruent The rap i st Congruent Therapl st 1 ncongruent Therapist Posi ti va Set \\ 3.33 *2 "1M D -1.50 X, 3.42 x 2 1M 0 + .25 X, 3-50 x 2 4.50 D = +1.00 I, 2.33 X 2 2.00 0 -.33 Neutral lat X, 2.17 h LIS D -.92 X, 2.42 0 +.75 X, 2.67 X 2 l& D .00 X, 2.08 x 2 0 +.42

PAGE 41

32 TABLE 7 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL CONCENTRATION MOOD Source ss df V F A (experimental vs. control) 1.81 1 1.81 .49 B (congruent vs incongruent) 6.59 1 6.59 un C (positive vs neutral) 1.12 1.12 .30 AB 4.75 4.75 1.28 AC 6.14 6.14 1.65 BC 3.87 3.87 1.04 ABC .45 .45 .12 error 323-02 SZ 3.71 Total 347.75 94 TABLE 7A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL CONCENTRATION MOOD Experimental Group Con tro l_Group_ Congruent Incongruent Congruent Incongruent Therapist Therapist Thera pist Therapist Positive X, 8.75 X, 10.08 X, 9.50 X, IO.33 Set X 2 £u31 X 2 9.67 X 2 8^0 X9-42 D +.17 D -.41 D -1.00 D -.91 Neutral Set I 11.08 *2 n -2S D +.17 X, = S.k2 0 m -I.I7 x 2 D 9.17 .00 x, 10.17 D -.hi

PAGE 42

33 TABLE 8 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL DEACTIVATION MOOD Source ss df V F A i txoe r I men ta 1 vs. control ) 7 82 7 82 1 67 B Iconoriipnt \/< I nrrtnarnAnt' 1
PAGE 43

34 TABLE 9 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL SOCIAL AFFECTION MOOD Source ss df V F A fcA ^fc Afc I AAA A* A 1 m \experi (i^nta i v 9 • con v ru i / 4/ 1 J 97 1 c H. 5 J ft 1 rr\nnfiufenf w c D VL^kJllyl Uclll V9 14 v/ 14 ^7 it. 2 "id C (posi ti ve vs. neutral) .03 • 03 .001 AB 19.62 19.62 3.14 AC j. ': 10.84 K74 3C 1.36 1.36 .22 ABC 22. 6a 22.60 3.62 error 54J. 18 6.24 Totai 639.15 94 *Signlf leant < .05. TABLE 9A PRETREATMENT MEANS (X.), POSTTREATMENT MEANS (x" 2 ) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL SOCIAL AFFECTION MOOD Experimental Grouo Control Ira m Congruent Therapi st 1 ncongruent Therapi st Congruent Therapi s% 1 ncongruent Therapi st Posi tl ve x, 6.17 X, 6.25 X, 6.08 X, 8.00 Set x 2 JJM D + T767 0 -1.25 x 2 = 0 -.58 D -.17 Neutral Set Xt 7.58 X~2 M2 D +.84 Xi 7.75 X 2 8.08 D +.33 x, 7.58 D -.S3 X, 6.42 x 2 "lilS D -1.17

PAGE 44

ss TABLE 10 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL ANXIETY MOOD Source ss df V A (experimental vs. control) 35.35 1 35-35 6.29* B (congruent vs incongruent) 4.27 1 4.27 .76 C (posi ti ve vs. neutral ) .00 1 .00 .00 AB 6.31 6.31 1.12 AC 4.00 4.00 .71 ti .07 .07 .01 ABC 8.33 8.33 1.48 error 488.92 6Z 5.62 Total 547.25 94 Significant < .05. TABLE 10A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL ANXIETY MOOD Experimental Grouo Control Group Congruent Therapist 1 ncongruent Therapi st Congruent Theraoljj 1 ncongruent Therapist; Positive Set X, 5.25 x 2 3J5 D -TTso X, 4.92 X, 5*08 D +7T5' X, 4.50 X 2 5-U8 d T77H X, 4.33 x 2 kj£ 0 .00 Neutral Set X, 5.84 X 2 4.42 0 -1.42 X, 3.83 X, 5.67 X 2 ^84 D +.17 X| 4.00 X 2 5.00 D +1.00

PAGE 45

36 TABLE 11 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL DEPRESSION MOOD Source ss df V V A (experimental vs. control) 3.04 1 3.04 .78 B (congruent vs. I n congruent) .00 1 .00 .00 C (posi tl ve v*. neutral) M M .13 AB .. .86 .22 AC 1.24 1.24 .32 BC -37 .37 .10 ABC 5-55 5-55 1.43 error 3.88 Total 39^.06 94 TABLE 11 A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (xJ AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL DEPRESSION MOOD Expert men tal GrouD Control Group Congruent Therapist Inconnruent Therapist Congruent Therapl st 1 ncongruent Therapist PosI tl ve Set X, 3.92 X 2 = 2J&Z 0 -1.25 X, 3.75 D -.83 ?! 3-58 x 2 3.66 D +.08 X, 3.75 x 2 2.83 D -.92 Neutral Set X, 5.33 X 2 4.42 0 -.91 X, 4.17 X2 ~U1 D -.84 X, 4.25 0 -.83 X, 3.08 D -.16

PAGE 46

37 TABLE 12 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL EGOTISM MOOD Source ss if 1 F A (experimental vs. control) 4.53 4.53 1.09 B (congruent vs. Incongruent) 2.10 | 2.10 .51 C (oosltlve vs. neutral) 3. 19 3.19 .77 AB .57 .57 .14 AC 11.58 11.58 2.80 BC 1.33 1.33 .32 ABC 2.60 2.60 .63 error 359.86 8Z 4.14 Total 385.76 94 TABLE 12A PRETREATMENT MEANS (Xj), POSTTREATMENT MEANS (Xo) AND MEANS CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE osriGr? sr. TP r macl egotism mood Experimental Group Control Group Conn r uen t Therapl st 1 ncongruent Tharapl st Congruent Therapl st Incongruent Therapist Posi tlve Set X, 3-08 x 2 2JL0 0 -.58 X| 3-25 X 2 2.08 D -1.17 X, 3.00 X 2 3.08 D +.08 Xj 2.67 D +.58 Neutral Set X, 2.58 D +.50 X, 2.50 X 2 2J6Z D +.17 X, 2.58 X 3.00 D +.42 X, 2.42 x 2 hJl D -.25

PAGE 47

TABLE 13 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL PLEASANTNESS MOOD Source ss df V F A (e^ )er i roe ita vs. control) 8.12 8.12 1.21 B (congruent vs. i nconqruent) 39.31 39.41 5.89* C (positive vs. neutral } 43.51 43.51 6.50* At .01 .01 .001 AC .02 .02 .003 BC 3.57 3-57 • 54 ABC 13.87 13.87 2.07 error 581. 76 iz 6.69 Total 690.17 94 Significant < .05. TABLE 13A PRETREATMENT MEANS (X.), POSTTREATMENT MEANS (X,) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL PLEASANTNESS MOOD Eweri mental Grouo Control Brouo Congruent Therapist 1 ncongruent Therapi st Congruent Therapist 1 ncongruent Therapist Posi tive Set Xj 4.25 X 2 6.67 0 +2.42 X, 4.33 x 2 itJi 0 .00 Xj 4.42 X 2 5.50 D +1.08 X, 4.33 X 2 iL^O D +.17 Neutral Set X, 4.17 X 2 4J2 0 .00 X, 4.75 X 2 itJ42. 0 -.33 X, 4.33 X4.42 D +.09 x, 4.17 x 2 1M 0 -1.50

PAGE 48

39 TABLE 14 SUMMARY TABLE OF THE ANALYSIS OF CO VARIANCE FOR THE MACL ACTIVATION MOOD Source ss df V F A (experimental vs. control) 2.90 1 2.90 .80 B (congruent vs. tncongruent) 1.80 1.80 .50 C (posi tl ve vs. neutral) .20 1 .20 .06 At, 5.58 5-58 1.54 AC .47 .47 .13 BC .00 .00 .00 ABC .12 .12 .03 error 314.94 £Z 3.62 Total 326.01 94 TABLE 14A PRETREATMENT MEANS (X.) POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL ACTIVATION MOOD Experimental Group Congruent Therapi st I ncongruent The rap I st Control Group Congruent Therapist I ncongruent Therapist Posi tl ve Set 4.25 • -.25 X| 4.00 E> -.33 X, 4.42 x 2 3.92 D -.50 Neutral Set Xj 3.75 x 2 LIS a -.25 X, 3.58 lag -.50 X, 3.92 X, 4.58 1*2 D -.34 D +.34

PAGE 49

4o TABLE 15 SUMMARY TABLE OF THE ANALYSIS OF CO VARIANCE FOR THE MACL NONCHALANCE MOOD Source a v f A (experimental vs. control) .87 ] .87 .23 B (congruent vs Incongruent) 1.43 1.43 .38 C (posi ti ve vs. neutral ) .17 I .17 .04 AB 3.52 3.52 .92 AC .24 ,2k .06 BC MM 8.46 2.22 ABC 37 .37 .10 error 331.90 8Z 3.81 Total 346.96 94 TABLE I5A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X,) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL NONCHALANCE MOOD In Mwlnant junta Contirol Group Congruent Therapist 1 ncongruent Therapl st Congruent Therapist 1 ncongruent Therapi st Posi tive Set X, 3.08 D = +.17 X, = 3.42 X 2 2^2 D = -.50 X, 3.83 X 2 3.33 0 -.50 X, 2.08 x 2 JU25. D -.33 Neutral Set X, 2.00 x 2 juaa D -.17 X, 3.08 X1,08 D .00 X, 3-58 x 2 1^8 0 -I .00 X, 2.00 h "i*58 0 +.58

PAGE 50

41 TABLE 16 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL SKEPTICISM MOOD Source ss df V r A (experimental vs. control) 1.98 1 1.98 .74 B (congruent vs. incongruent) .*9 1 .*9 .18 C (post tl ve vs. neutral) .10 1 .10 .04 Ho JO 00 AC 2.17 2.17 .81 BC .00 .00 .00 ABC .01 .01 .004 error 231.87 IZ 2.67 Total 236.98 94 TABLE I6A PRETREATMENT MEANS (X. ) POSTTREATMENT MEANS (X £ ) AND MEAN CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL SKEPTICISM MOOD Experimental Group Control Group Congruent The rap I st Incongruent The rap I st Congruent Therapist 1 ncongruent Therapist Posi ti ve X, 1.92 X, 2.75 X, 2.75 X| 1.67 Set x 2 -LZ5 0 -.17 X 2 2^ D -.25 X 2 2.33 D -.42 D +.08 Neutral X, 2.00 x, 2.17 X, 3.00 X 2 2J& D -.17 x, 1.75 Set H -Uil 0 -.26 x 2 2J2 D +.42

PAGE 51

TABLE 17 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE MACL STARTLE MOOD Source 11 at V r A (experimental vs. control) .03 .03 .06 B (congruent vs. incongruent) 13.47 13.47 26.94*** C (positive vs. neutral) 2.67 2.67 5.34* Aft rWi no • U*r o4 j*. 1 AC .37 .37 .74 BC 1.47 1 l.**7 2.94 ABC .03 1 .03 .06 error 43.70 8Z • 50 Total 61.78 94 ***Slgnif leant < .001. **Signi f icant < .05. TABLE 17A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE MACL STARTLE MOOD Experimental Groups Control Grou > Congruent Incongruent Therapist Therapist Congruent The rap 1 st 1 ncongruent Therapist Positive X, 1.42 X, 1.17 Set X 2 liOO X 2 1.83 D -.42 D +.66 *l H D 1.50 1.00 -.50 X, 1.25 x 2 1.67 0 +.42 Neutral X, = .92 X, .92 Set X 2 1.00 X 2 1.50 0 +.08 D +.58 II v 1.00 LJLZ +.17 X| 1.25 x 2 1.92 0 +.67

PAGE 52

43 TABLE 18 SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE ADJUSTMENT SCORE Source ss df V F A (experimental vs. control) 40.12 1 40.12 3.04 B (congruent vs. i ncongruent) .07 1 .07 .005 C (posl ti ve vs. neutral) 20.84 1 20.84 1.58 M .16 .16 .01 AC 33-34 33.34 2.52 BC 3.77 3.77 .29 ABC 9.98 9-98 • 76 error 1,149.70 8Z 13.21 Total I, 257.98 94 TABLE 18A PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X* 2 ) AND MEAN CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE DESIGN ON THE ADJUSTMENT SCORE Experimental Croup Control Group Congruent I ncongruent Congruent I ncongruent Therapist Therapist The rap I st Therapist Positive X, 45.42 X, 43.17 X, • 39.67 X, 44.83 Set X 2 45.75 X 2 42.42 X, 41.83 X, 47.25 D +.33 D +.75 D +2.16 0 +2.42 Neutral X, 42.00 X, 46.50 X, 37-17 Xj 49.58 Set X 2 41.58 X 2 47.00 X 2 37.50 X 2 49.58 D -.42 0 +.50 D +.33 D .00

PAGE 53

i*4 Changes after treatment on the AACL indicated that the Congruent Therapist, Experimental Group and the Incongruent Therapist, Control Group significantly reduced anxiety whan compared to the Incongruent Therapist, Experimental Group and the Congruent Therapist, Control Group (p ^ .01) (see Table 5). Similarly, the MACL Aggression mood Indicated that the Congruent Therapist, Experimental Group and the Incongruent Therapist, Control Group significantly reduced aggression when compared to the Incongruent Therapist, Experimental Group and the Congruent Therapist, Control Group (p ^ .01) (see Table 6). These interactions suggest that as measured on the AACL and the MACL mood of Aggression, the superiority of the Experimental Group treatment pertains only for those subjects who are given the congruent therapist Instructions, while Experimental Group subjects given the inoifigruent therapist instructions report "nontherapeutic" changes While only three dependent variables significantly distinguished between the Experimental Group and Control Group (in addition to two interaction effects) and the Adjustment Score Indicated no difference as predicted, no dependent variables significantly changed In a direction opposite to that predicted. Thus Hypothesis I appears to be supported. Hypothesis tt states that subjects, when offered a "psychotherapist" who is congru^r.c with their expectancies, will show more "positive therapeutic changes" on the dependent variables than subjects assigned a "psychotherapist" who is Incongruent with their expectations. This was found to be the case In the analysis of the main effects in four of the 14 dependent variables. The AACL showed that subjects given a therapist congruent with their expectations decreased anxiety after treatment

PAGE 54

significantly more than those subjects who were assigned a therapist incongruent with their expectations (who increased slightly in anxiety) (p ^ .001) (see Table 5). Similarly, the Congruent Group showed a significantly higher increase in the MACL mood of Pleasantness (p
PAGE 55

dependent variables. The Positive Set Group increased significantly less on the HACL mood of Startle after treatment than did the Neutral Set Group which increased more (p < .05) (see Table 17)* The Positive Set Group also reported a significantly higher increase on the HACL mood of Pleasantness (p ^ .05) than the Neutral Set Group which decreased reported pleasantness after treatment (see Table 13) • Only two of the 13 dependent variables which were expected to change significantly distinguished between the positive and neutral set groups. Of particular note is the fact that no significant changes were demonstrated on either the AACL or the HACL mood of Anxiety. However, as predicted, no change occurred on the Adjustment Score and no dependent variables changed significantly In the unpredlcted direction. Thus, the hypothesis appears supported. In summary, those dependent variables which demonstrated significant differences between one or more of the three independent variables were: the AACL and the HACL moods of Aggression, Social Affection, Deactivation, Anxiety, Pleasantness and Startle. Those dependent variables which did not significantly distinguish between any of the three independent variables were Dymond's Adjustment Score (as predicted) and the HACL moods of Concentration, Depression, Egotism, Skepticism, Nonchalance and Activation. As noted earlier, after Now! is factor analyzed the HACL he divided his 12 mood factors Into three groups of four moods each: those moods most clearly identified, those less clearly identified and those least clarly identified (see Table k) It is noteworthy that of the four moods in the group most clearly identified by Nowlis, three were found to

PAGE 56

distinguish significantly between the independent variables; of the four moods in the less clearly Identified group, two were found to distinguish significantly between the Independent variables; and of the four moods in the least clearly identified category, only one was found to distinguish significantly at ween the independent variables. The AACL, the scale with the greatest number of items, and probably, therefore, the tcale with the highest reliability, showed significant differences between two of three main effects (independent variables). Below is a summary table which on the vertical axis lists the three independent variables and their Interactions. On the horizontal axis is listed the 14 dependent variables and entered into the body of the table are the levels of significance at which a null hypothesis might e rejected. Although seven dependent variables did not reflect significant differences between any of the treatment conditions, there were changes worth noting in several of them. For example, the HftCL mood of Depression was found to decrease after treatment in all treatment conditions except one: the Positive Set, Congruent Therapist, Control Group condition where a small Increase in reported depression occurred (see Table HA). Evaluation by the Signs Test (Siegel, 195b) shows that the probability of seven out of eight cells decreasing by chance Is at less than the .035 level. This indicates that a reported decrease In depression tends to occur after treatment, regardless of the specific nature of the treatment within the study. The MACL mood of Concentration, similarly, did not result in any statistically significant differences between the treatment conditions. However, the Congruent Therapist, Experimental Group condition was the only treatment which described an increase In

PAGE 57

k8 TABLE 19 SUMMARY OF THE DEPENDENT VARIABLES AND THE SIGNIFICANCE LEVELS AT WHICH THEY DISTINGUISHED BETWEEN THE INDEPENDENT VARIABLES Now lis Clearly Identified Factors Nowl i s Less Clearly Identified Factors Nowl is Least Clearly Identified Factors AACL Aggr. Cone. Deact. So. Aff. Anx. Depr. Egot. Pleas. Acti v. Nonch. Skept. Start. A (experimental vs. control) B (congruent vs. i ncongruent) C (posi ti ve vs neutral) • 05 001 AB AC BC .01 .01 • 05 .05 05 05 • 05 ,001 • 05 ABC

PAGE 58

49 concentration (sea Table 7A). The MACL mood of Skepticism also displayed no statistically significant differences between the independent variables. However, the Incongruent Therapist, Control Group condition reported a slight increase in skepticism after treatment while all others reported a decrease in skepticism after treatment (see Table 16A). Hypothesis IV states that no relationship exists between differences in certain subject characteristics and changes in the dependent variables. This was evaluated by multiple regression technique. The subject differences under consideration were age, level of intelligence (Beta Intelligence Test, 1946), level of neuroticism (HPI, 1959) and level of introversion (HPI, 1959). Analysis of variance (Lindquist, 1953) was used to demonstrate that each of these four subject variables was randomly distributed across the eight treatment conditions. (The summary tables for these analyses on age and level of intelligence are presented in the Subjects section of the Method chapter. Below are the summary tables demonstrating the random distribution of introversion (Table 20) and neuroticism (Table 21). Pearson product -moment correlations were computed between each of these four variables and the change in scores from preto post treatment testing on each of the 14 dependent variables. This was done for each of the eight treatment conditions. This resulted in 448 correlations which are presented oelow in Tables 22, 23, 24, and 25. At the .05 level of significance, by chance alone wo would expect 24 significant correlations out of the total 443. In the present data, by frequency count, there are 26 correlations which are significant at the .05 level. This suggests that, over-all, the null hypothesis appears supported. It is difficult

PAGE 59

so TABLE 20 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANDOM DISTRIBUTION OF MPI INTROVERSION SCORES Source ss df V F A (experimental vs. control) 19.26 19.26 .34 B (congruent vs incongruent) 31.51 31.51 .55 C (positive vs. neutral) 33.84 • 33.84 .59 AB 52.51 52.51 .92 AC .02 .02 .00 BC 128.35 • 128.35 2.25 ABC 31.50 31.50 .55 w Cells 5.Q1?, 75 ft 56.96 Total 5.309.74 95 (An F of 3.95 is significant at the .05 level.)

PAGE 60

$1 TABLE 21 SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE RANDOM DISTRIBUTION OF MPI NEUROTICISM SCORES Sou rce ss at V m r A (experimental vs. control) 2.06 2.06 .01 B (congruent vs. incongruent) 408.38 408.38 2.04 C (positive vs. neutral) 135.40 135.40 .68 AB .37 .37 .00 AC 57.02 57.02 .28 BC 376.03 376.03 1 M ABC 63.37 63.37 .32 w Cel Is 12*600! 200.27 Total 18,665.96 95 (An F of 3.95 is significant at the .05 level.)

PAGE 61

52 to attribute a high degree of conclusiveness to results showing some consistency in the absence of significance. However, there are several observations which may be made and several trends worth Identifying. It Is important to note, also, that the complexity of both the data and the independent variables tends to reduce the meanlngfulness of interpretation. It will be noted In Table 22 that there Is a tendency for subjects, given a positive set and a congruent therapist and who score lower measured levels of intelligence, to decrease reported anxiety on the AACL after treatment. Conversely, this suggests that subjects with higher measured intelligence, when given a positive set and a congruent therapist, increase reported anxiety on the AACL after treatment. Similarly, subjects with lower measured intelligence, who are given a positive set and an incongruent therapist, seem to increase their scores on the AACL after treatment more than subjects with higher measured intelligence. It will be noted further, with the exception of the Experimental Group, Neutral Set, Congruent Therapist subjects (which showed an opposite trend), less bright subjects tend to increase the MACL mood of Activation after treatment. From these data, the conclusion may be tentatively advanced that people with lower intelligence seem to change in a therapeutic direction after certain types of placebo treatment. On the other hand, with the exception of the Experimental Group, Positive Set, Congruent Therapist, there appears to be a trend which suggests that subjects with higher measured Intelligence tend to increase more on the HACL mood of Social Affection after treatment than subjects who score lower on intelligence. This finding suggests a conclusion which is not In line with that indicated above from the analysis of the AACL and the MACL mood of

PAGE 62

53 TABLE 22 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRETO POSTTREATMENT TESTING ON EIGHT OF THE DEPENDENT VARIABLES AND SCORES ON THE BETA INTELLIGENCE TEST. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS Moods of the MACL Treatment Conditions AACL Aqqr Cone Deact. So. Aff. Anx. Depr. Eqot. Pleas. Acti v. Nonch. Skept. Start. Adj ustment Score Experimental Group Positive Set, Congruent Therapist + t, IU O Q zy t. H-Z ".II Sh 0 0 • zz 4, 5"7 to / + 09 zo rn • i>u Oil ZM+ o Ok ZH Positive Set, Incongruent Therapist -.51 +.05 +. 1 1 + .20 +.43 -.44 -.23 — • 15 +.39 -.23 +.52 -.18 -.61* + .48 Neutral Set, Congruent Therapist +.20 +.34 -.50 + .02 + .26 + .04 + .09 35 +.07 + .18 .00 + .01 + .52 -.50 Neutral Set, Incongruent Therapist -.13 -.39 +. 19 89* + .54* -.44 -.11 + 60* -.04 -.32 -.18 -.10 -.73* -.27 Control Group + 26 Positive Set, Congruent Therapist +.45 -.07 -.27 -.27 + .06 +.35 -.33 +. 02 +.31 -.27 -.27 + .37 -.01 Positive Set, Incongruent Therapist -.33 + .02 -.36 +.13 +.10 -.53* -.18 10 +.08 -.25 -.10 -.37 -.12 + .08 Neutral Set, Congruent Therapist + .05 -.32 -.13 -.05 + .25 + .31 + .09 18 -.24 -.02 -.17 + .01 -.22 -.09 Neutral Set, Incongruent Therapist + 1 1 + .41 +.18 +.06 + .09 + .07 -.05 28 -.15 -.38 -.26 -.26 + .07 + .16 ^Significant ^1 .05-

PAGE 63

5k Activation. That is, in this case, brighter subjects tend to increase Social Affection (personal comfort) as a result of the placebo situation, while in the former case, brighter subjects appear to increase anxiety and become less active. Of further interest is another tentative finding suggesting that brighter subjects, when given a congruent therapist, tend to Increase scores on the HACL Skepticism mood after treatment, but decrease skepticism after treatment when given an incongruent therapist. From Table 23 several interesting, but also tentative relationships may be discussed. In the Experimental Group, the more Introverted subjects who are ^iven the Positive Set, Congruent Therapist Instructions tend to increase reported Social Affection after treatment. However, the more introverted subjects who are given the Neutral Set, Incongruent Therapist instructions tend to decrease the HACL mood of Social Affection. An additional trend (with the exception of the Control Group, Neutral Set, Incongruent Therapist subjects) appears to be that introverted subjects, more than extroverted subjects, tend to decrease scores on the HACL mood of Skepticism after treatment. Several relationships presented in Table 2k may be tentatively discussed. It appears that in the positive set cells of the Experimental Group and in the neutral set cells of the Control Group the older the subject, the larger the increase in the HACL mood of Social Affection. Similarly, in the positive set cells of both the Experimental and Control Groups, the younger the subject the greater the increase In reported HACL Egotism mood after treatment. Another effect appears to exist in the relationships between age and change in anxiety after treatment. On both the AACL and the HACL Anxiety mood, it is noted that younger subjects

PAGE 64

55 TABLE 23 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRETO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND SCORES ON THE MPI INTROVERSION (LOW SCORES) -EXTROVERSI ON (HIGH SCORES). THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS Moods of the MACL Treatment Conditions AA.CL Aggr • Cone. Deact. So. Aff. Anx. Dsp r cqu l P 1 i=>ac r 1 cdb nC Z 1 V Nonch PL-,. j_ Start Adj ustment Score Experimental Group Positive Set, Congruent Therapist +.21 +.03 -.21 +. 10 54* + .01 +.20 +. 32 -.07 +.57* +.53* + .44 + .25 + .71* Positive Set, Incongruent Therapist +. 06 -.12 +.09 -.46 -.25 -.16 +.05 + .51 + .03 + .41 +.13 +.21 + .24 -.13 Neutral Set, Congruent Therapist +.07 + .02 -.1 1 -.33 -.14 +.03 -.06 + .14 -.31 -.11 + .29 + .40 -.13 + .26 Neutral Set, Incongruent Therapist -.02 -.31 +.09 -.1 1 +.60* +.41 -.28 -.36 + .02 + .01 +.01 + .01 -.51 + .01 Control Group Positive Set, Congruent Therapist +.12 -.36 -.52 .00 -.59* 1 i +.30 -.08 -.06 -.16 +.06 + .10 + .30 +.57* -.03 Positive Set, Incongruent Therapist -.07 +.52 +.08 +.08 -.40 -.05 +. 14 +.37 -.28 + .32 + .48 +.02 -.03 -.04 Neutral Set, Congruent Therapist +.31 +.04 -.05 +. 1 1 +.35 -.37 +.40 -.24 +.37 +.29 -.04 +.61* -.15 -.36 Neutral Set, Incongruent Therapist -.27 -.18 -.14 -.25 -.17 +.01 +. 1 1 +.29 -.17 -.65* + .15 -.33 -.19 -.30 *Significant ^ .05.

PAGE 65

56 TABLE 24 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRETO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND SUBJECTS' AGES. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS Moods of the MACL Treatment Conditions AACL Aqqn. Cone. Deact. So. Aff. Anx. Depr Eqot. Pleas. Acti v. Nonch. Skept. Start. Adj ustment Score Experimental Group Positive Set, Congruent Therapist +.07 +.07 +. 19 + .06 +.49 +. 1 1 -.19 -.36 -.35 -.41 -,56* + 14 + 14 <=>] Positive Set, Incongruent Therapist +.36 +.42 +.32 +.06 +.50 +.45 +.51 -.24 +.14 -.06 -.32 -.07 -.03 -.05 Neutral Set, Congruent Therapist + .51 +.01 -.34 + .13 -.22 +.32 -.20 +.30 +.03 -.05 + .06 +.52 + .50 Neutral Set, Incongruent Therapist +.03 +.23 -.10 .00 -.25 +.06 -.25 +.39 +.41 -.02 +.48 + .49 -.14 -.48 Control Group Positive Set, Congruent Therapist +.37 -.22 +.16 -.33 -.45 + .18 -.07 -.28 -.20 -.01 -.36 -.66* -.08 -.59* Positive Set, Incongruent Therapist +.25 +.15 +.20 -.43 -.21 +.26 -.19 -.18 +.21 +.24 -.22 -.21 -.29 -.15 Neutral Set, Congruent Therapist 60* +.02 +.36 +.12 + .31 +.23 +.02 +.28 +.32 +.30 -.01 + .05 -.24 + .41 Neutral Set, Incongruent Therapist -.12 -.29 + .22 -.31 +.18 -.25 +. 19 +.06 -.09 +.06 +.42 -.38 + 14 + .03 *Si gni f i cant K. 05.

PAGE 66

57 decrease reported anxiety after Experimental Group treatment more readily than do older subjects. This is especially so for the Positive Set, incongruent Therapist and Neutral Set, Congruent Therapist conditions. Eysenck defined his Neurotic! sm Scale as a measure extending along a highly neurotic to normal continuum, in the current study, the sample contains no "normals," therefore, serious questions may be raised concerning the meaningful ness of the neurotic! sm score. For example, does a low score on neuroticism in this study mean that a patient is psychotic? A biserial correlation (McNemar, 1955) between diagnosis (psychotic vs. nonpsychotic) and scores on the Neuroticism Scale results In a correlation of .57 (p 4. .001). This suggests that subjects with nonpsychotic diagnoses tend to score high on the Neuroticism Scale while subjects with psychotic diagnoses tend to score lower. However, this correlation accounts for only 32.49 per cent of the variance, and it is with extreme caution that interpretation of the neuroticism score data be made. As noted in Table 25, an apparent relationship exists between neuroticism scores and the NACL mood of Nonchalance. With the exception of the Experimental Group, Neutral Set, Congruent Therapist subjects, it appears that as neuroticism increases, nonchalance decreases after treatment. Another interesting relationship indicates that subjects scoring high on the Neuroticism Scale who are in the Positive Set, Congruent Therapist, Control Group decrease their MAC I mood of Startle after treatment; while subjects scoring high on the Neuroticism Scale who are in the Neutral Set, Irtcongruent Therapist, Control Group increase this Startle mood after treatment.

PAGE 67

58 TABLE 25 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRETO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND SCORES ON THE MPI NEUROT1CISM SCALE. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS """ Moods of the MACL Treatment Conditions AACL Aqqr. Cone. Deact. So. Aff. Anx. Depr. Eqot. PI eas Acti v. Nonch SkeDt Cf art Adj ustmen Sco re Experimental Group Positive Set, Congruent Therapist -.08 +.19 +.51 -. 12 +.52 +.17 -.43 -.46 -.36 -.39 -.43 -.29 -.24 56* Positive Set, Incongruent Therapist +.18 -.23 -.37 -.03 -.06 +.07 +.41 -.22 +.01 -.38 -.59* -.38 + .03 -.43 Neutral Set, Congruent Therapist -.24 -.41 +. 14 +.13 +.04 +.43 +.09 +.42 +.70* +.34 + .26 -.23 + .20 -.06 Neutral Set, Incongruent Therapist -.13 +.02 +.29 +. 12 -.03 +.56* +.21 +.12 +.45 +.42 -.07 +.48 + 12 -.51 Control Group Positive Set, Congruent Therapist -.01 +.03 + .52 -.17 -.14 +.08 +.03 -.16 +.19 -.18 -.17 -.69* -.10 Positive Set, Incongruent Therapist -.29 -.28 -.37 -.15 +.16 -.27 -.60* -.30 -. 12 -.41 -.28 -.02 -.30 + .42 Neutral Set, Congruent Therapist -.38 + .26 +.37 +.10 -.28 +.39 -.36 +.35 -.36 +.13 -.14 -•30 -.1 1 + .43 Neutral Set, Incongruent Therapist +.17 +.20 -.34 +.40 -.19 +.47 +.15 -.16 +.23 -.22 -.35 +.20 +. 56* + .04 *Significant ^.05-

PAGE 68

59 It is believed sufficiently important to stress once again the tentativeness of all relationships discussed under Hypothesis IV. It Is this writer's opinion that due to the low number of significant correlations, the data basically support the null hypothesis that no conclusively meaningful relationships exist between the dependent variables and the measured individual differences. It was believed that a subject's reported level of desire for psychotherapy would be related to his amount of verbalization when he was permitted to speak to the "psychotherapist." The tenability of this belief was evaluated by multiple regression technique utilizing only the Experimental Group and the correlations are presented below in Table 26. TABLE 26 MULTIPLE REGRESSION COEFFICIENTS BETWEEN AMOUNT OF SPEECH AND DESIRE FOR PSYCHOTHERAPY PRESENTED FOR EACH INDEPENDENT VARIABLE (EXPERIMENTAL DATA ONLY) Congruent Incongruent Positive Neutral — Therapl Theraol st Sej Set Amount of Speech .50 .02 .49 .08 vs. Desire for Psychotherapy n s 2k n s 2k n • 2k n z 2k Amount of verbalization, which was originally rated on a fourpoint continuum (see p. 26) was divided, for the purpose of this evaluation, Into two groups: (1) constant speakers and moderate speakers, and (2) little speakers and no speech. Desire for psychotherapy was rated on a three-point scale by each subject (see Appendix F).

PAGE 69

60 The correlations presented in Table 26 suggest that if a subject's desire for psychotherapy it high, his amount of verbalization is high i he is given a congruent therapist and/or a positive set. If, however, he is given an incongruent therapist and/or a neutral set there Is no relationship between his desire for psychotherapy and his amount of speech. Presented below in Table 27 are co-relationships between (I) amount of speech, (2) desire for psychotherapy; and (3) diagnosis on the one hand, and change scores from preto posttesting on the AACL and the moods most c'aarly identified by Nowlis on the other (data only for the Experimental Group). One Interesting conclusion which may be drawn i rom these data concerns the relationship between amount of speech and change scores on the AACL. It appears that subjects who speak the most increase anxiety from preto posttreatment testing. This is true in all treatment conditions and the correlation for all groups is +.30 (p < .05). Of further interest is the fact that level of desire for psychotherapy does not appear, by inspection, to appreciably influence the amount of change on these five dependent variables from preto posttreatment testing. Similarly, by inspection, diagnosis (psychotic or nonpsychotlc) does not •pp—r to be related to the amount of reported change In personal comfort after treatment as measured by the five dependent variables.

PAGE 70

ol ol Ui _l i I B p or ll < I m .-< a < o e H SC HZ Ui < o =• o — Ul UJ — I X Q 1— t— Q 2 5 • z a. — 3 S K" OS uj W U -J JJ< a. I Li VA CO X. -J < — r< a£ ui UJ < ft. o UI > X z I UI ft. 5 z ui ui < z — ii a m m -> uj _i ^{S UI (/I o 00 UI u. 1oc o t/ o H<_> X Z v> o UI < — UI UJ O o elf U. X u. UI o O V UUlJ -J UI w < V> < ft. UI > UI S£ 8 H a -j I -J H o.. E — j i2 a a 3 'J j 9 3 m CM 4a CA o CM CA O o CM o o o — 8 CM — o — o f 1 + + + 1 1 + 1 + 1 lA + r CM TV CM O • • + I -a -* o — • • + O JK & u %J >0 "i O 0 ~ X i/l y 0 c u >. c g o. co la 5 -Is i/ *J eM — y 5 & u 3 3 L 4J C U (/>•— CTiX C D 9 ?\ CM CO oo cn rs. r> o o UN O — ca — o — o + + + 1 • i + i S 3 0 u X *tO *> i/l O O ~ X (A v o 6 c u >. c *J ** CM — y C co C O) v. *J c C (/(>- O X c OHI 1*1 CM cm + + O O + + • + n. • K I E w x: u. 0 O *O — X UI *J £ 3 ~ UI O) in 4T ;v| I > — c I <7\ + I n o o • + + 8 I + v0 -tf 9 o +* r U 5 CL a w (A X MO *> y o C > C 3 ~ ui en I < a a J i z o + + o% O O • • • + + + y I a r I w y o C U JK c J<2 o. co a 5 — y c x*oo — n 5c a. o~ — Jo lA — O O vv c c I I y y uT C c 'c i i

PAGE 71

DISCUSSION As mentioned In the Introduction, the literature supporting the efficacy of psychotherapy has not demonstrated that personality and behavioral modifications which occur as a result of psychotherapy are entirely due to a psychotherapist's behavior and techniques (specific factors). The outcome of treatment, it was suggested, is a function of both the specific and the nonspecific (placebo) events occurring during psychotherapy, although, in the literature, the presence and the contribution of the placebo effect in psychotherapy appear to be entirely conjectural This study appears to demonstrate that the placebo effect in psychotherapy does exist. Subject-patients who spoke into a tape recorder for one session, believing that they were speaking to a psychotherapist, changed certain behaviors in therapeutic directions when compared to subjects who did not speak (see Hypothesis I). That Is, it has been shown that verbally reported patient behaviors such as anxiety and certain moods which would be expected to change as a result of therapist behavior in actual psychotherapy, also change, in fact, as a result of the placebo psychotherapy (nonspecific treatment) In which a therapist is absent. The group who spoke to the "psychotherapist for one session, when compared to the group who waited, increased their Social Affection mood on the MACL and reduced their level of anxiety as measured on both the AACL and the MACL Anxiety mood. 62

PAGE 72

63 As a result of these findings in the current study, it is proposed that the placebo effect does contribute to the outcome of psychotherapy, and that a measurement of the amount of contribution appears to have been made. However, it is important to question precisely what has been evaluated by a comparison between the Experimental Group (talkers) and the Control Group (waiters), if the subjects who spoke received placebo treatment and the subjects who waited received no treatment, the statistical comparison between these two groups was an accurate measurement of the placebo effect. However, it is believed that those In the Control Group, who waited, but were promised a therapist, and who changed on the dependent variables in a generally "nontherapeutic" direction, received some placebo benefit from the procedure (e.g., see Tables 1 1A and I3A). This finding does not Influence the conclusion that the placebo effect in psychotherapy does exist, however, the precise therapeutic benefit of the placebo, as measured in this study, has not oeen compared to the therapeutic benefit of no placebo treatment. Instead, the comparison between the Experimental and Control Groups appears to have been an analysis between a more effective (Experimental Group) and a questionably effective (Control Group) placebo treatment. In addition to congruence of expectations regarding the characteristics of the assigned psychotherapist and positive set regarding treatment outcome (see below), may other "psychologic mechanisms" through which the placebo effect takes place (Shapiro, 196*0 be Identified in order to account for the finding that this treatment did result In a placebo effect? Shapiro (1964) reports that throughout the history of medicine, "methods of depletion (which are currently accepted as placebos)

PAGE 73

were widely used (to effect cures): entities, cathartics, enemas, purges, stomachics, sweating, bleeding, leeching, cupping, starvation and dehydration. Methods of depletion and expulsion may relieve symptom by symbolically expelling bad thoughts and conflictual ego-alien impulses." In our more sophisticated and verbally oriented culture, Shapiro continues, "relief of symptoms may occur when the patient is able to express verbally conflictual and guilt-ridden thoughts and feelings in the free, nonjudging and accepting atmosphere of the doctor's office" (p. 81). In other words, verbal catharsis, it appears, may theoretically lie at the basis of the placebo effect in psychotherapy, just as physical -mechanical catharsis lay at the basis of medical treatment which today is widely looked upon as placebo. The treatment in this experiment is conducted in an obviously free, accepting and nonjudgmental atmosphere and each patient-subject may establish his own personal optimal level of catharsis. It would seem logical then, that catharsis may be a major underlying factor for the placebo effect in psychotherapy, at least in so far as outcome of psychotherapy is measured In this study. It Is interesting, in this regard, however, that the data indicate that a linear relationship does not exist between amount of speech (loosely, catharsis) during the "Interview" and degree of change on the dependent variables. In fact, in this context, although one significant correlation among the 25 computed would be expected to be significant by chance alone, the only significant correlation (.05 level) which does exist, suggests that the more a subject speaks, the greater the likelihood that his anxiety level (AACL) will increase. These correlation--, however, evaluate only the amount of speech in relation to the dependent variable changes after treatment. The

PAGE 74

analysis does not assess content of speech or nature of affect which are certainly extremely significant aspects of catharsis. In addition, this analysis evaluates only the linear relationships between amount of speech and degree of change on the dependent variables. The analysis does not deal with nonlinear relationships, and it Is conceivable that such relationships might exist. Shapiro (1964) reports that expressed desire for treatment is another important concept which reportedly Is basic in eliciting the placebo effect. That is, those subjects who are more highly motivated for treatment would be more responsive to placebo treatment. However, In this study, motivation of the subjects, as measured by their statements concerning their level of desire for psychotherapy, is not linearly related to the amount or nature of change after treatment as measured on the dependent variables. Interestingly, however, when subjects who report high motivation for psychotherapy are given a positive set concerning the outcome of the treatment and/or are offered a therapist who is congruent with their expectancies they tend to speak more. That is, greater reported desire for psychotherapy is related to a larger amount of verba) catharsis; however, it is unrelated to changes on the measures of mood and anxiety after placebo psychotherapy when a therapist Is absent. It appears, thus, that In this study, catharsis or degree of motivation, or both, have not been shown to underlie the placebo effect. Shapiro (1964), in his review of placebo literature, indicates that a great niunuer 0 f "psychologic mechanisms' have been proposed as being underlying characteristics of the placebo effect. Several are: catharsis,

PAGE 75

66 motivation, faith, learning and conditioning, trust, confidence and previous experience with healers. Shapiro (and Whitehorn, 1958) further indicate that no single characteristic of treatment hat bean consistently shown to underlie and to produce the placebo effect. The conflicting and inconclusive evidence in the literature is most conceivably due to the unclear nature of the placebo effect, differences of measurement of the effect from research to research and the different theoretical frameworks within which different experiments are planned. Thus far, the discussion has been concerned with the contrast between the Experimental Group (talkers) and the Control Group (waiters) and several of the characteristics of psychological treatment which nay determine the placebo effect. The conditions under which the placebo effect is most likely to take place or most likely to be lessened, are further clarified when attention is drawn to the way in which the treatment groups were divided within the present experiment. Analysis of Hypothesis II, indicates that patients who are offered a "psychotherapist" congruent with their expectations, report more "positive therapeutic changes" on measures of anxiety and moods than do patients who are assigned a psychotherapist who is incongruent with their expectations concerning the nature of the psychotherapist to be assigned. That is, the subject's perception of the treater or doctor is a significant aspect in encouraging or discouraging the placebo effect. Festinger (1957) and Festlnger and Bramel (Bachrach, 1962) propose that dissonance results when two cognitions which a person holds are inconsistent with each other according to the expectations of the person.

PAGE 76

67 Dissonance is said to be a motivating state which is comparable to other drive states, and just as hunger produces physical discomfort, dissonance results in psychological discomfort. When a patient-subject is instructed that the characteristics of his psychotherapist are incongruent with his personal expectations concerning this psychotherapist, the patientsubject experiences dissonance. This in turn, reduces the effectiveness of the placebo treatment, as is shown by the statistical analyses which indicate that the AACL and MACL moods of Startle and Deactivation decreased significantly more for the Congruent Therapist Group, while the MACL mood of Pleasantness increased significantly more in the Congruent Therapist Group. (It is noteworthy that although the Apfelbaum questionnaire [see Appendix E) instructs subjects to respond on the basis of their expecta tions, any individual subject may have completed the Items in terms of his ideal psychotherapist or in terms of a desired psychotherapist. There is no apparent way in which to evaluate these possibilities. If, however, any individual subject did respond in terms of an ideal or a desired psychotherapist, he may have received additional, or reduced, positive set, in addition to having the nature of the treater confirmed or not conf i rmed.) In addition to demonstrating that the patient's perception of the nature of the treater is one significant aspect of inducing a greater or lesser placebo effect, this research also suggests that the nature of the treatment situation itself is important. Analysis of Hypothesis III indicates that subjects who are told that the placebo psychotherapy will be extremely helpful to them (Positive Set Group) profit more from the

PAGE 77

u8 treatment than those subjects with whom positive treatment outcome is not discussed (Neutral Set Group). This conclusion is somewhat less certain than those drawn from hypotheses I and II. This Is so because only two of the dependent variables (MACL moods of Startle and Pleasantness) indicated that the positive set was significantly more effective in encouraging the placebo effect than the neutral set. The literature which served as the impetus for introducing this Independent variable (positive vs. neutral set) into the current study, apparently uniformly indicates that positive set or attitude toward a particular placebo treatment increases the placebo effect. (For example, see pp. 45-46 of the present study and also social psychological research on set and frames of reference: Newcomb, 1958, pp. 264-297; Maccoby et al .. 1958, p. 95.) Why hasn't the current study demonstrated more conclusively that a positive set concerning treatment outcome Is even more important in securing a placebo effect, as might be anticipated on the basis of the literature? Two explanations are suggested. On the one hand, it is possible that the wording of the two instructional sets (constructed by the author) were not sufficiently different and thus, two entirely separate groups concerning expectation of treatment outcome may not have been established. On the other hand, it will be noted (see Appendix A) that the Positive Set instructions emphasized the reputation and demand for the alleged psychotherapist, in addition to his willingness to help the subject. It is possible that instructions stressing this psychotherapist's professional reputation and the demand for his services did not increase the positive nature of the set for this particular group of subjects as a whole. Thus, portions of the Positive Set instructions may not have

PAGE 78

served the purpose for which they were intended. (By stressing characteristics of the psychotherapist, this aspect of the instructions may have had little influence on the level of set but may have had some vicarious influence on the congruence or incongruence of the psychotherapist for any particular subject.) In summarizing the results of the analysis of hypotheses It II and III, the following may be said: The placebo effect in psychotherapy does exist and through the technique of this study appears to be measurable. Further, the placebo effect Is apparently influenced by both the perceived characteristics of the treater, and the projected outcome of the treatment as reported to the patient. The influence of the perceived characteristics of the treater has been approached by uiysls of differences on the dependent variables between one group of subjects who were told that a therapist to be assigned to them will be congruent with their expectations, and another group who were told that the therapist will be incongruent with their expectations. The data apparently demonstrate that the group given the congruent therapist responds to the treatment with a greater positive placebo effect. The influence of the reported outcome of the treatment has been approached by analysis of differences on the dependent variables between one group of subjects who were given a positive set concerning treatment outcome and another group who were given a more neutral set. The data suggest that the group given the positive set tends to respond to treatment with a greater positive placebo effect. (However, congruent therapist instructions seem to be more instrumental in encouraging the placebo effect than the "positive set" i nst ructions.)

PAGE 79

70 Having found that offering a therapist who is congruent with a subject's expectations seems to encourage the placebo effect more than offering a positive set concerning treatment outcome may have Implications beyond the present study. There is a considerable body of research concerned with the matching of patients and therapists and the resulting relationships in psychotherapy (e.g., see Leery and Gill 1 19591) • In the present study It Is seen that matching a therapist with a patient's expectations does, in fact, increase the effectiveness of the treatment to a greater extent than offering a positive set concerning treatment outcome. This Is Indicated in this placebo situation, where no therapist actually exists; h ow ev er, It Is suggested that in psychotherapy perception of the psychotherapist may also be more valuable than belief about treatment outcome. That is, the nature of the relationship (in this study, a relationship in which the patient perceives his therapist as being congruent with his expectations) may be more essential In successful psychotherapy than statements concerning the value of the treatment Itself. An important finding of this study is that, as predicted, the Adjustment Score, which measures general level of psychological adjustment, does not change after the placebo interview. This is so regardless of the particular treatment (talk vs. wait, congruent therapist vs. incongruent therapist and positive set vs. neutral set) or combination of treatments in the study. This indicates that in a single treatment of placebo psychotherapy, no change In general level of psychological adjustment may take place. Had the placebo psychotherapy continued over a longer period of time would there have been changes In this Adjustment

PAGE 80

71 Score? That is, would extended placebo psychotherapy have produced changes In psychological adjustment similar to changes which occur after actual psychotherapy? (See Frank, 1961, and Rogers and Dymond, 195*, for evidence of changes in general level of psychological adjustment after psychotherapy.) It is difficult to speculate on precisely what may have happened to the subject's general level of psychological adjustment had the placebo treatment continued, although researchers such as Glasser and Whlttlow (1953, 195*0 submit that the effects of placebo treatment are not permanent and the gains from continued placebo treatment become less and less. What might have happened to moods and anxiety had the placebo psychotherapy continued? Martin et al (I960), In a study concerned with levels of therapist communications (discussed more fully in the Introduction), had subjects speak into a tape recorder for five 30-minute weekly sessions as if they were speaking to a psychotherapist. These authors demonstrate that although some initial positive therapeutic effect was indicated (lowering of anxiety), the procedure lost its positive therapeutic effectiveness as the number of sessions with the tape recorder Increased. Glasser and Whlttlow (1953, 195*0, In placebo drug studies, found that if subjects were given placebos over a period of time and required to answer questionnaires identifying the effects of these placebos, the positive placebo effect was demonstrated Initially, but the effect was less upon each successive evaluation. Findings such as these suggest that the placebo procedure used In the current study, if extended for more than one session, might not have continued to have positive therapeutic effect on mood and anxiety.

PAGE 81

72 Analysis of Hypothesis IV gives no clear picture of the contribution to the placebo effect of the individual patient differences in age* intelligence, introversion and neuroticism. This finding is not surprising in viewof the existing body of research which is equivocal and often contradictory in regard to this question (see Introduction). The evaluation of the data collected to explore this hypothesis resulted in borderline significance and lack of remarkable trends both within the four measured subject differences and within all of the dependent variables across individual subject differences. Even when tentative or apparent trends were identified (as discussed on pp. 49-59) t the complexity of the behaviors involved In the eight different combinations of the independent variables makes the meaningfulness of any interpretation of results doubtful. (In addition, the reader is reminded of the questionable use of the Neuroticism Scale or the HPI with this subject sample; see p. 57.) Pearson product -moment correlations were used to evaluate this hypothesis. These only Identify linear relationships between the dependent variables and the measured subject differences, and the possibility of a nonlinear relationship cannot be dismissed on the basis of the analysis performed. The problem of discussing the meaningfulness of the results and the inability to disclose significant trends in the data of this study apparently leads to the acceptance of the null hypotheses which state that, In the present sample, age, intelligence, introversion or neuroticism will not identify the type of subject who responds favorably to placebo psychotherapy. This is In keeping with the findings of Wolf et al (1957) who indicate that individual subject reactions to placebos are

PAGE 82

73 generally Inconsistent, unpredictable and not uniform (see also Hagans et al .. 1957. and Kurland, 1957). A pertinent issue may be raised concerning whether the subjects accepted the placebo treatment as a legitimate form of psychotherapy. There is no way in which to be certain of any conclusion in this regard. However, the nature of changes on the dependent variables certainly suggests that the subjects did perceive the procedure as actual psychotherapy. Moreover, it will be noted that the Skepticism mood on the MACL remained rather constant from preto posttreatment testing, and the only groups to report a small and insignificant increase in Skepticism, after having completed the procedure, were the Incongruent Therapist, Control Group subjects (see Table I6A). No subjects reported that they disbelieved the legitimacy of the treatment and the comments several of the subjects made to the examiner and into the tape recorder appear related to the issue of subjects accepting the treatment as real. For example, one subject in the Experimental, Neutral Set, Congruent Therapist Group reported the following Into the tape recorder during the placebo therapy: "I appreciate the privilege of being given a chance to express my feelings and views In this way. This is the first chance I've had to honestly and openly express my feelings and opinions and I appreciate it very, very much." It Is hoped that the placebo effects demonstrated in this treatment will be kept in mind In future research on the effects of specific methods in psychotherapy. It is further hoped that this demonstration will lead to the practice of separating effects common to all psychotherapy (e.g., nonspecific placebo effects) from those effects due to the

PAGE 83

specific aspects of the therapy being evaluated (e.g., therapist behaviors)

PAGE 84

APPENDICES

PAGE 85

APPENDIX A INSTRUCTIONS TO SUBJECTS

PAGE 86

INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS PRIOR TO ESTABLISHING THE TREATMENT CONDITIONS IN THE STUDY I'd like you men to do several things this afternoon. I think you will enjoy these tasks. You will learn something about yourself and may profit from what we do here for the next hour or so. (Distribute the personal information questionnaire.) I'd Ilka for you to answer these questions about yourselves. (Distribute the Maudsley Personality Inventory.) The next questionnaire is this one. (Read directions.) Begin. (Distribute the Anxiety Adjective Check List.) The next questionnaire is this one. (Read directions.) Please answer honestly the way you feel right now Begin. (Distribute the Hood Adjective Check List.) The next questionnaire is this one which asks you about your mood and the way you feel right now (Read directions.) Some of the items in this list are like items In the last list, but please answer them anyway. Be honest with yourself and just answer the way you feel rig ht now (Distribute the Modification of the Expectation Q-Sort.) This questionnaire is a little different. It says — (read directions). Now, don't think too long on any of these, but answer them as bast you can. (Distribute the Adjustment Score.) This questionnaire asks you how you feel about yourself and also about things in general. Read the instructions and work quickly INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS TO DETERMINE POSITIVE SET OR NEUTRAL SET Positive Set Instructions I can tell you now what this i s al 1 about. It's a real good opportunity for you men because you are being given a chance to talk with a great psychiatrist. Now, you can talk to him about anything you'd like 77

PAGE 87

7& relating to problems you've had, or problems you are having, or problem you can sea ahead of you In the future. This psychiatrist has such an excellent reputation that he is an extremely busy man, but he has offered to help you men to solve any problems you may have anyway. Because he is so busy he can't get here to the hospital so he suggested we do something a little differently because he is so interested in helping you. You will simply talk into a tape recorder, and the tape will be sent registered mail to him. I think this Is really great because he has helped so many, many people and I 'm sure he can help you with any problems you may possibly have. Neutral Set Instructions I can tell you now what this is all about. It is an opportunity for you men to talk with a psychiatrist. Now, you can talk to him about anything you'd like relating to problems you've had, or problems you are having, or problems you can see ahead of you in the future. This psychiatrist is an extremely busy man, and he is unable to get here to the hospital so he suggested we do something a little differently. You will simply talk into a tape recorder, and the tape will be sent registered mai I to him. INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS TO DETERMINE CONGRUENT THERAPIST OR INCONGRUENT THERAPIST Congruent Therapist Instructions I've been looking over your questionnaire on the way you would expect this psychotherapist or counselor to be, and I sae that this psychiatrist is the way you thought he would be. For example, Mr. __________ (subject seated on the Experimenter's right), he (read this subject's responses to I terns 7. 1^ and 29; sue Appendix E) And Mr. _________ (subject seated on the Exoerimenter s left), he (read this subject' s responses to items 1, 31 and 32; sea Appendix £). inconqruent Therapist Instructions I've been looking over your questionnaire on the way you would expect this psychotherapist or counselor to be, and I sea that this psychiatrist is not exactly the way you thought he would be. For example, Mr. (subject seated on the Experimenter's right), he (read a denial of the accuracy of this subject's responses to I terns 7, 1** and 29; see Appendix E). And Mr. (subject seated on the Experimenter's left), he (read a denial of the accuracy of this subject's responses to items 1, 31 and 32; sea Appendix £).

PAGE 88

lJ INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS AFTER HAVING ESTABLISHED THE FOUR TREATMENT CONDITIONS Mr. _____ (Control Group subject), I'd like for you to wait here for a while. Make yourself comfortable. You can look at some magazines, or rest, but please stay In this room. I'll be back in about forty-five minutes. Mr. (Experimental Group subject), please come with me. INSTRUCTIONS GIVEN TO THE EXPERIMENTAL GROUP SUBJECT AFTER CEING SEATED IN THE ROOM WITH THE TAPE RECORDER Now, this Is the tape recorder and the tape that will be sent to the psychiatrist. Just make yourself comfortable, and say whatever you would ll*e into this microphone. The controls on the machine are set, and please remain In your seat until I return. It's OK if there are pauses, and you don't have to talk all of the time. 1*11 be back in about fifty minutes to get the tap*. INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS AFTER THE EXPERIMENTAL GROUP SUBJECT IS REUNITED WITH THE CONTROL GROUP SUBJECT I'd like for you to retake several of these shorter tests. This will only take a few minutes more. (Distribute the Anxiety Adjective ChecK List.) This is the first one. You remember this one, it asks you to check the way in which you feel right now. Please answer honestly the way you feel right now (Distribute the Mood Adjective Check List.) This \s the next one. You remember this one, it asks you to mark the way you fuel right now. Please answer honestly the way you feel right now (Distribute the Adjustment Score.) This is the next one. Please work as quickly as you can. INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS AFTER THE PAIR HAS COMPLETED THE TESTS Mr. (Experimental Group subject), you may leave now. Mr. (Control Group subject), do you care to speak to the psychiatrist? (If the Control Group subject wishes to speak, he Is taken to the tape recorder; If, however, he declines to speak, he is dismissed at this time.)

PAGE 89

APPENDIX B SAMPLE ANXIETY ADJECTIVE CHECK LIST

PAGE 90

ROM Below you will find words which describe different kinds of feelings. Check the words which describe how you feel RIGHT NOW Some of the words may sound alike but we want you to check aH, of the words that describe your feelings. 1 _AFRAI D 21 GAY hi 2._ ___AGITATED 22. GLOOMY 1*2 r CnkCrUL 3. _ANGRY 23. GRIM to •J • PI EACAMT _BITTER 24. HAPPY Mi i\M 1 1 LCD 5. CALM 25. HELPLESS IAS 6.__ CHARMI NG 26. HOPELESS **o JCw-UIMl 7. CHEERFUL 27. 1 NSECURE 1*7 JCW 1 1 ncn 1 ML 8._ ^.COMPLAINING 28. JEALOUS hti HO 9. __C0NTENTED 29 JOYFUL 10. __C0NTRARY 30 Kl NOLV cn cm cuu 1 1 COOL 3 1 • r 1 5" • STEADY 12. __CR0SS 32. LONELY 1 citucn 1 < '3 „ ei teas 33. i All 1 1 1/\ LOVI NG 53_ —TENSE _£ASYGOING 3*.__ 5<*-_ TERRI FIEO 5.__ _F EARFUL 35. MEAN 55 THREATENED 16. _FEARLESS 36._ MERRY 56._ THOUGHTFUL 7-_ __FRETFUL 37-__ Ml SERABLE 57._ UNCONCERNED 18. _FRI ENuLY 38._ _N£RV0US 56. UNEASY '9.__ _FRIGHTENE0 39 _0VERC0NCERNE0 59_ __UPSET 20. _FURI0US Uo._ OVERWHELMED 60. WARM 61. WORRYING II

PAGE 91

APPENDIX C SAMPLE HOOD ADJECTIVE CHECK LIST

PAGE 92

Name Each of the kO words In the following list describes feelings or mood. Please use the list to describe your feelings at this moment Mark each word according to these Instructions: If the word definitely describes how you feel at the moment you read It, circle the double plus (++) to the right of Che word. For example, If the word Is calm and you are definitely reeling calm at the moment, circle the double plus as fol lows: calm + ? no (This means that you definitely feel calm at this moment.) If the word only slightly applies to your feelings at the moment, circle the single plus as follows: calm ++ 0 ? no (This means that you feel slightly calm at this moment.) If the word is not clear to you or If you cannot decide whether or not It describes your feelings at the moment, circle the question mark as fol lows: calm ++ + @ no (This means you can't decide whether you are calm.) If you clearly decide that the word does not apply to your feelings at the moment, circle the no as follows: calm ++ + ? (no) (This means that you are sure that you are not calm at the moment.) Work rapidly. Your first reaction is best. Work down the column before going to the next, mi nutes. Mark all the words. This should take only a few angry . ++ ? no concentrating ++ + ? no drowsy . ++ ? no affectionate ++ + ? no nonchalant .. ..+? apprehensive ++ ? no skeptical ++ + ? b 1 ue . ++ + ? no no no no no no no

PAGE 93

bo I d . earnest .. sluggish forgi vi ng clutched up lonely cocky i I ghthearted energetic playful .. suspicious startled defiant engaged in thought ++ + ++ + ++ + -H+ ++ + -H+ ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + no no m no no no no no no no no no no no pleased .. tired kindly fearful regretful egotistic overjoyed vigorous wi tty rebel llous serious .. warmhearted I nsecure self-centered ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ ++ + ++ + ++ + ++ +

PAGE 94

APPENDIX 0 SAMPLE ADJUSTMENT SCORE

PAGE 95

If the statement la true of you, circle 'True." If the statement is not true of you, circle False." 1. I am no one. Nothing seems to \m ma. True Falsa 2. I am optimistic. True False 3. I am a hostile person True False 4. Self-control Is no problem to me. True False 5. I tend to be on my guard with people who are somewhat more friendly than I had expected. True False 6. I try not to think about my problems. True False 7. I have an attractive personality. True False 8. I am a rational person. True Falsa y. I have a horror of failing In anything I want to accomplish. True False 10. I put on a false front. True False 11. I feel helpless. True False 12. I am a good mixer. True False 13* It is difficult to control my aggression. True False \k. I feel insecure wl thin myself. True False 1$. I really am disturbed. True Falsa 16. All you have to do is just insist with me, and I give in. True False 17* I have to protect myself with excuses, with rationalizing. True False 18. I have few values and standards of my own. True False 19. I take a positive attitude toward myself True Falsa bo

PAGE 96

&7 20. I am ambitious. True False 21. I don't trust my emotions. True False 22. I have the feeling that I am just not facing facts. True False 23* I shrink from facing a crisis or difficulty. True Falsa 2k. I am assertive. True False 25. i am tolerant. True False 26. I have initiative. True False 27* I express my emotions freely. True False 28. I can accept most social values and standards. True Falsa 29. I feel hopeless True False 30. It isoretty tough to be me. True False 31. I am contented. True False 32. My decisions are not my own. True False 33I am a failure. True False 34. I make strong demands on myself. True Falsa 35. I despise myself. True Falsa 36. I doubt my sexual power. True False 37* I am afraid of a full-fledged disagreement with a person. True Falsa 38. I am self-reliant. True False 33. I am sexually attractive. True False kQ. I am relaxed and nothing really bothers me. True Falsa kl. I am likable. True False k2. I am worthless. True False k3. I often kick myself for the things I do. True False kk. I am different from others. True Falsa

PAGE 97

45. 1 am poi sed. True Fal se 46. 1 am shy. True Fal se 47. 1 usually feel driven. True False 48. 1 can't seem to make up my mind one way or the other. True False 49. 1 feel emotionally mature. True False 50. 1 can usually live comfortably with people. True False 51. • • 1 understand myself. True Fal se 52. 1 am afraid of sex. True Fal se 53. 1 can usually make up my mind and stick to it. True False 54. 1 have a warm emotional relationship with others. True False 55. 1 usually like people. True Fal se 56. 1 am satisfied with myself. True False 57. 1 often feel humiliated. True False 58. 1 have a feeling of hopelessness. True False 59. 1 am intelligent. True False 60. 1 dislike my own sexuality. True False 61. My hardest battles are with myself. True False 62. 1 am a hard worker. True Fal se 63. 1 am disorganized. True False 64. 1 feel ti red out. True Fal se 65. 1 am responsible for my troubles. True Fal se JO. 1 am unrel i able. True Fa 1 se 67. 1 am a responsible person. True False 68. My personality is attractive to the opposite sex. True False 69. 1 want to give up trying to cope with the world. True False 70. 1 am confused. True False

PAGE 98

89 71. I feci adequate. True False 72. I am liked by most people who know me. True False 73 • I just don't respect myself. True False 74. I am impulsive. True False

PAGE 99

APPENDIX E SAMPLE OF THE MODIFICATION OF THE EXPECTATION Q-SORT

PAGE 100

If you were to have a psychotherapist or counselor to talk with about your problems, what would you expect him to be like? If the statement is true concerning how you would expect the counselor or psychotherapist to be, put a circle around "true." If the statement is false concerning how you would expect this counselor or psychotherapist to be, put a circle around "false." i True Fal M 2. Is careful not to let people waste his time.. True Fal M 3True Fal se k. Likes to have a hand In managing other people's True Fal M |. True Fa se 6. Blunt, straightforward, calls a spade a spade True Fa se 7True Fa se 8. True Fal se 9. True Fa se 10. Is likely to overestimate a person's abilities .. True Fa M It. True Fa se 12. 1* able to sense other people's feelings True Fal se 13. True Fa se \k. True Fal se 15. Expects the individual to shoulder his own reTrue Fa M 16. True Fa sc 17. Is likely to keep his irritations or resentments True Fa se 18. True Fa se 19. True Fa se 20. True Fa se 21. Hard to deceive, does not accept things at face value True Fa la22. True Fa ta 23. Is conscientious about duties and responsibilities True Fa M 2k. True Fa se 25. True Fa se 26. Reacts to most people in about the same way.. fret Fal se 27True Fa M 28. Tries to discover who's to blame for mistakes made True Fa M 29. True Fa se 30. True Fa se 31. Is troubled by the misfortunes of others True Fa se 32. True Fal se 33. Well adjusted, gets along well in the world.. True Fa se 34. Is quick to give encouragement and reassurance .. Has no trouble getting along with people, makes True Fa se 35. True Fa se 36. True Fa M 91

PAGE 101

APPENDIX F SAMPLE PERSONAL INFORMATION QUESTIONNAIRE

PAGE 102

NMM Ward: Aye: Number of years of school: Type of work before hospital: Marital status: Number of chi Idren: Are you service connected: % Number of times in a mental hospital: How long have you been In the hospital this time? If you had a chance to get psychotherapy or counseling would you want it: very much? a little? not at all? 93

PAGE 103

APPENDIX 6 SAMPLE LETTER SENT TO ALL SUBJECTS WHO SPOKE INTO THE TAPE RECORDER

PAGE 104

November 13. 1964 NAME Veterans Administration Hospital Ward 0 Gulfport, Mississippi Dear Mr. : Several weeks ago you were asked to take a series of tests and then talk about yourself into a tape recorder. We were studying a new way of having patients discuss their personal problems. What you said into the tape recorder and your test performance has been reviewed. If you have any questions or wish to discuss any aspect of what you did, please contact me at the Blloxi Veterans Administration Hospital, ID-21541, extension 283. In case I do not see you personally, I would like to express my sincere appreciation for your assistance in this exploratory study. With out your cooperation, and the cooperation of the other patients who participated, we would have been unable to evaluate this new approach. Sincerely yours. Herbert Goldstein, MA. Clinical Psychology Trainee J,

PAGE 105

APPENDIX H SUBJECT RAW DATA

PAGE 106

(Note: In Column 2 I represents Experimental Group, Congruent Therapist, Positive Set; 2 represents Experimental Group, Incongruent Therapist, Positive Set; 3 represents Experimental Group, Congruent Therapist, Neutral Set; k represents Experimental Group, Incongruent Therapist, Neutral Set; 5 represents Control Group, Congruent Therapist, Positive Set; 6 represents Control Group, Incongruent Therapist, Positive Set; 7 represents Control Group, Congruent Therapist, Neutral Set; 8 represents Control Group, Incongruent Therapist, Neutral Set. Columns 7 through 2Q represent changes in the dependent variables from preto posttreatment testing. A change score of 0 Is represented by 09. Negative change scores are represented by increasing numbers and positive change scores are represented uy decreasing numbers. That is, a change score of -1 is represented by 10, a change score of -2 is represented by II, etc. A change score of +1 is represented by 08, a change score of +2 Is represented uy 07, etc. In Column 21 01 represents Tittle, or no speech during treatment, and 02 represents a moderate amount or constant speech during treatment. In Column 22, 01 represents the greatest stated desire for psychotherapy, 02 represents a moderate stated desire for psychotherapy and 03 represents a stated lack of desire for psychotherapy. In Column 2\ 01 represents a nonpsychotlc diagnosis and 02 represents a psychotic diagnosis.) 37

PAGE 107

Subj No Treat G rouD 1 0 l\j p 1 1 m IXCU 1 \J AAPL Adj otu [ c Annr Cone ueac t So Art An 1 1 107 28 04 33 1 c 1 p OQ OQ uy flQ uy no uy u/ 2 1 1 07 2k u^. 97 z / 1 0 1 u 07 1 9 1 z 1 1 uy no uy ao 09 3 1 076 28 30 j u 30 1 3 OA OR OR uo 1 3 1 3 AT 07 4 1 099 w j j 1 3 46 43 OQ 14 1 1 1 1 OS uo 1 z UO 07 1 1 1 3 28 34 48 07 U/ 1 9 1 Z no. uy no. 09 A A 09 6 1 088 20 48 41 1 1 16 1 0 1 1 OA uo 1 1 1 1 UH1 n 1 U 7 i 1 099 3? 9A 1 3 1 J OQ uy 1 n 1 u uo 1 U n£ Uo 13 8 1 J 1 9 1 9Q z;? 3ft ?R 1 1 1 1 OQ uy 1 n 1 U no no 09 09 aO Oo Q 1 093 1 9 40 ^u UP, zu 1 1 1 n 1 u no uy l n 1 U 02 1 7 10 1 102 9Q R 1 P 1 1 9 OQ uy 1 9 t z no uy 1 z n "7 07 1 0 ] ] 1 088 28 30 37 1 1 1 1 07 u/ 1 9 1 Z U5 Uo no UO 1 n 1 0 12 1 078 34 08 18 OA 1 9 Z 1 9 1 z 1 9 1 Z 1 9 1 Z 1 ii 1 *t 13 2 089 24 12 37 04 05 05 06 1 0 05 07 u / 14 2 099 14 40 43 12 1 3 1 ] 0Q Uj7 0Q 07 u/ 1 0 1 u 15 2 088 36 26 29 06 18 09 08 1 9 1 3 OR uo 16 2 1 19 18 44 40 12 08 12 1 1 1 1 07 u/ 1 9 1 z 1 7 2 1 AO 1 02 0 1 2422 22 12 06 09 12 05 1 1 16 18 2 099 14 06 32 09 09 09 08 07 1 1 07 19 2 106 10 44 45 05 13 06 09 06 09 09 20 2 109 22 30 43 05 04 05 07 07 06 08 21 2 103 12 38 43 06 09 08 12 09 12 05 98 Depr Eqot Pleas Aeti v Noneh Skep t Start Amt of —1 \j 1 — >— — 1 1 Des P ;vrhn 09 08 01 10 07 08 09 02 03 09 uz 09 09 08 09 09 09 09 02 03 0? 10 09 09 08 05 09 14 01 01 09 U*L 13 1 1 08 13 12 13 1 1 01 01 0? 09 10 09 09 06 09 09 01 0? 12 15 06 10 1 1 1 1 09 01 01 01 u 1 11 08 07 07 08 10 09 01 Ul 02 09 08 09 09 1 1 10 07 u / 09 U^7 01 0 1 0 1 u 1 09 08 03 08 10 10 1 0 01 01 01 u 1 1 1 11 10 08 12 07 08 01 01 09 14 08 07 08 10 06 09 02 03 0? 09 09 02 06 06 1 1 09 02 01 02 08 06 09 06 09 06 06 02 02 02 1 1 09 03 10 1 1 14 09 01 01 01 09 06 09 09 1 1 09 09 02 01 01 09 09 08 09 09 09 09 02 02 01 1 £ 1 0 1 n 1 U l n 1 U 1 1 E 2 A "7 07 14 09 01 02 02 12 11 12 09 09 07 09 01 02 02 09 12 09 12 09 09 02 03 01 08 06 06 07 07 08 08 02 02 02 10 11 11 09 09 11 09 02 01 02

PAGE 108

99 >ubj No Treat Group 1 .Q,. 1 ntro Neuro Aqe ft ft Al AALL Adj Score Aqq.r Cone Deact So ATT Dep r tqot D 1 a ^ r r 1 eas nC 1 1 V Ci/pn f j La 1 c Amt of Soeech Des Psycho Di aq 22 2 095 26 24 22 aQ Oo l n 1 U 1 4 1 3, 1 1 1 1 1 R 1 0 no uy no uy l n 1 u OQ uy On UO riQ On 07 u / 01 02 02 23 2 112 18 18 40 09 A ~7 07 A 1 07 Uo 1 n 1 U no uy 1 u 1 i 1 1 1 0 07 u/ OQ uy OQ uy 1 0 01 01 02 24 2 089 18 48 44 00 15 1 A 1 0 1 0 1 z uy 1 5 n c Up uo 1 A 1 0 OQ uy 1 c 1 p OQ uy OA 01 01 01 25 3 098 30 42 45 13 09 1 9 1 Z uy 1 j uo 1 9 1 9 07 u / 07 07 u / 07 1 1 0Q 02 02 02 26 3 1 04 40 1 0 45 1 0 l a 1 U no Uo fin uy HQ uy 1 n 1 u no uy no uy OR uo In 1 O 0Q uy 0Q uy 0Q 0Q 01 01 02 27 3 077 18 36 45 07 07 09 09 An uy 1 1 '3 no uy no uy u/ uo 1 9 OQ uy 10 0Q 02 02 28 3 090 21 20 45 09 AA 09 no uy no uy no uy no uy no uy no uy no uy 1 9 0Q 12 0Q 09 01 01 02 29 3 104 15 28 35 13 1 1 Ub no oy no uy Up no uy 1 1 j 1 n 1 u 07 OQ uy 1 0 0Q U P 01 03 02 30 3 109 1 3 38 I. 1 HI 07 1 u no. uy no no uy o£ uo no uy no uy no uy 1 1 I 1 0Q uy 0Q 0Q 01 01 01 31 3 124 19 42 58 1 0 1 1 nQ Uo AA 09 Uo nR uo 1 1 1 n 1 u nR UO uo 07 u/ 0Q uy OQ up 08 01 01 01 32 3 093 24 12 41 1 2 09 uy AO. uy 1 L 1 4 nR uo 1 1 1 I 1 1 i I 1 1 i 1 1 9 0Q uy 0Q 0Q 02 02 02 33 3 1 1 1 30 38 1.0 HO 1 3 u/ no uy HQ Uy 1 9 no uy 1 i 1 1 no 0Q 0 1 12 0Q 06 09 01 03 01 34 3 077 1 0 18 38 31 19 nR. Uo l n 1 u u/ no uy uo 1 0 1 \j 07 06 07 u / 08 08 10 01 01 02 35 3 093 14 42 f, 1 41 14 a/' 06 13 09 no uy 1 1 1 1 1 n 1 u 1 1 1 1 u / 1 n 0Q uy 1 0 1 0 08 01 01 01 36 3 1 08 1 1 44 20 14 1 c 1 b no uy no uy 1 1 1 1 no uy 1 9 u / 1 0 0Q up 0Q 10 13 09 01 02 01 37 4 1 02 26 1 1 /1 n 40 l 0 1 2 i n i U UO 1 9 1 Z no uy no uy 1 9 no uy 0 Q 0Q 12 06 09 09 01 02 02 38 4 098 19 34 J. r 45 1 1 09 1 1 1 1 no uy 1 n 1 u 07 u/ no uy 1 9 uy 06 0Q 10 10 09 01 02 02 39 4 093 14 08 37 05 03 U4 1 5 nR uo Ik 1 H 1 9 1 Z 1 I 1 1 OQ uy 1 9 0Q 10 12 07 02 02 02 4o 4 097 12 28 35 12 05 1 1 n£ Ub no uy nQ uy uo 1 9 1 Z no uy On UO OQ uy 0Q uy 0Q 06 01 03 01 41 4 108 23 13 40 1 1 07 1 1 1 A 1 U nn uy n"7 u/ no uy no uy OQ uy OQ Uy 1 0 1 u OQ uy 0Q U P 0Q 02 02 01 42 4 101 38 34 40 08 1 1 09 10 09 08 1 1 08 1 0 1 0 09 1 A 1 0 An 09 AA 09 no UZ n 1 U 1 09 UZ 43 4 106 40 00 30 10 05 09 09 09 06 11 09 09 1 1 09 08 09 09 02 03 02 44 4 087 10 43 42 10 10 03 1 1 07 12 06 07 09 09 09 07 07 07 01 01 01 —

PAGE 109

>ubj No Treat G roup 1 1 ntro Neuro Aqe AACL Adj Score Aggr Cone Deac 45 4 088 24 26 43 09 10 09 10 07 46 4 102 26 30 24 10 08 10 10 09 47 4 116 17 32 49 09 12 07 10 10 48 4 111 14 16 36 09 12 09 10 10 49 5 108 26 42 30 06 03 1 1 11 09 50 5 092 10 44 44 09 09 09 09 09 51 5 107 29 25 55 06 10 10 10 17 52 5 079 28 36 42 12 08 09 08 1 1 53 5 36 10 09 09 08 11 05 54 5 083 23 35 37 10 10 08 07 07 55 5 118 22 35 31 12 06 07 09 10 56 5 119 30 24 31 08 10 1 1 17 09 57 5 101 25 17 29 10 01 07 10 1 1 58 5 129 13 48 39 07 09 06 07 09 59 5 098 28 11 43 11 04 06 12 04 60 5 092 18 42 28 15 03 04 09 06 61 6 093 22 46 43 10 00 10 09 09 62 091 26 32 44 05 09 07 1 1 09 63 6 098 20 35 34 07 00 09 09 08 64 6 102 26 38 41 13 00 09 1 1 14 65 6 073 2b 1 O no nn uy UD uo HQ 66 6 081 12 42 41 08 05 14 10 08 67 6 087 28 10 32 10 04 11 1 1 09 100 So Amt of Des Aff Anx Depr Eqot Pleas Activ Nonch Skept Start Speech Psycho Diaq 09 09 1 1 10 10 09 08 08 08 06 10 07 09 1 1 09 12 10 09 08 11 10 05 09 09 08 07 10 09 10 13 09 10 1 1 11 12 09 08 09 07 07 10 09 10 09 09 08 09 08 09 10 09 12 12 11 10 06 09 11 11 07 13 13 09 09 07 1 1 09 10 08 09 09 1 1 14 11 09 08 06 12 09 10 09 1 1 09 06 16 09 08 07 10 10 1 1 09 12 1 1 08 09 09 09 09 09 02 09 09 10 09 11 04 09 09 10 09 09 03 09 09 09 09 09 09 10 09 02 13 09 10 10 14 08 09 07 07 12 07 09 08 03 12 08 02 09 09 09 09 10 13 07 1 1 10 09 09 09 09 09 13 07 13 09 14 10 09 10 10 07 12 09 10 10 11 09 09 09 07 14 11 06 08 07 10 09 08 09 09 09 09 08 09 07 09 09 05 06 09 08 09 09 14 10 08 1 1 09 09 07 11 10 08 09 06 02 01 01 02 02 02 02 02 01 03 02

PAGE 110

101 Subj Treat Adj So Amt of Des No Group I Intro Neuro Age AACL Score Aggr. Cone Deact Aff Anx Depr Eqot Pleas Activ Nonch Skept Start Speech Psycho Diag 68 6 100 22 39 35 12 16 10 13 08 08 10 13 08 06 09 06 07 09 69 6 098 18 18 kl 08 11 09 09 09 09 10 09 09 05 08 u 09 09 70 6 087 32 06 37 07 12 09 08 06 09 05 06 07 08 03 09 09 09 71 6 103 18 22 kk 09 1 1 09 10 08 10 08 09 09 08 09 09 08 08 72 6 117 28 ko 30 10 05 09 10 05 09 1 1 09 09 10 1 1 11 09 09 73 7 097 10 kk k5 12 06 12 08 09 09 08 10 08 06 1 1 12 10 09 Ik 7 100 13 22 38 10 07 10 1 1 15 10 07 1 1 09 1 1 09 06 10 09 75 7 109 29 15 35 ok 09 08 09 09 10 09 07 09 11 11 13 08 09 76 7 081 21 16 38 09 10 09 10 10 10 12 1 1 09 10 11 13 11 06 77 7 108 36 02 38 09 10 12 09 09 09 09 09 09 06 09 09 09 09 78 7 093 2k k8 kk 10 ok 08 08 10 10 08 13 09 10 06 11 10 09 79 7 116 19 k2 k3 10 10 09 09 08 09 07 06 09 12 09 1 1 09 09 80 7 090 32 26 k3 10 12 07 08 12 10 10 06 07 09 09 1 1 06 09 81 7 103 17 k2 ko 1 1 08 09 1 1 11 09 11 17 09 09 09 09 10 09 82 7 076 3k 08 k2 09 10 09 10 08 09 10 08 10 07 08 07 08 10 83 7 103 1 1 k3 35 09 12 09 09 09 1 1 09 10 09 1 1 09 10 09 09 8k 7 086 ]k k2 k] 07 06 06 06 07 12 06 10 06 11 11 08 10 09 85 8 086 16 08 32 08 01 08 09 05 09 09 09 10 09 ok 09 07 09 86 8 110 2k 06 37 05 13 09 09 09 09 07 09 09 12 12 09 09 09 87 8 080 27 06 LlLl 11 12 12 11 11 12 12 09 06 13 10 ok 13 11 88 8 095 27 22 k\ 09 06 09 10 09 1 1 09 09 09 1 1 11 09 09 09 89 8 105 2k k8 ko 08 13 09 09 09 15 09 10 12 13 12 09 09 06 90 8 089 23 k2 ok 08 09 ]k 03 06 00 OS 06 03 06 09 Ok 06 I

PAGE 111

102 Subj Treat ~ Adj So Amt of Des No Group I Intro Neuro Age AACL Score Agar. Cone Deact Aff Anx Depr Eqot Pleas Activ Nonch Skept Start Speech Psycho Diaq 91 8 089 23 02 kk 10 10 07 07 09 07 09 09 09 09 07 09 09 09 Q9 8 \2k 28 kk k\ 10 ok 07 09 06 09 07 08 09 10 10 09 11 09 93 8 115 18 19 53 08 07 09 09 09 09 09 09 09 11 07 07 09 09 3k 8 095 23 22 m 1 1 10 09 06 11 11 11 11 13 13 07 09 10 05 95 8 101 23 28 31 10 09 07 1 1 08 09 08 12 10 10 09 09 10 09 96 8 089 2k 2k 32 09 15 09 09 15 06 09 09 12 09 09 03 09

PAGE 112

BIBLIOGRAPHY Abramson, H. A., Jarvik, M. E. Levlne, A., Kaufman, H. R. and Hi rsch, H. W. Lysergic acid diethylamide (LSD-25). XV. The effects produced by substitution of a tap water placebo. J. psycho 1 .. 1955, 4£ 367-383. Apfelbaum, B. Dimensions of transference in psychotherapy. Berkeley, Calif.: Univ. of California Press, 1958. Boryatta, E. F. The new principle of psychotherapy. J. clin. psycho 1.. •959, JLi, 330-339. Dyraond, R. F. An adjustment score for C-sorts. J. of consult, psvchol .. 1953, J2. 339-342. Dymond, R. Adjustment changes over therapy from self-sorts, in Rogers, C. R. and Dymond, R. (eds.) Psychotherapy and personality change. Chicago, Illinois: Univ. of Chicago Press, 1954, 76-85. Eysenck, H. J. The effects of psychotherapy: an evaluation. J. con sult, psvchol 1952, JO, 319-324. Eysenck, H. J. Hanual of the Mauds lev personality inventory London, England: Univ. of London Press, 1959. Eysenck, H. J. Handbook of abnormal psvcholocy. New York, New York: Basic Books, If6t< Fenlchel, 0. The psychoanalytic theory of neurosis New York, New York: Norton and Company, 1945. Festinger, L. A theory of cognitive dissonance Evanston, 111.: Row, Peterson, 1957. Festinger, L. and Bramel, D. The reactions of humans to cognitive dissonance, in Bachrach, A. J. (ed.) Experimental foundations of clinical psychology. New York, New York: Basic Books, 1962. Hscher, H. K. and Dlin, B. M. The dynamics of placebo therapy: a clinical study. Am. J. med. scl .. 1956, £&> 504-512. Foulds, G. Clinical research In psychiatry. J. ment. scl .. 1958, 104. 259-265. 103

PAGE 113

101 Frank, J. 0. Some effects of expectancy and influence in psychotherapy, in Hasserroan, J. H. and Moreno, J. I. (eds.) Progress in psychotherapy. New York, Mew York: Grune and Stratton, 1958. Frank, J. 0. persuasion and heal i oq Baltimore, Md.: Johns Hopkins Press, 1961. Glasser, E. M. and Whittlow, G. C. Evidence for a nonspecific mechanism of habituation. J. physio! .. 1953, 122. 43-M*. Glasser, E. H. and Whittlow, G. C. Experimental errors in clinical trials. Clin, scl .. 1954, JJL, Number 2, 199-210. Glideman, L. H., Gantt, W. H. and Teitelbaum, H. A. Some implications of conditional reflex studies for placebo research. Am. J. psych i at .. 1957, Jil, 1103-H07. Gourlay, N. Covariance analysis and its applications in psychological research. Brit. J. statis psycho! .. 1953, & 25-33. Hagans, J. A., Doering, C. R. Ashley, F. W. and Wolf, S. The therapeutic experiment. J. lab, clin. med .. 1957, 282-285. Hofling, C. K. The place of placebos in medical practice. Gen, pract .. 1955, JJL, 103-118. Hongingf ield, G. The placebo effect— a review of nonspecific factors in treatment. Cooperative studies in psychiatry. 55 Nov., 1963. Jaco, E. G. Patients, physicians and illnesses. New York, New York: Free Press, 1958. Joyce, C. R. Consistent differences In individual reactions to drugs and dummies. Brit. J. of Pharmacol .. 1959, Jj*, 512-521. Kellogg, C. £., Norton, N. W., Lindner, R. H. and Gurvitz, M. Rev! sed beta examination: manual New York, New York: Psychological Corporation, 19^6. Kometsky, C. and Humphries, 0. Relationship between effects of a number of centrally acting drugs and personality. A.M. A. arch, neural Pivchtat-. 1957, 11> 325-327. Kurland, A. A. The drug placebo— its psychodynamic and conditional reflex action. Behav. scl 1957, £ 101-110. Kurland, A. A. The placebo, in Masserman, J. H. and Moreno, J. I. (eds.) Progress In psychotherapy X> N*" York, New York: Grune and Stratton, 1958.

PAGE 114

105 Lasagna, L., Hosteler, F., Von Felsinger, J. H. and Beecher, H. K. A study of the placebo response. Am. J. of med .. 1954, Ijfci 770-779 Laverty, S. G. Sodium amytal and extroversion. J. of neurol. neurosurq. and psychic, 1958, 2L 50-54. Leary, T. and Gill, tf. The dimensions and a measure of the process of psychotherapy: a system for the analysis of the content of clinical evaluations and patient-therapist verbalizations, in Rubinstein, E. A. and Parloff, M. B. (eds.) Research in psychotherapy Washington, 0. C: Am Psychol. Assn., 1959, 62-95. Leiberman, R. An analysis of the placebo phenomenon. J. chron dtsea ., 1964. 15, 761-785. Lindquist, E. F. Design and analysis of experiments in psychology and education.. Be s ton, Mass.: Houghton Mifflin, 1953. Locke, J. Locke's worths ,,. London, England: Davidson, 1823, i 275* Lyerly, S. B., Ross, S., Krugman, A. D. and Clyde, D. J. Drugs and placebos: the effects of instructions upon performance and mood under amphetamine sulphate and chloral hydrate. Psychol, rep .. 1962, 12, 383-392. McNemar, Q. P s ychological statistics. Mew York, New York: Wl ley and Sons, 1955. Maccoby, E. E., Newcomb, T. M. and Hartley, E. L. (eds.). Readings in social psychology New York, New York: Holt and Company, 1958. Martin, B., Lundy, R. M. and Lewin, M. H. Verbal and G.S.R. responses In experimental interviews as a function of three degrees of "therapist" communication. 4. of ab. and soc. psycho! .. I960, o Q. 234-240. Newcomb, T. M. S ocial psychology New York, New York: Holt and Company, 1958. Nowlls.V. Description and analysis of mood. N. V. acad. of sci. annals 1956, 6i, 345-355. Nowlis, V. Some studies of the influence of films on mood and attitude. Technical report number Research project number NR 171-342, Jan., I960. Nowlis, V. Methods for studying mood changes produced by drugs. Rev psvchol. appl .. 1961, JJL. 373-386. Ray, W. S. Ajn introduction to experimental design New York, New York: Macmi 1 1 an Company 1 960

PAGE 115

10b Rogers, C. ft. Client centered therapy Boston, Mass.: Houghton Mifflin, 1954. Rogers, C. R. and Dymond, R. Psychotherapy and personality change Chicago, Illinois: Univ. of Chicago Press, 1954a. Rogers, C. R. On becoming a person Boston, Mass.: Houghton Mifflin, 1961. Rosenthal, 0. and Frank, J. D. Psychopathology and the placebo effect. Psvchol. bull .. 1956, 52, 294-302. Shapiro, A. K. Factors contributing to the placebo effect. Am. J. of psvchoth .. 1964, !§,, supplement 1, 73-88. Shoben, E. H. Some observations on psychotherapy and the learning process, In Mowrer, 0. H. (ed.) Psychotherapy, theory and research New York, New York: Ronald Press, 1953. Siegel, S. Nonparametri c statistics Mew York, New York: McGraw-Hill, 1956. Slack, C. W. Experimenter-subject psychotherapy: a new method of introducing intensive office treatment for unreachable cases. Ment. hvg .. I960, f& 238-256. Strupp, H. H. Psychotherapy. Ann, rev, psycho) ,.. 1962, JjJ, 445-478. TibbeUs, R. W. and Hawklngs, J. R. The placebo response. J. of ment. sci .. 1956, IPJk 60-66. Uhlenruth, ".. H., Canter, A., Neustadt, J. 0. and Payson, H. E. The symptomatic relief of anxiety with meprobanate, phenobarbital and placebo. Am. J. osvchiat.. 1959. Hi, 905-910. Whltehorn, J. C. Psychiatric Implications of the "placebo effect." Am.. J. psvchlat .. 1958, 114, 662-664. Wolf, S., Ooering, C. R., Clark, M. L. and Hagans, J. A. Chance distribution and the placebo "reactor." J. lab, clin. reed .. 1957 49 837-841 Wolf, S. The pharmacology of placebos. Pharmacol, rey 1959, 1 1 698-702. Zuckerman, M. The development of an affect adjective check list for the measurement of anxiety. J, consult, psycho 1 .. i960, 2it> 457-462. Zuckerman, H. and Blase, 0. V. Replication and further data on the affect adjective check list measure of anxiety. J. consult, psycho] .. 1962, 2jb, 291.

PAGE 116

BIOGRAPHICAL NOTE Herbert Goldstein was born on July 20, 1936 in New York City, New York. He graduated from Allentown High School, Allentown, Pennsylvania, in June, 1954. He received the Bachelor of Arts degree from the University of Florida In June, 1958. He served in the United States Army from September, 1958 to September, I960. He entered the Graduate School of the University of Florida in September, I960. In December, 1962 he received the Master of Arts degree with a major in psychology and a minor in educational foundations. While in graduate school in the University of Florida he held graduate assistant ships in the Psychology Department, In the Reading Laboratory and Clinic and a cl inl clanship at the University Counseling Center. He served as Psychology Trainee In the Gulfport-Bi loxl Mississippi Veterans Administration Hospitals from January, 1964 until December, 1964. He Is currently a USVRA fellow completing his predoctoral internship at the J. Hillis Miller Health Center.

PAGE 117

This dissertation was prepared under the direction of the chairman of the candidate's supervisory committee and has been approved by all members of that committee. It was submitted to the Dean of the College of Arts and Sciences and to the Graduate Council, and was approved as partial fulfillment of the requirements for the degree of Doctor of Phi losophy April 24, 1963 Dean, Graduate School Supervisory Committee:


55
TABLE 23
PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND
SCORES ON THE MPI INTROVERSION (LOW SCORES)-EXTROVERSION (HIGH SCORES). THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS
Moods of the MACL
So. Adjustment
Treatment Conditions
AA.CL
Aggr.
Cone.
Deact.
Aff. Anx.
Depr.
Eqot-
Pleas.
Activ.
Nonch.
Skept.
Start.
Score
Experimental Group
Positive Set, Congruent Therapist
+.21
+.03
-.21
+. 1 0
-.54* +.01
+ .20
+.32
-.07
+ .57*
+ .53*
+ .44
+ .25
+ .71*
Positive Set, Incongruent Therapist
+. 06
-. 12
+ 09
-.46
vO
1
LTV
CM
+.05
+.51
+ .03
+ .41
+.13
+.21
+ .24
-.13
Neutral Set, Congruent Therapist
+.07
+ .02
-.1 1
-.33
-.14 +.03
-.06
+ .14
-.31
-.11
+ .29
+ .40
-.13
+ .26
Neutral Set, Incongruent Therapist
-.02
-.31
+.09
-.11
+.60* +.41
-.28
-.36
+.02
+ .01
+ .01
+ .01
-.51
+ .01
Control Group
Positive Set, Congruent Therapist
+. 12
-.36
-.52
.00
*
-.59* +.30
00
0
1
-.06
-.16
+ 06
+ .10
+ .30
+.57*
-.03
Positive Set, Incongruent Therapist
-.07
+.52
+ .08
+. 08
LA
O
O
-3r
1
+. 14
+.37
-.28
+ .32
+ 48
+ .02
-.03
-.04
Neutral Set, Congruent Therapist
+ .31
+.04
-.05
+. 11
+.35 -.37
+.40
-.24
+.37
+ .29
-.04
+ .61*
-.15
-.36
Neutral Set, Incongruent Therapist
-.27
-.18
-.14
-.25
-.17 +.01
+. 11
+.29
-.17
-. 65*
+. 15
-33
-.19
-.30
*Significant ^ .05.


APPENDIX C
SAMPLE MOOD ADJECTIVE CHECK LIST


67
Dissonance is said to be a motivating state which is comparable to other
drive states, and just as hunger produces physical discomfort, dissonance
results in psychological discomfort. When a patient-subject is instructed
that the characteristics of his psychotherapist are incongruent with his
personal expectations concerning this psychotherapist, the patient-
subject experiences dissonance. This in turn, reduces the effectiveness
of the placebo treatment, as is shown by the statistical analyses which
indicate that the AACL and MACL moods of Startle and Deactivation de
creased significantly more for the Congruent Therapist Group, while the
MACL mood of Pleasantness increased significantly more in the Congruent
Therapist Group.
(It is noteworthy that although the Apfelbaum questionnaire [see
Appendix E) instructs subjects to respond on the basis of their expecta
tions. any individual subject may have completed the items in terms of
his ideal psychotherapist or in terms of a desired psychotherapist.
There is no apparent way in which to evaluate these possibilities. If,
however, any individual subject did respond In terms of an ideal or a de
sired psychotherapist, he may have received additional, or reduced, posi
tive set, in addition to having the nature of the treater confirmed or
not confIrmed.)
In addition to demonstrating that the patient's perception of the
nature of the treater Is one significant aspect of inducing a greater or
lesser placebo effect, this research also suggests that the nature of the
treatment situation itself is important. Analysis of Hypothesis III in
dicates that subjects who are told that the placebo psychotherapy will be
extremely helpful to them (Positive Set Group) profit more from the


105
Lasagna, L., Hosteler, F., Von Felsinger, J. H. and Beecher, H. K. A
study of the placebo response. Am. J. of med.. 1954, jj, 770-779
Laverty, S. G. Sodium amytal and extroversion. J. of neurol. neurosurg.
and aaifitLtt.* 1958, L 50-54.
Leary, T. and Gill, H, The dimensions and a measure of the process of
psychotherapy: a system for the analysis of the content of
clinical evaluations and patient-therapist verbalizations, in
Rubinstein, E. A. and Parloff, H. B. (eds.) Research In psycho
therapy. Washington, 0. C.: Am Psychol. Assn., 1959. 62-95.
Leiberman, R. An analysis of the placebo phenomenon. J. chron disea..
1964, JLi, 761-785.
Lindquist, E. F. Design and analysis of experiments in psychology apA
education. Bcston, Mass.: Houghton Mifflin, 1953.
Locke, J. Locke's works. London, England: Davidson, 1823, 275.
Lyerly, S. B., Ross, S., Krugman, A. D. and Clyde, D. J. Drugs and pla
cebos: the effects of instructions upon performance and mood
under amphetamine sulphate aid chloral hydrate. Psvchol. rep..
1962, 383-392.
McNemar, Q. Psychological statistics. New York, New York: Wiley and
Sons, 1955.
Maccoby, E. E., Newcomb, T. H. and Hartley, E. L. (eds.). Readings in
social psychology. New York, New York: Holt and Company, 1958.
Martin, B., Lundy, R. M. and Lewin, M. H. Verbal and G.S.R. responses in
experimental Interviews as a function of three degrees of "ther
apist" communication. J. of ab. and soc. psycho!.. I960, bO.
234-240.
Newcomb, T. M. Social psychology. New York, New York: Holt and Company,
1958.
Nowlls.V. Description and analysis of mood. N. Y. acad. of sci. annals.
1956, 6*, 345-355.
Nowlis, V. Some studies of the influence of films on mood and attitude.
Technical report number 7. Research project number NR 171 34Z,
Jan., i960.
Nowlis, V. Methods for studying mood changes produced by drugs. Rev.
asvcho 1, app_l.. 1961, jJL, 373-386.
Ray, W. S. An introduction to experimental design. New York, New York:
Macro!11an Company, 1960.


44
Changes after treatment on the AACL indicated that the Congruent Thera
pist, Experimental Group and the Incongruent Therapist, Control Group
significantly reduced anxiety when compared to the Incongruent Therapist,
Experimental Group and the Congruent Therapist, Control Group (p ^ .01)
(see Table 5). Similarly, the MACL Aggression mood indicated that the
Congruent Therapist, Experimental Group and the Incongruent Therapist,
Control Group significantly reduced aggression when compared to the In
congruent Therapist, Experimental Group and the Congruent Therapist, Con
trol Group (p 4 *01) (see Table 6). These interactions suggest that as
measured on the AACL and the MACL mood of Aggression, the superiority of
the Experimental Group treatment pertains only for those subjects who are
given the congruent therapist Instructions, while Experimental Group sub
jects given the incongruent therapist instructions report "nontherapeutlc"
changes.
While only three dependent variables significantly distinguished
between the Experimental Group and Control Group (in addition to two in
teraction effects) and the Adjustment Score Indicated no difference as
predicted, no dependent variables significantly changed In a direction
opposite to that predicted. Thus Hypothesis I appears to be supported.
Hypothesis it states that subjects, when offered a "psychothera
pist" who is congrccoc with their expectancies, will show more "positive
therapeutic changes" on the dependent variables than subjects assigned a
"psychotherapist" who is incongruent with their expectations. This was
found to be the case in the analysis of the main effects in four of the
14 dependent variables. The AACL showed that subjects given a therapist
congruent with their expectations decreased anxiety after treatment


101
Subj Treat Adj So Amt of Des
No
Group
i .a.
1 ntro
Neuro
Aqe
AACL
Score
Aqqr.
Cone
Deact
Aff
Anx
Dep r
Eqot
Pleas
Acti v
Nonch
Skept
Start Speech Psycho Diaq
68
6
100
22
39
35
12
16
10
13
08
08
10
13
08
06
09
06
07
09
69
6
098
18
18
47
08
11
09
09
09
09
10
09
09
05
08
11
09
09
70
6
087
32
06
37
07
12
09
08
06
09
05
06
07
08
03
09
09
09
71
6
103
18
22
44
09
1 1
09
10
08
10
08
09
09
08
09
09
08
08
72
6
117
28
40
30
10
05
09
10
05
09
11
09
09
10
1 1
11
09
09
73
7
097
10
44
45
12
06
12
08
09
09
08
10
08
06
11
12
10
09
74
7
100
13
22
38
10
07
10
1 1
15
10
07
11
09
11
09
06
10
09
75
7
109
29
15
35
04
09
08
09
09
10
09
07
09
11
11
13
08
09
76
7
081
21
16
38
09
10
09
10
10
10
12
11
09
10
11
13
11
06
77
7
108
36
02
38
09
10
12
09
09
09
09
09
09
06
09
09
09
09
78
7
093
24
48
44
10
04
08
08
10
10
08
13
09
10
06
11
10
09
79
7
116
19
42
43
10
10
09
09
08
09
07
06
09
12
09
11
09
09
80
7
090
32
26
43
10
12
07
08
12
10
10
06
07
09
09
11
06
09
81
7
103
17
42
40
1 1
08
09
11
1 1
09
11
17
09
09
09
09
10
09
82
7
076
34
08
42
09
10
09
10
08
09
10
08
10
07
08
07
08
10
83
7
103
1 1
43
35
09
12
09
09
09
11
09
10
09
11
09
10
09
09
84
7
086
14
42
41
07
06
06
06
07
12
06
10
06
11
11
08
10
09
85
8
086
16
08
32
08
01
08
09
05
09
09
09
10
09
04
09
07
09
86
8
110
24
06
37
05
13
09
09
09
09
07
09
09
12
12
09
09
09
87
8
080
27
06
44
11
12
12
1 1
11
12
12
09
06
13
10
04
13
11
88
8
095
27
22
41
09
06
09
10
09
1 1
09
09
09
1 1
1 1
09
09
09
89
8
105
24
48
40
08
13
09
09
09
15
09
10
12
13
12
09
09
06
90
8
089
23
42
41
04
08
09
14
03
06
00
06
06
03
06
09
04
06


17
adjustment (The Adjustment Score) were administered twice to each subject
--once prior to and once immediately after the treatment. Changes In
scores on these tests were used to evaluate changes as described in the
four hypotheses.
(I) The Anxiety Adjective Check List (AACL) (see Appendix B).
The Anxiety Adjective Check List (AACL) (i960) has been shown by
Zuckerman to be a "quick measure of anxiety level." It was chosen for
this study because of its demonstrated validity and reliability and the
fact that the instructions may be modified to permit sensitive evaluation
of changes in the level of manifest anxiety over short periods of time.
The check list is an empirically developed pool of 61 adjectives
with varying affective connotations. Twenty-one adjectives were identi
fied, on the basis of an item analysis, as discriminating between psychi
atric patients rated high on anxiety and normal control subjects. Of
these items, eleven are anxiety negative adjectives and ten are anxiety
positive. Subjects may obtain a score from 0 to 21 on the check list.
Zuckerman has developed two forms. One instructs subjects to denote how
they feel "In general' and the other asks how they fee "today." Al
though standardized on college students, Zuckerman reports no differences
in performance due to age, sex or level of education.
in two studies (i960, 1962) Zuckerman reports internal and test-
retest reiiabiIities for the "In general" form to be significant at
greater than the .001 level. However, the "today" form, although simi-
arly internally reliable, had law test-retest reliability (r .68,
o ^ ,00i). These results were anticipated and suggest that the today
form is sensitive to short-term fluctuations in anxiety. Validity was


November 13* 1964
NAME
Veterans Administration Hospital
Ward D
Gulfport, Mississippi
Dear Mr. :
Several weeks ago you were asked to take a series of tests and
then talk about yourself into a tape recorder. We were studying a new
way of having patients discuss their personal problems.
What you said into the tape recorder and your test performance
has been reviewed. If you have any questions or wish to discuss any
aspect of what you did, please contact me at the Biloxi Veterans Admin-
1strat ion Hospital, ID-21541, extension 283.
In case I do not see you personally, I would like to express my
sincere appreciation for your assistance in this exploratory study. With
out your cooperation, and the cooperation of the other patients who par
ticipated, we would have been unable to evaluate this new approach.
Sincerely yours,
Herbert Goldstein, MA.
Clinical Psychology Trainee
95


4
TABLE 15
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL NONCHALANCE MOOD
Source
df
V
F
A (experimental vs. control)
.87
1
.87
.23
B (congruent vs. incongruent)
1.43
1
1.43
.38
C (positive vs. neutral)
.17
1
.17
.04
AB
3.52
1
3.52
.92
AC
.24
1
.24
.06
BC
8.46
1
8.46
2.22
ABC
.37
1
.37
.10
error
331. SO
IZ
3.81
Total
346.96
94
TABLE I5A
PRETREATMENT MEANS (X,), POSTTREATMENT MEANS (X,) AND MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL NONCHALANCE MOOD
Experiment Group Control Group
Congruent Incongruent Congruent Incongruent
Therapist
Therapist
Therapist
Therapist
Post tive
Set
X, 3.08
X2 3.2S
D +.17
X, 3.42
*2 -
D -.50
X, 3.83
x2-Ui
D -.50
X] 2.08
x2 US.
D -.33
Neutral
Set
X, 2.00
x2 1>83
0 -.17
Xj a 3.08
x2 1*08
D .00
X, 3.58
*2 1^8
D -1.00
X, 2.00
*2 U&
D +.58


29
f
By inspection, the data appear to have significant homogeneity of
variance. Norton (as cited by Lindquist, 1953) demonstrates that unless
the heterogeneity of variance is so extreme that it is readily apparent
upon inspection, the effect upon the F distribution will be negligible.
No formal tests of homogeneity of variance were conducted. Statistical
significance was set at the .05 level prior to the analysis. Below are
the summary tables of the analyses of covariance for each of the depend
ent variables (Tables 5 through 18). In addition, the mean change scores
frompre-to posttesting in each treatment condition are presented for
each of the dependent variables (Tables 5A through 18A).
Hypothesis I states that the Experimental Group (placebo treat
ment group) will show more "positive therapeutic changes" in the AACL and
the moods of the MACL than the Control Group (wait group). This predic
tion was supported in the analyses of the main effects of three of the
fourteen dependent variables. On the MCL, the Experimental Group showed
a greater decrease in reported anxiety after treatment than did the Con
trol Group (p ^ .05) (see Table 5)* On the MACL Social Affection mood,
the difference was significant between the increase in Social Affection
for the Experimental Group after treatment and the decrease after treat
ment in Social Affection for the Control Group (p^ .05) (see Table 9).
Similarly, the Anxiety mood on the MACL decreased after treatment for the
Experimental Group and increased for the Control Group (p^.05) (see
Table 10).
Two dependent variables Indicated that an interaction exists be
tween the experimental vs. control treatment (Hypothesis l) and the con
gruent vs. incongruent therapist treatment (see below. Hypotheses II).


59
It is believed sufficiently Important to stress once again the
tentativeness of all relationships discussed under Hypothesis IV. It is
this writer's opinion that due to the low number of significant correla
tions, the data basically support the null hypothesis that no conclu
sively meaningful relationships exist between the dependent variables and
the measured individual differences.
It was believed that a subject's reported level of desire for
psychotherapy would be related to his amount of verbalization when he was
permitted to speak to the "psychotherapist." The tenability of this be
lief was evaluated by multiple regression technique utilizing only the
Experimental Group and the correlations are presented below in Table 26.
TABLE 26
HULTIPLE REGRESSION COEFFICIENTS BETWEEN AMOUNT OF SPEECH ANO
DESIRE FOR PSYCHOTHERAPY PRESENTED FOR EACH INDEPENDENT
VARIABLE (EXPERIMENTAL DATA ONLY)
Congruent
Therapist
Incongruent
Therapist
Positive
Set
Neutral
Set
Amount of
Speech
.50
.02
.49
.08
vs.
Desire for
Psychotherapy
n s 24
o
c
s

e
n s 24
Amount of verbalization, which was originally rated on a four-
point continuum (see p. 26) was divided, for the purpose of this evalua
tion, Into two groups: (1) constant speakers and moderate speakers, and
(2) little speakers and no speech. Desire for psychotherapy was rated on
a three-point scale by each subject (see Appendix F).


4!
TABLE 16
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL SKEPTICISM MOOD
Source
$s
4#
V
r
A (experimental vs. control)
1.98
l
1.98
.74
B (congruent vs. Incongruent)
.49
l
.49
.18
C (positive vs. neutral)
.10
l
.10
.04
AB
.36
l
.36
.13
AC
2.17
l
2.17
.81
BC
.00
l
.00
.00
ABC
.01
1
.01
.004
error
ULM
liZ
2.67
Total
236.98
94
TABLE I6A
PRETREATMENT MEANS (X.), POSTTREATMENT MEANS (X,) AN0 MEAN
CHANGE SCORES (0) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL SKEPTICISM MOOD
Experimental Group
Control
Group
Congruent
Therapist
Incongruent
Theraplst
Congruent
Theraplst
1ncongruent
Therapist
Posi tive
X, 1.92
X, 2.75
X, 2.75
X, 1.67
Set
x2 105
0 -.17
x2 2*50
D -.25
X2 -
D -.42
X2 1.75
D +.08
Neutral
X, 2.00
X, 2.17
X, 3.00
x2 iOl
D -.17
X, = 1.75
Set
X2 1.55
D = -.42
X2 1.91
0 -.26
X2 2J2
0 +.42


METHOD
The study involved 96 subjects Mho Mere placed into eight treat*
roent conditions as diagrammed below in Tabie i.
TABLE i
THE EXPERIMENTAL DESIGN
Experimental Group
Control
Group
Congruent
Therapist
1nstructions
1ncongruent
Therapist
Instructions
Congruent
Therapist
Instructions
1ncongruent
Therapi st
1nstructions
Posi ti ve
set
concerning
treatment
outcome
n a 12
n 12
n 12
n 12
Neutral
set
concern!ng
treatment
outcome
n 12
n 12
n 12
n 12
n 48
n 48
The technique and procedure of establishing these treatment conditions,
and the way in which the data were collected are fully explained in the
Procedure section. The tests which were administered to the subjects and
the use to which the test information was put in the design are described
In the Means of Measurement section.
13


22
Dymond (1954) demonstrated that a significant difference existed
between the mean adjustment scores on the Q-sort of subjects in two months
of client-centered psychotherapy as compared to control subjects who
waited a comparable length of time (p ^.01). Dymond reports that the
test-retest reliability of this control group is +.86. Validity was es
tablished in two ways: (I) rank order correlations between self-ideal
correlations and the Q-sort adjustment score of clients before therapy
began was +.83 and the rank order of these same subjects after therapy
was +.92; (2) each therapist rated the success of his therapy for each
client and these ratings correlated at better than the 5 per cent level
with the subject's own scores on the 0-sort.
For the purpose of the present study, the 74 items on the Q-sort
were randomly organized into a questionnaire. The subject is asked to
mark an item "true" if it pertains to him and mark "false" if the item
does not pertain to him. A total of the positive items marked true and
negative items marked false constitute a subject's adjustment score.
The AACL, MACL and the Adjustment Score were administered twice
to all subjects, once prior to and once immediately after the treatment.
The change In scores frompre- to posttreatment testing constitute meas
ured changes in manifest anxiety, 12 independent moods and general level
of psychological adjustment.
(4) Ite .Mauds 1 ffiL.Pa.riffiHfll Uy I nyent;ory (HP.Q .
In order to evaluate part of Hypothesis IV, the HPI (Eysenck,
1959, 1961) was administered to obtain subject differences in level of
neurotic!sm and degree of introversion. "Extroversion, as opposed to In
troversion, refers to the out-going, uninhibited, social proclivities of


18
evaluated by administrations of the today form to a class in general psy
chology. The check list was given on 10 nonexamination days and three
examination days. The examination-day administrations resulted In a sig
nificantly higher reported level of anxiety than the nonexamination-day
administrations (p<.0005). The today form, with instructions modified
to read "right now," will be used in this study.
(2) The Hood Adjective Check List (MACL) (see Appendix C).
Nowlis (1956, 19, 19I) assumed mood to be a multidimensional
characterization of a person's feeling or behavior which is accessible to
seif report. The MACL resulted from a series of seven factor analytic
studies by him (i960). In those seven studies Nowlis demonstrated that
40 adjectives, from his original pool of 200, have relatively consistent
and high loadings on twelve separate factors. These 40 adjectives con
stitute the MACL which Is scored separately for each of the twelve fac
tors (which are the identified moods). Further, the twelve factors
(moods) are divided into three groups: those "factors most consistently
identified," those "factors identified fairly consistently" and "factors
tentatively identified." Below are the twelve mood factors and the par
ticular adjectives constituting the factors.
Certain specific moods on the MACL have particular relevance for
the purpose of this study because they would be expected to change in a
certain direction as a result of successful psychotherapy. Aggression
and Deactivation would be expected to decrease, while Social Affection
would be expected to increase. Similarly, after successful psychotherapy
the moods of Anxiety and Depression would be expected to decrease and
Pleasantness and Activation would be expected to increase. At a lower


PLACEBO PSYCHOTHERAPY AND CHANGE IN
ANXIETY, MOOD AND ADJUSTMENT
By
HERBERT GOLDSTEIN
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
April, 1965


INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS PRIOR TO
ESTABLISHING THE TREATMENT CONDITIONS IN THE STUDY
I'd like you men to do several things this afternoon. I think
you will enjoy these tasks. You will learn something about yourself and
may profIt from what we do here for the next hour or so.
(Distribute the personal Information questionnaire.)
I'd like for you to answer these questions about yourselves.
(Distribute the Maudsley Personality Inventory.)
The next questionnaire Is this one. (Read directions.) Begin.
(Distribute the Anxiety Adjective Check List.)
The next questionnaire is this one. (Read directions.) Please
answer honestly the way you feel right now. Begin.
(Distribute the Mood Adjective Check List.)
The next questionnaire is this one which asks you about your mood
and the way you feel right now. (Read directions.) Some of the items in
this list are like items in the last list, but please answer them anyway.
Be honest with yourself and just answer the way you feel right now.
(Distribute the Modification of the Expectation Q-$ort.)
This questionnaire is a little different. It says(read direc
tions). Now, don't think too long on any of these, but answer them as
best you can.
(Distribute the Adjustment Score.)
This questionnaire asks you ho* you feel about yourself and also
about things in general. Read the instructions and work quickly
INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS TO DETERMINE
POSITIVE SET OR NEUTRAL SET
Positlve Set Instructions
I can tell you now what this is all about. It's a real good op
portunity for you men because you are being given a chance to talk with a
great psychiatrist. Now, you can talk to him about anything you'd like
77


48
TABLE 19
SUMMARY OF THE DEPENDENT VARIABLES AND THE SIGNIFICANCE LEVELS AT WHICH THEY DISTINGUISHED
BETWEEN THE INDEPENDENT VARIABLES
Now 1 Is C1ea r1y
1dentified Factors
Nowlis Less Clearly
1 dent!f1ed Factors
Nowlis Least Clearly
1dentified Factors
AACL
Aggr. Cone. Deact.
So.
Aff.
Anx. Depr. Eqot.
Pleas.
Activ. Nonch. Skept. Start.
A (experimental
vs. control)
05
05
05
B (congruent vs.
incongruent)
.001
.05
05
.001
C (positive vs.
neutral)
05
.05
AB
.01
.01
AC
BC
ABC


36
TABLE 11
SUMMARY TABLE OF THE ANALYSIS OF COVARIANCE FOR THE
MACL DEPRESSION MOOD
Source
ss
df
V
V
A (experimental vs. control)
3.04
1
3.04
.78
B (congruent vs. incongruent)
.00
1
.00
.00
C (positive vs. neutral)
.49
1
.49
.13
AB
.86
1
.86
.22
AC
1.24
1
1.24
.32
BC
.37
1
.37
.10
ABC
5-55
1
5-55
1.43
error
317,51
8Z
3.88
Total
394.06
94
TABLE 11A
PRETREATMENT MEANS (X.), POSTTREATMENT MEANS (X,) ANO MEAN
CHANGE SCORES (D) FOR EACH TREATMENT CONDITION OF THE
DESIGN ON THE MACL DEPRESSION MOOD
Experimental Group
Control
Group
Congruent
Therap1st
1nconnruent
Therapist
Congruent
Therapist
1ncongruent
Therapist
Posltlve
Set
X, 3.92
D -1.25
X, 3.75
*2-1^22
D -.83
X, 3.58
X2 -i,66
D 4.08
X, 3.75
X2 2^81
D -.92
Neutral
Set
X, 5.33
X*2 4.42
0 -.91
X, 4.17
*2 L31
D -.84
X, 4.25
*2 IM
0 -.83
X, 3.08
X2 2^22
D -.16


51
TABLE 21
SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE
RANDOM DISTRIBUTION OF MPI NEUROTICISM SCORES
Source
ss
m
V
F
A (experimental vs. control)
2.06
i
2.06
.01
B (congruent vs. incongruent)
408.38
i
408.38
2.04
C (positive vs. neutral)
135.40
i
135.40
.68
AB
.37
i
.37
.00
AC
57.02
i
57.02
.28
BC
376.03
i
376.03
1.88
ABC
63.37
i
63.37
.32
w Cells
IZi&ldi
&
200.27
Total
18,665.96
95
(An F of 3*95 is significant at the .05 level.)


APPENDICES


66
motivation, faith, learning and cerxiitioning, trust, confidence and
previous experience with healers.
Shapiro (and Whitehorn, 1953) further indicate that no single
characteristic of treatment has been consistently shown to underlie and
to produce the placebo effect. The conflicting and inconclusive evidence
in the literature Is most conceivably due to the unclear nature of the
placebo effect, differences of measurement of the effect from research to
research and the different theoretical frameworks within which different
experiments are planned.
Thus far, the discussion has been concerned with the contrast be*
tween the Experimental Group (talkers) and the Control Group (waiters)
and several of the characteristics of psychological treatment which may
determine the placebo effect. The conditions under which the placebo ef
fect is most likely to take place or most likely to be lessened, are
further clarified when attention is drawn to the way in which the treat
ment groups were divided within the present experiment. Analysis of
Hypothesis II, indicates that patients who are offered a "psychothera
pist" congruent with their expectations, report more "positive thera
peutic changes" on measures of anxiety and moods than do patients who are
assigned a psychotherapist who is incongruent with their expectations
concerning the nature of the psychotherapist to be assigned. That is,
the subject's perception of the treater or doctor is a significant aspect
in encouraging or discouraging the placebo effect.
Festinger (1957) and Festlnger and Bramel (Bachrach, 1962) pro
pose that dissonance results when two cognitions which a person holds are
inconsistent with each other according to the expectations of the person.


6o
The correlations presented in Table 26 suggest that if a subject's
desire for psychotherapy is high, his amount of verbalization is high if
he is given a congruent therapist and/or a positive set. If, however, he
is given an incongruent therapist and/or a neutral set there Is no rela
tionship between his desire for psychotherapy and his amount of speech.
Presented below in Table 27 are co-relationships between (I)
amount of speech; (2) desire for psychotherapy; and (3) diagnosis on the
one hand, and change scores from pre- to posttesting on the AACL and the
moods most clearly identified by Now!is on the other (data only for the
ExperI mental Group).
One interesting conclusion which may be drawn rrom these data
concerns the relationship between amount of speech and change scores on
the AACL. It appears that subjects who speak the most Increase anxiety
from pre- to posttreatment testing. This Is true in all treatment con
ditions and the correlation for ail groups is +.30 (p < .05). Of
further interest Is the fact that level of desire for psychotherapy does
not appear, by inspection, to appreciably Influence the amount of change
on these five dependent variables from pre- to posttreatment testing.
Similarly, by inspection, diagnosis (psychotic or nonpsychotic) does not
appear to be related to the amount of reported change In personal comfort
after treatment as measured by the five dependent variables.


63
As a result of these findings in the current study, it is pro
posed that the placebo effect does contribute to the outcome of psycho
therapy, and that a measurement of the amount of contribution appears to
have been made. However, it is important to question precisely what has
been evaluated by a comparison between the Experimental Group (talkers)
and the Control Group (waiters). If the subjects who spoke received pla
cebo treatment and the subjects who waited received no treatment, the
statistical comparison between these two groups was an accurate measure
ment of the placebo effect. However, it is believed that those in the
Control Group, who waited, but were promised a therapist, and who changed
on the dependent variables in a generally "nontherapeutic" direction, re
ceived some placebo benefit from the procedure (e.g., see Tables I1A and
I3A). This finding does not Tnftuence the conclusion that the placebo
effect in psychotherapy does exist, however, the precise therapeutic bene
fit of the placebo, as measured in this study, has not oeen compared to
the therapeutic benefit of no placebo treatment. Instead, the comparison
between the Experimental and Control Groups appears to have been an anal
ysis between a more effective (Experimental Group) and a questionably ef
fective (Control Group) placebo treatment.
In addition to congruence of expectations regarding the charac
teristics of the assigned psychotherapist and positive set regarding
treatment outcome (see below), may other "psychologic mechanisms" through
which the placebo effect takes place (Shapiro, 1964) be identified in
order to account for the finding that this treatment did result In a pla
cebo effect? Shapiro (1964) reports that throughout the history of medi
cine, "methods of depletion (which are currently accepted as placebos)


54
Activation. That is, in this case, brighter subjects tend to increase
Social Affection (personal comfort) as a result of the placebo situation,
while In the former case, brighter subjects appear to increase anxiety
and become less active. Of further interest Is another tentative finding
suggesting that brighter subjects, when given a congruent therapist, tend
to Increase scores on the MACL Skepticism mood after treatment, but de
crease skepticism after treatment when given an incongruent therapist.
From Table 23 several interesting, but also tentative relation
ships may be discussed. In the Experimental Group, the more Introverted
subjects who are *lven the Positive Set, Congruent Therapist Instructions
tend to increase reported Social Affection after treatment. However, the
more introverted subjects who are given the Neutral Set, Incongruent
Therapist instructions tend to decrease the MACL mood of Social Affection.
An additional trend (with the exception of the Control Group, Neutral Set,
Incongruent Therapist subjects) appears to be that introverted subjects,
more than extroverted subjects, tend to decrease scores on the MACL mood
of Skepticism after treatment.
Several relationships presented in Table 24 may be tentatively
discussed. It appears that in the positive set cells of the Experimental
Group and in the neutral set cells of the Control Group the older the
subject, the larger the increase in the MACL mood of Social Affection.
Similarly, In the positive set cells of both the Experimental and Control
Groups, the younger the subject the greater the increase in reported MACL
Egotism mood after treatment. Another effect appears to exist in the re
lationships between age and change in anxiety after treatment. On both
the AACL and the MACL Anxiety mood, it Is noted that younger subjects


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E5157SL47_O0DVBS INGEST_TIME 2014-11-24T22:01:57Z PACKAGE AA00026419_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


5
from an investigation of actual psychotherapy is a complex, if not impos
sible task. Thus, a demonstration of the genuine effects of psychother
apy free of placebo effects seems never to have been made.
in this study an attempt is made to Investigate a form of "psy
chotherapeutic encounter11 in which specific factors could not contribute
to the outcome of treatment. That is, any changes which accrue as a re
sult of the treatment are not due to a psychotherapist's behavior, tech
nique or manipulations during treatment. Change, tnstead, must oe at
tributed to the presence of the placebo factor and must be referred to as
a placebo effect.
A psychotherapeutic encounter" can take place when a patient re
ports to a tape recorder for one "therapy session" believing a psycho
therapist will listen and respond to his talking. Measures of change in
such "affective comfort" (Shoben, 1953) as moods and anxiety taken imme
diately after this treatment will indicate the effect of the treatment.
Borgatta (1959) suggests, satirically, that this technique be used to
demonstrate the relative ineffectiveness of actual psychotherapy. Slack
(I960), on the other hand, demonstrated that therapeutically inaccessible
juvenile delinquents could be introduced to psychotherapy by initially
having them speak into a tape recorder. Eventually these juveniles were
introduced to the therapist who had been giving them feedback from the
tapes, and actual psychotherapy was initiated. Martin, Lundy and Lewin
(I960) evaluated the reinforcing effects of three degrees of therapist
communication on the affectively toned verbalizations of their subjects.
The group intended to have virtually no communication from the therapist
spoke into a tape recorder as If they were speaking to a psychotherapist.


53
TABLE 22
PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EIGHT OF THE DEPENDENT VARIABLES AND
SCORES ON THE BETA INTELLIGENCE TEST. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS
Moods of the MACL
So. Adjustment
Treatment Conditions
AACL
Aqar.
Cone.
Deact.
Aff.
Anx.
Depr.
Eqot.
PIeas.
Activ.
Nonch.
Skept.
Start.
Score
Experimental Group
Positive Set, Congruent Therapist
+ .36
+ .10
-.29
+ .42
-.11
+ .22
+ .37
+ .09
O
Csl
l
i
VJ1
O
I
N>
-P-
+ .31
+. 49
-.24
Pos i ti ve Set,
Incongruent Therapist
-.51
+.05
+ 1 1
+ .20
+. 43
-.44
-.23
-.15
+.39
cr\
Csl
I
+.52
-.18
-.61*
+ .48
Neutral Set,
Congruent Therapist
+.20
+. 34
-.50
+ .02
+ .26
+. 04
+ .09
-.35
+.07
+.18
.00
+ .01
+ .52
-.50
Neutral Set,
Incongruent Therapist
-.13
-.39
+ .19
-. 89*
+.54* -.44
-.1 1
+ .60*
-.04
-.32
-.18
-.10
-.73*
-.21
Control Group
Pos i tive Set,
Congruent Therapist
+. 1+5
-.07
-.27
-.21
+.06 +.35
-.33
+.02
+.31
-.27
-.21
+ .26
+ .37
-.01
Positive Set,
Incongruent Therapist
-.33
+ .02
-.36
+. 13
+.10 -.53*
-.18
-.10
+ .08
-.25
-.10
-.37
-.12
+ .08
Neutral Set,
Congruent Therapist
+ .05
-.32
-.13
-.05
+ .25
+ .31
+ .09
-.18
-.24
-.02
-.17
+ .01
-.22
-.09
Neutral Set,
Incongruent Therapist
+ 11
+.41
+.18
+. 06
+. 09
+ .07
-.05
-.28
-.15
OO
-.26
26
+ .07
+ .16
*Significant ^ .05-


If you were to have a psychotherapist or counselor to talk with
about your problems, what would you expect him to be like? If the state
ment is true concerning how you would expect the counselor or psycho
therapist to be, put a circle around "true." If the statement is false
concerning how you would expect this counselor or psychotherapist to be,
put a circle around "false."
1.
Calm, easygoing
True
False
2.
Is careful not to let people waste his time..
True
False
3.
Looks for the good points In people
True
False
4.
Likes to have a hand in managing other people's
affairs *
True
Fal se
5.
Is concerned with what's right
True
False
6.
Blunt, straightforward, calls a spade a spade
True
False
7.
Hard to get to know
True
Fal se
8.
is likely to give advice and guidance
True
False
9.
Cares what other people think of him
True
False
10.
Is likely to overestimate a person's abilities ..
True
False
11.
Is Indulgent, forgiving
True
Fal se
12.
I* able to sense other people's feelings
True
Fal se
13.
Is careful not to upset others
True
False
14.
Judges the behavior of others. .. ..
True
Fal se
15.
Expects the individual to shoulder his own re-
sponsibl11 ties
True
False
16.
Is logical, sticks to the facts .. .. ..
True
False
17-
Is likely to keep his irritations or resentments
to himself
True
False
18.
Is gentle, tender
True
Fal se
19.
Self-satisfied
True
False
20.
Never makes people feet uncomfortable ..
True
False
21.
Hard to deceive, does not accept things at face value
True
False
22.
Businesslike
True
False
23.
Is conscientious about duties and responsibilities
True
False
24.
Likes to do a good job ..
True
Fal se
25.
Is not emotional
True
Fal se
26.
Reacts to most people in about the same way..
True
False
27.
Sympathetic
True
False
28.
Tries to discover who's to blame for mistakes made
True
False
29.
Is able to change his opinions easily
True
False
30.
Diplomatic
True
Fal se
31.
Is troubled by the misfortunes of others
True
False
32.
Persuasive
True
False
33.
Well adjusted, gets along well in the world..
True
False
34.
Is quick to qive encouragement and reassurance ..
True
False
35.
Has no trouble getting along with people, makes
friends easily
True
Fal se
36.
Critical, not easily impressed
True
False


(Note: in Column 2. I represents Experimental Group, Congruent
Therapist, Positive Set; 2 represents Experimental Group, Incongruent
Therapist, Positive Set; 3 represents Experimental Group, Congruent Ther
apist, Neutral Set; 4 represents Experimental Group, Incongruent Thera
pist, Neutral Set; 5 represents Control Group, Congruent Therapist, Posi
tive Set; 6 represents Control Group, Incongruent Therapist, Positive
Set; 7 represents Control Group, Congruent Therapist, Neutral Set; 8 rep
resents Control Group, Incongruent Therapist, Neutral Set. Columns 7
through 20 represent changes In the dependent variables from pre- to
posttreatment testing. A change score of 0 is represented by 09. Nega
tive change scores are represented by increasing numbers and positive
change scores are represented by decreasing numbers. That is, a change
score of -1 is represented by 10, a change score of -2 is represented by
II, etc. A change score of +1 is represented by 08, a change score of +2
is represented by 07. etc. In Column 21. 01 represents Tittle, or no
speech during treatment, and 02 represents a moderate amount or constant
speech during treatment. In Column 22. 01 represents the greatest stated
desire for psychotherapy, 02 represents a moderate stated desire for psy
chotherapy and 03 represents a stated lack of desire for psychotherapy.
In Column 23. 01 represents a nonpsychotic diagnosis and 02 represents a
psychotic diagnosis.)
97


16
TABLE 3
SUMMARY TABLE OF AN ANALYSIS OF VARIANCE DEMONSTRATING THE
RANDOM DISTRIBUTION OF I.Q.
Source
ss
df
V
F
A (experimental vs. control)
96.00
1
96.00
.62
B (congruent vs. incongruent)
.37
1
.37
.00
C (positive vs. neutral)
2.04
1
2.04
.01
AB
165.38
1
165.38
1.07
AC
3.37
1
3.37
.02
BC
66.67
1
66.67
.43
ABC
140.17
1
140.17
.90
w Cells
H-657-83
SSL
155.20
Total
14,131.83
95
(An F of 3*95 Is significant at the .OS level.)
reason to expect that educational level, length of hospitalization and
diagnosis are randomly distributed across treatment conditions. The psy-
chotic patients all have shown good remission of their symptoms and all
patients serving as subjects in the study were judged by a psychiatrist
to be sufficiently psychologically sound to return to the conmunity and
be employed.
Means of Measurement
Two tests upon which change in feelings of personal comfort can
be monitored (the Anxiety Adjective Check List and the Mood Adjective
Check List), and one test evaluating general level of psychological


TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS II
LIST OF TABLES Iv
INTRODUCTION . I
The Present Study 4
METHOD. 13
Subjects 14
Means of Measurement 16
Procedure 24
RESULTS 28
DISCUSSION 62
APPENDICES 75
A INSTRUCTIONS TO SUBJECTS 76
B SAMPLE ANXIETY ADJECTIVE CHECK LIST 80
C SAMPLE MOOD ADJECTIVE CHECK LIST 82
D SAMPLE ADJUSTMENT SCORE 85
E SAMPLE OF THE MODIFICATION OF THE EXPECTATION
Q-SORT 90
F SAMPLE PERSONAL INFORMATION QUESTIONNAIRE 92
G SAMPLE LETTER SENT TO ALL SUBJECTS WHO SPOKE
INTO THE TAPE RECORDER 94
H SUBJECT RAW DATA 96
BIBLIOGRAPHY 103
Hi


102
Subj
No
T reat
Group
i .a.
1 ntro
Neuro
Aqe
AACL
Adj
Score
Aggr
Cone
Deact
So
Aff
Anx
Depr
Eqot
Pleas
Acti v
Nonch
Skept
Amt of
Start Speech
Des
Psycho Diaq
91
8
089
23
02
44
10
10
07
07
09
07
09
09
09
09
07
09
09
09
92
8
124
28
44
41
10
04
07
09
06
09
07
08
09
10
10
09
11
09
93
8
115
18
19
53
08
07
09
09
09
09
09
09
09
11
07
07
09
09
94
8
095
23
22
46
11
10
09
06
1 1
11
11
1 1
13
13
07
09
10
05
95
8
101
23
28
31
10
09
07
11
08
09
08
12
10
10
09
09
10
09
96
8
089
24
24
32
09
15
09
09
09
15
06
09
09
12
09
09
03
09


LIST OF TABLESContinued
Table Page
27 MULTIPLE REGRESSION COEFFICIENTS BETWEEN THE
VARIABLES LISTED ON THE VERTICAL AXIS AND FIVE
DEPENDENT VARIABLE CHANGE SCORES FOR THE EX
PERIMENTAL GROUP. THE DATA ARE PRESENTED
SEPARATELY FOR EACH OF THE INDEPENDENT VARIA
BLES IN ADDITION TO THE TOTAL EXPERIMENTAL
GROUP 61
vili


26
verbalizations and to observe that the Control Group subject remained in
his room. Record was made of the amount of the subject's speech, and in
several cases the content of the subject's verbalizations. With a stop
watch, amount of speech was rated as follows: constant speaker, no longer
than 3~m¡nute pauses; moderate speaker, no longer than 5~minute pauses;
little speaker, at least some verbalization; no speaker, says nothing.
At the end of 45 minutes the Experimenter turned off the tape re
corder and guided the Experimental Group subject back into the testing
room. Both subjects were then asked to "retake several of the shorter
tests." The following were readministered:
7. Anxiety Adjective Check List
8. Mood Adjective Check List
9. The Adjustment Score
At this point, the collection of the data was completed. The Ex
perimental Group subject was dismissed. If the subject asked about feed
back, he was told that he would receive some type of feedback within sev
eral weeks. He was also encouraged not to tell anyone the nature of what
had happened because the Experimenter would be unable to see everyone for
this project. The Control Group subject was then asked If he wished to
"speak to the psychiatrist." If he agreed, he was given 45 minutes with
the tape recorder, however, If he declined, he was dismissed. Twenty-one
of the 48 Control Group subjects chose to speak.
If any subject discussed suicidal or homicidal content, mentioned
intensely pressing issues or made special requests, the Experimenter
promptly called this Information to the attention of the ward psychiatrist.
This occurred in twelve cases.


78
relating to problems you've had, or problems you are having, or problems
you can see ahead of you In the future* This psychiatrist has such an
excellent reputation that he is an extremely busy man, but he has offered
to help you men to solve any problems you may have anyway. Because he is
so busy he can't get here to the hospital so he suggested we do something
a little differently because he is so interested in helping you. You
will siftply talk Into a tape recorder, and the tape will be sent regis
tered mail to him. I think this Is really great because he has helped so
many, many people and I'm sure he can help you with any problems you may
possibly have.
Neutral Set instructions
i can tell you now what this is all about. It is an opportunity
for you men to talk with a psychiatrist. Now, you can talk to him about
anything you'd like relating to problems you've had, or problems you are
having, or problems you can see ahead of you In the future. This psy
chiatrist is an extremely busy man, and he Is unable to get here to the
hospital so he suggested we do something a little differently. You will
simply talk into a tape recorder, and the tape will be sent registered
mai1 to him.
INSTRUCTIONS GIVEN TO EACH PAIR OF SUBJECTS TO DETERMINE
CONGRUENT THERAPIST OR INCONGRUENT THERAPIST
Congruent Therapist Instructions
I've been looking over your questionnaire on the way you would
expect this psychotherapist or counselor to be, and I see that this psy
chiatrist Is the way you thought he would be. For example, Mr.
(subject seated on the Experimenter's right), he (read this subject's re
sponses to I terns 7. 14 and 29; see Appendix E). And Mr. (sub
ject seated on the Experimenter's left), he (read this subject's responses
to items 1, 31 and 32; see Appendix E).
Inconqruent Therapist instructions
I've been looking over your questionnaire on the way you would
expect this psychotherapist or counselor to be, and I see that this psy
chiatrist is not exactly the way you thought he would be. For example,
Mr. (subject seated on the Experimenter's right), he (read a denial
of the accuracy of this subject's responses to items 7, 14 and 29; see
Appendix E). And Mr. (subject seated on the Experimenter's left),
he (read a denial of the accuracy of this subject's responses to items I,
31 and 32; see Appendix E).


ACKNOWLEDGMENTS
This study would not have been completed had It not been for the
foresight, wisdom and encouragement of many people to whom the writer is
indebted and grateful.
A special measure of gratitude is due Dr. Audrey S. Schumacher,
chairman of the supervisory committee, for her genuine interest, her un
tiring patience and her constant support and guidance. The assistance
and encouragement of the remainder of the supervisory committee, Dr. R.
J. Anderson, Dr. 8. Barger, Dr. J. J. Wright and Or. V. A. Hines, is
warmly accepted and greatly appreciated.
The writer wishes to acknowledge the staff and office workers of
the Gulfport, Mississippi Veterans' Administration Hospital who gave un
sparingly of their time, their cooperation and their resources. A par
ticular note of gratitude is extended to three people who specifically
asked to remain unmentioned, but without whom the progress of this study
would certainly have been impeded.
The writer wishes to thank his parents for their faith, for their
Investment and for wanting someday to say: "there's my son, the Doctor."
Most of all, the writer thanks his wife, Alee, for her under
standing, her support, her guidance and her remarkable patience which
sustained the writer in his efforts to complete this study.
ii


104
Frank, J. D. Son affects of expectancy and influence in psychotherapy,
in Hasserraan, J. H. and Horeno, J. I. (eds.) Progress in psycho
therapy. New York, New York: Grue and Stratton, 1958.
Frank, J. D. Persuasion and healing. Baltimore, Md.: Johns Hopkins
Press, 1961.
Glasser, E. H. and Whittlow, G. C. Evidence for a nonspecific mechanism
of habituation. J. physiol.. 1953. 122. 43-44.
Glasser, E. H. and Whittlow, G. C. Experimental errors in clinical
trials. Clin, scl.. 1954, H, Number 2, 199-210.
Glideman, L. H., Gantt, W. H. and Teitelbaum, H. A. Some implications of
conditional reflex studies for placebo research. Am. J. psychiat..
1957. JLL. 1103-1107.
i : 1 )
Gourlay, N. Covariance analysis and its applications in psychological
research. Brit- J, 1953. &. 25-33.
Hagans. J. A., Doering, C. R., Ashley, F. W. and Wolf, S. The thera
peutic experiment. J. lab, clin, med.. 1957. itii, 282-285.
Hofling, C. K. The place of placebos in medical practice. Gen, pract..
1955, iL 103-118.
Hongingfield, G. The placebo effecta review of nonspecific factors in
treatment. Cooperative studies in psychiatry. Nov., 1963.
Jaco, E. G. Patients, physicians and illnesses. New York, New York:
Free Press, 1958.
Joyce, C. R. Consistent differences in Individual reactions to drugs and
dummies. Brit. J. of Pharmacol.. 1959. 14 512-521.
Kellogg, C. £., Horton, N. W., Lindner, R. H. and Gurvitz, H. Revi sed
beta examination: manual. New York, New York: Psychological
Corporation, 1946.
Kornetsky, C. and Humphries, 0. Relationship between effects of a number
of centrally acting drugs and personality. A.H.A. arch, neural.
psychiat.. 1957, 22. 325-327.
Kurland, A. A. The drug placeboIts psychodynamic and conditional re
flex action. Behav. sci.. 1957. 2, 101-110.
Kurland, A. A. The placebo, in Masserman, J. H. and Horeno, J. i. (eds.)
Progress in psychotherapy. New York, New York: Grue and
Stratton, 1958.


56
TABLE 24
PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND
SUBJECTS1 AGES. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS
Moods
of the
MACL
Treatment Conditions AACL
Aggn.
Cone.
Deact.
So.
Aff.
Anx.
Depr.
Eqot.
Pleas.
Acti v.
Nonch.
Skept.
Start.
Adj ustment
Score
Experimental Group
Positive Set, Congruent Therapist +.07
+ 07
+ 19
+ .06
+.49
+. 1 1
-.19
-.36
-.35
-.41
-.56*
+ .14
+ 14
-.51
Positive Set, Incongruent Therapist +.36
+.42
+ .32
+. 06
+. 50
+.45
+ .51
-.24
+. 14
-.06
-.32
-.07
-.03
-.05
Neutral Set, Congruent Therapist +.51
+.01
-.34
+ .13
-.22
+.32
-.20
+.30
+. 03
-.05
+ 06
+.52
+ .50
+ .50
Neutral Set, Incongruent Therapist +.03
+.23
-.10
.00
-.25
+. 06
-.25
+.39
+. 41
-.02
+.48
+.49
-. 14
-.48
Control Group
Positive Set, Congruent Therapist +.37
-.22
+. 16
-.33
-.45
+ .18
-.07
-.28
-.20
-.01
-.36
-.66*
-.08
-.59*
Positive Set, Incongruent Therapist +.25
+.15
+ .20
-.43
-.21
+ 26
-.19
-.18
+. 21
+.24
-.22
-.21
-.29
-.15
Neutral Set, Congruent Therapist -.60*
+.02
+. 36
+. 12
+ .31
+.23
+.02
+. 28
+.32
+.30
-.01
+ .05
-.24
+ .41
Neutral Set, Incongruent Therapist -.12
-.29
+ .22
-.31
+. 18
-.25
+.19
+. 06
-.09
+ .06
+ .42
-.38
+ .14
+ .03
'^Significant .05.


98
Subj Treat Adj So Amt of Des
No
Group
i .a.
1 ntro
Neuro
Aqe
AACL
Score
Agg.r
Cone
Deact
Aff
Anx
Dep r
Eqot
Pleas
ActI v
Noneh
Skept
Start
Speech
Psycho
Di aq
1
1
107
28
04
33
15
09
09
09
09
07
12
09
08
01
10
07
08
09
02
03
02
2
1
107
24
02
27
10
07
12
1 1
09
09
09
09
09
08
09
09
09
09
02
03
02
3
1
076
28
30
30
13
06
08
08
13
07
07
10
09
09
08
05
09
14
01
01
02
4
1
099
13
46
43
09
14
11
08
12
08
07
13
11
08
13
12
13
1 1
01
01
02
5
1
113
28
34
48
12
07
09
12
09
09
09
09
10
09
09
06
09
09
t
01
02
01
6
1
088
20
48
41
16
10
11
06
1 1
04
10
12
15
06
10
1 1
11
09
01
01
01
7
1
099
36
32
26
13
09
10
08
10
06
13
1 1
08
07
07
08
10
09
1
01
02
02
8
1
121
29
36
38
1 1
09
10
09
09
09
08
08
09
09
1 1
10
07
09
!
01
01
01
9
1
093
12
40
48
20
1 1
10
09
10
02
17
09
08
03
08
10
10
10
01
01
01
10
1
102
29
25
51
12
09
12
09
12
07
10
11
11
10
08
12
07
08
01
01
02
11
1
088
28
30
37
1 1
07
12
05
06
08
10
14
08
07
08
10
06
09
02
03
02
12
1
078
34
08
18
15
06
12
12
12
12
14
09
09
02
06
06
11
09
02
01
02
13
2
089
24
12
37
04
05
05
06
10
05
07
08
06
09
06
09
06
06
02
02
02
14
2
099
14
40
43
12
13
1 1
09
09
07
10
11
09
03
10
11
14
09
01
01
01
15
2
088
36
26
29
06
18
09
08
12
13
08
09
06
09
09
11
09
09
02
01
01
16
2
119
18
44
40
12
08
12
09
1 1
07
12
09
09
08
09
09
09
09
02
02
01
17
2
102
24
22
22
12
06
09
12
05
11
16
16
10
10
12
07
14
09
01
02
02
18
2
099
14
06
32
09
09
09
08
07
1 1
07
12
11
12
09
09
07
09
01
02
02
19
2
106
10
44
45
05
13
06
09
06
09
09
09
12
09
15
12
09
09
02
03
01
20
2
109
22
30
43
05
04
05
07
07
06
08
08
06
06
07
07
08
08
02
02
02
21
2
103
12
38
43
06
09
08
12
09
12
05
10
11
11
09
09
11
09
02
01
02


8
However, despite the logical consistency of Eysenck's specula
tions, the literature concerning the placebo responsiveness of neurotics
and introverts is equivocal. For example, Fischer and Oiin (1956), using
psychiatric patients, demonstrated that neurotics were the most respon
sive subjects in a drug placebo situation. On the other hand, Kurland
(1958) disputes this conclusion on the basts of his finding that no dif
ference exists between the placebo reactivity of psychotics and nonpsy-
chotics. Joyce (1959) in a drug study with medical students, indicates
that extroverts are the most responsive subjects, while Laverty (1958)
demonstrates that introverted subjects are the most responsive in drug
treatment situations.
In this study, age, intelligence, introversion and neurotic!sm
are the patient characteristics whose contribution to the placebo effect
will be evaluated.
(2) The doctor-
The nature of the treater is reportedly an important factor for
encouraging or discouraging placebo effects (Shapiro, 1964; Uhlenruth
et al.. 1959). For example, Glldeman et al. (1957) demonstrated that a
male physician gained a 70 per cent cure rate in a treatment for which a
female nurse could achieve only a 25 per cent cure rate. He also has
shown that if the administering physician Is perceived as an expert
healer, the response of the patient to treatment will be greater than if
the physician Is believed to be a medical quack. These effects were ob
tained, conceivably, because the patient maintains certain cognitive ex
pectancies about the nature of treaters, and when these expectancies are


INTRODUCTION
The extensive literature dealing with processes and systems of
psychotherapy generally makes the explicit assumption that behavioral
changes following psychotherapeutic encounter are a function of the in*
tended efforts of the psychotherapist. That is, behavioral modification
results from the direct and specific efforts, manipulations or techniques
of the psychotherapist. Within the framework of Roger's theory (1961),
for example, the therapist approves and accepts the client and as a re*
suit the client comes to accept himself. The patient in psychoanalysis
achieves personal satisfaction and comfort as the psychoanalyst Inter*
prets the patient's repressed psychosexual conflicts and the patient es*
tablishes a mature sexual adjustment (Fenichel, 1945) Shoben (1953) de
scribes the goal of psychotherapy as the alleviation of symptoms in addi
tion to the increase in a patient's affective comfort. Thus, In all psy
chotherapy, "specific factors" (therapist bohsviors) aJl^geily produce
intended effects (personal comfort of the patient).
Eysenck (1952, 1961) states, and Rosenthal and Frank suggest
(1956) that in addition to the intended efforts of the psychotherapist
(specific factors), there are other, nonspecific factors in psychotherapy
which significantly contribute to the outcome of treatment. "Nonspecific
factors" are loosely defined as those "placebo effects," situational
events and conditions in addition to the intended efforts of the therapist
which reportedly function in all treatment situations (Shapiro, 1964).
They serve to cloud interpretation of "real" effects of psychological
I


Name:
Ward:
Age:
Number of years of school:
Type of work before hospital:
Harital status: Number of children:
Are you service connected: %
Number of times in a mental hospital:
How long have you been In the hospital this time?
If you had a chance to get psychotherapy or counseling would you want it:
very much?
a little?
not at all?
93


58
TABLE 25
PEARSON PRODUCT-MOMENT CORRELATIONS BETWEEN SUBJECTS' CHANGE SCORES FROM PRE- TO POSTTREATMENT TESTING ON EACH OF THE DEPENDENT VARIABLES AND
SCORES ON THE MPI NEUROTICISM SCALE. THE DATA ARE PRESENTED FOR EACH OF THE EIGHT TREATMENT CONDITIONS
Moods
of the
MACL
Treatment Conditions
AACL
Jtefl.Lv
Cone.
Deact.
So.
Aff.
Anx.
Depr.
Eqot.
Pleas.
Acti v.
Nonch.
Skept.
Start.
Adj ustment
Score
Experimental Group
Positive Set, Congruent Therapist
-.08
+. 19
+. 51
-.12
+.52
+.17
-.43
-.46
-.36
-.39
-.43
-.29
-.24
-.56*
Positive Set, Incongruent Therapist
+.18
-.23
-.37
-.03
-. 06
+.07
+. 41
-.22
+. 01
-.38
-.59*
-.38
+ .03
-.43
Neutral Set, Congruent Therapist
-.24
-.41
+. 14
+. 13
+. 04
+. 43
+.09
+.42
+ .70*
+. 34
+ .26
-.23
+ .20
-.06
Neutral Set, Incongruent Therapist
-.13
+.02
+.29
+ .12
-.03
+.56*
+.21
+. 12
+ .45
+.42
-.07
+ .48
+ 12
-.51
Control Group
Positive Set, Congruent Therapist
-.01
+.03
+ .52
-.17
Hr.
-.14
+. 08
+.03
-.16
+ 19
-.18
-.17
-. 69*
-.10
Positive Set, Incongruent Therapist
-.29
-.28
-.37
-.15
+. 16
-.27
-. 60*
-.30
-.12
-.41
-.28
-.02
-.30
+ .42
Neutral Set, Congruent Therapist
-.38
+ 26
+.37
+.10
-.28
+.39
1
LO
+.35
-.36
+. 13
-.14
-.30
-.1 I
+ .43
Neutral Set, Incongruent Therapist
+.17
+. 20
-.34
+. 40
-.19
+.47
+. 15
-.16
+.23
-.22
-.35
+.20
+. 56*
+. 04
''Significant ^.05.


Name
Below you will find words which describe different kinds of feel
ings. Check the words which describe how you feel RIGHT NOW. Some of
the words may sound alike but we want you to check al 1 of the words that
describe your feelings.
l._
AFRAI0
21.
GAY
41.
PANICKY
2.
AGITATED
22.
GLOOMY
42.
PEACEFUL
3-
ANGRY
23.
GRIM
1

PLEASANT
4.__
BITTER
24.
HAPPY
44.
RATTLED
5-_
CALM
25-
HELPLESS
45._
SAD
6.
CHARMING
26.
HOPELESS
46.
SECURE
7.
CHEERFUL
27-
1NSECURE
47-
SENTIMENTAL
8.
COMPLAINING
28.
JEALOUS
48.
SERIOUS
9-
CONTENTED
29-
JOYFUL
49. _
SHAKEY
IQ.__
CONTRARY
SjJ
o

1
KlNDLY
50.
SOLEMN
l-
COOL
31.
LIGHT-HEARTED
51.
STEADY
I2._
CR0SS
32.
LONELY
52.
TENDER
13.
..DESPERATE
33.
LOVING
53-
TENSE
!4.__
EASY-GOING
34.
MAD
54.
TERRIFI£D
*5-
FEARFUL
35.
MEAN
55.
THREATENED
16.
..FEARLESS
36.
MERRY
56.
THOUGHTFUL
17.
^FRETFUL
37.
MlSERABLE
57.
UNCONCERNED
18.
FRIENOLY
38.
NERVOUS
58.
UNEASY
I9.__
_FRIGHTENED
39.
OVERCONCERNEO
59.
UPSET
20.
^FURIOUS
40.
OVERWHELMED
60.
WARM
61.
WORRYING
81


49
concentration (see Table 7A). The HACL mood of Skepticism also displayed
no statistically significant differences between the independent varia
bles. However, the Incongruent Therapist, Control Group condition re
ported a slight increase in skepticism after treatment while all others
reported a decrease in skepticism after treatment (see Table 16A).
Hypothesis IV states that no relationship exists between differ
ences in certain subject characteristics and changes in the dependent
variables. This was evaluated by multiple regression technique. The
subject differences under consideration were age, level of intelligence
(Beta Intelligence Test, 1946), level of neuroticism (HPI, 1959) and
level of introversion (HPI, 1959). Analysis of variance (Lindquist,
1953) was used to demonstrate that each of these four subject variables
was randomly distributed across the eight treatment conditions. (The
summary tables for these analyses on age and level of intelligence are
presented in the Subjects section of the Method chapter. Below are the
summary tables demonstrating the random distribution of introversion
(Table 20) and neuroticism (Table 21).
Pearson product-moment correlations were computed between each of
these four variables and the change in scores fromp re- to posttreatment
testing on each of the 14 dependent variables. This was done for each of
the eight treatment conditions. This resulted in 448 correlations which
are presented below in Tables 22, 23, 24, and 25. At the .05 level of
significance, by chance alone wo would expect 24 significant correlations
out of the total 448. In the present data, by frequency count, there are
26 correlations which are significant at the .05 level. This suggests
that, over-all, the null hypothesis appears supported. It is difficult