EFFECTS OF VARIATION IN AWARENESS AND
REINFORCEMENT IN THE CONDITIONING
OF VERBAL BEHAVIOR
EDWARD GEORGE SCHLEIMER
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
UNIVERSITY OF FLORIDA
3 1262 08552 2299IIII i I
3 1262 08552 2299
The author is indebted to Dr. Richard J. Aderson, Chailran of
the supervisory committee for his great help, timel and tactful
intervention, and eneouragemt in the preparation of this paper.
Special thanks are due also to Dr. Louis D. Cohen, Clinical Director
of the Paychology traniing progrc for his astute criticism and
comentary. The assistance of Dr. Marvin Shaw, Dr. Be BnBarger and
Dr. George R. Bartlett, the remaining members of the graduate com-
mittee, is warmly acknowledged.
The patience, understanding and endurance shown through the
long months of work on this project by the author's wife, Barbara,
is deeply appreciated.
The help of Bill Nettleton of the Tulane Bio-edical Computing
Center at New Orleans, in analysis of parts of the data, was grate-
fully and thanklly received.
TABLE OF CONTENTS
LIST OF TABLES
LIST OF FTB TES
Analyses of Variance
Acquisition I-Verbal Xntelligence
Suama7r of Results
A SHIPLET-mNATFORD VOCABUIARI WORD TEST
B ORDER OF TREATMENTS DURIM C CONDITIONING PERIODS
C ORDER OF VERB APPEARANCE DURD? CONDITIONING
D ANALYSIS OF VARIANCE StUMARI TABLES
LIST OF TABLES
1 Coaparison () Between Tretatmentroup
Levels to Determine Matching of Levels on Age
and Verbal Intelligence. 18
2 Sample Format for Analysis of Variane Sumary
3 Smarwy of F-Ratios Obtained in Analyses of
Variance for Acquisition I. 21
S Sutmary of F-Ratios Obtained for the Analyses
of Variance"in Acquisition II. 28
$ Sumary of F-Ratios Obtained in the Analyses
of Verbal Intelligence as a Subjects Variable. 36
6 Sumary of m-Ratios Obtained in the Analysis
of Age as a Subjects Variable. 4O
7 Stmmary of Interactions Which Fail to Produce
Statistically Significant Results or Meaningful
8 Su~nary of the Effects of the Treatment
Variables on the Dependent Measures of Response
LIST OF FIGURES
1 Effects of Response Class (IWe or le,They) and
Type Reinforcement by Successive Blocks of Data
During Acquisition I. 22
2 Effects of Awareness and Type Reinforcement by
Successive Blocks of Data During Acquisition I. 23
3 Effect of the Triple Interaction of Awareness,
Response Class, and Type Reinforcement During
Acquisition *I 23
4 Effects of Awareness and Type Reinforcement on
Response Duration by Successive Blocks of Data
During Acquisition I. 24
5 Effect of Awareness on the Variability of Units '
During Successive Acquisition I Blocks. 24
6 Effects of Response Class and Type of Reinforce-
ment on Rate of Conditioning During Acquisition II. 29
7 Effect of the Interaction of Response Class and
Type Reinforcement on Rate of Conditioning During
Acquisition II 29
8 Effects of Awareness and Type of Reinforcement on
the Duration of the Response During Acquisition II. 30
9 Effect of Interaction Between Awareness and Type
Reinforcement on Response Units During Acquisi-
tion II. 30
10 Effect of Interaction Between Awareness and
Response Class on the Variability of Response
Latency During Acquisition II. 30
LIST OF FmURES (Continued)
Effect of Interaction Between Awareness and VI
on Rate of Conditioning During Acquisition I.
Effect of Interaction Between VI and Type
Reinforcement on Conditioning During Acquisition
Effect of Interaction Between Awareness and VI
on Response Duration During Acquisition I.,
Effect of Interaction Between Awareness and VI
on Response Unite During Acquisition I.
Effect of the Triple Interaction
Age and Type of Reinforcement on
tioning During Acquisition I.
Effect of the Triple Interaction
Age and Type of Reinforcement on
During Acquisition I.
Rate of Condi-
Effect of Interaction Between Age and Type
Reinforcement on the Variability of Response Units
During Acquisition I.
It has been made clear (Dulany, 1962; Eriksen, 1962; Spielberger,
1962) that "awareness" is a significant and crucial variable in any
experiment concerned with verbal operant conditioning. Within the
typical verbal conditioning framework, awareness has been defined as
a subject's ability to verbalize the relationship between a contingent
stimulus and the response which it follows* Conceptual definitions of
"awareness," however, are not made specific in the literature. The
meaning of awareness, apart from the operational context in which it
usually appears, is closely allied to the meaning of terms such as
"cognizance" or "insight." As such, it is purely a hypothetical con-
struct. All of these terms refer to a state of the organism wherein
information is available to a degree which will permit more or less
definitive purposeful action to occur. The strength of the action
potential generated by this information corresponds to what is meant
by a subject's degree of awareness. The extent of awareness or insight
can only be measured in terms of the change in behavioral activity
which may result as the awareness or insight is acquired.
Considerable controversy has been stimulated by the issue of
whether or not awareness is necessary in order for conditioning to
occur. The series of experiments reported by Spielberger (1962) has
demonstrated that the behavior change which occurs in the verbal condi.
tioning paradigm occurs only in the presence of some degree of aware-
ness of the reinforcement contingency. Within the larger conceptual
framework envisioned by Dulany (1962), a subject in these experiments
must have some notion of what is expected of him, first forming par-
tially or wholly correct hypotheses regarding the reinforcement he
receives, then forming hypotheses regarding the behavior which may be
expected of him, and finally deciding whether or not to conform to
these expectations behaviorally. If, for one reason or another, the
subject's hypotheses are incorrect, or if he understands the test
requirements perfectly but does not desire to "behave," there will be
no evidence of conditioning.
The greater proportion of earlier verbal conditioning studies
reported conditioning to have occurred without awareness (Krasner, 1958).
Articles often cited were those of Greenspoon (1955) and Taffel (1955),
in which the instrumental conditioning of plural nouns and first person
pronouns was said to have occurred without the subjects' knowledge of
the reinforcement contingency. Their awareness, defined by the subjects'
introspective reports of the associations made during conditioning which
led them to respond in a particular fashion, was most often assessed by
a brief Post-Conditioning Interview (PCI). This interpretation,
learning without awareness, was acceptable to these authors without
further analysis presumably because it was consistent with the belief
of earlier learning theorists. For example, Thoradike's "Law of Effect"
stated that "...a satisfying after-effect could strengthen the connec-
tion which it followed and to which it belonged in cases where the
learner did not know what the connection was..." (Thorndike and Rock, 1934).
More recently, however, Kimble (1961) suggests that only voluntary or
emitted responses may be conditioned instrumentally, whereas involun.
tary responses (elicited responses and reflexes) may be conditioned
classically. This latter frame of reference seems more suitable in
view of the results obtained by Dulany (1962) and Spielberger (1962).
According to Levin (1961), those experimenters who reported instru.
mental conditioning of verbal behavior without awareness erred in their
technique of assessment for awareness. Referring specifically to the
brief PCI, he reported that more extensive questioning would reveal
the existence of correlated behavioral hypotheses, i.e., some degree
of awareness, which would account for the learning shown in these
experiments. Several other studies, which examine more closely the
concept of awareness, may be cited as supportive evidence for this
interpretation of the relationship of awareness to learning
(Spielberger, Berger, and Howard, 1963; Ekman, Krasner, and Ullman,
1963; Simkins, 1963; Cohen, Greenbaum, and Mausson, 1963).
As Simkins (1963) points out, however, any attempts to explain
learning solely in terms of differences in awareness as it has been
defined, are futile. In his study, differential learning resulted not
from the variation of awareness alone, but from the co-variation of
reinforcement, awareness, and the particular sort of response which was
required. Spielberger et a. (1963) and Cohen et al (1963) both discuss
awareness with reference to its part in forming the mediating hypotheses
necessary for conditioning to occur.
Kimble (1962) has pointed out that the typical verbal conditioning
procedure is one in which the subject's awareness of the requirements
of the task is undergoing modification. Thus, the verbal conditioning
paradigm is an appropriate method to study means of controlling this
awareness and assessing the new behavior potentials which result.
The approach closely follows a current definition of learning (Kimble,
1961, pp. 2-8) as "...a relatively permanent change in behavior
potentiality which occurs as a result of reinforced practice...."
Awareness is one of the major constructs under consideration in
this research, and an attempt is made to clarify further its functions
in verbal learning. This experiment is conceived of as one of several
needed to find additional dimensions and clarify interrelations which
will increase the usefulness of awareness as a construct or mediational
process for learning theory, and eventually increase its usefulness
also as a practical tool for behavior modification.
The second major variable considered in this experiment is the
verbal behavior of the experimenter which serves as the reinforcement
for the verbal behavior of the subjects. The outcomes of interest
here deal with the interactions between awareness and varied reinfor-
cers in the production of behavior change. The efficacy of awareness
as a "conditioner," for example, may vary with systematic variation in
the verbal behavior of the experimenter. Likewise, the effect of a
particular experimental reinforcer may change as a subject's degree of
awareness varies. Besides providing answers to questions such as these,
this approach can be expected to raise new questions and uncover rela-
tionships which may be followed-up through subsequent experiments. The
approach through the verbal operant paradigm has an extremely good poten.
tial as an aid in the development of a science of behavior modification
(Greenspoon, 1962; Krasner, 1961; Luborsky and Strupp, 1961).
Several experiments have been done which seek to analyze the func.
tion of awareness as a mediating process in learning. Cohen et al.(1963)
state that "...awareness may be a mechanism whereby the subject can
increase or inhibit his responses depending upon his motivation to
respond..." (p. 419).
In the study by Ekman et al.(1963), awareness was found to distinguish
two groups in rate of learning when "set" was varied independently.
Aware subjects who could respond in a socially desirable manner showed
more learning than aware subjects who could respond only in a less
desirable manner. The same differentiation was not achieved when the
subjects were not-aware*
Other studies in the general area of "Influencing" have undertaken
to study the reinforcers used in producing learning, and have not been
concerned with awareness directly. For example, Ring and Kelley (1963)
contrast Augmentation and Reduction as modes of influence. The compari-
son is basically one between praise and reproof as means of producing
opinion change. Little or no difference was found between these rein-
forcers when the standard to which the subjects were to conform was a
plausible one. When the standard was implausible, Reduction (reproof)
was most effective. The awareness criterion as used in verbal condi*
tioning studies may conceivably serve to make the conditioning task a
more plausible one, and the same relationship may be expected to exist.
That is, variation in reinforcement may be expected to have less effect
on conditioning when subjects are aware of the requirements of the task.
An earlier study of Simkins (1961) showed that variation in the experi-
menter's attitude, as reflected in praising or criticizing the work of
subjects, was significantly related to conditioning. Critical treat-
ment resulted in greater conditioning of hostile verbs. Awareness was
not assessed in this study. Gewirts and Baer (1958) showed that prais-
ing subjects prior to the conditioning task resulted in less learning
than complete deprivation of social reinforcement as a prior-to-condi-
tioning treatment. On the other hand, in a subsidiary finding of an
experiment by Merbaum (1963), subjects who "liked" their treatment
conditioned to a greater extent. Also Ashby, Ford, Guerney, and
Querney (1957) found that a reflective type of treatment produced
significantly less guardedness and covert resistance than did a more
direct and interpretative leading type of treatment approach. The
reflective treatment may have had fewer connotations of criticism to
the subjects who participated, and their behavior reflected this
The question raised in these and other experiments (see reviews
by Krasner, 19611 and Greenspoon, 1962) deals directly or indirectly
with the mediating properties of varied reinforcers. There is nothing
in the literature, however, which undertakes to vary both reinforce.
ment and awareness as controlled interacting independent variables in
an Analysis of Variance design, and observe the effects upon the con.
ditioned response which results from this interaction* In the present
study, the effects of two contrasting types of verbal reinforcement are
examined and, in addition, awareness is manipulated independently at
Awareness* Awareness is examined in relation to its effect on
the rate of conditioning, and is examined also regarding its effect
on the quality of the responses produced* Since there are two levels
of awareness in the experiment, the hypothsees generated are these
First, there will be a difference in the rate of conditioning produced
by variation in the degree of awareness, with greater awareness pro.
during the higher rate of conditioning and second, there will be a
difference in the quality of the response produced by variation in the
degree of awareness* The higher level of awareness of the contingency
and the response class, in other words, may be expected to result in
a higher rate and different quality of learning than the lower level.
Directional hypotheses regarding the quality of the response are not
offered because of the lack of previous research evidence in this area.
Reinforcement. Reinforcement is also examined with regard to its
effect on both the rate of conditioning and the quality of the condi-
tioned response. Hence, the hypotheses generated are these: First,
there will be a difference in rate of conditioning produced by varia-
tion in reinforcement frae positive to negative and second, there will
be a difference in the quality of the conditioned response as reinforce.
ment varies froe positive to negative* For the most part, previous
research suggests that negative reinforcement will be the more effective
treatment when rate of conditioning is the dependent variable. Direc-
tional hypotheses regarding response quality are not offered.
Awareness and Reinforcement. The interaction effects of these
two variables are also examined with reference to the rate and quality
of the conditioned response. It is expected that awareness will have
different effects upon conditioning depending upon the type of rein-
forcement used to elicit the response. This is purely a research
hypothesis since no previous research has investigated this relation-
ship directly. With reference to the dependent variable of rate of
conditioning, however, previous research would suggest that the combina-
tion of increased awareness and negative reinforcement would produce
the highest rate of conditioning*
Age. Rate and quality of conditioning are examined as age varies,
the experimental sample of subjects being divided for this purpose at
the median for age* Differential effectiveness of awareness and rein-
forcement may be found when the experimental samples of subjects differ
in age. This is a non-directional research hypothesis.
Verbal Intelligence. The rate and the quality of the conditioned
response are analysed with respect to variation in the verbal (vocabulary)
intelligence of the subjects. Here also, the conditioned response may
be expected to vary in rate and in quality when the experimental samples
differ on the criterion of intelligence. No directional hypotheses are
made here because of the interactions which may occur. However, as a
general effect, an increase in intelligence would presumably be asso-
ciated with a higher rate of conditioning and a "higher quality" response.
Ninety-six (96) male nursing assistants at the Gulfport lNeuro
psychiatric Division of the Bilexi, Mississippi, Veterans Administra.
tion Center comprised the experimental subjects (%s). The Ss have an
average age of 38.3 years with a range frea 20 to 59 years, and a
converted verbal IQ average of 103 with a range from 69 to 137. Both
the age and the verbal intelligence distributions are moderately flat.
tened, with that for age appearing bimodal, peaked at age 30 and again
at age 40-45.
A modified Taffel-type sentence construction task is utilised.
A basic assumption in the design, of course, is that evidence of
learning will not be produced when the Ss are not aware of the reina
forcement contingency or the response class (Dulany, 1962)o The S
must make hypotheses which are to some degree correct regarding both
of these in order for learning to be demonstrated. Reinforcement
contingency refers to the verbal behavior of the experimenter (B)
which follows immediately after a subject's response, and is inter-
pretable by S as dependent upon some characteristic of that response.
Response Class refers to those specific utterances which have been
selected by 1 as the behavior to be reinforced Awareness is defined
in terms of a S's ability to verbalize the relationship between the
contingent stimulus (E1s verbal reinforcement) and the response which
it follows (Q's utterance). The ability of a S to verbalize such a
relationship is manipulable as an independent variable by preocondit
tioning instruction (Letchworth and Wishner, 19631 Saikins, 1963).
Awareness is manipulated in this manner herein Further, the ability
has been shown to have a significant relation to learning (the frequency
of use of the correct response class). There are degrees or levels of
this ability, and the levels correspond closely to the degrees of condi.
tioning which result. The levels have been assessed fram the verbal
reports of Sa, and the reports have typically taken the form of corre.
lated hypotheses or partially correct subjective inferences regarding
the nature of the reinforcement contingency and the response class.
As already mentioned, the development of this ability may be hastened
by independent manipulation* A S may be given both reinforcement
hypotheses and behavioral hypotheses through pre-conditioning instruc-
tions. These hypotheses will make the task of recognition of the
reinforcement contingency and the response class a simpler one.
Verbal reinforcement was given by on a 100 per cent fixed
schedule. The reinforcement was "positive" for half the Sa, and
"negative" for the remaining half. The positive reinforcing statements
made by 1 were 1. "Unbhn, that's good"; 2. "Real good one!; 3. "That's
fine"; 4. "That's a geod neW; 5. "That's a good one too." with the
appropriate inflection, read these reinforcers from a list, giving one
statement for each correct response emitted by S. The five statements
chosen as negatively reinforcing were these 1. "That's not so good";
2. "You can do better"; 3. "Kind of poor"; 4. "Not very goode';
5. "Kind of bad." These reinforcers were used consecutively, with an
inflection appropriate to the content, whenever an incorrect response
was emitted by S. The positive reinforcers were chosen to convey i's
satisfaction with Sa' performance, and the negative reinforcers, E's
dissatisfaction. The verbal reports of pilot and experimental Ss con-
firmed that the positive reinforcers were experienced as praise and
the negative reinforcers as criticism.
Following the conditioning period, the age of each S was obtained,
and the Shipley*Hartford Vocabulary Scale was administered. This scale
is a 40-item multiple choice word test which takes no more than 10 mine
utes to complete. Appendix A presents this scale in its entirety. No
time limit was imposed and Ss were instructed to guess.
Thus far, the independent variables of awareness and reinforce-
meat have been discussed. The dependent variable referred to most
often has been the rate of conditioning, defined as the frequency of
usage of the selected response class. The responses are assessed also
in terms of their quality. Quality herein refers to three measurable
aspects of the verbal response* These areas Response Latency; Response
Duration; and Response Units (the number of words in the response).
Both the means and the variances of these measures may be assessed
individually as dependent variables in the same manner as is rate of
conditioning. These time and number measures of the conditioned res-
ponse correspond closely to the Interaction Chronograph variables of
"activity" and "silences" (Matarasso and Saslow, 1958), and they may
be expected to permit a close analysis of the interaction effects of
awareness and reinforcement in verbal conditioning.
Response Class was first person pronouns (I, We) for half the gs,
and third person pronouns (He, They) for the remaining half. A S for
example, might be assigned to the treatments awaree" "positive rein.
forcementl" and "response class I, We." In this instance, he would
receive instructions producing the higher level of awareness, and would
receive positive reinforcement (one of B's statements) on every occa-
sion that he constructed a sentence beginning with the pronoun "I" or
"We." He would receive no response from R if he began his sentences
with "He" or "They."
There are a total of 176 stimulus cards, the first 16 of which
comprise the operant measure for each of the dependent variables. Of
the remaining 160 cards, the first five blocks, totalling 80 cards, are
arbitrarily termed Acquisition 1, and the remaining five blocks, also
totalling 80 cards, termed Acquisition Il. In reality, Acquisition 1
is a quasimextinction period during which treatment conditions are var*
ied for control purposes. The distinction is made because at the half.
way point the treatments of Type Reinforcement and Response Class are
reversed for half the Ss. As examples from Appendix 3, Ss number 15,
47, and 79 receive positive reinforcement for "I" and "We" during
Acquisition I, and are changed to negative reinforcement for "I" and
"We" during Acquisition II. Is 21, 53, and 85, on the other hand,
receive positive reinforcement for "I" and "We" at first, and are then
changed to positive reinforcement for "He" and "They." Finally, Sa
27, 59, and 91 begin with positive reinforcement for "I" and "We,"
and are changed to negative reinforcement for "He" and "They."
One of the psychiatric unit's counseling offices was used as the
experimental room* One E the author, interviewed all Ss. In addition
to the usual office furniture, a three by five foot hardwood table
was added. S and S were seated at opposite ends of the table, separ.
ated by. a low screen. The purpose of the screen was to shield from
SO' view the notations which I made on his check sheets, and to cover
the motions E made while operating the timing apparatus. Thus, an
attempt was made to reduce reinforcing agents other than K's verbal
behavior to a minimum. Those which were uncontrollable, e.g*, dis*
tracing noises or temporary variation in the placement of room fur.
nishingsp were expected to have random effects on conditioning.
The deck of three by five inch cards used for presentation of the
stimulus words was positioned in front of S. E's materials consisted
of the list of verbal reinforcers, two stopwatches of the type cali-
brated in fifths of a second, and tally sheets for recording Ss'
responses* Approximately two-thirds of the Ss took part in the experi-
ment between the hours of one and three pa.me and the remaining third
between the hours of seven p.mn and twelve midnight*. ach 3 was seen
individually for approximately 60 minutes. The conditioning periods
used 40.50 minutes of this time and the preoconditioning instructions
and post-conditioning interview and testing used the remainder*
Ss were required to construct sentences from the material printed
on three by five inch white index cards. A neutral past-tense verb was
centered in the top third of each card, and the four subject pronouns
were centered across the bottom thirds The order of the pronouns was
assigned randomly to control for position effects. The verbs used were
established as emotionally neutral using the criteria employed in an
experiment by Binder (1955). The verb list was modified through use
with a number of pilot Si. A list of the verbs in order of their
appearance during conditioning is given in Appendix Co
All -s received these general instructions. These were read aloud
by I as S examined another copy taped on the table before him. "This
is an experiment on making-up sentences. Each card before you has a
verb on it, followed underneath by four pronouns. You are to choose
whichever pronoun you prefer to begin each sentence, and then use the
verb on the card in completing the sentence." then pointed out the
verb and the pronouns, saying: "This word up here is the verb and
these four down here are the pronouns. You are to pick one of these
four for the beginning of each sentence, and then use the verb to help
you in making-up that sentence." Since a small portion, perhaps one-
fourth of the nursing assistants who served as Ss, were somewhat appre-
hensive about the experimental situation, thinking that it was perhaps
a job evaluation, the following note was added to the instructions
which the Ss had before thems "Note: This is a research project of mine
which has the hospital Research Board's approval. The project has no
relation whatever to your jobs, and no results will be reported to the
Nursing Service. You are identified by number only, not by name, and
no personal questions are asked."
This was sufficient instruction prior to the first block of
trials. The first block of 16 cards established the operant level of
response on the dependent variables for each 8J The 10 remaining
blocks of 16 cards were separated one from the other with blank slips
of paper, providing an occasional check by R on the progress of condi-
tioning, and allowing an occasional temporary rest for the Ss. Follow-
ing the completion of the first block of 16 trials, additional
instructions were given. Half the Ss received one set of instructions,
and the remaining half received the second set. These instructions
differentiate the levels of awareness. The higher level is achieved
through the instructions labLed "Aware," and the lower level, through
the instructions labeld "non-Aware." Instructions for the Aware treat-
ment group were as follows: "All the remaining cards are to be done in
the same manner, except that now there is a pussle or problem to be
solved. Once in awhile, I will make some comments or remarks after
your sentences, and these comments should help you to figure out how
the problem can be solved. Your Job is first to figure out Just how
my replies are connected to your sentences, and then show me that you
understand that connection by the way you make-up your sentences from
then en. The stopwatches and timing have nothing to do with the prob-
lem, and I am not interested in the correctness of your sentences, in
other words the English, the grammar, or sentence structure that you
use. Here is a hints the solution will be found only in the way you
use the typed words on the cards. If you use the words the right way,
you will find out why I am commenting the way I do." The instructions
for the non-Aware treatment group were exactly the same with the excep-
tion of the last two sentences. These two sentences were omitted from
the instructions given to the non-Aware Ss.
Returning for a moment to the discussion of the operant condi-
tioning paradigm, information dealing with the reinforcement continue
agency is given to the Aware 8s. In other words, they know that there
is a connection between the way they make-up their sentences and the
responses given by I. In addition, they are also given some idea of
the appropriate response class through the "...here is a hint..."
portion of the instructions. Knowledge of both the reinforcement
contingency and the response class was found in previous studies among
those Ss who conditioned the most rapidly* By contrast, the non-Aware
Ss are given information dealing solely with the reinforcement contin-
gency, and the hypotheses which they might form regarding the response
class are left entirely unstructured*
The instructions were read aloud to Ss twice, and then the card
upon which they were printed was given to Ss for their perusal and for
reference during conditioning* There were frequent questions from Ss
regarding the portion of the instructions which dealt with the rein.
forcement contingency. The questions were answered by repetition of
parts of the same instructions, and in all cases this seemed sufficient
to satisfy Ss.
Since the design required that Ss be assigned randomly to the
various treatment groups as they arrived for the experiment, an attempt
was made to ascertain whether or not this random assignment resulted in
treatment groups which were matched on the variables of Age and Verbal
Intelligence (VI), Accordingly, t was used to determine whether or not
the differences between the means of the two levels of each of the three
treatments, awareness, reinforcement, and response class, were signify*
canto Table I summaries this analysis. It may be seen, for example,
that the mean ages of the Aware and non-Aware groups were 37.5 and 39.0
years, respectively, and that this is not a significant difference by
Sttest. None of the values o t obtained in these comparisons reaches
statistical significance. The largest value of t obtained among the
comparisons was .96, yielding a probability of occurrence of almost one
time in five. Hence, it may be assumed that the random assignment of Ss
resulted in an approximate matching of treatment groups on the variables
of Age and VI.
The Pearson Product-Mement Correlation Coefficient (r) was used to
determine the correlation between Age and VI in the experimental sample.
There is a nonsignificant tendency (r = +.06, .28) for VI to increase
with Age in this sample of Ss.
C4 N C14
M < IS*0
4 @ C
If3 5 & t
Analyses of Variance
A three-dimensional factorial design was used in the analysis of
these data* The design and computational fomat followed that presented
by Lindquist (1953, pp. 240*253). The analyses of Awareness, Response
Class, and Type Reinforcement constitute a three-factor (A x B x C) ex.
perlment, and the analyses of the Age and VI variables constitute a
Treatments z Treatments x Subjects experiment. All of the FPratios preo
sented result frem these three-factor analyses in which the within-cell
variance is used, as the best estimate of the population variance, for
the error term. Because of the large number of analyses performed, an
individual Analysis of Variance summary table will not be reproduced in
the text for each of the comparisons which were made. Instead, four
tables of the F-ratios obtained will be used in explication of the
results (Tables 3, 4, 5, and 6). A sample Analysis of Variance summary
table from which these ratios are derived is reproduced in Table 2.
The individual summary tables are given in their complete form in
Table 2 remains essentially the same throughout all the analyses
with the exception that the variables of Age and VI are substituted for
Response Class in the Treatments x Treatments x Subjects portions of
The analyses for Acquisition I utilised combined data, that is,
each score used was an average of the 80 individual trials for each _
during this conditioning period. The Feratios presented in Table 3,
then, do not represent the differences obtained from block to block of
trials during the first conditioning period, but rather represent the
overall mean differences between levels of the three independent variables.
SAMPLE FORMAT FOR ANALYSIS OF VARIANCE SUMMARY TABLES
Source df SS MS F
Awareness (Awr) 1
Response Class (Rea) 1
Type Reinforcement (Ta) 1
Aws xRla 1
Av8a xTra 1
Rea x Tra 1
Awa x Ra x Tra 1
Within 88 *
Underlined IFratios in Table 3 either are statistically significant
values, or are discussed because the graphed data have meaningful
theoretical implications. Bach of the ratios was obtained in an analy-
sis similar to that presented in Table 2.
The first dependent variable listed in Table 3, Response Frequency
(rate of conditioning), yielded three significant Faratios when analysed.
The first of these was for levels of awareness--Aware Ss gave a signifil
cantly greater number of responses of the correct response class (Q -
5.09, P 4L.05)) That is, Aware Ss conditioned significantly more than
non*Aware es. The main effect of Response Class was also significant
( 8.55, P -e.01), a result which was not expected, with conditioning
proceeding at a greater rate when the response class was the third per-
son pronouns Be, They, rather than the first person pronouns I, We.
Figure 1 portrays this effect in conjunction with the effects produced
by variation in reinforcement fra positive to negative. The main
effect of reinforcement was also significant ( -( 9.47, -*c.01), with
negative reinforcement (criticism) producing greater responsivity with
u t2 4 s.I.
w l * *
- I s
I I I h
5z a a
the correct response lass. The awenes and reinforemnt effects
12 C '" ^
Figure 1 Effects of Response Class (I, We or He, They) and Type
Reinforcement by successive blocks of data during Acquisition I
the correct response class, The awareness and reinforcement effects
are presented in succeeding acquisition blocks in Figure 2.
None of the remaining F-ratios in Table 3 reached statistical
significance One of them, that for the triple interaction in the
dependent variable of the mean latency of the response, however,
approaches significance (F 3.17, P .083). Response latency refers
to the span of time which elapses after presentation of the stimulus
card and before S begins his response. The meaning of the triple inter-
action may be stated in this manners Increasing So' level of awareness
has a non-significant tendency to increase the differential effective-
ness of positive as opposed to negative reinforcement; and the two rein-
forcers have opposing effects on I, We as contrasted with He, They, as
the response class. Reference to Figure 3 will provide a graphic
1 2 3 4 5
Figure 2 Effects of Awareness and Type Reinforcement by successive
blocks of data during Acquisition I
Figure 3 Effect of the triple interaction of Awareness, Response
class, and Type Reinforcement during Acquisition I
F = 3.17, P = .083
1.7 5 "--~ . .- . "-^
Figure 4 Effects of Awareness and Type Reinforcement on Response
Duration by successive blocks of data during AcquiSition I
1 2 3 4
Figure 5 Effect of Awareness on the variability
successive Acquisition I blocks
of Units during
presentation of this interaction. As may be observed, there is little
difference in the effectiveness of positive or negative reinforcement
when Sa are onmAware (nA) HBowever, variation in reinforcement does
tend to produce some difference in response latencies when 8s are Aware
(A), and these effects are opposite when Response Class changes from
I, We to Re, They. Interpretation of this particular graph is made more
difficult by the fact that negative reinforcement for the particular
response class as represented in the diagram constitutes negative rein-
forcement for the opposing response class. That is, referring to the
left portion of Figure 3 where the response class is I, We, negative
reinforcement here refers to criticism on each occasion when the S
responds with ie or They* This interaction can probably best be sumar.
ised by the following statement. Among Aware as, but not among non-
Aware Ss, the threat of criticism tends to produce more rapid respond-
ing when the responses being criticized are first person rather than
third person pronouns. As stated, this relationship approaches sigani
finance at P = .083.
For the dependent variable of Response Duration, none of the F-
ratios reached statistical significance. However, the interaction
between Awareness and Type Reinforcement has some meaning when the data
are plotted by succeeding Acquisition blocks (See Figure 4). Here, it
may be seen that the Aware Ss who receive positive reinforcement have a
consistent though not significant tendency to lengthen the duration of
their responses. The Aware Ss who received negative reinforcement, on
the other hand, gave responses which are consistently but not signifi-
cantly shorter in duration (L 2*48, PX .13)*
The dependent variable of the variance of the response units is
included because awareness has a consistent though nonesignificant
effect (Q = 1.64, f 4<25) on the variability of response units.
That is, increasing awareness tended to increase the variability in
the number of words So used in their sentences (See Figure 5).
The analyses of variance during Acquisition II differ in two main
respects from those of Acquisition I. One of the treatment groups,
Response Class, is changed so that the "treatment levels," rather than
being I, We pronouns on the one hand, and He, They on the other, become
the same pronouns or different pronouns. In other words, the treatment
"Response Class" during Acquisition II refers to whether Spare rein*
forced for the same response class as during Acquisition I or a new and
opposite response class. This sort of substitution is feasible for the
most part because, as bornesout by the Acquisition I analysis, there were
no significant differences found among the six dependent variables
measuring quality of the conditioned response between those Ss who re*
sponded with I, We, and those who responded with He, They. Rate of
conditioning, however, was differentiated by Response Class during
Acquisition I. Therefore, on this one dependent variable, Response
Frequency, in which there was a significant difference between levels
of the Response Class treatment, additional statistical tests are per.
formed in the Acquisition II analysis to determine the effect of varia.
tion in the response class pronouns.
Acquisition II differs in a second respect also from Acquisition I,
in that the experimental manipulation of reinforcement as a treatment
is altered* In Acquisition II, half of the Sf from both reinforcement
groups are now changed to the contrasting type of verbal reinforcement.
This procedure opens to analysis a further dimension--the effect on the
conditioned response of an abrupt change to a new, contrasting type of
verbal reinforcement with the Awareness treatments held constant.
The F ratios computed in the analyses of variance for Acquisition
II are reproduced in Table 4 in the same manner as those for Acquisition
I in Table 3. The ratios which reach statistical significance and those
which suggest meaningful relationships are underlined*
With reference to Table 4, the analysis of variance for Response
Frequency (rate of conditioniag) yielded two significant main effects
and an interesting though non-significant interaction effect. Aware.
ness no longer differentiated the groups on rate of conditioning. The
main effect of Response Class was highly significant (L = 14.19, P
(.301). Those Ss who were reinforced for responding with the same re-
spoase class as during the five Acquisition I blocks gave significantly
more responses of the correct response class than did those Ss whose
reinforcement forced them to change to the opposing or different response
The second significant main effect was that of Type Reinforcement
(l 19.84, Z = 4.001). Those is who received negative reinforcement
conditioned at a significantly higher rate during Acquisition II than
those Ss who received positive reinforcement. These results correspond
with those obtained during the first five blocks of trials (Acquisi-
tion 1). The relative effectiveness of the levels of reinforcement
may be seen in Figure 6.
The interesting but non-significant interaction was that between
Type Reinforcement and Response Class (L 3.04, <*09). Reference
to Figure 7 suggests that negative reinforcement tends to have a rela-
q 9 *
S5 ; 1 a .
4| Q 4CO I<
6j n4 *a *
* 0 N *
I S |S| 2 w"
* . ,*I1*
*IN0NU * I
zds s s
0iffer^~i i-)'-0- -- - - --
D- Scn .i-)
Acquisition II Blocks
Figure 6 Effects of Response Class and Type of Reinforcement on rate of
conditioning during Acquisition II
Figure 7 Effect of the interaction of Response Class and Type Reinforce-
ment on rate of conditioning during Acquisition II
o A -
1 2 3 4 5
Acquisition II Blocks
Effects of Awareness and Type of Reinforcement on the dura-
the response during Acquisition II
Figure 9 Effect of interaction
between Awareness and Type
Reinforcement on Response Units
during Acquisition II
Figure 10 Effect of interaction
between Awareness and Response
Class on the variability of
Response Latency during Acquisi-
F = .61 P =.11F 2.7, .1
F = 2.61, P = .11
F = 2.67, P = .11
tively greater effectiveness when applied to the Sa who were required
to change their response class to avoid criticism than it did on these
So who were not required to change their response class.
In other words,
when Ss were forced to change to a new response class, negative rein.
forcement showed a nonesignificant tendency to be relatively more effect.
tive than positive reinforcement.
An attempt was made to analyse further the meaning of the signifi-
cant main effect of Type Reinforcement. Accordingly, Ss were compared
on their response frequency during Acquisition IX by whether their rein.
forcement during Acquisition I had been positive or negative* Those fs
who had previously been reinforced negatively criticizedd), subsequently
showed somewhat but not significantly more responses of the correct re.
sponse class (greater conditioning) than these Ss who had previously
been reinforced positively (t a 1.14, X = .13)0
In an effort to determine more exactly the meaning of the main
effect of response class, two additional "t".tests were computed. The
first of these compares the rate of conditioning of those So who re-
sponded with 1, We, with those who responded with He, They. This com-
parison is made because of the significant main effect of Response
Class during Acquisition I, wherein third person pronouns were more
easily conditioned than first person pronouns. If this relationship
continued into Acquisition IIX the Same versus Different classification
for the levels of the treatment "Response Class" during Acquisition X1
might be less clear than would be expected from the highly significant
, which was obtained. As it turns out, those S whose response class
was He, They, rather than I, We, were somewhat but not significantly
more easily conditioned ( =- 1.46, P <.08). Thus, this moderately
lower conditionability of I, We pronouns as opposed to He, They pro.
nouns is not felt to detract from the meaningfulness of the Acquisition
II treatment of Response Class. A '"J".test was used also in the compar-
ison between conditioning rates produced by negative reinforcement of
first person pronouns as opposed to negative reinforcement of third per-
son pronouns. It was found that negative reinforcement of first person
pronouns is somewhat but not significantly more effective than negative
reinforcement for third person pronouns (L = 1.24, P. ==.11).
Among the dependent variables dealing with the quality of the re-
sponse, only one value of F reaches statistical significance. This one
deals with the variability in the length of time S. spent in responding
(Variance Response Duration). Positive reinforcement led to a signifi-
cant increase in the variability of the duration of the response, and
negative reinforcement led to decreased variability ( -= 5.13, LP <.05).
With reference to Table 4, interestingly enough, the reverse tendency is
suggested by the data dealing with the variability of response latency.
Here, those Ss who received negative reinforcement are somewhat but not
significantly more variable in the length of time it takes them to begin
their responses (L 2.25, L <.15).
Regarding Mean Response Latency, those Ss who are reinforced nega-
tively have a non-significant tendency to wait longer before beginning
their responses. Also, there is a non-significant interaction between
Awareness and Response Class (Q -= 1.72, ( <.20). Bere, the data sug-
gest that Aware Ss have a long mean latency of response when a new
(different) response class is required, whereas the non-Aware *s' mean
latency is equally as long as when the response class is the same as it
was throughout Acquisition I*
Figure 8 portrays the non-significant interaction between Awareness
and Type Reinforcement by successive blocks during Acquisition II. Here
it will be noted that the Aware Ss tend to be relatively less affected
by variation in reinforcement from positive to negative (L = 3.48
Li <*07). Comparison of this graph with Figure 4 will point out the
reversal which has occurred in the effect of the two opposing reinforcers
on Response Duration* In Acquisition I the Aware Ss tended to be most
affected by variation in reinforcement, whereas in Acquisition II the
non-Aware S assumed this effect. Though the F-ratios obtained did not
reach statistical significance, the data would suggest that as condi.
tioning progresses the differential effect of levels of reinforcement
on Response Duration decreases.
Continuing with reference to Table 4, another interesting though
nonesignificant interaction was obtained, this one in the mean number
of units in the is' responses. This interaction (F = 2.61j = .rI),
portrayed in Figure 9, suggests that negative reinforcement tends to
produce a high number of units per response among the Aware Ss, whereas
positive reinforcement tends to produce higher unit responsiveness among
the non*Aware Ss. This interaction occurs in Acquisition II, but was
not observed during the first five blocks of trials (Acquisition I).
The last of the F-ratios of Table 4 to be discussed is that for
Variance Response Latency, in which there was a non-significant inter-
action between Awareness and Response Class ( = 2.67, = .11).
As may be seen in Figure 10. the variability of latency tends to in-
crease for Aware Sa when the response class reinforced becomes a diff-
erent one than that experienced during Acquisition I. The task is
essentially a new one for the Ss in Acquisition II, and the Aware and
non-Aware Ss are differentiated in variability of response latency in
a manner similar to their differentiation on variability of response
units (See Figure 3). In Acquisition II, when response class remains
the same, the Aware now tead to be lower in variability than the
nonAvware Spe The relatively high variability of the nonoAware Ss when
their response class remains the same, as may be observed on the right
margin of Figure 10, was accounted-for by high variability among these
Sa during the early blocks of trials during Acquisition II. The meaning
is consistent for both the Aware and the non-Aware Ss here, in that the
data suggest a steady and consistent, though non-significant, decrease
in variability on certain of the dependent variables as conditioning
Acquisition I-.Verbal Intelligence (ShipleylHartford Vocabulary Scale)
For the investigation of the effects of Verbal Intelligence (VI)
in this sample of So, the three-dimensional factorial design was main-
tained. Awareness and Type Reinforcement were continued as the treat.
ment variables, and in place of Response Class, VI was substituted as
a Subjects variable. Matching the cells by number of So for intelli-
gence (high scale score 27-39; low scale score = 13*26) reduced the
number of Ss per cell to eleven, giving a total N of 88. The summary
of Table 2 remains essentially the same except that the within vari-
ance df becomes 80, and the total f 87 Examination of the effects
of VI as a variable was confined to the trials during Acquisition I.
The reason for this was that Response Class as a treatment variable
during Acquisition II, where its levels were same or different, showed
that Response Class was involved in several meaningful interactions*
1 So numbered 5, 7, 27, 30, 32, 42, 46, and 79 were emitted from
the analysis using a table of random numbers. This provided an equal
number (22) of S from both the high and the low Verbal Intelligence
range for the treatment levels of Awareness and Type Reinforcement.
However, Response Class as it was defined in Acquisition I (I, We or
He, They), was not involved in interactions either with Awareness or
Type Reinforcement. Thus, not only sample sise (which prohibits a
four-way analysis of variance), but also the complexity of the Acquisi.
tion II interactions forbade examination of VI as a factor during
Table 5 suamarises the analysis of VI and its interactions with
Awareness and Type Reinforcement. Again, those F-ratios which are sig-
nificant or those which suggest meaningful relationships are underlined.
Beginning with the dependent variable of Response Frequency, it may
be seen that the main effects of both Awareness and Type Reinforcement
are again significant. Here, those Ss who are more aware of the rein-
forcement contingency and the response class, and those Ss who received
negative reinforcement, condition at the most rapid rate. Classification
of the Ss by VI, however, now results in two additional interactions
which have not appeared before* The first of these is between Aware-
ness and VI (See Figure 11). As may be seen, a low VI does not differ-
entiate the Aware from the non-Aware Ss on rate of conditioning. How.
ever, high V% does, the high VI Aware Ss conditioning much more rapidly,
and the high VI non-Aware gs much less rapidly than So of low VI. This
relationship is clarified further by reference to Figure 12 which por.
trays the interaction between VI and Type Reinforcement. Here, it may
be seen that variation in reinforcement has greater differential effec-
tiveness among low VI Ss than it does aaong high. The general effect of
reinforcement is clear also, and appears as it has previously. Negative
reinforcement is more effective than positive in increasing the rate of
conditioning. Both these interactions are significant beyond the .05
3 1 1 fo g
I 41 f| |
-1 0 0 0
* * .
** S. *9
'S -H C)
~-c 0 r1
C -A 0 0 0
Uo u ou o r
V C c V
-4 0- C C 1
4 1 (om l
II '3 ii
t- C c
0 C .0 C
\ to u
o Q) W C* .7
suzui asuodsad :3 :3
I o 1
.\ 0 O 0
0 C Ja
suods ,n. c o
< -. 0
i M ^i*r
Awareness and VI interact again in a meaningful fashion when the
dependent variable is the mean duration of the Ss' responses (See Fig.
ure 13). Here, VI has differential effects across the two levels of
awareness. That is, high VI Aware Ss and low VI non-Aware Ss have sig-
nificantly greater mean durations of their responses than do the low VI
Aware or the high VI non-Aware *s. In effect, then, awareness differen-
tiates high and low VI Ss on the length of time they take to respond-.
increased awareness producing longer durations among high VI Ss, and
shorter durations among low VI Sa.
The meaning of this relationship is expanded by reference to Fig-
ure 14, in which the interaction between VI and the mean number of units
per response approaches statistical significance ( 3.49, L <.07).
This suggests that awareness may differentiate high and low VI Ss on the
number of units in their responses. Increasing fs' awareness has a non-
significant tendency to increase the high VI Ss' Mean Response Units,
but has little or no effect on the Mean Response Units of the low VI Ss,
if anything-tending to decrease it slightly.
The main effect of VI is highly significant with reference to Mean
Response Units, with the high VI S3 giving a greater number of units
(words) in their responses.
The experimental sample was divided at the median for age with the
following levels resulting low, 20-38 years; and high, 39-59 years.
The total N in this analysis was reduced to 80.2
2 Ss numbered 7, 12, 17, 18, 23, 27, 29, 30, 31, 33, 37, 38, 56,
68, 70, and 74 were emitted from the analysis by use of a table of ran.
dam numbers in order to provide an equal cell "n" (20) in both levels
of each of the two treatment variables.
Reference to Table 6 provides a summary of the effects of Age and
its interactions with Awareness and Type Reinforcement on four depen.
dent variables. Beginning again with Response Frequency or rate of con-
ditioning as the dependent variable, it may be seen that both Aware.
ness and Type Reinforcement show the same main effects as they have in
the previous analyses. Increased awareness and negative reinforcement
both lead to an increase in the rate of conditioning*
The data also show a very slight tendency (! = 1.35, P (.25)
for there to be a three-way interaction between Awareness, Age, and
Type Reinforcement. Though the probability of this interaction occurr-
ing by chance is large, the possible implications for further research
warrant some consideration of the suggested relationships Figure 15
presents in graphical form this three-way interaction. As may be seen,
Age tends to differentiate the effectiveness of Awareness as a means of
increasing rate of conditioning* However, only s in the low age group
(20.38 years) show this tendency to increase their rate of conditioning
when they are more aware. So in the high age group (39-59 years) do not
show a eemparable increase in conditioning with increasing awareness*
Age has a non-significant effect on Mean Response Latency ( = 2.25,
P = .16). Here, the high age group tends to be somewhat slower in re-
sponding than the lower age group.
With regard to Mean Response Duration, Age adds a statistically
significant three-way interaction to the meaning of Awareness and Type
Reinforcement as treatment variables. Figure 16 presents this inter-
action, comparing the interactions of awareness and reinforcement at the
two age levels* Here again, awareness differentiates the effect of
positive or negative reinforcement on the length of time spent in
j ~ ^
h s *
a* Cd 0
o1 4 C4 8 ,
Figure 15 Effect of the triple interaction of Awareness, Age and Type
of Reinforcement on rate of conditioning during Acquisition I
F = 1.35, P = .25
Figure 16 Effect of the triple interaction of Awareness, Age and Type
of Reinforcement on Response Duration during Acquisition I
F =4.92, P = <.05
Figure 17 Effect of interaction between Age and Type Reinforcement on
the variability of Response Units during Acquisition I
F = 2.66, P = .11
responding in a much more striking fashion among the low age Ss than
among the high age SL. Specifically, in the high age group, increasing
awareness acts to increase Mean Response Duration in both the positive
and negative types of reinforcement. In the low age group, however,
increasing awareness causes a marked decrease in Mean Response Duration
when reinforcement is negative and a marked increase in duration when
reinforcement is positive.
The non-significaoe of Age and Type Reinforcement when the depend*
dent variable is the variability in the number of units per response
(Variance Response Units) is reproduced in Figure 17. As may be seen,
Age once again tends to differentiate the effectiveness of the two types
of reinforcement (F 2*66, 11). In this case, negative reinforce-
ment has a no-nsignificant tendency to reduce variability, and positive
reinforcement increase variability of the number of units per response,
only among Sa of the low age group.
In discussing the results obtained in these analyses, an emphasis
has been placed upon the significant or near-significant Foratios which
were obtained. It may be noted, for example, that though there are
seven dependent variables measured in the experiment, none of the
summary tables present the full expanse of data which hae been treated.
In some cases, the lack of results pertaining to particular dependent
variables might be expected to have equally as important implications
as do the results which reach statistical significance. With this in
mind, Table 7 is used to sumarise the relationships which failed to
show statistical significance. In the table, the "W" in column 1 and
row 6 indicates that variation in awareness accounted for no change in
the variability of response duration.
N M M
K N M K K M
' iu a i
^ | I : 5
S e 5 J
As may be observed in Table 7, Type Reinforcement as a treatment
variable had some main or interactive effect with all seven of the de-
pendent variables. Awareness as a treatment variable closely follows
reinforcement in overall effectiveness, producing change in six of the
seven dependent variables, and failing only in producing an effect on
the variability of the response durations. As may be seen, Response
Class (the contrast between first and third person pronouns as a treat-
ment) is extremely limited in meaning, having its effect only upon the
initial rate of conditioning* VI and Age correspond fairly closely in
being important sources of variability in this experiment. however, in
these data, variation n VI bears no relation to Mean Response Latency
or the variabilities of Latency, Duration, and Units. Likewise, varia-
tion in Age bears no relationship to Mean Response Units or the varia-
bilities of Response Latency and Duration.
From this summary it can be seen that the independent variables of
Awareness, Type Reiaforcement, Age, and VI are all extremely important
sources of variance in the dependent variables examined. Not only with
reference to the rate of conditioning, where one of these variables,
Awareness, has been shown many times over by other experimenters to be
a significant source of variation, but also among the dependent vari-
ables chosen to clarify and expand the meaning of the "quality" of the
response, here, with few exceptions, these four independent variables
are found to be inextricably associated with the conditioning process.
Only two of the comparisons noted in Table 7, for example, appear to
have any face value. It may prove important to know, as these data
suggest, that VI bears no relation to Mean Response Latency, and that
Age is unrelated to Mean Response Units.
Summary of Results
The results presented in the foregoing sections were organized
according to the overall results during each conditioning period.
A statement of the effects of each of the independent variables sep.
arately would be of value in clarifying further the role each of these
variables assumed in the conditioning process in this experiment.
Awareness. Awareness is directly related to rate of conditioning
This relationship occurred several times in the foregoing analyses, with
statistical significance being reached consistently. Awareness inter*
acts with VI to a degree reaching statistical significance with refer-
ence to both rate of conditioning and the mean duration of the response.
Increased awareness caused a corresponding increase in rate of condi-
tioning among high VI -s, but had no effect whatever on the rate of
conditioning of ts in the low VI range. This relationship is an impor-
tant one because the data suggest that lack of awareness actually pro.
duces a decrement in the rate of conditioning among the high VI -s.
Regarding Mean Response Duration, Awareness produces opposite effects on
the high and low VI groups, raising the Mean Response Duration for the
high VI Ss, and shortening the duration for the low VI Ss. Finally,
Awareness interacts significantly with both Age and Type Reinforcement
with reference to the Mean Response Duration. Here, Awareness differ*
entiates the effectiveness of the opposing reinforcers at the low age
level, but fails at the high age level*
The remaining effects of awareness on the conditioning process
failed to reach statistical significance, but suggested several meaning-
ful relationships. With regard to rate of conditioning. Awareness com-
bines with Age and Type Reinforcement in a manner which suggests that
awareness is effective in differentiating between the two types of
reinforcement only within certain limits imposed by the ages of the Ss.
Awareness combines with Response Class and Type Reinforcement to produce
a reduction in Mean Response Latency when first person pronouns are
criticized. Further, an increased Mean Response Latency tends to occur
when Aware Ss are required to adapt to a change in response class during
Acquisition II. Awareness combines with Type Reinforcement in tending
to produce an increase in Mean Response Duration when reinforcement is
positive, and a decrease in duration when reinforcement is negative.
This particular tendency reverses itself as conditioning progresses)
and, as nonoAware Ss begin to approach asymptotic conditioning level,
their mean durations are now higher for the positive reinforcement
group and lower for the negative reinforcement group than the mean dura-
tions of the Aware Ss which became stabilized earlier. Awareness combines
with Type Reinforcement to produce opposing effects on Mean Response
Units, tending to increase the number of units when reinforcement is
negative and decrease units when reinforcement is positive. Further,
Awareness tends to decrease Mean Response Units among low VI Ss, and
increase the units among high VI Js. Awareness tends to increase the
variability of response latency when a new and different response class
is learned, and decrease the variability of latency when the same re-
sponse class is continued. Awareness tends to increase the variability
of response units. This effect is slight and occurs only at the begin-
ning of the conditioning process.
Response Class. Response Class has one main effect. Third person
pronouns are significantly more easily conditioned than first person
pronouns in this sample of Ss.
Type Reinforcement. Negative reinforcement increases rate of
conditioning to a significantly greater extent than does positive rein.
forcemento This relationship is found throughout the various analyses.
Late in the conditioning process, positive reinforcement causes a sig-
nificant increase and negative reinforcement a significant decrease in
the variability of response duration. Regarding rate of conditioning,
reinforcement combines with VI in that negative reinforcement is sig-
nificantly more effective on low VI Ss. Positive reinforcement signi-
ficantly increases Mean Response Duration as Ss increase in awareness,
whereas negative reinforcement decreases the duration with increasing
awareness. This effect occurs only in the low Age group.
The remaining relationships do not reach statistical significance.
When Awareness is present, negative reinforcement for first person pro.
nouns tends to decrease Mean Response Latency. Negative reinforcement
tends to increase Mean Response Units with increasing awareness, whereas
positive reinforcement tends to decrease Mean Response Units with in-
creasing awareness* Negative reinforcement tends to increase the varia-
bility of response latency. The variability of response units was de-
creased by negative reinforcement and increased by positive reinforce-
ment when Ss were in the low age group.
Verbal Intelligence. VI is directly and significantly related to
Mean Response Units when sa are Aware. High VI produces a significantly
greater rate of conditioning when Ss are Aware, and a significantly
lower rate when Ss are non-Aware. Low VI results in a significantly
greater rate of conditioning when Ss are reinforced negatively, and a
significantly lower rate of conditioning when Ss are reinforced posi-
Age. In this sample. Age is involved in only one statistically
significant effect. Age produces a significant interaction between
Awareness and Type Reinforcement when Response Duration is the cri-
terion. Negative reinforcement reduces duration with increasing aware*
ness, while positive reinforcement increases duration as awareness in.
creases. This effect occurs only in the low age range. The two remain.
ing effects in which Age takes part are similar in that age variation
permits interaction effects to occur between Awareness and Type Rein.
foreement. Both these effects suggest that only Ss in the low age range
vary in rate and quality of conditioning as Awareness and Type Rein.
The treatment of awareness as an independent variable In this
study is seen as an extremely important experimental manipulation.
As was the case here, an attempt was successfully made to distinguish
between two otherwise matched experimental groups on the awareness cri-
terion. The distinction was made on the basis of the quality of the
information the groups received prior to the learning task. The group
which received information dealing with the reinforcement contingency
and information concerning the response class differed remarkably on
many measurable conditioned response criteria from the group which had
information dealing with the reinforcement contingency alone. The
learning task as represented in this study is only one of many possible
learning problems, and is a relatively easy problem solving task which
proceeds by successive elimination of non-reainfrced responses* How-
ever, the results strongly support the contentions of several of the
authors already cited e**g,* Dulany (1962), Kimble (1962), Spielberger
(1962), that the cluster of hypotheses and intentions which form a
subject's frame of reference in a learning task of this sort is a frame
of reference which must be manipulated and controlled if its effect upon
the conditioned response is to be predictable.
As stated, awareness was found to be manipulable enough to produce
measurable differences in the rate of conditioning and in the quality
of the conditioned response in this study. The differences are
attributable in part to the higher level of awareness produced by
instructions dealing with both the contingency and the response class.
Spielberger's (1962) PCI was used with 61 of the 96 subjects, and the
responses to these questions supported the inference that verbalised in-
sight would occur more rapidly among Aware subjects. The dependent varia-
bles which have been examined, dealing with those characteristics of the
conditioned response termed "response quality," have particular signifi-
cance. It has been demonstrated many times that an increase in aware-
ness accompanies an increase in the rate of conditioning. Here, however,
"response quality" attempts an approach to those more elusive factors in
learning, the motivational and intentional variables. For example, in
Dulany's (1962) theoretical network, behavioral intentions play as great
a theoretical role as do the relatively easily manipulable behavioral
hypotheses of which a subject's frame of reference is composed.
The dependent measures examined herein were the length and amount
of variability produced in the latency, the duration, and the units of
the conditioned response. The choice of these particular variables was
made because time and frequency measures of experimenter-subject inter-
action behavior have high reliability (Matarasso, 1962). Latency, dura-
tion, and units correspond closely to the variables of "silences" and
"activityt.,which accounted for the major portion of the total variance
among "time" measures of interaction behavior in a factor analytic
study by natarazso and Saslow (1958).
Before proceeding with the discussion of the relation of these
variables to the motivational and intentional properties of the inter-
actions in this experiment, the research hypotheses regarding rein-
forcement, age, and verbal intelligence will be considered.
The research hypotheses regarding type of reinforcement were sub-
stantiated. Positive and negative reinforcement, as defined herein,
differed in their effects upon both the rate of conditioning and the
quality of the conditioned response. The hypotheses regarding the
interaction effects of awareness and reinforcement were also supported.
Here, the effectiveness and the nature of the conditioned response
resulting from the use of a particular class of reinforced, varied
according to the level of awareness or frame of reference which had
been imposed upon the subjects in a group. The major present finding,
that negative reinforcement is more effective than positive in increas-
ing the rate of conditioning regardless of awareness level, substan.
tiates the hypothesis made and finds some support in the literature.
A "failure" group showed significantly more conditioning than a
"success" group in a study by Kanfer and Karas (1959), and Kanfer, Bass,
and Guyett (1963) found that subjects were more influenced when the
experimenter disagreed with their opinions*
The results of these studies as well as the present one, however,
do not allow a general statement to be made regarding the presumed
greater efficacy of negative reinforcement. In the present study, for
example, the subjects' age and intelligence became quite important
factors in determining the characteristics of the responses produced
by differing reinforcers. It was noted, for example, that negative
reinforcement was a great deal more effective conditioner than positive
for subjects of the low intelligence range, but that the difference was
only slight when subjects were of the higher intelligence group. In a
like manner, negative and positive reinforcement had opposing effects
upon response duration when subjects were of the age group which cam-
pared with that of the experimenter, whereas their effects were similar
when the age group was appreciably higher than that of the experimenter
giving the praise and criticism. The suggestion here, of course, is
that not only do subject variables complicate analysis of the response,
but that experimenter variables may have to be dealt with simultan-
eously. A clear example is provided in a study by Binder, McConnell,
and Sjoholb (1957), wherein subjects were more easily conditioned to
use hostile words when the experimenter was a five foot, ninty pound
female than when he was a six foot five inch, two-hundred twenty pound
ex-Marine. Another example is provided by Sapolsky's (1960) finding
that when subject and experimenter were "incompatible," conditioning
failed to occur-ountil after the experimenter left the roome
In the present study, again, the results imply that experimenter
characteristics played a role in changing the effectiveness of negative
reinforcement when applied to the differing age groups. The effect may
be interpreted in at least two ways, both of which have similar impli-
cations for the discussion of motivation and intention which succeeds
this section. It is possible that the experimenter was less effective
in his criticism of the high age group. This may have been true
because of societal respect for elders which resulted in a subtle
softening of the criticism; or, perhaps because elders are less inclined
to remain attentive to juniors in age, and be particularly attentive or
inattentive to criticism. Regarding the low age group, it might be less
likely that the experimenter's criticism would be softened, and perhaps
more likely that this group (whose mean age was lower than that of the
experimenter) would remain attentive to the criticism. These hypothe-
ses may be tested experimentally to determine if the age and perhaps
also the professional position of the experimenter served to confound
the results dealing with age and intelligence. On the other hand,
the effects (referring to Figures 12, 16, and 17) may be real, and not
confounded by experimenter variables.
As may be assumed from this discussion, the research hypotheses
dealing with subjects' age and verbal intelligence were supported by
the results. Both these subject variables were sources of variation
in rate of conditioning and in the quality of the conditioned response.
The dependent variables dealing with response quality are now
examined with reference to the mediating properties of awareness and
reinforcement as they were treated in this design. A summary is pro*
vided in Table 8 which includes all those relationships which attained
statistical significance in this sample. Also included are several
results which suggested trends but did not reach statistical signifit
chance. Two columns are provided for each major independent variable.
Columns A and B under "Aware" refer to Aware and non-Aware subjects
respectively. For "Reinforcement," A and B refer to positive and nega-
tive respectively. For both the Age and the Verbal Intelligence varia-
bles, A and B refer to the high and low groups on that variable. The
dependent variables listed at the left margin were increased (I) or
decreased (D) by the action of the treatment level. Underlined items
reached statistical significance.
Several assumptions must be made in attempting to determine the
meaning of these dependent variables with regard to their use in assess*
ing the motivational and intentional properties of the trftatment varia.
bles. For the purposes of this discussion, a treatment is said to have
desirable effects from the point of view of behavior modification if it
causes the subjects to accomplish exactly what the experimenter wants
them to accomplish, and if they expend their best effort to produce a
response which qualitatively approaches the limit of their ability. For
convenience, this desirable result of treatment will be termed "Task-
Oriented Effort" (TO~).
Ml 4! ~
64 P41 a
U~ U 4
B' 41 4 1
i a I r jE II
From the data analyses of the results section, it is felt that
the following assumptions regarding the meaning of the dependent varia-
bles with reference to TOE are justified. All the following signify an
increase in TOE An increase in Mean Response Latency! an increase in
Mean Response Duration; an increase in Variance Response Duration; an
increase in Mean Response Units; and an increase in Variance Response
With reference to Table 8 now, it may be noted that those treat-
ment groups higher on the Awareness criterion, in all eases reported,
gave responses indicative of a greater task involvement, i.e., greater
TOE. The relationships examined which seem to warrant such an assump-
tion are testable experimentally, and should be subjected to experi-
mental scrutiny in subsequent research. Concerning aware.iess as
mediating process, the results here suggest that increased motivation
and increased intent to respond accompany increased Awareness in the
verbal conditioning situation examined.
With reference to the levels of reinforcement used as treatment,
three of the four measures suggest that positive reinforcement increases
the subjects' TOR and negative reinforcement decreases it. In one case,
however, negative reinforcement served to increase TOE. This occurred
when the unit responsivity of Aware subjects increased when criticism
was used as the reinforced.3
The effects of Age on TOR are not discussed because of the con-
founding which may have occurred as a result of characteristics of the
experimenter* During the experiment proper, for example, it was noted
3 These subjects may have received more positive reinforcement
from the successful avoidance of criticism than those who were directly
praised for the correct response.
that a great degree of variation in rapport between experimenter and
subject occurred, and that at least part of this seemed to be indepen.
dent of the treatment administered* It is not unlikely that certain
personality variables, not measured in this study, accounted for much
of the individual differences in responsiveness to a situation of this
The effect of Verbal Intelligence on TOE may also be confounded
to an unknown degree by experimenter characteristics. However, the
analysis suggests that subjects of the higher intelligence range
demonstrate an increase in TOR while those of the low range fail to
do so. An important dimension of verbal learning situations appears
to be open for analysis here. Specifically, the variation in mediating
responses which seem to be produced in part by variations in the sub-
ject variables of age and verbal intelligence, as well as that produced
directly by variation in treatment, appears to be amenable to analysis
utilizing the verbal conditioning paradigm in conjunction with variables
assessing response quality.
Several important questions come into focus, and several expert
mental designs are suggested as a result of this research. Those of
most interest to the author deal with the interaction effects of sub.
ject and experimenter variables in the diadic interview or problem
solving situation under varied experimental conditions. The fact that
such variables may be controlled and examined experimentally gains in
importance because of the implications such definitive control has for
behavior modification. The verbal learning paradigm has been considered
by many as a close analogue to the behavior modification which is under-
taken by professional clinicians in the psychotherapy situation. Future
research using a paradigm similar to that which has been explicated in
this research, may be expected to provide an important source of tech-
nical skill for the clinician whose job it is to induce psychological
growth in those who come to him for professional help.
In summary, the hypotheses related to awareness as a treatment
variable were confirmed. An increase in the subjects' awareness was
found to correspond with an increase in rate of conditioning. Also,
awareness was found to be a major factor in the production of variation
in the "quality" of the response. Analysis of the measures of response
quality as indicators of differential mediational processes, suggests
that an increase in awareness may be accompanied by an increase in the
motivation and intention to respond in a task-centered manner.
Reinforcement was found to be significantly related to rate of
conditioning. The hypothesis that negative reinforcement is more effec-
tive than positive in increasing the rate of conditioning was confirmed.
Also, variation in type of reinforcement was found to be accompanied to
a significant degree by variation in the quality of the conditioned
response. Analysis of the effects of varied reinforcement upon response
quality suggests that positive reinforcement or praise results in a
greater tendency to respond in a task-centered manner.
Variation in age has no effect upon rate of conditioning in this
sample. The data suggest that unspecified experimenter variables may
account in part for this lack of results. Age was significantly related
to one measure of response quality, but it was felt that the data did
not warrant a statement being made regarding the mediating properties
of this subject variable.
Verbal intelligence was found to be significantly related both to
the rate of conditioning and to the quality of the conditioned response.
Higher intelligence produced a greater rate of conditioning* More Ia*
portantly, high verbal intelligence was associated with changes in re.
sponse quality which suggest increased motivation and intent to respond
in a task-centered manner.
This research examined the effects of several variables on the
conditioning process in a two-person problem solving situation. The
task involved selection of pronouns for use as the subject pronouns
of sentences which were to be constructed. One of the variables ex.
mined, type of reinforcement, was contingent upon selection of a par-
ticular class of pronouns. The conditioned response was analysed along
several dimensions of quality as well as on the dimension of frequency of
occurrence. The effects on the conditioned response were examined across
the four major treatment variables of: level of awareness; type of rein-
forcements subject's agoe and subject's verbal intelligence. The results
obtained suggest that all four of these treatment variables are closely
related to the behavior change which occurs* The results were discussed
with emphasis placed on the inadvisability of conducting research in
verbal learning without some attempt being made to control the effects
of these treatments* The role of thei-vatious treatments in producing
mediating hypotheses was discussed, and a beginning attempt at measure-
ment of the mediational processes was made through analysis of response
quality and Task-Oriented Effort. Finally, the importance of further
research efforts using the verbal learning paradigm was stressed, with
attention being directed to the direct analogy to the behavior modifi-
cation produced in psychotherapy*
Ashby, J. D., Ford, D. H., Guerney, B. G. Jr., and Guerney, Louise F.
Effects on clients of a reflective and a leading type of psychotherapy.
Psychol. Monogr., 1957, 71, Whole No. 453.
Binder, A., McConnell, D., & SJohola, N. A. Verbal conditioning as a
function of experimenter characteristics. J. abnorm. soc. Psychol.,
1957, 55, 309-314.
Cohen, A. R., Greenbaum, C. W., and Mausson, Helge H. Commitment to
social deprivation and verbal conditioning. J. abnorm. soc. Psychol.,
1963. 67, 410.421.
Dulany, D. B. Jr. The place of hypotheses and intentions an analysis of
verbal control in verbal conditioning. 1. Pers., 1962, 30(2), 102-129*
Ekman, P., Krasner, L., and Ullman, L. P. Interaction of set and awareness
as determinants of response to verbal conditioning. J. abnorm, soc.
Psychol. 1963, 66, 387*389.
Briksen, C. W. Figuents, fantasies, and follies: a search for the
subconscious mind* J. Pers., 1962, 30(2), 3.26.
Gewirtz, J. L., and Baer, D. Me Deprivation and satiation of social
reinforcers as drive conditions. J. abnorm. soc. Psychol., 1958,
Greenspoon, J. The reinforcing effect of two spoken sounds on the
frequency of two responses. Amer. J. Psychol., 1955, 68, 409-416.
Greenspoon, J. Verbal conditioning and clinical psychology. (In)
Bachrach (ed.) Experimental foundations of clinical psychology. New
York Basic Books, 1962.
Kanfer, F. H., Bass, B. M., and Guyett, 1. Dyadic speech patterns,
orientation, and social reinforcement. J. Consult. Psychol., 1963,
Kanfer, F. H., and Karas, Shirley C. Prior experimenter-subject interaction
and verbal conditioning. Psychol. Rap., 1959, 5, 345-353.
Kimble, G. A. Hilgard & Marquis' Conditioning and learning. (2nd ed.)
New Yorks Appleton-Century-Crofts, 1961.
Kimble, G. A. Classical conditioning and the problem of awareness.
J. Pers., 1962, 30(2), 27-45.
Krasner, L. Studies of the conditioning of verbal behavior. Psychol.
Bull., 1958, 55, 148-170.
Krasner, L. The therapist as a social reinforcement machine. (In)
Strupp & Luborsky (Ids.) Research in Psychotherapy. Vol. II, APAs
Washington, DC, 1961.
Letchworth, G. E. and Wishner, J. Studies in efficiency: verbal
conditioning as a function of degree of task centering, a replication.
J. abnorm. soc. Psychol. 1963, 67, 282-286.
Levin. S. M. The effects of awareness on verbal conditioning. J. exp.
Psychol., 1961, 61, 67-75.
Lindquist, E. F. Design and analysis of experiments in psychology and
education. Boston: Houghton Mifflin, 1953.
Luborsky, L. and Strupp, H. Research problems in psychotherapy: a three-
year follow-up. (In) Strupp & Luborsky (Ids.) Research in Psychotherapy,
Vol. II, APAs Washington, DC, 1961.
Mataraszo, J. D. Prescribed behavior therapy: suggestions from interview
research. (In) Bachrach (Ed.) Experimental foundations of clinical
psychology. New York: Basic Books, 1961.
Matarasso, J. D. and Saslow, G. Factor analysis of interview interaction
behavior. J. consult. Psychol., 1958, 22, 419-429.
Merbaum, M. The conditioning of effective self-references by three
classes of generalised reinforcers. J. Pers., 1963, 31# 179-191.
Ring, K. and Kelley, H. H. A comparison of augmentation and reduction
as modes of influence. J. abnorm. soc. Psychol., 1963, 66 95-102.
Sapolsky, A. Effect of interpersonal relationships upon verbal condition-
ing. J. abnorm. soc. Psychol., 1960, 60. 241-246.
Simkins, L. Effects of examiner attitudes and type of reinforcement
on the conditioning of hostile verbs. J. Pers., 1961, 29, 380-395.
Simkins, L. Instructions as discriminative stimuli in verbal condition-
ing and awareness. J. abnorm. soc. Psychol., 1963, 66, 213-219.
Spielberger, C. D. The role of awareness in verbal conditioning. J.
Pers., 1962, 30(2), 73-101.
Spielberger, C. D., Berger, A., and Howard, Kay. Conditioning of
verbal behavior as a function of awareness, need for social approval,
and motivation to receive reinforcement. J. abnorm. soc. Psychol.,
1963, 67, 241-246.
Taffel, C. Anxiety and the conditioning of verbal behavior. J.
abnorm. soc. Psychol., 1955, 51 496.501.
Thorndike, E. L. and Rock, R. T. Jr., Learning without awareness of
what is being learned or intent to learn it. J exp. Psychol.,
1934, 17, 121-134.
SRIPLIYEtARTFORD VOCABULARY WORD TIST
draw, eat, speak, sleep
allow, sew, cut, drive
forgive, pound, divide, tall
pin, eraser, sofa, glass
swi, recall, number, defy
drink, erase, fall, think
silvery, titled, young, dreadful
swift, auddy, lsfy, hearty
graen, obvious, sceptical, afraid
conductor, officer, book, pretender
deserve, distrust, fight, separate
vleame, fix, stir, enchant
defy, excite, signify, bicker
red, sharp, uninformed, precise
submerge, strengthen, vent, deaden
length, fame, head, loyalty
yield, buy, associate, tell
bright, large, speedy, low
laughter, speed, grace, malice
stolen, pointed, remade, soiled
tease, belittle, cut, waste
drunk, ballast, heading, ape
help, turn, strip, bewilder
humorous, paltry, fervid, plain
reduce, strew, inform, delight
eat, lament, dominate, cure
senator, inhabitantafish, atom
dispossess, intrude, rally, pledge
charm orphan, diago, pond
untidy, involatile, rigid, sparse
dried, notched, aimed, blunt
moldy, loose, supple, convex
mitigate, direct, pertain, abuse
appropriate, intend, revoke, maintain
brush, hole, building, lute
maniacal, curious, devout, complaining
outcast, priest, lentil, locker
waken, ensue, incites placate
rashness, timidity, desire, kindness
vain, sound, first, level
__ _C___ __
ORDER OF TREArIMNTS DURING CONDITIONING PERIODS
SAcquisitien I Acquisition II
Subjects Aa ,am Tra .Ra Tra
- -- --
OVER OF VERB APPEARANCE DURING CONDITIONING
Operant Acquisition I Acquisition II
o-- .- -- ... .
ANALYSIS OF VARIANCE SaMAR1I TABLES
Peaponse Frequency-Acquisition I
Source df SS MS F
An -T 36.75 367.75 .09
Rea 1 61.76 61.76 8.5$
Ta 1 68.34 68.34 9.17
AmaxERa 1 3.01 3.01 .42
AvaxTra 1 7.60 7.60 1.0$
RcazTra 1 *92 .92 .13
AwaxRcaxTnr 1 2.10 2.10 .29
Within 88 63t.94 7.22 -
Total 95 81$,2 8.58 -
Mean Response Latency-4cquisition I
Source df SS MS F
Aw 1 2.16 2.16 731.
Rca 1 .40 .40 .06
Tra 1 .02 .02 .00
AwnazRa 1 .03 .03 .00
AwaxTra 1 7.37 7.37 1.15
RcaxTra 1 6.83 6.83 1.06
AwaxRcxTra 1 20.35 20.35 3.17
Within 88 5$6.07 6.41 -
Total 95 601.22 6.33 -
Mean Response Duration-Aequisition I
Source df SS MS F
Ava T .06 .06 722
Rca 1 .00 .00 .00
Tra 1 .15 .15 .56
AwaxRca 1 .09 .09 .33
AvaxTra 1 .67 .67 2.48
RcaxTra 1 .00 .00 .00
AwaxRcaxTra, 1 .02 .02 .07
Within 88 2.16 .27 -
Total 95 2$.15 -
APPEDIX D (Contined)
Mean Response Units-Acquisition I
Source df 85 MS F
Awa T .67 .67 .69
Rea 1 .11 .11 .11
Tra 1 .8 .81* .87
AwaxRoa 1 .03 .03 .03
AwaxTa 1 .74 .74 .76
RcaxTzd 1 .08 .08 .06
AwaRcasxTra 1 .05 .05 .0
Within 88 85.75 .97 -
Total 95 88.27 -
Variance Response Unite-Acquisition I
Source df w8 MS F
Ava T 1.38 1.38 176
Rea 1 .21 .21 .25
Tra 1 .23 .23 .27
AvaxRca 1 .36 .36 .43
wazTra 1 .3b .34 .4o
RcaxTra 1 .19 .19 .23
AwcaxBaxTra 1 .2 .2t .29
Within 88 7k.20 .8 -
Total 95 77.15 -
Response Frequency-Acquisition II
Source df 88 MS 7
Awa T 4.60 k.60 .28
Rea 1 231.26 231.26 1k.19
Tra 1 323.40 323.40 19.8k
AvaxRia 1 13.95 13.95 .86
AvaxTra 1 5.32 5.32 .33
ReaxTra 1 49.60 49.60 3.0o
Aamxcaxtra 1 11.07 11.07 .68
Within 88 1k34.76 16.30
Total 95 2073.96 -
Mean Response Latency-4cquisition 11
Source df SS MS F
Awa T .38 .38 7ok
Rca 1 .1S .5 .0
Tra 1 12.87 12.87 1.*2
AvaxRca 1 15.6k 15.61 1.72
AwvaTra 1 .8U .8k .09
RcaxTra 1 3.11 3.11 *.3
AwaxRcaxTra 1 .63 .63 .07
Within 88 799.90 9.09 -
Total 95 833.82 -
APPMDIX D (Continued)
Mean Response Duration-Acquisition II
Mean Response Unite-Acquisition II
Variance Response latency-Acquisition II
Variance Response Duration-Acquisition II
~__ _____ I_________~_~I___ _II__I___ ____ ~~~__ ~____~_________ _1 3___1__1_ ___~_
APPUDIC D (Continued)
Response Frequency--Acquisition I-Verbal Intelligence
Souree df SS MS F
A T 13.68 68 38 o$5
VI 1 .35 .35 .0o
Tra 1 83.26 83.26 11.53
AvxVI 1 35.13 35.13 4.87
AwaxTra 1 1.75 1.75 .24
VIxTra 1 36.63 31.63 4.80
AwaxVIxTra 1 .96 .96 .13
Within 80 577.88 7.22
Total 87 777.61 -
Mean Response Duration-Acquisition I-Verbal Intelligence
Souree df SS S F
Aa .02 .02 .07
VI 1 .09 .09 .33
TTra 1 .18 .18 .67
A vaVI 1 1.,7 1.17 5.4?
Awaxtra 1 .75 .75 2.78
V-TxTra 1 ,10 .10 .37
AwaxVIxTra 1 .01 .01 .0o
Within 80 21.23 *27 -
Total 87 23.85 -
Mean Response Unite-Acquisition II-Verbal Intelligence
Source df SS MS F
Awa 1 .80 .80 .90
VI 1 6.51 6.51 7.31
Tra 1 1.77 1.77 1.99
AmaxVI 1 3.11 3.11 3.49
AwaxTra 1 1.02 1.02 1.15
VIATra 1 .94 .91 1.06
AwaxVIxTra 1 .58 .58 .65
Within 80 71.55 .89
Total 87 86.28 -
Response Frequency-Acquisition I-*Age
Source df SS MS F
Ave 1 38.92 38.92 1.78
Age 1 .09 .09 .01
Tra 1 68.08 68.08 8.35
AwaxAge 1 11.01 11.01 1.3y
AvaxTra 1 3.13 3.13 .38
AgeTra 1 .06 .06 .01
AuaxAgeTB r 1 11.01 11.01 1.35
Within 72 586.60 8.15 -
Total 79 718.90 -
APPENDIX D (Continued)
Mean Response Latency-Acquisition I-Age
Source df SS MS F
Awa 1 7.*5 7.%5 1.25
Age 1 13.63 13.63 2.25
Tra 1 1.09 1.09 .18
AmaxAge 1 .88 .88 .
AwaxTra 1 .13 .13 .02
Agelxra 1 .38 .38 .06
AwaxJgexTra 1 .00 .00 .00
Within 72 435.67 6.05
Total 79 459.31 -
Mean Response Duration-Acquisition I-Age
Source df SS MS
Aw 1 .30 .30 1.20
Age 1 .00 .00 .00
Tra 1 .27 .27 1.08
AwaxAge 1 .03 .03 .12
AwaxuTa 1 .81 .81 3.24
AgexTra 1 .09 .09 .36
AwaxgexTra 1 1.23 1.23 .92
Within 72 18.15 .25 -
Total 79 20.88 -
Variance Response Units-Acquisition I-Age
Source df 8 MS F
Awa 1 90 .90 f1
Age 1 .01 .01 .01
Tra 1 1.01 1.01 1.28
AwaxAge 1 .00 .00 .00
AwaxTra 1 .70 .70 .89
AgexTra 1 2.10 2.10 2.66
AwaAgexzTra 1 .04 .o0 .0
Within 72 57.18 .79 -
Total 79 61.94 -
Edward George Shhleiner was born December 30, 1933, in
Flushing, New York, He attended grade school at New York Public School
107, and high school at Bayside, and Phelps, New York, graduating from
the Phelps High School in June, 1951. He enrolled at the University of
Florida in the Fall of that year and remained until enlisting in the United
States Navy as a Naval Aviation Cadet. Upon his discharge he re-entered
the University of Florida and received the Bachelor of Arts degree with
a major in Psychology in June, 1960* He continued his studies toward an
advanced degree and received the degree of Master of Arts in Psychology
in June, 1962. He worked as a Psychology Trainee at the Veterans Adminis-
tration hospital in Bay Pines, Florida, until September, 1962, when he
returned to the University of Florida to continue his studies. While
attending school he served as a Clinician in the University Counseling
Center. Since September, 1963, he has been an Intern in Psychology at
the Veterans Administration Hospital in Gulfport, Mississippi. He expects
to receive the degree of Doctor of Philosophy in Psychology in April, 1965.
Edward George Schleimer is married to the former Barbara Joan
Fischer and is the father of one child.
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 Philosophy.
April 24, 1965
Dean, Collge of/Ar s and Sciences
Dean, Graduate School