Title: Behavior charting as an adjunct to the dyadic counseling relationship
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00098383/00001
 Material Information
Title: Behavior charting as an adjunct to the dyadic counseling relationship
Physical Description: x, 115 leaves. : ; 28 cm.
Language: English
Creator: Davis, Jannar William, 1944-
Publication Date: 1972
Copyright Date: 1972
 Subjects
Subject: Counseling   ( lcsh )
Behavior modification   ( lcsh )
Counselor Education thesis Ph. D
Dissertations, Academic -- Counselor Education -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis -- University of Florida.
Bibliography: Bibliography: leaves 109-113.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00098383
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000577264
oclc - 13944042
notis - ADA4959

Downloads

This item has the following downloads:

PDF ( 3 MBs ) ( PDF )


Full Text














BEHAVIOR CHARTING AS AN ADJUNCT
TO THE DYADIC COUNSELING RELATIONSHIP














By

JANNAR WILLIAM DAVIS


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
1972
















ACKNO1ITLEDGMENTS


I would like to express my gratitude to the following

people who were instrumental in the realization of this

study:

Dr. Robert O. Stripling, chairman of my doctoral

committee, for his continuing guidance culminating four

years of interest and support during my graduate career.

Dr. Henry S. Pennypacker, for providing the spark of

an idea, as well as a rationale, for this research.

Dr. Hal G. Lewis, for sharing his broad knowledge

and intellectual curiosity which underlie this research.

Sue, my wife, for her faith in me and her encourage-

ment in completing this study.















TABLE OF CONTENTS


Page


ACKNOWLEDGMENTS .. . . . . .

LIST OF TABLES . . . . .

LIST OF FIGURES . . . . . .


ABSTRACT.


I INTRODUCTION . ....
Purpose of the Study.
Rationale of the Study.


* . 1
. . 4
. . 6
6


II REVIEW OF THE LITERATURE. . . . .
Introduction . . . . . .
The Experimental Analysis of Behavior
Precise Behavior Management .. ...
The Counseling Relationship . . .
Feedback .. . . . . . .
Description of the Instruments. . .


METHOD . . . . . . . .
Design . . . . . . .
The Sample . . . . . .
Procedure . . . . . . .
Selection and Training of Counselors.
Selection and Training of Raters. .

RESULTS AND ANALYSIS OF THE DATA. ...
Client Self-Referent Behavioral
Change . . . . . . .
Recorded Client Behavior Changes. .
Reliability . . . . . . .


9
9
11
23
S 30
35
S 38


S 43
43
S 44
S 45
48
50

52

53
55
63


iii


CHAPTER


V

vi

vii


III








IV


. o o o o .


. o o o o .






TABLE OF CONTENTS (CONTINUED)


CHAPTER

V SUMMARY, CONCLUSIONS, AND IMPLICATIONS. .
Summary . . . . . . . .
Conclusions . . . . . . .
Implications . . . . . .

APPENDICES

A MATERIAL FOR RATERS . . . . . .

B FREQUENCIES OF RATED BEHAVIORS BY TREATMENT
PAIRS . . . . . . . . . .

BIBLIOGRAPHY . . . . . . . . . .

BIOGRAPHICAL SKETCH . . . . . . . . .


Page

90
90
94
96




101


104

109

114














LIST OF TABLES

LBLE Page

1 SELF-REFERENT SCORES OF CLIENTS BY
TREATMENT . . . . . . . .. 54

2 INDEPENDENT t TESTS OF LAST-SESSION
BEHAVIOR FREQUENCIES BY TREATMENT . 56

3 CLIENT BEHAVIOR ACCELERATIONS BY TREATMENT. 57

4 THE RELATIONSHIP BETWEEN COMFORT PAIR
COMPONENTS BY CLIENTS . . . . .. 61

5 PEARSON PRODUCT-MOMENT COEFFICIENTS OF
RELIABILITY BETWEEN TWO INDEPENDENT RATERS. 85

6 PEARSON PRODUCT-MOMENT COEFFICIENTS OF
INTRARATER RELIABILITY . . . . ... 86

7 CORRELATION BETWEEN COUNSELOR AND RATER
EVALUATIONS OF THE EXPERIMENTAL CLIENT'S
IN-SESSION PUT UPS AND PUT DOWNS. . . 88















LIST OF FIGURES


FIGURE

1


COMPARISON OF PUT UPS AND PUT DOWNS FOR
TREATMENT PAIR 1 . . . . .

COMPARISON OF PRESENT AND PAST TENSE
STATEMENTS FOR TREATMENT PAIR 1 . .

COMPARISON OF PUT UPS AND PUT DOWNS FOR
TREATMENT PAIR 2 . . . . .

COMPARISON OF PRESENT AND PAST TENSE
STATEMENTS FOR TREATMENT PAIR 2 . .

COMPARISON OF PUT UPS AND PUT DOWNS FOR
TREATMENT PAIR 3 . . . . .

COMPARISON OF PRESENT AND PAST TENSE
STATEMENTS FOR TREATMENT PAIR 3 .

COMPARISON OF PUT UPS AND PUT DOWNS FOR
TREATMENT PAIR 4 . . . . .

COMPARISON OF PRESENT AND PAST TENSE
STATEMENTS FOR TREATMENT PAIR 4 . .

COMPARISON OF PUT UPS AND PUT DOWNS FOR
TREATMENT PAIR 5 . . . . . .

COMPARISON OF PRESENT AND PAST TENSE
STATEMENTS FOR TREATMENT PAIR 5 .


Page


S. 67



. 69



S. 71



S. 73



75



S. 77



S. 79



S. 81



83






Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy

BEHAVIOR CHARTING AS AN ADJUNCT
TO THE DYADIC COUNSELING RELATIONSHIP

By

Jannar William Davis

August, 1972

Chairman: Dr. Robert O. Stripling
Major Department: Counselor Education

This study investigated the effect of behavior charting,

as an adjunct to counseling, upon changes in client behavior

during counseling. Behavior charting refers to the pro-

cedure of counting one or more specified behaviors over a

period of time and plotting their frequencies on Lindsley's

Standard Behavior Chart.

The control condition required that each counselor

engage in his customary therapeutic procedures with a ran-

domly assigned client.

Under the experimental condition, the counselor used

behavior charting as an adjunct to his usual counseling

approach with a second client. The client charted three

behaviors outside the session: positive statements about

self and others (put ups), negative statements about self

and others (put downs), and a third behavior agreed upon by


vii






counselor and client as relevant to the client's concern.

The counselor charted the client's in-session put ups and

put downs. The counselor discussed all charted behaviors

with the experimental client during each session. External,

independent raters evaluated the taped counseling sessions

of each experimental and control client for put ups, put

downs, statements in the present tense, and statements in

the past tense.

The hypotheses stated that clients whose counseling

included behavior charting would show greater change than

clients who received counseling without behavior charting

in the following behaviors:

a. self-referent statements measured by a pre-
and post-treatment q-sort;

b. a higher frequency of put ups;

c. a lower frequency of put downs;

d. a higher frequency of present tense statements;

e. a lower frequency of past tense statements.

The hypotheses, as well as the chosen client behaviors,

were based upon a Rogerian model of psychotherapy which

assumes that the client's concept of self and others becomes

more positive during counseling while his negative concept

of self and others decreases; the client also exists

increasingly in the present rather than the past.


viii






Data for statistical and graphic analyses were avail-

able from five treatment pairs, each comprising a control

and an experimental client. Clients, selected from the

population of a university college of education, par-

ticipated in six sessions of individual counseling over a

three-week period. Counselors were advanced doctoral stu-

dents in a program of counselor education. Raters were

students from the departments of psychology and counselor

education.

Statistical procedures revealed no significant differ-

ences between the experimental and control groups in the

frequencies of any recorded behavior, including self-

referent statements measured by the Dymond Q-Adjustment

Score in pre- and post-counseling administrations. However,

a study of plotted data showed individual differences in the

rate of change in all behaviors between the experimental

and control client of each treatment pair. Fifty percent

of the predicted behavior changes were in the direction

specified by hypothesis 2, a result at the level of chance.

An unpredicted conclusion of the data analysis was the de-

celeration of all observed client behaviors on 70 percent

of the occasions for behavior change.

Each client behavior frequency represented the geo-

metric mean frequency of two independent behavior counts.

ix






Interrater reliability averaged .65, while intrarater

reliability reflected a high degree of stability in the

mean correlation of .95 for eight of ten raters. The

reliability of counselor-rater evaluations of an experimental

client's put ups and put downs varied greatly, ranging from

.97 to -.59. Counselors experienced difficulty in counting

client behaviors when their interaction with the client

became intense.

Both counselors and clients found behavior charting to

be valuable in heightening the client's awareness of behav-

iors and facilitating positive behavior change. Future

research may determine the types of clients and behaviors

for which a systematic form of behavior counting, recording,

and analysis is best suited.













CHAPTER I


INTRODUCTION


Counseling, as a mode of human interaction which has

gained the status of a profession during this century, is

puzzling to its supporters and antagonists alike. A mass

of anecdotal data, as well as a gathering body of research,

supports the effectiveness of counseling in changing the

lives of some clients on some occasions. The testimony of

grateful clients and scholarly journal reports bear witness

that counseling at times meets an ultimate criterion of a

profession: it makes a significant difference in the lives

of those it serves. Just as medicine and law can literally

save a life, so can counseling.

Yet counseling, unlike the medical and legal professions,

has not clearly defined its role with respect to client out-

comes, or the means by which desired changes can be effected

in the life of a client. Counselors, for example, seem to

agree that their principal object is the facilitation of

positive client change or growth. Client "growth," however,

is a vague outcome usually defined as a varying collection

1







of qualitative attitudes or dispositions, rather than a set

of specified, measurable behavior changes. The lack of pre-

cision in defining "growth" has tended to obscure the analy-

sis of the specific behavior of counselor and client during

the counseling session, characteristically their only occa-

sion for interaction. Moreover, an implied functional re-

lationship between counselor behavior and client behavior

has not been made explicit. Precisely what happens during

counseling to promote client change, as well as why such

events modify client behavior and attitudes, is the subject

of current research in behavior change.

The forms or structure of counseling, if not the con-

tent and results, can be defined. "Counseling" is a generic

label for "a special kind of interaction" between persons,

designated counselor-and client, who communicate directly

through verbal and non-verbal exchanges. Such an interac-

tion is instrumental in producing a "change of state" in the

client (Pepinsky and Karst, 1964). At the same time, the

interaction gives "both members of the dyad reinforcing power

over the other's behavior . they have become mutually

significant people" (Blocher, 1967). Client and counselor

each modifies his behavior in relation to the other, but the

focus of change is on the client, who is expected to gen-

eralize any new behavior or attitudes beyond the counseling

interview to the "real world" of his everyday environment.







Since counseling has traditionally been offered by

middle-class therapists to middle-class and upper-class

individuals (Hollingshead and Redlich, 1958), successful

client outcome seems to rest, in part, on an initial simi-

larity of client and counselor. "By and large, counselors

using traditional 'insight oriented' approaches to treatment

do and probably should continue to select clients much

like themselves; that is, clients who are middle class,

achievement oriented, verbal and well-socialized" (Blocher,

1967).

A strong expectancy in the efficacy of counseling has

not dispelled the specter of Hebb's (1949) or Eysenck's

(1952) research. The "spontaneous remission" of psycho-

pathological symptoms in control groups of patients without

benefit of psychotherapy cast doubt that counseling was in

fact better than no treatment at all. A decade later, Bixler

echoed the earlier criticisms of counseling and its support-

ing research. That client and counselor believe in the

counseling process is not sufficient; demonstrated and

predictable changes in client behavior and attitude must

become evident.

Until these workers can offer evidence that their
techniques are more effective than placebos, we
must look upon their contribution as no more than
that. Faith and ad hominem arguments about the
'worth and dignity and integrity' of man may be







temporarily persuasive but they are very poor sub-
stitutes for evidence. (Bixler, 1963)

Unsubstantiated theory in counseling has been accom-

panied by a similar lack of definition in counseling prac-

tice. Speaking of public school counselors, Ohlsen con-

cluded that

failure to define the unique nature of their services,
the conditions under which these services can best
be provided, and their willingness to permit employees
to assign them inappropriate duties have been serious
problems for counselors. (Ohlsen, 1969)

A vagueness in counseling theory and practice led to over-

optimistic claims of success which could not be supported

in divergent areas of social concern such as race relations,

poverty, and a fuller utilization of academic talent. The

discrepancy between expectations of counseling and the

lagging efforts of the profession to establish standards

for the role and preparation of counselors became apparent

in the latter 1960s (Stripling, 1967). Neither the tra-

ditional practice nor research of counseling has substanti-

ated the claimed impact of that "special kind of interac-

tion" between client and counselor.


Purpose of the Study

This experimental study was planned to investigate the

effect of behavior charting, as an adjunct to counseling,

upon changes in client behavior. Behavior charting is the







procedure of counting a specified recurrent behavior over

a given period of time, and plotting the frequency of that

behavior on a graph. The Standard Behavior Chart (SBC)

developed by Lindsley (1964) was used to record chosen

client behavior frequencies.

This research was designed to compare clients who par-

ticipated in behavior charting during counseling to clients

who received counseling without behavior charting. Cumula-

tive client change over the treatment period was indicated

by a pre- and post-treatment administration of the Q-Adjust-

ment Score (AS) (Dymond, 1954). The AS requires a q-sorting

by the client of a set of self-referent statements. Inde-

pendent raters counted the frequency of the following verbal

behaviors which occurred during each counseling session:

(a) positive client statements about self and others, (b)

negative client statements about self and others, (c) client

verbal expressions in the present tense, and (d) client

verbal statements in the past tense. These verbaliza-

tions represent specific observable behaviors based on a

Rogerian model of psychotherapy (Rogers, 1961). Coun-

seling is interpreted to bring about the following at-

titudinal changes inferred from the behavioral pinpoints

above: (a) the client's concept of self and concept of

others becomes more positive while (b) the client's







negative concept of self and others decreases and (c) the

client exists increasingly in the here-and-now rather than

the past.

The following hypotheses were tested:

1. Clients whose counseling includes behavior charting

will show greater change in self-referent statements than

clients receiving counseling without behavior charting, as

measured by the AS.

2. Clients whose counseling includes behavior charting

will show greater behavior change during counseling than

clients receiving counseling without behavior charting in the

following specific behaviors:

a. a higher frequency of positive statements about
self and others

b. a lower frequency of negative statements about
self and others

c. a higher frequency of verbal expressions in the
present tense

d. a lower frequency of verbal expressions in the
past tense.


Rationale of the Study

Any client "growth" or changes in attitudes which occur

during the process of counseling will be evident as overt be-

havior changes in the client. These overt changes include the

client's verbal behavior about himself and others in his







environment. "When we speak of a self-concept, we may

simply be talking about the aggregate of sentences [one]

says to himself (and others) about himself" (Homme, de Baca,

Cottingham and Homme, 1968).

The behavior changes will be observable notonly to

the counselor and client, but to the raters who evaluate the

client over the period of counseling. External raters will

be able to indicate changes in the four specified client

behaviors objectively in terms of their accelerating or

decelerating frequency of occurrence.

An alteration of the environment produces behavior

change (Skinner, 1953: 1971). Counseling, in the present

study, represents an attempt to modify the antecedent and

consequent events which are functionally related to the

client's behavior. Counseling becomes a means of evaluating

the effect of a client's environment so that alterations by

counselor and client may lead to and support changes in the

behavior of the client.

If counseling produces behavior change, the changes

will become evident through the recording of specified client

responses over the period of counseling. Counseling with

and without behavior charting, as two differing environmental

contingencies of client behavior, should produce observable

differential changes in behaviors. Such changes can be




8

detected by a continuous record of client behavior manifested

during each counseling session. Whether or not differences

in behavior are present, plotting the response frequencies

of designated client behaviors over all sessions will reveal

a pattern of counselor-client interaction during counseling.














CHAPTER II


REVIEW OF THE LITERATURE


Introduction

A review of research reveals an abundance of studies

concerning the role and function of counseling. Especially

prominent is status-type research achieved by survey or

questionnaire. Litwack (1968) highlights the need for

experimental research directed at the process of counseling

to supplement post hoc assessment of outcomes. "There is

S. a paucity of research that investigates methodology

and success of either counseling or counselor education.

The human relationship in counseling is still in the early

stages of investigative effort." The scarcity of definitive

experimental research stems from two sources. First, coun-

seling traditionally has looked to global inferences of

attitude and feeling such as "self-actualization" or

"improved self-concept." Moreover, counseling research may

have attempted an inappropriate comparison between client

changes promoted by counseling and changes brought about by

the passage of time, without counseling.







Krumboltz (1967) states that the counseling profession

has not been served by an assumption that "some generality

known as counseling is generally effective for general

clients." In fact, research has not unequivocally demon-

strated that counseling is better than no treatment at all.

Part of the problem, according to Rice (1965), "arises from

treating phychotherapy as a homogeneous 'treatment' variable

in outcome designs, an assumption that ignores the tremendous

variation which we know to exist even within a single the-

oretical orientation."

Characteristic of research which attempted to demon-

strate a wide range of beneficial client outcomes from

"counseling," as compared to no therapeutic treatment,

are studies by Cartwright and Vogel (1960), and Barrows

(1971). Cartwright and Vogel compared changes of 30

neurotic clients on the Minnesota Multiphasic Person-

ality Inventory, the Thematic Apperception Test,

and the Q-Adjustment Score during matched periods of

therapy and no therapy. The measures showed conflicting

results. Predicted outcomes for the therapy condition,

including greater client change with longer time in therapy,

and greater improvement with more experienced therapists,

were only partially supported. Less client improvement over

longer waiting periods without therapy could not be




11

confirmed; the authors indicate that "spontaneous remission"

may have accounted for part of the unexpected client im-

provement with time.

Barrows (1971) investigated the effect of audiotape

playback in secondary school counseling. Assuming the

homogeneity of "secondary school counseling," he predicted

that counseling with or without audiotape playback would

produce greater improvement in counseled students' classroom

behavior than would no counseling. The results were not

significant. About 55 percent of the counseled students

improved their classroom behavior, as judged by their

teachers, compared with 33 percent of the control group

of students who received no counseling. Barrows seems to

have confounded the outcome by asking teachers to indicate

which students from the total sample showed improvement.

Although the teachers did not know which students of the

sample had been counseled, they know that all subjects were

involved in an experiment. Heightened teacher sensitivity

to any behavior improvement, judged in retrospect, cannot

be discounted as a source of error variation.


The Experimental Analysis of Behavior

A recent shift in research has been from an "among-

treatments" design, comparing counseling as a unitary






variable to the effects of time, to a "within-treatment"

design, comparing different methods and techniques of coun-

seling. The change in focus follows from a renewed cog-

nizance of the diversity of client problems, approached

through individually tailored treatments.

The experimental analysis of behavior has direct impli-

cations for the practice and theory of counseling, since it

provides a systematic means of an experimental manipulation

and review of counselor and client process variables. Rice

foreshadowed the necessity for a synthesis of scientific

method and the study of human behavior found in the current

approach known as the experimental analysis of behavior.

"The most fruitful approach to a field in which there is

uncertainty even about the variable to be measured or con-

trolled is that of a systematic, quantifiable, naturalistic

observation" (Butler, Rice, and Wagstaff, 1963).

Skinner (1966) has described several characteristics

of an "experimental analysis of behavior." Basic to the

experimental approach is a record of the frequency of re-

peated instances of an operant, a class of behaviors which

have a particular observable effect on the environment. The

operant, as the behavior of interest, is the dependent

variable which varies with the manipulation of independent

antecedent and subsequent events in the organism's environ-






ment. Accordingly, the task of the investigator is to

discover all the environmental variables which affect the

operant. The experimental control of these variables, once

discerned through a recording of the behavior of interest

and its accompanying events, is preferred to a post hoc

statistical evaluation. While the number of organisms

studied is usually smaller in this longitudinal approach

than in statistical cross-section approaches, the length of

study is usually longer.

An applied study of behavior assumes that observable

responses manifest prior learning experiences. The empiri-

cal-inductive approach thus proceeds from the empirical

evidence of maladaptive behavior to a labelling of these

responses, rather than proceeding from an a priori theory of

behavior to a diagnosis of the client (McDaniel, 1966).

The behavior of concern is recorded systematically accord-

ing to its frequency of occurrence in the context of ante-

cedent and subsequent environmental events. Modification

or relearning of behavior can then emerge through environ-

mental changes based on the charted responses (Bijou, 1966).

Counseling and the experimental analysis of behavior

share a common goal to facilitate the realization of the

client's individual potential. Whether conceptualized as

an increased awareness and acceptance of organismic experi-






ence (Rogers, 1961) or as increased skill in the manipula-

tion of one's own environment which controls behavior

(Skinner, 1953), such "self-control" is a legitimate end of

therapy. An ordered analysis of behavior can provide a

powerful tool for facilitating the client's acquisition of

self-control.

"Outreach" counseling as a helping relationship is

supported by the contemporary involvement of the experimental

analysis of behavior. Outreach counseling takes place in

the client's environment whenever possible--in the home,

special education class, prison, or mental institution--

and actively seeks to modify that environment for the

benefit of the client. A current index to journals in

education reveals that behavioral counseling techniques in

outreach counseling include relearning in public speaking,

stuttering, interracial behaviors, soiling by an elementary

school student, and risk-taking behavior. The literature

offers profuse examples of outreach relationships which occur

with the client in his own environment; involves parents,

teachers, and peers significant to the client in that

environment; and utilizes the counselor-client interaction

to examine current behavior and feelings as a basis for

desired client change.

A comprehensive experimental environment for learning




15

was created for a group of 16 student-inmates of the National

Training School for Boys during 1965 (Cohen, Filipczak and

Bis, 1967). The authors designed a learning environment

which included potential behavior reinforcers such as a gen-

eral classroom, a self-study area, a conference room, and a

student recreation lounge with a store. A token economy of

immediate extrinsic rewards (points exchanged for money) was

implemented to achieve measurable improvement in educational

and social behaviors. Academic progress in reading, language

skills, and mathematics resulted in praise and attention

from the staff, leading to acceptable social behavior more'

like that of public school students. Programed texts and

individualized instruction allowed each student to work at

his own pace. After he reached 90 percent accuracy on each

unit of study, the student began the next unit. The boys

were paid daily for the achievement points earned. Payment

allowed the students to enter the recreation lounge where

money could be spent for food or other desired items from

the store. The success of the experimental learning en-

vironment is confirmed by a group Scholastic Aptitude Test

score gain of one grade level in reading, arithmetic, and

spelling, and a half-year gain in language skills over a

nine-month period for student-inmates who had been termed

failures in public school programs.




16

Phillips (1968) used token reinforcement procedures to

shape a variety of social, self-care, and academic behaviors

in a home-style treatment program for "pre-delinquent" boys.

The three boys involved were rewarded with points for

appropriate academic behaviors such as achieving 75 percent

accuracy on assignments in arithmetic and reading. "Fines,"

or loss of awarded points, were levied for inappropriate

social behaviors including tardiness, aggressive statements

to peers, and incorrect grammar. Points could be exchanged

weekly for an allowance, use of a bicycle, television

viewing, and permission to stay up past bedtime. Phillips

reported the treatment program to be practical, economical,

and effective in the home-type environment during the

three-month experimental period.

The studies by Cohen et al. and Phillips employed

tangible, extrinsic rewards to provide environmental con-

tingencies for behavior previously unaffected by largely

unavailable natural consequences such as teacher praise or

good grades. The behavior of the student-inmates became

more nearly "normal" as it was shaped to respond to the

natural contingencies usually overlooked in describing a

person as "self-disciplined" or "motivated." A

deliberate imposition of artificial consequences can play an

important part in transforming behavior from "maladjusted"




17

to "adaptive" through controlled learning environments such

as those described above.

Schaefer and Martin (1966) focused on the pervasive

problem of "apathy" among patients in mental insti-

tutions. Assuming that the institutional environment re-

inforced "apathy," the researchers modified the environment

to support alternate behaviors which could not be termed

apathetic. From a group of 40 females who showed varying

degrees of apathy, as noted in their medical records, 20

were randomly assigned to the treatment condition of a token

economy, while the remainder served as a control group who

received no treatment beyond their usual interaction with

the hospital staff. Baseline data were collected on all

subjects who were assigned daily "apathy scores" based on

the severity of their observed isolation and the limitations

of their behavior. Experimental subjects were reinforced

over a three-month period for personal hygiene (e.g.,

showering, hair combing, use of cosmetics), social interaction

(e.g., speaking up in group therapy sessions, playing cards

with other patients), and adequate performance of assigned

housekeeping chores. The rationale behind all reinforced be-

haviors was that the patient who emits such behaviors

"cared" about people and objects in his surroundings and

could not be called apathetic or uncaring. The experimental







group, although equivalent to the control group in mean

initial apathy score at the end of the first month, showed

significant within-group and between-group differences in

behavior at the end of the second and third months of treat-

ment. They showed a steady and significant decrement in

apathy scores compared to their own initial mean ard to the

means of the control group for corresponding time periods.

The main value of the study lies in a comparison between

conventional hospital treatment, which had not changed

"apathy" in the patients during the experimental period,

and the deliberate application of reinforcers which changed

behavior significantly in three months (p < .005 between

overall experimental and control group means).

A 20-month study of 60 institutionalized chronic

schizophrenics conducted by Atthowe and Krasner (1968)

lends additional credence to the usefulness of a systematic

contingency program. Tokens, established as positive

reinforcers during a three-month shaping period, promoted

significant gains in attendance at group activities, re-

quests and use of passes to leave the ward, and discharge

rate during the following 11 months. A six-month baseline

period gave the researchers initial data on patient be-

haviors and potential reinforcing events. The authors

found the token ecomony to be effective as a controlled







environment designed specifically for patient gains in

"normal" behavior, contrasted to the usual institutional

environment which served to maintain or intensify mal-

adaptive behavior. The contingent reinforcement program

represented by a token economy was successful in "combating

institutionalism, increasing initiative, responsibility,

and social interaction, and in putting control of patient

behavior in the hands of the patient."

Although outreach counseling emphasizing behavior

change has been chiefly concerned with segments of the popu-

lation regarded by society as "abnormal," recent research

concerns "normal," yet distressing client problems such as

stuttering, overeating, or marital difficulties.

Webster (1970) devised a behavioral approach to the

elimination of stuttering. Clients were eight severe stut-

terers, ages 15 through 47, who had previously experienced

little success in speech therapy. Each client advanced

through the fluency shaping program at a rate based on his

performance during each phase of the program. Following a

three-day baseline period when observers counted the fre-

quency of speech blocks during reading sessions, each client

learned to detect accurately his own stuttering responses as

he read aloud. This "self-definition" of speech blocking

involved the client actively in the fluency shaping process







and provided direct feedback as a foundation for ultimate

self-control of stuttering. Continuous playback of client

speech through a delayed auditory feedback technique prompted

discrimination of speech blocks and helped establish self-

maintained fluent, yet slow speech. The speech rate in-

creased with laboratory conversation practice. Clients

generated large amounts of spontaneous speech as they

described magazine pictures during sessions of gradually

increasing length. Pictures were then faded out, leaving

smooth conversation without visual prompting. The effec-

tiveness of the brief (10-40 hours) systematic fluency pro-

gram is indicated by a followup study 10 months later. Only

one client reported more than five blocks per day; after

several hours of retraining he reported no further stuttering.

A program reported by Stuart (1967) applied the experi-

mental analysis of behavior to the problem of overeating.

The objective of the program, like Webster's program to

eliminate stuttering, was the client's ultimate control of

his own behavior. Thirty-minute treatment sessions were

scheduled three times weekly over a four- to five-week

period for eight female clients judged obese by their

physicians. During these sessions Stuart instituted a number

of environment changes to aid client discrimination and

control of overeating. Each client kept a food data sheet







to record the quantity and circumstances of food intake,

and a weight range sheet which showed the effect of food in-

take on body weight measured four times daily. An awareness

of the circumstances and consequences of overeating was

heightened further by the client's manipulation of his eat-

ing environment. The client interrupted his meals for sev-

eral minutes to reinforce his ability to control his own rate

of eating. He also made eating a "pure" experience without

coupling it with such stimuli to overeating as watching tele-

vision or reading in bed. Only foods which required prepara-

tion were used, one portion at a time. All such environ-

mental changes were designed to make the client aware of,

and thereby able to control, the various aspects of his

eating behavior. A followup 12 months after the beginning

of the program showed that the group of clients had lost an

average of a pound per week during that year, although fluc-

tuations in weight were common. Every client had lost at

least 30 pounds during the program of weight reduction.

Client interviews at the end of the year revealed no "symptom

substitution" for overeating. In fact, three clients who

were heavy smokers reported a substantial reduction in

smoking through application of the principles of their

eating control program.

Goldiamond (1965) summarized the consultative role of

the counselor implicit in the experimental analysis of







behavior. The professional helper acts as a consultant to

the client for a problem which can be expressed as a collec-

tion of behaviors that the client will be instrumental in

modifying. Goldiamond, as consultant, cited cases in which

he had instructed the client to arrange certain environmental

changes; in other cases, he had trained the client in a

functional analysis of behavior so that the client could

determine for himself which changes to arrange. Weekly ses-

sions afforded a time for the client to report the effect

on his behavior of attempts to change the environment, and

to formulate further changes.

A marital case illustrates the consultative relationship.

The behaviors of concern were the client's alternate shouting

and sulking for hours on end in the presence of his wife, who

had been sexually unfaithful two years earlier. Goldiamond

suggested procedures for stimulus change combined with new

contingencies for shouting and sulking. The husband was to

take his wife out for dinner once a week, and during that

occasion he was to speak with the normal courtesy and con-

sideration as he would give to any dinner partner. The

husband could also sulk as long as he wished, but only

while seated on his "sulking stool" in the garage. The

client kept daily records of his sulking and civilized

conversation with his wife. At the end of 10 weeks, "there






was no sulking in the garage and the partners were able to

commune."


Precise Behavior Management

The practical application of the experimental analysis

of behavior to a modification of the behaver's environment

is known as precise behavior management. Lindsley (Duncan,

1971a) contrasted precise behavior management to behavior

modification. Although both are derived from techniques

of operant conditioning, precise behavior management

emphasizes natural consequences of behavior, such as the

teacher's praise or censure, rather than artificial ex-

trinsic reinforcers such as tokens or candy. Extrinsic

reinforcers or punishers may be used to promote behaviors

of sufficient frequency and consistency to come under

the control of consequences already a part of the en-

vironment. But such artificial events are acknowledged as

behavior prosthetics, supportive tools to be replaced by

events naturally occurring in the behaver's surroundings.

The emphasis on ultimate natural consequences is balanced

by an interest in natural antecedents of behavior. While

behavior modification seems to strive for experimental

control of the circumstances under which given behavior

occurs, precise behavior management looks for natural






events which cue certain behaviors. These antecedents,

once determined as causal events through observation, can

then be manipulated in the ongoing situation.

Pennypacker (1972) amplified the differences between

precise behavior management and operant techniques. First,

the focus of precise behavior management is human, rather

than animal behavior. Operant conditioning of animals

served as a means to establish specific, consistent responses

which could be observed under different treatment conditions.

For example, operant conditioning techniques can institute

bar-pressing behavior in rats. That behavior is observed

to vary under different environmental circumstances, such

as presentation of food or administration of shock, con-

tingent on bar-pressing. Precise behavior management also

employs operant conditioning to establish new behaviors in

humans, but the purpose is not mere objectivity in record-

ing specific responses. The acquisition of new or modified

behavior is a learning experience designed to supplant

maladaptive behavior. The change in behavior may be a

relatively discrete observable event, such as talking with

peers of the opposite sex, or it may be a collection of

behaviors from which an attitude or tendency such as

"apathy" or "aggression" is inferred.







Operant conditioning of laboratory animals, besides

specifying uniform recordable responses, attempts a strict

control of the experimental environment through physical

isolation so that a single independent variable, e.g.,

presentation of shock, can be manipulated. All other con-

current environmental stimuli are regarded as potential

sources of error variance. Precise behavior management,

however, utilizes the person's natural environment, where

the behavior is taking place, with the sum of all interact-

ing environmental events impinging on behavior. Again,

strict experimental control of the environment is subjugated

to the functioning of the human involved. The objective

of an experimental approach to the study of behavior is not

the imposition of a mechanistic, impersonal technology but

the implementation of planned behavior changes tailored to

the individual.

Alper and White (1971) discussed pertinent charac-

teristics of precise behavior management. First, it

employs a common language system based on a precise de-

scription of behaviors and events which are observable and

countable. The connotative confusion of such inferred

labels as "aggression," "hyperactivity," or "lack of

motivation" is reduced by pinpointing the group of be-

haviors which have fostered the labels. Moreover, the




26

specific behaviors observed are not "healthy" or "unhealthy,"

but are qualified as adaptive or maladaptive to the parti-

cular situation in which they occur. Precise behavior man-

agement furthers an understanding of the behaver through ac-

tual observation and precise specification of the behavior

and surroundings. The behaver is considered an active parti-

cipant in the process of environmental control as he "learns

how to learn." The behaver and manager can both observe

behavior, and the behaver can record the frequency of chosen

responses as various attempts are made to alter the environ-

ment.

Bradfield (1970) delineated the components of a system

of precise behavior management. The first is a precise pin-

pointing of behavior as a "movement cycle." The movement

cycle is a unit of behavior with an accurately observable be-

ginning and end. Once completed, the cyclic behavior may be

repeated or replaced by a different behavior. For example,

the movement cycle for a child's out-of-seat behavior may be

defined as getting up from the desk,walking about the room,

and returning to a sitting position at the desk. The child

is then free to arise and return again, or to perform other

behavior such as talking with a classmate while seated.

The behavior, once described as a movement cycle, is

observed and recorded as a frequency, derived from a count of






the cycles over a given period of time. Behavior charting,

the second component, is the "feedback" loop of the system.

It provides the behaver and contingency manager with knowl-

edge of the relationship of behavior and its circumstances.

The final component of the precise behavior management

system is the Behavior Bank, an international computer bank

which stores successful behavior management projects. By

mid-1971, over 11,000 projects from the fields of education,

psychology, social work, medicine,and nursing had been

deposited. About 60 percent of these projects concerned

applications to special education (Koenig, 1971).

Johnston and Pennypacker (1971) have used "precision

teaching," precise behavior management in the context of

the classroom, with psychology classes at the University

of Florida. Their methods challenge a number of tradi-

tional concepts in education. First is the notion that

only a fraction of the students can achieve a high grade

through intraclass competition. Every student who

met clearly specified performance criteria achieved the grade

of A. Over 90 percent of the students in the class, whose

enrollment averaged 75 students per quarter for the two

years reported, reached the criteria designated as "A"

performance. That is, most students could discuss the con-

tent of the course with the fluency and detailed knowledge






of an expert. The pervasive concept of education which

separates the teaching and learning functions between

professor and student was also disputed. "Contingency

managers," students who had earned an "A" in the course

previously, each shared a one-to-one relationship with five

or six class members in the "performance session." The

contingency manager met with his protege (class member)

twice or three times weekly, for less than half an hour

each time, to allow the protege to perform verbally on text

and lecture material. The student discussed any errors with

his contingency manager and plotted his rate of progress on

a Standard Behavior.Chart. Student, contingency manager,

and instructor could quickly determine the student's past

performance, project his future performance under similar

circumstances, or compare his performance with that of any

other class member. The continuous direct record of relevant

academic behavior reflected the sequential evaluation of the

student in relation to the announced standards of achieve-

ment as criterion rates of correct, incorrect, and irrelevant

answers. The instructor was free to present lectures with-

out the burden of evaluating 75 different students whose

behavior might otherwise have been sampled through one or

two hour-long tests during the quarter. Neither did he

need to speak on all relevant textual material. The nature







of the contingency manager-protege schedule required the

student to read and show verbal proficiency over short

segments of the text before performance on the next seg-

ment was attempted. Johnston and Pennypacker, in summary,

have implemented a learning model based on behavioral

principles which support an active involvement of the

learner in achieving specified criteria through continuous

performance and evaluation of academic behavior.

Precision teaching has achieved success in classes for

children disabled by emotional and physical handicaps.

Roberts (1971) has recounted her experiences with a class

of trainable mentally retarded children. She constructed

a program of flash cards containing functional stimuli such

as days of the week, traffic signals, and units of money.

Reinforcement for each student was contingent upon a gradual

increase in his rate of correct responses.

Duncan (1971b) applied precision teaching to thoughts

or "inner behavior." A three-and-one-half-year-old girl

who counted and charted her selfish thoughts as well as

selfish actions during kindergarten substantially reduced

the frequency of selfish actions while maintaining the level

of selfish thoughts. A 12-year-old girl counted and charted

angry feelings and angry outbursts toward her younger

brother. Counting accelerated her thoughts and outbursts.




30

She then selected potentially reinforcing contingencies for

a decrease in angry behavior. For 15 or less outbursts

daily, she arranged to receive a dime; for 10 or less, 20

cents. Outbursts decelerated during these contingencies and

continued to decrease when she returned to counting only.

Similarly, angry thoughts dropped when she told each one to

her teacher, and remained low when she resumed counting only.


The Counseling Relationship

The focus of this writer's research is an application

of the rationale and methodology of precise behavior manage-

ment to the counseling relationship, conceived as a learning

environment. This section reviews the literature concerning

pertinent aspects of this application to counseling..

Several researchers have examined the nature of the

successful counseling relationship through a study of coun-

selor and client characteristics. Blocher (1967), drawing

on past studies, has noted the similarity between "success-

ful" clients and their counselors. Both, compared to

samples of "adults in general," tended to display greater

anxiety and self-dissatisfaction (Lorr, Katz, and Rubenstein,

1958), have greater needs for achievement, and show a

greater willingness to communicate problems and feelings to







others (Hiler, 1958). They also tended to belong to the

middle and upper social classes (Bailey, Warshaw, and Eichler,

1959), to have more formal education (Hiler, 1958), and

higher measured intelligence (Stieper and Wiener, 1965).

"Generally, the good prospect for counseling . is

described as not very disturbed, well motivated, well

educated and having good personal resources" (Stieper and

Wiener, 1965).

Pepinsky and Karst (1964) have proposed a process of

"convergence" to explain an increasing similarity between a

therapist and clients termed "successful" by the therapist.

Convergence implies that the client will increasingly modify

his behavior toward his counselor's behavior as witnessed

during the interview. "The key to the [convergence] process

is the ability of the therapist to influence the client

either to change his beliefs or to modify his behavior

S. ." through the counselor's use of a "psychological

grammar." This grammar comprises "categories of belief

and action" presented by the counselor to the client, as

well as "rules of belief and action" which the client

learns to use during the counseling relationship. "Success-

ful" client change, largely determined by the counselor,

seems to depend heavily on the client's ability to model the

counselor's verbal behavior and to infer standards of be-

havior acceptable beyond the session.




32

Reinforcement is a vital aspect in the convergence of

counselor and client. The experimental analysis of behavior

incorporates the use of reinforcing events to promote be-

havior change. "Insight" therapy also seems to utilize

principles of reinforcement in establishing necessary con-

ditions for therapy. Truax (1966) analyzed a taped session

of Carl Rogers and client, and found evidence of selective

counselor reinforcement of certain client behaviors. Meas-

ured on scales constructed by Truax, the therapist was

observed to be more empathic, warm, and accepting, and less

directive when the client expressed himself in a style

similar to that of the therapist. The following classes of

client behavior were positively reinforced by attention and

verbal response from the therapist: (a) learning dis-

criminations about self and feelings, (b) expressions of

insight, (c) verbal expressions similar in style to

therapist's expressions, and (d) awareness of a problem.

Client discrimination, expressions of insight, and aware-

ness of a problem were inferred from the client's verbal

responses, while similarity of expression was observed

directly.

Allen (1967) concurred with Truax's observation of

therapeutic reinforcement:







It would appear that much more might be accomplished
if it were clearly acknowledged that the success
of the client-centered counselor turns on the fact
that most human beings are more likely to repeat
behaviors that are rather regularly succeeded by
sensitive expressions of warmth and understanding
by another positively regarded person than those
that are not.

Truax and Carkhuff (1965) experimentally manipulated

counselor levels of empathy and positive regard during a

one-hour interview in which three therapists each met a

different schizophrenic patient. The therapist attempted to

provide high levels of empathy and regard during a 20-minute

baseline period, to lower the conditions during the middle

period, and to renew the higher levels during the last 20

minutes. Client levels of empathy, regard, and depth of

intrapersonal exploration were observed to vary directly

with the therapist's manipulations of empathy and positive

regard. Truax and Carkhuff concluded that the findings

suggested "a causal relationship between the level of some

therapist-offered conditions and some of the patients'

therapy behavior."

The reinforcing properties of the therapist were fur-

ther discriminated in a study by Holder, Carkhuff and Berenson

(1967). A group of clients were assessed on scales for

empathy, respect, genuineness, concreteness, and depth of

self-exploration. The three highest and three lowest ranked







clients were interviewed by a counselor who manipulated

the conditions of empathy, warmth, and genuineness during

the interview. Twenty-minute segments of high, low, and

high counselor conditions were found to affect directly the

levels of empathy, genuineness, and warmth of the lower

functioning clients. The higher functioning clients did

not vary significantly from their level of functioning at

the start of the interview. This finding qualifies the

earlier study of Truax and Carkhuff (1965). Client level

of functioning, as well as that of the counselor, is an

important variable in dyadic reinforcement.

Just as the counselor can selectively reinforce client

behaviors under different levels of empathy, genuineness,

and warmth, the client can differentially reinforce coun-

selor behaviors. Carkhuff and Alexik (1967) found that

counselors who rated high in empathy, warmth, and genuine-

ness maintained their initially high level of function even

though coached clients manipulated their level of self-

exploration. Low-level counselors were dependent on the

client level of self-exploration, and lowered their "core -

conditions" of empathy, warmth, and genuineness during the

middle third of the interview.







Feedback

The counseling relationship, whether it fosters new

insights leading to behavior change or, conversely, the

acquisition of new behaviors which facilitate a change of

feelings, may be viewed as a learning situation. "Inner"

and "outer" behaviors are modified in successful counseling.

Feedback, or knowledge of results, forms a significant link

between initial overt or covert learner response to stimuli

and a response modification through perception of the ef-

fects of the initial response. Azrin, Rubin, O'Brien et al.

(1968) used a loud tone as immediate corrective feedback

for slouching behavior, defined individually for each of 25

adult subjects. The researchers designed a lightweight

cloth harness with a micro-switch and buzzer which sounded

each time the subject rounded his shoulders beyond the

critical point of adjustment for his harness. The results

showed the feedback procedure effective in reducing slouch-

ing for all subjects during every period in which the

harness was worn. The reduction in slouching averaged 86

percent during activation of the feedback mechanism. Immedi-

ate sensory feedback required the active participation of

the subject to change his posture, rather than passive

reliance on a rigid brace. In addition, the feedback

apparatus encouraged the continuous use of relevant muscles

rather than their sporadic use during an exercise period.






Greenspoon and Forman (1956) explored the effect of

visual feedback on learning a motor task. Five groups of

eight undergraduate students were given several trials to

draw a three-inch line while blindfolded. Four groups

received feedback by seeing the results of each trial after

a delay of 0, 10, 20, or 30 seconds. The fifth group

received no feedback between trials. Differences in

accuracy between feedback groupswere significant, with in-

creasingly accurate lines drawn after decreasing delays of

feedback. The group receiving feedback 30 seconds after

each trial was found superior in accuracy to the group which

got no feedback at all.

Chapanis (1965) designed an experiment to separate the

effects of a varying amount, rather than a delay of feed-

back for a repetitious task. Sixteen undergraduate college

students, divided evenly into four treatment groups, punched

digits on a teletype tape for an hour daily for 24 days.

Treatment conditions were no feedback, when the digit

counter was not connected; minimal feedback through a cumula-

tive count, when the digit counter was never reset to 0;

feedback through actual count, when the digit counter was

set to 0 for each hour session; and maximum feedback with

actual digit count and subject's log entries of the

count during every 15 minutes of each session. Chapanis







predicted that knowledge of digits punched would be a

reinforcer for the subject, and that the greater the knowl-

edge of results, the higher the subject's performance in

punching digits. The prediction of performance differences

between groups was not supported. However, Chapanis' con-

clusion that feedback may not improve performance cannot be

generalized from his study since he did not test the arranged

event of digit count knowledge before the experimental

treatment to determine if it was in fact a positive rein-

forcer or accelerator of digit punching. The outcome of

his experiment confirmed that knowledge of digits punched

was not a reinforcer producing measurable differences in

performance. The chief reinforcer for the monotonous

task was likely the termination of the hourly session. An

interval schedule dependent on the passing of time was

prepotent over a ratio schedule dependent on response

frequency.

Baller and Lower (1971) noted that professionals in

the role of learning facilitator often fail to give precise,

immediate feedback despite an impressive body of evidence

supporting knowledge of results as a crucial element in

learning. Counselors, as well as teachers, seem to have

neglected the value of feedback. Baller and Lower posit

feedback as a "basic human need" based on their observation




38

of man's goal-oriented behavior "motivated by the expectancy

of obtaining a particular reward when he engages in certain

behaviors." Johnston and Pennypacker (1971), citing the

role of feedback in precision teaching, stated that

typically, the assessment process is viewed as
divorced from the teaching process. This seems to
sacrifice the opportunity for an immediate, con-
tinuous measure of the process of teaching as well
as to ignore the instructional possibilities in-
herent in assessment situations . .

Lang (1966) has summarized the rationale underlying the use

of feedback in a learning environment, particularly counsel-

ing. "Treatment is facilitated if clients have knowledge

of results and are aware of the principles involved in the

therapeutic task."


Description of the Instruments

The Standard Behavior Chart (SBC), developed by Lindsley

(1964), has been described by Stephanie Bates, a kindergarten

student who had learned to plot accurately some of her own

behaviors (Bates and Bates, 1971). A paraphrase of her

description follows:

The frequency of an observed behavior is plotted at the

intersection of a vertical line, which shows the day of the

week, and a horizontal line, which shows how often the be-

havior occurs. Some frequency points fall between the

printed frequency lines on a day line. The chart




39

accommodates 140 consecutive calendar days marked according

to simple plotting conventions. Days on which the

behavior occurred and was counted are "rated days,"

marked with a point at.the appropriate frequency. "Ig-

nored days," during which the behavior happened but was

not counted, are marked by a line connecting the rated days

on either side. Days on which the behavior did not have a

chance to occur are "no chance" days, and receive no mark

on the chart.

When Stephanie counted her smiles to classmates,

she placed a point on the chart for a rated day. When

she did not count her smiles at kindergarten, she waited

until another rated day had been plotted, and drew a

line between the last rated day and the present one. On

weekends, when she did not attend kindergarten, she did

not mark Saturday or Sunday at all.

The SBC utilizes frequency, or rate, as the measure of

behavior. Johnston and Pennypacker (1971) have observed

that all behavior occurs in some quantity over some period

of time. The expression of these two characteristics in

a single statistic, frequency, is "the most useful and

sensitive unit by which to measure many kinds of behavior."

The SBC thus avoids the limitations of time-sampling

techniques. Samples of behavior taken at discrete intervals,






no matter how closely spaced, place an artificial ceiling

on a behavior occurring more frequently than the selected

interval. For example, a recording at two-second intervals

of facial tics occurring at a rate of one per second will

show a restricted frequency of 30 instead of 60 tics per

minute. The frequency of behavior happening less often

than the time sample interval will vary directly with the

length of the interval. That is, tics of one-second

duration occurring five times per minute have an average

frequency of .17 (5 occurrences/30 chances for observation)

per minute on a two-second observation schedule, and at an

average frequency of .42 (5 occurrences/12 chances for

observation) per minute at a five-second recording interval.

Lindsley has reported that the SBC increased communi-

cation at least 10 times among a group of teachers sharing

behavior data (Duncan, 1971a). When teachers made their

own charts, about 30 minutes were required for them to

make their data clear to the others. Standardizing data

and terminology through use of the SBC provided an ade-

quate range for recording behavior and reduced communica-

tion time to about three minutes. The six-cycle semi-

logarithmic design has separate scales or cycles for

behavior occurring from a thousand times a minute to once

every thousand minutes. This feature insures that the







types of behavior measured are not biased by the nature

of the chart (Duncan, 1971a).

The diversity of charted behaviors, as well as their

temporal range, has increased during the last half-decade.

Although behavior charting was developed to record external

behavior measurable by any number of observers, "inner"

behaviors have also been recorded by the behaver. Success

thoughts, anxiety feelings, joy, love, and compassion have

been charted. Lindsley believes that charting may be "one

of the few sensitive techniques that we have to keep track

of these inner thoughts, feelings, and urges" (Duncan,

1971a). Home (1965) has referred to thoughts and feelings

as "coverants," or covert operants which manipulate en-

vironment contingencies but are observable to others only

through the verbal report of the behaver.

The SBC, in sum, provides a systematic and compre-

hensive method to record specified behaviors of diverse

frequency and complexity with ease. Most importantly, the

learner or behaver receives direct feedback on the effect

of his behavior on the environment, as well as the impact

of the environment on his behavior.

The Q-Adjustment Score (AS) was derived by Dymond (1954)

from the Q-sort test of self-concept constructed by Butler and

Haigh (1954). To obtain the adjustment criterion, two well-




42

trained practicing clinical psychologists divided the Butler-

Haigh set of 100 self-referent statements into two categories,

"unlike a well-adjusted person" and "like a well-adjusted

person." The distributions of the two psychologists differed

on only two of the 100 items. Both judges agreed that 26

items were irrelevant. The resulting set of 74 items was

sorted by four judges into two equal piles, "like me" and

"unlike me" for the well-adjusted person. The composite

self-description of the well-adjusted person included 37

"positive adjustment indicators" in the "like me" category,

and 37 "negative adjustment indicators" in the "unlike me"

category. A person's resemblance to this ideal adjust-

ment type is computed easily by counting the number of the

74 items the subject places in the same category as that

assigned to the hypothetical well-adjusted person.

The reliability of the AS was estimated by the test-

retest method for a control group (n = 23) not exposed to

therapy. Reliability over the two-month period was .86.

Test validity was determined through a correlation of AS

scores and the Butler-Haigh Q-sort of self-concept.

The rank-order correlation between the instruments for the

23 control subjects was .83, while the rank-order correla-

tion after a two-month period was .92. The AS thus evi-

dences satisfactory validity in its high agreement with

the Butler-Haigh Q-sort of self-concept.














CHAPTER III


METHOD


Design

A multiple subject, dual treatment design was used

to generate data concerning client behaviors during coun-

seling. Fourteen clients (nine female, five male) were

randomly assigned to seven counselors (four female, three

male) for six 50-minute counseling sessions distributed over

a three-week period. Each counselor was randomly assigned

t'..'o clients who were then assigned at random to the ex-

perimental or the control condition. Thus, each counselor

served as his own control by interacting with one client

under each treatment condition.

The control condition stipulated that the counselor

employ his customary therapeutic procedures to facilitate

client change in attitude and/or behavior. The experimental

condition added the treatment variable of behavior charting,

and stipulated that the counselor employ behavior charting

as an adjunct to his customary therapeutic procedures. In

the experimental condition, the counselor charted the




44

client's positive statements about self and others (put ups)

and his negative statements about self and others (put

downs) during each counseling session. The client charted

put ups, put downs, and a third behavior agreed upon by

counselor and client as relevant to the client's concern,

outside the counseling session.

The design of this study purposely incorporated a

variety of counselor and client opportunities for behavior

charting. The consideration of several behaviors and oc-

casions for charting by both counselor and client afforded

an opportunity to obtain differences between the control

and the experimental procedures.


The Sample

Clients were drawn from the undergraduate and graduate

populations enrolled in the College of Education at the

University of Florida. The following criteria were used

to select a pool of clients for this study:

(a) The client was not currently participating
in an individual counseling relationship.

(b) The client indicated that he wished to dis-
cuss a legitimate personal concern of a
social, educational, or vocational nature.

(c) The client agreed to remain in counseling
for at least six sessions over a period of
three weeks, with provision for additional
sessions if desired by the client.







All clients entered counseling voluntarily with the

understanding that they were participating in a study to

explore different modes of counseling. Three prospective

clients withdrew from the study before their first coun-

seling session, while another client withdrew after the

first session. Two clients whose concerns were judged

insufficient by both counselor and client to warrant six

counseling sessions were terminated after their first

session. A replacement for each terminated client was

immediately selected from the client pool.


Procedure

At the beginning of the first counseling session under

each treatment, the counselor administered the AS card

set of self-referent statements as a pretest measure to each

control and each experimental client. The following instruc-

tions were given by the counselor:

This deck of cards contains statements you might
make about yourself. Sort the cards into a
"like me" stack and an "unlike me" stack. If
a statement is true about you more often than
not, place it in the "like me" stack. If a
statement is not true about you more often
than not, place it in the "unlike me" stack.
Place each card in the stack which you honestly
feel describes you best at present. Please
begin.

After collecting the sorted cards, the counselor arranged

a schedule of twice-weekly sessions at times convenient to

both client and counselor.







In the control condition, the counselor initiated a

counseling relationship with his client, using his

customary approach for the remainder of the first session

through the final session. However, an additional statement

was delivered by the counselor during the first session

to each experimental client:

Keeping a record of some of your behaviors,
including feelings, may help you to clarify those
behaviors, and to change them if you wish.
I would like for you to count three behaviors
outside our counseling sessions, beginning
today. "Put ups" are the positive statements
you make about yourself or others, while "put
downs" are negative statements you make about
yourself or others. Keep a separate count
of the put ups and put downs that you actually
say. Positive or negative thoughts are not
counted unless they are expressed aloud. I'd
also like for you to count any behavior that
we feel is relevant to your concern. It can
be a behavior that others can observe, or it
may be a thought or feeling you have.

The counselor gave the client an opportunity to ask any

questions about the behaviors to be counted. After the

counselor and client had chosen a relevant behavior which

the client could identify accurately as it recurred, the

counselor instructed the client in procedures of behavior

charting. Copies of the Standard Behavior Chart and the

Rate Computation Sheet were introduced as instructions

were given to the client. The following instructions pre-

pared by the researcher were closely paraphrased by






counselors during the first session under the experimental

condition:

The behavior we've chosen is first counted.
Keep a daily tally on a card that you carry with
you. Make a mark after each occurrence of the
behavior [counselor demonstrates], or keep
count on a wrist counter [demonstrates] Also
keep track of the time during which the behavior
happened. You can count a behavior all day, or
for part of a day. If the behavior happens only
in certain situations, then mark the time that
you're in those situations.
Use the Rate Computation Sheet [show to
client] to get the frequency of behavior.
Divide the behavior movements you counted by
the time over which you counted to get fre-
quency [demonstrate .
Plot the frequency on the Standard Behavior
Chart [show to client Place a point on the
vertical line for the day you're plotting, at
the frequency shown by the horizontal lines
[demonstrate]. You have just plotted a rated
day. An ignored day, when the behavior occurs
but you don't count it, will have a line through
it connecting the points for rated days on either
side. A no chance day, when the behavior doesn't
occur, gets nothing. Leave it blank.
The record floor is a line which puts a lower
boundary on the behavior we've plotted. It shows
the frequency of the behavior if it had occurred
only once during the time you counted [demon-
strate on chart .

The counselor answered any questions about procedures of

behavior charting.

During each session, the counselor counted client

put ups and put downs and shared their frequencies with

the client at the end of the session. The client shared

the frequencies of his put ups, put downs, and relevant

behavior charted during the time since the last session.






All frequency data were plotted on the Standard Behavior

Chart in the following manner: client put ups and put downs

counted by the counselor during the session were plotted in

contrasting symbols or colors on a single chart. The

experimental client similarly plotted on a single chart

his own put ups and put downs which occurred outside the

session. By plotting two imcompatible behaviors such as

put ups and put downs on single charts, counselor and

client could examine the daily changes in the relationship

of these behaviors, both in and out of session. On a

separate chart, the client plotted the behavior chosen by

him and his counselor as relevant to his concern.

At the conclusion of the last session under both the

experimental and the control conditions, the counselor

administered the AS card set of self-referent statements

to each client as a posttest measure. The instructions

given to the client were identical to those preceding the

pretest administration.


Selection and Training of Counselors

The counselors for this study were seven advanced doc-

toral students enrolled in graduate programs of counselor

education or clinical psychology at the University of Florida.

Each counselor had completed at least two years of super-

vised experience in individual counseling prior to the







study. Two counselors had used behavior charting, while

the remainder had no previous experience with the technique.

The counselors represented a variety of approaches within

the client-centered tradition of counseling.

The counselors were trained by the researcher prior to

the treatment period in the principles and techniques of

behavior charting. During a two-hour training session, the

researcher demonstrated charting procedures with the aid of

a plastic transparency of the Standard Behavior Chart.

Counselors next listened to a number of three-minute taped

segments of counseling sessions to gain skill in detecting

and discriminating client put ups and put downs. The

counselors and researcher were free to stop the tape when-

ever one or more persons detected a put up or a put down.

A consensus on the classification of each positive or

negative statement was reached before the researcher ad-

vanced the tape. Finally, a brief simulation of a counsel-

ing session allowed the counselors further practice in the

identification of client put ups and put downs. Two train-

ees role-played a "counselor" and his "client," while

the other trainees independently counted put ups and put

downs. Additional instruction in the methodology of be-

havior charting was available from the researcher

throughout the treatment period. The researcher had no




50

direct contact with any client under either treatment condi-

tion, but acted only as consultant to the counselors.


Selection and Training of Raters

Raters were initially selected from undergraduate

and graduate students who received psychology course credit

for participation in the study. Additional raters were

drawn from volunteers in the graduate program of counselor

education at the University of Florida.

All 14 raters were trained by the researcher during the

first two weeks of the treatment period. The training

session comprised rater practice in the identification

of client put ups, put downs, statements in the present,

and statements in the past. Each rater was provided with

a set of definitions for the four client behaviors (Appendix

A). Next, raters listened to a series of three-minute taped

segments of counseling sessions taken from sources other

than this study. The first segment was played twice, with

ample opportunity for raters or the researcher to stop the

tape so that put ups and put downs could be identified and

discussed. For the second taped segment, raters were

instructed to count client statements in the past and

present tense, as well as put ups and put downs. A replay

and discussion of the client behaviors followed. A third




51

taped segment of client interaction was similarly rated.

Finally, each rater independently counted the four specified

client behaviors in a seven-minute calibration tape designed

to assess intra-rater reliability in behavior counting. To

provide a check on an individual rater's consistency over

the period of the study, each rater listened to the calibra-

tion tape three times: before rating any tapes, after rating

six tapes, and after rating all twelve assigned tapes.














CHAPTER IV


RESULTS AND ANALYSIS OF THE DATA


This study has examined the effects of a treatment

variable, counseling with and without behavior charting,

upon the outcome variable of client behavior. A statistical

analysis provides a summary comparison of the control and

experimental groups, while a graphic analysis furnishes

an evaluation of behavior change during counseling for each

of the five sets of clients paired by treatments.

Two of the original seven blocks of treatment pairs

were lost for purposes of statistical and graphic compar-

ison. A client who withdrew from school during the second

week of the study was not replaced because of the short

time remaining in the study. Another client chose to

withdraw after the fifth session, when the rater evaluating

the taped counseling sessions inadvertently violated the

client's confidentiality. In each case, the treatment

counterpart of the withdrawn client was not used for data

analysis because of the randomized block design which

matched subjects.







Client Self-Referent Behavioral Change

Hypothesis 1 states that "clients whose counseling

includes behavior charting will show greater change in

self-referent statements than clients receiving counseling

without behavior charting, as measured by the AS." The AS,

a q-sort of 74 self-referent statements, was administered

as a pre- and posttest measure of cumulative change in

self-referent behavior during counseling. An analysis of

covariance (Table 1) revealed no significant difference,

at the .05 level, between the experimental and control

posttest scores adjusted for the variability of the pretest

scores.

A study of the distribution of the AS raw scores

supports the statistical evaluation. The scores within

and across groups, with the exception of a single control

score, are relatively stable and homogeneous, forming a

cluster near the maximum score of 74 which indicates a

hypothetical ideal of self-adjustment. Two possible defi-

ciencies of the AS measure may offer an explanation of

these results. First, the statements seem to possess a

high face validity which would allow a client to score as

"well-adjusted" in self-concept regardless of his true

feelings. Moreover, the AS provided a limited sample

of client behavior: some twenty minutes compared to about







TABLE 1

SELF-REFERENT SCORES OF CLIENTS BY TREATMENT


Control


Posttest

67

59

70

63

50


Source of Variance

Treatment

Blocks

Residual

Total


Source of Variance

Treatment


Pretes

54

64

55

71

68


Experimental
t Posttest

54

60

60

68

72


Analysis of Covariance
Sum of Squares
XX YY XY

28.9 2.5 8.5

478.6 60.6 105.9

771.5 387.0 250.5

1279.0 450.1 364.9


Adjusted SS

.026


3 F 1/.31

.026 .0003 n.s.


Blocks


Residual


305.665


101.888


Total


Pretest

69

31

69

63

63







six hours of recorded client behavior which included the

self-referent behaviors of put ups and put downs. A later

discussion of the frequencies and changes in rate of put

ups and put downs for each client during counseling will

furnish a more comprehensive and accurate measure of self-

referent behavior.


Recorded Client Behavior Changes

Hypothesis 2 concerns client change in four specified

behaviors: put ups, put downs, verbal expressions in the

present tense, and verbal statements in the past tense.

The experimental treatment of counseling with behavior

charting was predicted to effect a higher frequency of

put ups and present tense statements and a lower frequency

of put downs and past tense statements than the control

condition of counseling without behavior charting.

Independent t tests (Table 2) were conducted to compare the

experimental and control group means of last-session

frequencies for the four evaluated behaviors. No comparison

of group means was significant. However, a comparison of

behavior change in clients taken by treatment pairs (Table

3) reveals 10 acceleration changes in 20 possible comparisons

of like behaviors in the directions predicted by hypothesis

2. In behavior charting, an acceleration is defined as














TABLE 2

INDEPENDENT t TESTS OF LAST-SESSION
BEHAVIOR FREQUENCIES BY TREATMENT


x x
Experimental Control df t

lut ups .15 .15 8 .000 n.s.

'ut downs .14 .09 8 .394 n.s.

'resent statements 1.20 1.43 8 .212 n.s.

?ast statements .49 .75 8 .226 n.s.









TABLE 3


CLIENT BEHAVIOR ACCELERATIONS BY TREATMENT


Behavior Counselor Treatment
Experimental Control


Put ups






Put downs






Present statements






Past statements


.484
.889
1.172
.768a
.900a

.404
.834b
1.111
.270b
.778

.629
1.145
.650
.705
1.077c

1.110
1.396d
1.084d
.633d
.822


.954
.960
2.766
.706
.344

.391
.981
.538
.698
.597

1.247
.657
.840
.894
.644

1.291
1.721
2.004
.656
.529


a
Experimental
predicted by
bExperimental
predicted by
CExperimental
predicted by
dExperimental
predicted by


acceleration >
hypothesis 2a.
acceleration <
hypothesis 2b.
acceleration >
hypothesis 2c.
acceleration <
hypothesis 2d.


control acceleration,

control acceleration,

control acceleration,

control acceleration,





58

the rate of change in behavior during a week. The accelera-

tion is found by the formula

behavior frequency of day (X + 7)
= acceleration.
behavior frequency of day X

Table 3 also presents data correlated with the

assumption of a Rogerian counseling model which states that

counseling produces an increase in put ups and present

statements, and a decrease in put downs and past statements.

Only one pair of clients (pair 3) experienced an acceleration

of put ups during counseling; the other pairs showed a

deceleration in put ups (indicated by an acceleration <1),

although two experimental clients (4 and 5) decelerated less

than their control partners. The relative behavior change

of experimental clients 4 and 5 conforms to the direction of

change predicted in hypothesis 2a. All but one client expe-

rienced a deceleration in put downs in accord with the

Rogerian model of change in counseling. Experimental clients

2 and 4, who show a greater deceleration in put downs than

their control partners, fulfill the expected direction of

behavior change in hypothesis 2b. Statements in the present

tense accelerated for two experimental clients (4 and 5), as

predicted by hypothesis 2c, when they were compared to their

counterparts under the control treatment. Past statements

decelerated, as assumed by the Rogerian model of counseling,







for treatment pairs 4 and 5. However, the celebration for

experimental clients 1, 2, and 3 showed a relative decrease

compared to the rates of change for their control treatment

partners. This result agrees with the prediction of

hypothesis 2c.

Each counselor was assigned a control and an experi-

mental client as a check for the effects of individual

counselors on client behavior. Table 3 illustrates the

consistency of counselor influence upon the behavior of

both members of the treatment pair. For every counselor,

the celebration rate of put ups for both clients of the

treatment pair was in the same direction; that is, both

clients of pairs 1, 2, 4, and 5 decelerated in put ups,

while both members of pair 3 accelerated. Similarly, both

members of pairs 1, 2, and 3 accelerated in past statements

while each member of pairs 4 and 5 decelerated. Four

treatment sets of clients decelerated in put downs, and two

sets decelerated in present statements. The variability

among counselors, in sum, seemed to have a greater effect

upon client change than the treatment condition. A major

effect of this counselor variability by treatment pairs

of clients was the deceleration of all recorded client

behaviors in 28 of 40 (70 percent) of the opportunities for

behavior change, regardless of the predicted direction of

change based upon the Rogerian model of counseling.






A corollary which may be inferred from hypothesis 2

states that for a given client, the effect of counseling

may be judged by the magnitude of the difference between

two incompatible behaviors called a "comfort pair." Table

4 presents the relationship between a client's put ups and

put downs, as well as his present and past tense statements.

The "multiplier" represents that number by which the accele-

ration of one component of the comfort pair must be multi-

plied to obtain the acceleration of other component. In

mathematical terms,

acceleration of component 1
multiplier = acceleration of component 2.

In Table 4, the acceleration of put ups is divided by the

acceleration of put downs; similarly, the present tense

statements' acceleration is divided by that of the past

tense statements. The greater the magnitude of the multi-

plier, the greater the difference in celebration rates for

the behavioral components of the comfort pair. A multi-

plier greater than 1 indicates that put ups or present

statements increased at a relatively higher acceleration

than their respective comfort pair member, put downs or

past statements. A multiplier which approaches 1 shows a

fixed relationship between comfort pair members, while a

multiplier less than 1 indicates that put downs or past











- r- '- 0 o --i ,-4 %D
In OM 0 D -4 r-


o o
*o*
-W -4
m CN













Ln r-


-4













0: 0















0 -'
m00


00
In r






















0- -I
*l











-t 00






.-
r-i











--i4~
*-4


CO
00













CO

o 00













r- o














0 %
o o




00 \D


0U wU U C uu OU
NN m m :t T Am LA


- CN














M CN
CN m





















10 >D









4











00
* m




























O


r-


















o on
(N .
O *
















00 '














cDn


4J



C-



ra Q)





0
.-I











-)i
x *c







U U

0 0

O
OO
,







statements increased more rapidly than put ups or present

statements. All five experimental clients have a multiplier

which indicates a higher relative acceleration of put ups

than put downs. However, only two experimental clients show

a multiplier of greater than 1 for the acceleration of

present statements compared to past statements. Likewise,

the multipliers for two control clients fit the Rogerain

model which predicts that present statements will increase

at a greater rate than past statements. Three of five con-

trol clients have multipliers which indicate that their put

ups increased at a higher rate than their put downs. Only

two control clients show a similar relationship of present

to past statements. By counselors, who appear to have in-

fluenced clients more than the treatment condition, three

pairs of clients show growth (the multiplier> 1) in the

relationship of put ups to put downs. Two pairs show growth

in the ratio of present to past statement accelerations.

The reported behavior changes occurred over a three-week

period. A longer period of observation seems necessary to

confirm the tentative differences in direction of recorded

behaviors from the directions predicted by the Rogerian

model.

The raw frequency and derived celebration lines are

plotted on modified Standard Behavior Charts. Clients are





63

blocked by treatment pairs for each counselor. Each figure

comprises two panels, abbreviated versions of the SBC with

space for recording 60 rather than the customary 140 days.

The range of frequencies has not been changed. In each

panel the frequencies are plotted in chronological order of

the days on which counseling sessions were held. The

celebration lines plotted for each client in a panel were

calculated by the method of least squares upon the logarithms

of the frequencies to provide a line of best fit for the

logarithmic cycles of the SBC. The first panel of Figure

1 compares the experimental and control client for counselor

1 on put ups, a component of the comfort pair: put ups/put

downs. The second panel compares the clients for counselor

1 on put downs, while Figure 2 portrays the distribution

of experimental and control frequencies for the second

comfort pair: present tense statements/past tense statements.

The five treatment pairs of clients are represented on all

rated behaviors in Figures 1-10.


Reliability

The rationale for this study states that any changes in

client behavior during counseling should be observable to

external raters who can record objectively these behavior

changes. The interpretation of frequency data involves two

























H


E4
z



E-


P:
0

ri,











rLQ
0
H






Ez~
U,
04






0

z
0


H


'C
0





65



o


0 0
0 10



o0


0 _-- 0
-- 0 bi
"
0

O







uu
0


-



-o


00
OO 0o On - w g o
0 o 0 --9 o o








O 0
to
OD
o
w
0o
U' )

0


En








-0

0 OU


S ...* U S.t o:,: O -
go O 0 -- 0 O
0L 0 O *

31nNIW Ll3d S.LN3'.'13AOW
















,-I

H











4D
0








CN


Z
H









Z
Ul

w




Ka









I4

0


0
H
c-t~

04

u
U






































O o o n
0 -
Ooo. L Ou


.1 .. 1 I I
0
0
0 L


1 .i.i i I ., I i I i l I
0 LO 0 0
00 0
-O


c


3.LnNI I U8:3d SLN3.':-AO'#4


(1
C
SU 14
ti E 2
(O a J





0 oUIu
00
0 o4
o

x-
0 li














w
o



- Q
iO













.-o
4J
U)













- o U



z
0 0
U

0
0 0







,,-
0 ( b
=





S. )


* 0


0 O


i


























H



04













Em




H
E-1




E-4


0

0
13

44





E-4
P4





H
0



U









3 z



"I
1_ / a-


0 0
0 o


|II| I 1 I in ; I i -;i if i n .IllI l I I .1.. i I I

0 '
*


2 O
0


LJ
-o





(o



V)


OO 0 o On 4 -- o O
0 0 0 0 0 l
OO 0O" O
0 0 NI0

31nNIi U3ld S.LI 3YI3AOi


I _ _


i
















N








z
-I







0
H















H




U,
H E-
z











E-4












0

0
IC
U
ul
H















U)








m







o o


10
-. .3


0 0




0
Oz






So
00 0o0 0





0 0 0 0 n w
C -








\\ : 0 00
S\ :- O


O
to c q

C,


'U,
-0
7 3
: (I)







0








O- U-




0I C) : 0 U
00 o :O 01
0
-j



LI
O I0





















m

H











0
F94






rN






O

z




0

aU
H

0E
0
u





73




012










S0 0
.x-
o cU


lo o y

OoL

o


u3
S-O






0 / 0 .
O












7 z





0 O
0 0
:i
00 0 0











00 00 tn 0. -- 0 LJ
0 o o o





:-o o J
















3LnNIVI JcJd S. 3YJ-AOVI
E =2
ro










Oo \ : 0 0






O n O

311NIVJ l U3d SINY3YI3AOW ri

0
















en














E
0


Hl
E1

U ,









U3

















H
E-
z





E










TI4
0
0
CL













































0 0 0 OL -- n to L
0 0 0
L --"


o0 0 0 0 O
)0 O 0
rC O -


3.1nlNII U3d S.IN3'i]3AOPI


lcu


o 0




0

0
-j
La
0


C. 2
4J
4-10
91


-I


-I-


0




0:
A


















I






J
O
I'
)










4..2


0


V,

C





L




LL
0








UL
4J















:0
(7


:^C L
a


i ~ ~ ~ ~ 1 I --I I . Is I I I .


I -- -


I


I






















H











H
P










rrz









0
z
0
CJ2









H


u-




77



0
oi
o








00
-_ J
:



o


0 0
0 0 0 0


S O


Ooo O q o
0- 0 0
(D-
00 O O



w


-0 "


.0
0 x
---.----------------: o .. In
-~o









-LI.
1-I
LI
: y

/^




OO :0 -
- o
311nNlVi tl3d Slri13W3AOW i





















H



E-4


E





E-4





E-4


KC
E-4











~zn
H-

E-4


a4




EA












0
U
0
U)





79





4,
0 E

oo
(0



o j





.
0 o o
0


10 3











On






.o
F -0









0 c





o





0 0 : -- 000
S 0 .





C)
c Z
CCy







O 03
: G u J lt
OO On I












O a lJ




31nNIA d3d IN3Y SAOVi
CO 0







00

C' ,.3d S I























U-)









E-4







E-4:



a~








CL4
H


U)








0

U)
H


04

U




81




o .


o





cM
1c
(0















OO O o Ol
o -
.0 I!















OI
oo









o z
S-


















0 0 c3
0-0 l







OO OO Oo i O c,
OO O -- o O
O -






SI;
0









I~ -i

















Ln

H
















E-4
mc
H







En



'cc
O H





ul




z
E-4 U



Ha -









El)
W
zz



H
CL,
D4
0






z


0
U)
H












S1.


-;.'; i r I ,iili I I i I..ij l i ">iiill l i I ,1.11 i' i ,.a; i i I i
0o 0 o 0 --n .O L
0o -- o
O0 0

0 U') _


' rr-, 0 -Iigi i i . ~ : ,-.Trl- I --. 1
0 00 -O to
00
O


311NIVi U3d S.IlEIJ3AO


0

o g
z


0


OIJ




0 t


0

W
(U)


-i

Lee
: cu

:- o




X-J

0

0

0


t)

E 2
1) LJ

u 0
4-n
U)


p -_ r-i~ C


00
00
0 U


f 1






aspects of objectivity: the agreement of independent

raters, and the consistency of a single rater over time.

Interrater and intrarater reliability provide measures of

these facets of objectivity. Table 5 presents Pearson

product-moment coefficients showing the correlation be-

tween two independent ratings of a series of counseling

sessions for the same client. Each rater was assigned the

tapes of a control client of one counselor, and tapes of an

experimental client of another counselor, to prevent rater

bias by the style of counselor interaction. Each correla-

tion, unless otherwise noted in Table 5, is based upon a

comparison of 24 behavior frequencies (4 behaviors X 6

sessions) evaluated by each rater. A mean positive correla-

tion of .656, with a range from .975 to .282, is reported in

Table 5. The variability of independent ratings was resolved

by plotting the geometric mean of each pair of behavior

frequencies as a single frequency point. Underlying this

practice was the assumption that the true behavior frequency

lay somewhere between the two observed frequencies for the

same behavior. The geometric mean is the proper measure

of central tendency for ratio data.

Intrarater reliability, or the consistency of evalua-

tion over time, is given in Table 6. Each rater evaluated

a seven-minute calibration tape at the following times:








TABLE 5

PEARSON PRODUCT-MOMENT COEFFICIENTS OF RELIABILITY
BETWEEN TWO INDEPENDENT RATERS


Client
Rated


Coefficient of
Correlation (r)


.282
.773a


.975b
.303


.643c
.496


.727
.837


.670
.854


.656


a
Observed frequencies include the rerating by a single rater
of one tape inaudible to the other independent rater.
Based on four taped sessions. Session 2 was lost because of
a recorder malfunction; session 5 was accidentally erased by
the researcher.
CBased on taped sessions 1-3. One rater became unavailable
to evaluate the remaining sessions.








TABLE 6

PEARSON PRODUCT-MOMENT COEFFICIENTS
OF INTRARATER RELIABILITY


Correlation b Coefficient c
r 12 r13 r
Rater 12 13 23


.936

.938

.645

.910

.261


.849

.986

.848

.961

.897


.882

.945

.943

.975

.000


.901d


.701

.886

.878

.609

.807


.761

.985

.785

.170

.737


.982

.934

.980

.132

.991


12 = correlation between calibration pre-sessions rating
b and mid-sessions rating
br3 = correlation between calibration pre-sessions rating
and post-sessions rating
cr12 = correlation between calibration mid-sessions rating
and post-sessions rating
Rater omitted calibration mid-sessions rating
eRater became unavailable at end of study


.892e




87

before evaluating any tapes from this study, after evaluating

half the tapes, and after completing the evaluation of all

twelve tapes (6 sessionsX 2 clients). The correlations were

calculated for the relationships of the calibration tape

pre-counseling rating to the calibration tape mid- and

post-counseling ratings. The high correlation of the mid-

and post-counseling calibrations illustrate the stability

of intraraters evaluations after experience had been gained

by evaluating the first six tapes. The test-retest relia-

bility of eight of 12 raters approximates or exceeds .90,

while the remaining raters show a mean slight correlation

of rating consistency. Rater 5, whose post-counseling

rating is highly consistent with his original rating of the

calibration tape, seemed to recover the initial criteria

for evaluating client behaviors after a drift from the

evaluation standards reflected by a correlation of .000

between calibration ratings 2 and 3.

The data available from counselors on their counts of

the experimental client's in-session put ups and put downs

have been correlated with the geometric mean of the

corresponding independent raters' evaluations (Table 7).

The high variability of the correlations, ranging from -.59

to .97, supports the spontaneous comment of every counselor

that counting while counseling was a difficult task.




















TABLE 7

CORRELATION BETWEEN COUNSELOR AND RATER EVALUATIONS
OF THE EXPERIMENTAL CLIENT'S IN-SESSION
PUT UPS AND PUT DOWNS


Experimental Correlation
Client Rateda Behavior Coefficient (r)

le put ups -.21
put downs .57

2e put ups .65
put downs .97

3e put ups .08
put downs -.59


Counselor evaluations
available


of experimental clients 4e and 5e not




89

Counselors related that their attempts to count were

sporadic; that is, when the interaction with the client

became intense, the counselor either forgot to count

client put ups and put downs or chose not to "distract"

the client. The counselor likely felt the intrusion more

than the client on such occasions.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs