llllINHIQHIIHJIllllllllllllllUNlllllUlllHlllllllllHl

1293 10409 7989

 

This is to certify that the
thesis entitled

THE USE OF MODELING TO TEACH
EMPATHY T0 NURSING STUDENTS

presented by

Janice Mary Layton

has been accepted towards fulfillment
of the requirements for

 

Ptho degree incounseling.

Personnel Services, and Educational
Psychology (Major in Educational
Psychology)

My (ya/AM

Majoresr/rpofso a

Date July 7’ 1978

0-7 639

 

OVERDUE FINES ARE 25¢ PER DAY
PER ITEM

Return to book drop to remove
this checkout from your record.

 

" 56

 

 

 

 

 

 

 

 

 

 

 

 

THE USE OF MODELING TO TEACH
EMPATHY TO NURSING STUDENTS

By

Janice Mary Layton

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Counseling, Personnel Services, and
Educational Psychology

1978

ABSTRACT

THE USE OF MODELING TO TEACH
EMPATHY TO NURSING STUDENTS

By

Janice Mary Layton

The primary purpose of this study was to compare the effective-
ness of different modeling conditions in teaching empathy to nursing
students. A secondary purpose was to examine the construct validity of
the instruments used. The participants were junior and senior
baccalaureate nursing students who volunteered for the project. Seventy
students were randomly assigned by class to one of five groups. Fifty-
six students completed the project.

The treatment consisted of modeling, labeling, and covert
rehearsal conditions, which were combined to form four treatment
groups—-modeling-only, modeling-labeling, modeling-rehearsal, and
modeling-labeling-rehearsal. All treatments were videotaped. The fifth
group was a no-treatment control.

A repeated measures design was employed with posttesting
immediately following the treatment and three weeks later. The
followhup posttest was for memory and was considered an important aspect
of this study, as most studies of modeling and empathy do not include

follow-up.

Janice Mary Layton
The dependent variable, empathy, was measured in three ways.

Immediately after treatment, students completed the Empathy Test--Form I,

 

a cognitive test of principles of empathy constructed for this project.
Students then interviewed a simulated client. The interview was evaluated

in two ways. Clients completed the Barrett-Lennard Relationship Inventory,

 

and the interview was audiotaped for later rating using Carkhuff's

Empathic Understanding in Interpersonal Processes Scale. The second

 

posttesting three weeks later was the same, except that an alternate

form of the Empathy_Test (Form 11) was used. The use of an interview

 

posttest was an important element in the study, since most studies use
only written posttests, and those that used interviews showed no
treatment effect for modeling alone.

It was hypothesized that the treatment groups would perform
better than the control group, and that the elements of labeling and
rehearsal would further improve performance. In relation to follow-up
testing, it was hypothesized that all groups would retain their level
of behavior three weeks later. Because the most serious problem in
doing research on empathy seems to be its measurement, a third
hypothesis related to construct validity was included, stating that
the measures used would be significantly correlated with each other.

Data were analyzed using multivariate analysis for repeated
measures, as well as for the two testing periods separately. There were
no treatment effects for the group as a whole. However, when data were
analyzed separately for junior and senior students, the results for junior
students showed a treatment effect on two of the instruments--the
Carkhuff Scale and Empathy Test II. There was also an interaction of treat-

ment and time, with treatment groups improving on the second posttesting.

Janice Mary Layton

To counteract the effect of memory loss, students were given
instructions to practice what they learned between the two testing
sessions. It was hypothesized that the students' practicing constituted
a condition of overt rehearsal and accounted for the improvement seen
at the second testing. Scheffé post-hoc pair comparisons showed that
only the two groups receiving covert rehearsal performed significantly
better than the control group. This finding is consistent with other
research on modeling, which indicates that covert plus overt rehearsal
produces the most change in behavior.

Several explanations are possible as to why juniors but not
seniors benefited from the modeling treatment. Since juniors' average
performance was lower than seniors', perhaps the treatment is better
for beginning students. Juniors may also have felt more of a need and
therefore been more motivated to learn. Seniors, because they have more
experience, may be set in their interview style and need different
learning experiences than those provided.

The hypothesis about correlations between the instruments
used was confirmed for some of the instruments, thus lending support
to their construct validity. The Carkhuff Scale was correlated with

itself for time one and time two, with the Barrett-Lennard Relationship

 

Inventory for both tastings, and with Empathy Test II.

 

To the memory of my grandmother,
Isabelle M. Sauers,

who was a special person in my life.

ii

ACKNOWLEDGMENTS

I wish to express particular gratitude to Dr. Walter Hapkiewicz
who, as my academic advisor and dissertation chairperson, helped me at
every stage of my doctoral program. His understanding, advice, and
support were always available when needed, and his tact and humor made
my life as a doctoral student more bearable. He has been a true friend.

Special thanks are also due to the members of my doctoral
corrmittee, Dr. Patricia Busk, Dr. Robert McKinley, and Dr. Ann Olmsted.
Their interest in my learning, the time they spent helping me, and
their example as knowledgeable and dedicated professional persons is
deeply appreciated.

I also want to thank two members of the Educational Psychology
faculty--Dr. Lee Shulman and Dr. Stephen Yelon--who helped me on
different occasions and whose accomplishments were an inspiration to me.

Doing my dissertation would have been impossible without the
help of Dr. Isabelle Payne and the faculty, staff, and students of the
School of Nursing at Michigan State University. Dr. Payne offered the
facilities of the school, enabling me to make videotapes, recruit
students, and collect data. Many faculty members assisted me in
. recruiting students and encouraged me by expressing interest as I
carried out the study. To the three faculty who participated in the
videotapes--Anita Chesney, Sue Emmert, and JoAnn Westrick--I am
especially grateful. The students who were participants in the study
made data collection enjoyable by their interest in the project and

iii

their enthusiasm for learning.

To Holly Holdman, who helped me in many ways, I owe a special
debt. She participated in the videotapes, recruited and trained
simulated clients, and assisted in the collection of data. She also
made many helpful suggestions and gave freely of her support and
encouragement. Completing this dissertation would have been much more
difficult without her help.

Finally, I want to note that this investigation was partially
supported by a Health Resources Administration National Research Service

Award (No. 1 F31 NU05083-Ol) from the Division of Nursing.

iv

TABLE OF CONTENTS

Page

1. PURPOSE, OVERVIEW, AND ORGANIZATION OF THE STUDY ....... l
Purpose and Overview of Study ............... l
Organization of Study ................... 2

II. REVIEW OF RELATED LITERATURE ................. 3
Empathy .......................... 3
Definition and Importance of Empathy .......... 3

Construct Validity ................... 6

Tests Measuring Empathy ................ l0

Empathy and Nursing .................. l4

Modeling ......................... l8

The Modeling Paradigm ................. 18

Modeling and Interpersonal Behavior .......... 21

Modeling and Empathy .................. 24

III. METHODS ........................... 32
Design .......................... 32
Hypotheses ........................ 33
Measurement of the Dependent Variable ........... 34
Empathy Test ...................... 35

Subjects ......................... 37
Procedures ........................ 38
Videotape Production .................. 38

Modeling ...................... 39

Labeling ...................... 41

Rehearsal ...................... 41

Subject Recruitment, Assignment, and Attrition ..... 43

Training Simulated Clients ............... 44

Data Collection .................... 44

Rating the Audiotaped Interviews ............ 46

Page

IV. RESULTS AND DISCUSSION .................... 50
Results .......................... 50
Modeling and Empathy .................. 50
Total Group Results ................. 50
Results for Juniors and Seniors ........... 54
Construct Validity ................... 63
Feedback Sheet ..................... 64
Discussion ........................ 66
Modeling and Empathy .................. 66
Results for Junior and Senior Students ....... 66
Interaction of Treatment and Time .......... 68
The Remaining Hypotheses .............. 72
The Instruments in Retrospect ............ 74
Construct Validity ................... 78
V. SUMMARY AND CONCLUSIONS ................... 82
Summary .......................... 82
Conclusions. . . . 1 .................... 85
Modeling and Empathy .................. 85
Nursing Education ................... 87
Construct Validity ................... 88
APPPENDICES
APPENDIX
A. Empathic Understanding In Interpersonal Processes: A Scale
For Measurement ...................... 89
B. Barrett-Lennard Relationship Inventory ............ 94
C. Empathy Test - Form I .................... 96
D. Empathy Test - Form 11 .................... 99
E. Narration for Videotape "Empathy in Nursing" ......... 103
F. Empathy Teaching Project Consent Fonn ............ 109
G. Directions for Students: Empathy Teaching Project -

Session I ......................... 110

vi

APPENDIX Page

H. Directions for Students: Empathy Teaching Project —

Session II ......................... III

I. Feedback Sheet ........................ ll2

J. Rules for Rating Accurate Empathy in Special Cases ...... ll3

K. Scores for Each Subject on Each Variable by Groups ...... ll4
LIST OF REFERENCES .......................... ll7

vii

Table

10.
11.

12.

13.

I4.

15.

16.

LIST OF TABLES

Experimental Design and Number of Subjects in Each Group. . . .

Item Analysis Data for Form I and Form II of the Empathy
Test ............................

Second Item Analysis on Form I and Form II of the Empathy
Test ............................

Length of Videotape Components and Length of Each Section
Within Each Component ....................

Group Means and Standard Deviations on Each Posttest for
Time l and Time 2 ......................

Multivariate and Univariate Analysis of Treatment Effects . . .

Multivariate and Univariate Analysis of Repeated Measures
Effects ...........................

Multivariate and Univariate Analysis of Interaction of
Treatment and Time .....................

Multivariate and Univariate Treatment Effects for Time T. . . .
Multivariate and Univariate Treatment Effects for Time 2. . . .

Junior Students' Group Means and Standard Deviations on Each
Posttest for Time 1 and Time 2 ...............

Senior Students' Group Means and Standard Deviations on
Each Posttest for Time 1 and Time 2 .............

Junior Students: Multivariate and Univariate Analysis of
Treatment Effects ......................

Junior Students: Multivariate and Univariate Analysis of
Repeated Measures Effects ..................

Junior Students: Multivariate and Univariate Analysis of
Interaction of Treatment and Time ..............

Junior Students: Multivariate and Univariate Effects for
Time 1 ...........................

viii

Page
33

36

37

42

52
53

53

53
54
54

55

56

57

57

58

Table
17.

18.

19.

20.

21.

22.

23.

24.

25.

Junior Students: Multivariate and Univariate Effects for
Time 2 ............................

Senior Students: Multivariate and Univariate Analysis of
Treatment Effects ......................

Senior Students: Multivariate and Univariate Analysis of
Repeated Measures Effects ..................

Senior Students: Multivariate and Univariate Analysis of
Interaction of Treatment and Time ..............

l
Scheffe Post-Hoc Comparisons for Junior Students on the
Carkhuff Scale ........................

Scheffe Post-Hoc Comparisons for Junior Students on the
Carkhuff Scale for Time 2 ..................

Scheffe Post-Hoc Comparisons for Junior Students on the
Empathy Test for Time 2 ...................

Correlations of the Three Measures Used for Time l and
Time 2 ............................

Summary of Students' Responses on the Feedback Sheet ......

ix

Page

59

59

6O

6O

61

62

63

64
65

LIST OF FIGURES

Figure Page

l. Component processes governing observational learning in

in the social learning analysis .............. l9
2. Graphic representation of junior students' scores on the

Carkhuff Scale over Time T and Time 2 for the control and

treatment groups ...... . . . . . . ......... 7O
3. Graphic representation of junior students' scores on the

Carkhuff Scale over Time l and Time 2 for the five

groups .......................... 7l

CHAPTER I

PURPOSE, OVERVIEW, AND ORGANIZATION OF THE STUDY

Empathy is a quality needed by all nurses. The foundation of
good nursing practice is the ability to understand and communicate
with patients and their families. With greater and greater emphasis
being placed upon the expanded functions of the professional nurse in
such roles as primary care nursing, health maintenance, long term care,
and independent practice, empathy and related interpersonal skills are
becoming increasingly important. Without this basic skill, the other
skills and abilities the nurse possesses are weakened in effectiveness.

Over the years, nursing educators have used a variety of
methods to teach empathy to nursing students, but the success of these
methods has rarely been tested empirically. One method that has had
little systematic use in nursing, but that has been successfully applied

in teaching empathy to counseling students and other groups, is modeling.

Purpose and Overview of Study
The purpose of this study is to assess the effectiveness of
modeling in teaching empathy to nursing students. Using variations of
modeling conditions, empathy will be demonstrated via videotape to
junior and senior baccalaureate nursing students. The variations of
modeling to be tested are modeling-only, modeling-labeling, modeling-
rehearsal, and modeling-labeling-rehearsal. There are, then, four

treatment groups plus a control group that will receive no treatment.

1

Each student's learning will be measured in three ways: a written

posttest, constructed for this project (Empathy_Test), an interview with

 

a simulated client that will be audiotaped and rated using Carkhuff's

Empathic Understanding_in Interpersonal Processes Scale, and the Barrett-

 

Lennard Relationship Inventory that is completed by the simulated

 

clients. All three measures will be taken immediately after the student
views the videotape and three weeks later.

Two important aspects of this study are the use of simulated
clients to test learning and follow-up testing. Most studies examining
the effectiveness of modeling to teach empathy are deficient in one or
both of these elements. Nearly all studies reviewed used subjects'
responses to written client situations as a posttest rather than actual
or simulated interviews. While written situations provide maximum
internal validity or control of the experimental task, they are deficient
in external validity or generalization to real-world tasks.

Likewise, most studies test learning immediately after treatment
but do no follow-up or testing for memory. Both generalizability and
retention of learning are crucial issues to those involved in teaching

empathy to students, and hence are given a central role in this study.

Organization of Study

 

The following chapter is a review of relevant literature and
research on the topics of modeling and empathy. Chapter III reports
the methods used in the study. Chapter IV presents and discusses the
results, and the final chapter outlines the summary and conclusions of

the study.

CHAPTER II

REVIEW OF RELATED LITERATURE
The theoretical basis for this study is formed by the convergence
of two bodies of literature--the first on empathy and the second on
modeling. Theoretical and measurement issues are discussed, as well as

pertinent research studies pertaining to empathy and modeling.

am

In relation to empathy, the specific issues to be discussed are
the definition and importance of empathy, construct validity, tests

measuring empathy, and empathy and nursing.

Definition and Importance of Empathy

Most of the current writing and research on the role of empathy
in therapeutic relationships stems from the work of Carl Rogers and his
associates. In his early writings, Rogers (l957) identifies empathy,
together with warmth and genuineness, as the three qualities of a
therapist that form the necessary and sufficient conditions for
therapeutic growth to occur.

Rogers' (1959) early definition of empathy, which formed part
of his theory of client-centered therapy, is as follows:

The state of empathy, or being empathic, is to perceive

the internal frame of reference of another with accuracy

and with the emotional components and meanings which pertain

thereto as if one were the person, but without ever losing
the "as if" condition. (p. 2T0)

Rogers' (1975) later definitions encompass the above, except that now
he views empathy as a process rather than a state, and he adds the
element of communicating one's understanding to the client and checking
with him on the accuracy of one's perceptions. Traux and Carkhuff,
(l967) early co-workers of Rogrs, put forth a similar definition:

"Accurate empathy involves both the therapist's sensitivity to current

 

feelings and his verbal facility to communicate this understanding in a

 

a language attuned to the client's current feelings" (p. 46). On the
importance of empathy, Rogers (1975) states:

Over the years... the research evidence keeps piling up, and
it points strongly to the conclusion that a high degree of
empathy in a relationship is possibly thg_most potent and
certainly one of the most potent factors in bringing

about change and learning. (p.3)

Carkhuff, (l969a) who has developed a comprehensive human
relations development model that is used extensively in a variety of
settings, mirrors Rogers' position on the importance of empathy.

Empathy is the key ingredient of helping. Its explicit

communication, particularly during early phases of helping,

is critical. Without an empathic understanding of the
helpee's world and his difficulties as he sees them,

there is no basis for helping. (p. l73)

Rogers, Gendlin, Kiesler, and Traux (l967) note that certain
qualities of the therapist are stressed by writers espousing a number
of therapeutic approaches. Traux and Carkhuff (l967) likewise state:

Although most theorists concern themselves in their

writings with discussing the client, three characteristics

of an effective counselor emerge from the divergent viewpoints:

(1) An effective therapist is integrated, nondefensive,
and authentic or genuine in his therapeutic encounters.

(2) An effective therapist can provide a non-threatening,
safe, trusting or secure atmosphere by his acceptance,
unconditional positive regard, love or nonpossessive warmth
for the client.

 

(3) An effective therapist is able to "be with," "grasp the
meaning of," or accurately and empathically understand the

client on a moment-by-moment basis (p. l).

Empathy, genuineness, and warmth, then, although key concepts
in the work of Rogers and his associates, are not limited in their
theoretical import only to client-centered therapy. Rather, they are
seen as essential to any therapeutic encounter and in fact as basic to
all growth-producing relationships. As Truax and Carkhuff (l967) state:

Research seems consistently to find empathy, warmth,

and genuineness characteristic of human encounters that

change people--for the better.... The finding that most

human encounters can indeed be for better or for worse

suggests promising leads for research into the prevention

rather than just the treatment of psychological disturbance....

The implications hold not only for the training and

functioning of psychotherapists, but also for the training

of teachers and educators, marriage partners, employers

and supervisors, and parents (pp. l4l-l42).

And, we might add, of nurses.

Research evidence from Rogers and Traux and Carkhuff supports
the importance of empathy in therapeutic relationships. Rogers et al
(1967) studied a number of facets of the therapeutic process. One aspect
of their research was to correlate empathy, process (experiencing). and
outcome measures in schizophrenic patients receiving psychotherapy. They
found positive correlations between judge's ratings of empathy and
process measures, between patient's perceptions of empathy and process,
and between judges' ratings and patients' perceptions of empathy.
Therapists' perceptions of their own empathy were negatively correlated
with judges' ratings and process. Process measures were positively
correlated with many outcome measures, and patients receiving the highest
level of empathy as rated by judges showed the greatest reduction in

symptoms.

Traux and Carkhuff (l967) reported a series of studies on the
relationship between empathy and outcome. They found the level of
empathy to be significantly higher for successful than for unsuccessful
therapy cases. They also pointed out that low empathy from therapists
correlates with deterioration on outcome measures. This particular
finding is of interest in that it sheds some light on a longstanding
controversy about whether or not psychotherapy is effective. What Traux
and Carkhuff note is that if outcome measures from all persons receiving
therapy are averaged, the measures are not different from control groups
receiving no therapy. It is this type of finding that led some
(cof. Eysenck, 1960) to conclude that therapy is not effective. The
reason why the average results are not different, however, is that while
some persons in therapy improve, others deteriorate. Traux and Carkhuff
conclude, therefore, not that psychotherapy is ineffective, but that
therapeutic relationships can be for better or for worse.

The research findings of Rogers et al. and Traux and Carkhuff,
then, support the assertion that empathy is important in therapeutic
relationships, and that teaching empathy to nurses and nursing students

is a worthwhile educational goal.

Construct Validity,

 

Probably the most serious problem in doing research on empathy
is the issue of construct validity. This section will provide some
background information on construct validity, followed by a discussion of
some research on the construct validity of tests measuring empathy.

The American Psychological Association in its Standards for

 

educational andpsychologjgal tests (1974) defines a construct as

"a theoretical idea developed to explain and organize some aspects of

existing knowledge" (p. 29).

Magnusson, (1966) in his discussion of the validity of tests
developed to measure constructs, states that one begins with a
logically defined variable, in this case empathy.

This variable is included as a logical construct in a

system of concepts, in which all of the concepts logically

belong, and where the relationships are explained by a

theory. From this theory certain practical consequences

can be derived about the outcome of the test under certain

conditions. These consequences can be tested. If the

result is what was expected in a series of such tests, the

test is said to have construct validity for the variable

tested (p. 130).

The test constitutes an operational definition of a construct. Validity
is a quality of a test designed to measure a construct, and the purpose
of establishing construct validity is to develop theories that explain
and predict behavior.

Unlike criterion validity, construct validity cannot be derived
from a single testing or represented by a single value. Construct
validity is established only from a series of testings. Magnusson (1966)
points out that the determination of construct validity is an application
of the deductive method of scientific research. "This evaluation takes
place in accordance with the classical procedure: theory--deduction--
hypothesis--experimental testing--data which falsify or verify the
hypothesis" (p. 131).

Magnusson further states that systematic variance in tests
of construct validity can be broken down into two components--
variance due to relevant characteristics of the individuals tested
and variance due to properties of the method of testing used. Only the
first type of variance is explainable by the construct. Variance due

to method, if it is present, is a spurious inflation of correlations

between tests intended to measure a construct. Campbell and Fiske's (1959)

multitrait-multimethod approach to establishing validity is an attempt
to separate variance due to traits from variance due to methods. Their
approach consists of measuring each of several traits or variables by
each of several methods. The results of testing are presented in a
matrix that displays the following correlations: (1) each variable with
each other variable regardless of method, (2) different variables with
the same method, (3) different variables with different methods, and

(4) the same variables with different methods. The correlations of the
same variables with different methods are validity coefficients.

There are four requirements for satisfactory validation. The
first is that correlations of the same variable measured by different
methods (validity coefficients) must be significantly greater than zero,
a requirement known as convergent validity. Requirements two, three,
and four are necessary for discriminant validity. Two is that the
validity coefficients must be greater than correlations of different
variables measured by the same method. The third states that validity
coefficients of a given variable must be greater than correlations of
this variable with any other variable measured by any method. The
fourth requirement, which Magnusson (1966) describes as impossible to
maintain rigorously because of the difficulty in judging the effect of
unreliability on large matrices, is that with the same and different
methods, the magnitude of correlation coefficients between different
variables must follow the same pattern.

Cronbach and Meehl (1955) discuss a number of ways of testing
construct validity. One way is to study differences between groups
of individuals who, according to the theory being tested, can be expected
to differ on a given variable. Another method is to measure changes

brought about by an experimental study designed to test the theory. A

third way is through correlations of tests designed to measure the
same variable, provided the correlations are not inflated by variance
due to similar methods. A fourth way is to measure item or subtest
correlations of single factor tests.

Cronbach (1971) points out that most studies of construct
validity lead to intermediate rather than clearcut positive or negative
conclusions, and that the emphasis in construct validation should be
on the strength of each relationship rather than on statistical
significance. He also notes some reasons why high correlations might
not be obtained. One is restriction of range if one is testing groups
in which the individuals do not vary a great deal on the trait being
measured. Correlations can be expected to vary from group to group
both because of restriction of range and because of the number of
variables affecting performance in any given situation.

Hempel (1966) cautions against a strictly operationalist view
of identifying a construct with one way of measuring it. Even in physical
sciences, measurements are made in different ways depending on the
circumstances. He notes, for instance, that "laws and theoretical
principles afford a large variety of ways for measuring distance" (p. 93).
The measurement used and the accuracy required are theoretical decisions.
As theories develop over time, operational criteria are refined and
modified. Hempel suggests an analogy about the relationship between the
theory and operational definitions.

The process might be compared to building a bridge across

a river by putting it first on pontoons or on temporary

supports sunk into the river bottom, then using the bridge

as a platform for improving and perhaps even shifting the

foundations, and then again adjusting and expanding the

superstructure, in order to develop an increasingly well-
grounded and structurally sound total system. Scientific

laws and theories may be based on data obtained by means

of initially adopted operational criteria, but they will
not fit these data exactly (p. 96).

10

Continuing, Hempel emphasizes the relationship of constructs and theory.

Empirical import... is not the only desideratum for
scientific concepts: systematic import is another
indispensable requirement-—so much so that the empirical
interpretation of theoretical concepts may be changed in
the interest of enhancing the systematic power of the
theoretical network. In scientific inquiry, concept
formation and theory formation must go hand in hand

(pp. 96-96).

Cronbach (1971) also cautions against considering constructs
apart from theory.

The most serious criticism to be made of programs of

construct validation is that some of them are haphazard

accumulations of data rather than genuine efforts at

scientific reasoning. Merely to catalog relations

between the test under study and a variety of other

variables is to provide a do-it-yourself kit for the

reader, who is left to work out his own interpretive

theory. Construct validation should start with a

reasonably definite statement of the proposed inter-

pretation. That interpretation will suggest what

evidence is most worth collecting to demonstrate

convergence of indicators (p. 483).
Tests Measuring Empathy

With this discussion of construct validity as background, the
next issue is the status of tests measuring empathy. Kurtz (1970) and
Kurtz and Grummon (1972) correlated six different empathy measures with
counseling processes and outcomes to clarify the construct of empathy and
to determine which operational measures of empathy are most useful in
counseling research. Kurtz notes that few studies compareialternative
ways of measuring empathy. "The state of the measurement of empathy
seems to be one of confusion, with the various approaches developing
independently and with few attempts toward synthesis" (p. 50). This
seems a serious lack, considering the necessity of convergent and

discriminant analysis in establishing construct validity.

11

Kurtz identified four methods of measuring empathy-- (1) pre-
dictive, which involves predicting another's responses, (2) situational,
in which the counselor responds to a standardized situation, (3) per-
ceived, in which clients and counselors use a questionnaire to rate the
counselor's empathy, and (4) tape-rated or judged empathy, which involves
having trained judges rate the counselor's empathy.

Kurtz used six measures of empathy-~Kagan's Affective Sensitivity

 

Scale (a situation test), two predictive tests-- the Interpersonal Check-

 

list and the Kelly Rep. Predictions, Carkhuff's Empathic Understanding in

Interpersonal Processes Scale ( a tape-rating or judged method), and the

 

client and counselor forms of the Barrett-Lennard Relationship Inventory

 

(a questionnaire measure of perceived empathy). In addition, he also used
three measures of counseling process and six outcome measures. Correlations
among the empathy measures, between the empathy measures and the process
outcome measures, and between the process and outcome measures were

examined using a correlation matrix. A least-squares multiple regression
equation was also calculated for each outcome measure to discern with
empathy measures best predicted outcome. There were thirty-one subjects

in Kurtz' study.

The correlation of client-perceived empathy taken after the third
interview was significantly correlated with that taken at the end of therapy
(r=.66. p.<.OOl). Client-perceived empathy after the third interview was
also positively correlated with tape-judged empathy and approached signi-
ficance (r=.31. p.<.10). However, client-perceived empathy taken near the
end of therapy was not correlated with tape-judged empathy (r=.OO). None
of the other correlations between empathy measures approached significance.

When the correlations between empathy and process measures were

investigated, only the relationship of tape-judged empathy and self

12

exploration was significant (r=.47, p<.Ol). None of the process and
outcome measures were significantly correlated. The correlations between
different outcome measures were all positive and many were significant.

Most of the correlations between empathy and outcome measures
were not significant and many were negative. The notable exception was
client-perceived empathy after the third interview, which was significantly
correlated with several outcome measures. Tape-judged empathy was
positively correlated with all outcome measures, but only one correlation
was statistically significant. When multiple regression equations were
calculated, client-perceived empathy was the best predictor, accounting
for 44 percent of the variance on clinical evaluations of the M.M.P.I.
and 30 percent of the variance of the composite outcome measure. Tape-
judged and client-perceived empathy combined produced a multiple
correlation of .54 (p<.02) with the Tennessee total positive score,
accounting for 30 percent of the variance. No other combinations of
empathy measures predicted any outcome score at the .10 level of
probability or less.

Of their data, Kurtz and Grummon (1972) state that the six
empathy instruments, thought to measure a single construct, are in
fact measuring different variables. The one exception is the moderate
relationship between client-perceived and tape-judged empathy. This
latter finding is consistent with the findings of Rogers et al. (1967)
who also found a relationship between tape-judged and client-perceived
empathy.

Kurtz and Grummon suggest that predictive and therapist-perceived
measures are not useful in psychotherapy research. Likewise, they state

that the situation-empathy measure employed (Kagan's Affective Sensitiv-

 

ity Scale), while it may be of value in training, is not useful for

13

research on counseling.
In conclusion, Kurtz and Grummon (1972) indicate that ambiguity
still exists regarding the construct of empathy.

The fundamental question raised by these data is whether
anyone has been able to measure therapist empathy
successfully. With the exception of some correlation between
tape-judged and client-perceived empathy, neither the present
study nor those examined were able to establish construct
validity.... There is little doubt that there is a relation-
ship, and perhaps a very substantial relationship, between
what has been called empathy and therapy outcomes. But
present empathy measures may in fact be tapping other aspects
of the therapist's behavior and the therapeutic relationship
which accounts for these findings (p. 114).

Langer (1972) notes there is confounding of instrument and
person in the measurement of empathy. That is, the Carkhuff Scale is

always used by judges and the client form of the Relationship Inventopy

 

is completed only by clients. In a follow-up of the Kurtz and Grummon
study, Langer attempted to separate the effects of instrument and
rater by having the Relationship Inventory used by both clients and judges.
He also attempted to clarify some ambiguity in the findings of the
Kurtz and Grummon study by obtaining a more complete sampling of tape-
rated empathy. Judge-rated empathy using both the Carkhuff Scale and

the Relationship Inventory were made on the same interviews but by differ-

 

ent judges.
Results of Langer's study indicate that correlations between

client-perceived and tape-rated empathy on the Relationship Inventory

 

were significant (r=.46 and r=.42, p<.05), but those between tape-rated
Relationshipplnventory and tape-rated Carkhuff Scale (r=.ll) were not.
Langer concludes that clients and judges agree about empathy when they
use the same instrument, but that independent judges do not agree when
they use different instruments. He also found no significant relation-

ships between the tape-rated Relationship Inventory and the outcome

 

14

measures. Also, with a larger sample of Carkhuff-rated tapes, no
relationship was found between the tape-rated Carkhuff Scale and the
client-perceived Relationship Inventory. These findings support the
conclusion of Kurtz and Grummon that different empathy instruments are
measuring different variables, and that the construct validity of empathy
remains ambiguous. However, a qualification must be added to the

Langer study. His interrater reliability was only moderate on both

the tape-judged Relationship Inventory (r=.63) and the tape-judged

 

Carkhuff Scale (r=.74). Low correlations imply a large error factor,
and thus the results should be considered tentative.

From these studies, client-perceived empathy emerges as the
best predictor of therapy outcome. The results of predictive studies
of tape-judged empathy and outcome, and the results of the correlation
between tape-judged and client-perceived empathy are mixed. The second
best empathy measure appears to be tape-judged using the Carkhuff Scale.
In the present study, the client form of the Barrett-Lennard and the
Carkhuff Scale are both used to measure the empathy of the students

during the interview posttest.

Empathy and Nursing

In a review of the literature on therapeutic effectiveness of
counseling by nursing personnel, Peitchinis (1972) says that until the
last decade there were few controlled studies of the nurse-patient
relationship. The importance of communication with patients began to
be emphasized in the fifties, with a substantial increase in the volume
of articles in the sixties. Most of these articles were not research
reports, however. Peitchinis has done several correlational studies

relating personality traits of nurses (as measured by the Edwards

15

Personal Preference Scale) to the qualities of empathy, warmth, and
genuineness, the Rogers triad of therapeutic conditions. She also
reviews a number of similar studies on personality traits. Her rather
gloomy conclusion is that nurses are low in therapeutic effectiveness.
She does see hope, though, in the literature and in nursing curricula.
Kalisch (1971) conducted an experiment to develop empathy in
nursing students. This study is of special interest here, because role
modeling was a component of the treatment. Unfortunately, the modeling
was not well defined or separated from the other components. It is
significant, nevertheless, that Kalisch considered modeling of sufficient
importance to include in her study. The other components of the treatment
were didactic training, role playing, and experiential training. The
control groups received lectures and discussions on human behavior.
The experimental groups performed significantly better on four of the
seven dependent variables (significant differences on Truaxu,Accurate

Empathy Scale, helper form of the Relationship Inventory, predictive

 

 

empathy with a patient, and empathy ratings by clinical instructors: no

significant differences on helpee form of the Relationship Inventory,

 

predictive empathy with a classmate, and the Empathy Test for a

 

generalized other.) Six-week follow-up using the Accurate Empathy Scale

 

showed the experimental groups continued to perform significantly better.
This study by Kalisch is worthy of note on several accounts. It is one

of the few experimental studies designed to increase empathy in nursing
students, it measured empathy in several ways, one of which was assessment
of students' clinical performance, and it included follow-up testing,

an element lacking in most studies of empathy.

Empathy training based on the human-relations models of Carkhuff

and Gazda was used by LaMonica, Carew, Winder, Haase, and Blanchard (1976)

16

to increase the empathy level of staff nurses. The training focused on
perceiving and responding with empathy and included both didactic and
experiential elements. A pretest-posttest design was used with one
experimental and two control groups, the second control group being used
to assess the effects of test-retest and time. Nurses in the experimental
group and first control group were those who scored less than 2.0 on

Carkhuff's Index of Communication as measured by the Carkhuff Scale.

 

Posttest results showed a significant increase in empathy in the
experimental group. The authors add, however, that while results were
significant, only three subjects were functioning at a minimally
facilitative level (3.0) and that more training was probably needed
to bring all nurses to an adequate level. Two limitations mentioned
were that performance was not measured in the clinical setting and there
was no long term assessment of effects. Although the treatment did not
include modeling, its importance was acknowledgedirlthe discussion:
"The human-relations-model of helping... should be role modeled by
deans, educators, and administrators.... No one can maintain a helpful
caring attitude in an institution that mitigates against it" (p. 450).
Rosendahl (1973) attempted to explicate the contribution of
empathy, warmth, and genuineness by teachers to the growth in self-
actualization of nursing students. Using a regression equation, the
author related the predictor variables--empathy, warmth, and genuineness
as measured by Traux and Carkhuff's Relationship Opestionnaire to the
criterion variables--time competence and inner-directed support as measured

by subscales of the Personal Orientation Inventory. A relationship was

 

found for inner-directed support but not for time competence. Although
this study does not directly use the modeling of empathy, the idea of

modeling underlies the attempt to use the teacher-student relationship

17

for certain learning goals.

Mansfield (1973) analyzed behaviors associated with empathy.
Six interviews by an expert psychiatric nurse were videotaped and rated
using Truax' Accurate Empathy Scale. Mansfield then categorized and
described the nurse's verbal and nonverbal behavior during the inter-
views. In a similar study, Stetler (1977) analyzed nurse-patient
interactions with the aim of describing communicative behaviors of
empathic and non-empathic nurses. Thirty-two nurses, judged to be high

or low in empathy on the Barrett-Lennard Relationship Inventory, inter-

 

viewed simulated patients. The audiotaped interview data were rated
using an instrument designed by Stetler to analyze verbal and vocal
behaviors. On most of the categories, there were no differences between
the high and low empathy groups. Stetler suggests the factors she
measured are not the ones that contribute to client perception of
empathy. Both of these exploratory studies are of interest in their
attempts to discover what behaviors contribute to empathy. With further
studies of this nature, better behavioral measures as well as improved
ways of teaching empathy might eventually be developed.

This concludes the section on issues and research related to
empathy. The importance of empathy was underscored by a review of some
studies relating empathy to therapeutic processes and outcomes. The
construct validity of tests of empathy was discussed to provide some
insight into the problems of measuring empathy. Finally, some studies
of empathy and nursing were reviewed. Although empathy and other
interpersonal skills are acknowledged by nurses and nursing educators
as important in nurse-patient relationships, the number of studies on
empathy is quite small, a finding supporting the need for more research

in this area.

18

Modeling

The modeling section of this review will discuss the modeling
paradigm, followed by a brief review of the application of modeling to
various intepersonal behaviors and a discussion of research applications

of modeling in teaching empathy.

The Modeling Paradigm

 

The person most responsible for the widespread use of modeling
techniques is Bandura (1977). He says that most human behavior is
learned through modeling and that complex behavior is best learned in
this way. He added a cognitive component to the behaviorist models of
learning and, in so doing, changed the emphasis regarding which variables
are important in the learning and performance of behavior.

Bandura separates the learning of behavior from its performance,
maintaining that stimuli and reinforcers increase the likelihood of
behaviors occurring because of their cognitive interpretation. His
position is not inconsistent with most behaviorists, but he emphasizes
cognitive variables rather than observable behavior. He states
there are four sets of processes or variables involved in learning and
performance--attention, retention, motor reproduction, and motivation.
(See Figure l for a diagram of Bandura's model.)

Attention is basic to learning. The learner must perceive
the significant features of the behavior to be learned, a process
affected by such variables as the distinctiveness and complexity of the
cues, the prestige of the model, and the value of the modeled behavior
to the learner. Incentive or motivation is also important to learning

and can, in fact, overcome individual variations in observers or models.

19

Amm .a .Naap .ataecmm sate emsaae<v

.mmmxpmcm mcwccmm_ _mwoom on» cw mcwccmop chowum>comno mcwcgm>om mommmooca “cocoasou p mgzmmd

pcmaoocowcwmmuepmm
acoEmogoecwmm mzowcmow>

ucmsoocomcwom chcouxm

xomnuood aumcaoo<

m:o_poauocnom
eo cowum>cmmno epom

momcoamom «cocoasou
mo xuwpwnmppm><

macswp_amaau Pauwmsga

mecmogmm Lopez
Pmmcmogmm oppoasxm
cowum~wcmmco m>wpwcmou

meteou UPPOQEAm

acmeoocomcmom omen

pom szuaoocoa

Fo>o4 Femaoc<

mmwuwomqmu zgomcmm
moPHmFLQuomcmgu co>eomno

o:_m> szoppoczm

mocmpm>oca

sumxopneou

oocopm> o>ruooee<

mmmcm>wuocwpmmo

P_=ewpm m=s_meoz

 

mmmmmuomm
4<zomh<>mhoz

mmmmmuoma
onhoaoomamm «Che:

mmmmmuoma onpzmhmm

mmmmmuoma
4<onhzuhh<

20

Retention is the heart of Bandura's model, for it is here
that cognitive processing or coding and organization is included.

In order for observers to profit from the behavior of

models when they are no longer present to provide direction,

the response patterns must be represented in memory in

symbolic form. Through the medium of symbols, transitory

modeling experiences can be maintained in permanent memory.

It is the advanced capacity for symbolization that enables

humans to learn most of their behavior by observation.

(Bandura, 1977, p. 25)

Two elements that enhance cognitive coding or organization are labeling
and rehearsal. Studies of both children and adults show that observers
who code modeled activities with words or labels retain behavior better
than those who do not. For example, in a study on the learning of hand
signals for the deaf, Gerst (1971) notes that verbal labels and imaginal
responses were both superior to concrete verbal descriptions of modeled
responses. Verbal labels were also found to be superior for retention.

Friederich and Stein (1975) studied the effects of verbal
labeling and role playing in increasing helping behavior in kindergarten
children. Labeling was found to be more effective for verbal measures
of learning and role playing for nonverbal measures; a sex by measure
interaction was also noted. The authors suggested that labeling may be
more important in learning, while role playing is more important for
performance.

Bandura (1969) states that covert rehearsal or practice is often
as effective as overt rehearsal and is more effective for highly
symbolic tasks. In a review article, Friedman (1972) concurs on the
effectiveness of covert behavioral rehearsal. In a study carried out
to enhance assertive behavior, Friedman found that covert was as

effective as overt rehearsal. He also notes that rehearsal aided

retention of assertive behavior on two-week follow-up, but adds that

21

evidence about maintenance of behavior is sparse. When Jeffery (1976)
compared overt and covert rehearsal, he found that the highest level
of learning was achieved by covert followed by overt rehearsal.

Because of the importance of labeling and rehearsal to learning
through modeling, both of these elements are included in the present
study. Bandura (1972) suggests that retention is aided not only by
coding and rehearsal, but by interaction of these components. Their

effects will, therefore, be studied both singly and in combination.

Modeling and Interpersonal Behavior

 

Before focusing on the specific application of modeling to
teach empathy, some studies using modeling to teach various interpersonal
behaviors will be reviewed. Modeling has potential in teaching many
clinical and counseling skills; the studies reviewed here exemplify some
of these specific applications.

A series of studies by Ivey, Normington, Miller, Morrill, and
Haase (1968) used video modeling as one component of a teaching technique
they call microcounseling to teach attending behavior, reflection, and
summarizing of feeling to beginning counselors. Microcounseling involves
reading prepared materials, viewing video models, cue discrimination,
videofeedback, supervision with feedback, and role playing. The training
program was designed to model and reinforce the desired behavior, but
the modeling component was not separately evaluated. Interview posttests
revealed significant differences after one to two hours of training.

In what seems to be the first study to examine the effects of
modeling in teaching counseling skills, Eisenberg and Delaney (1970)
compared modeling and reinforcement, in combination and separately,
to increase counselor tacting response leads (a counselor statement that

elicits specific rather than abstract descriptions from clients). With

22

subjects responding to videotaped client statements, modeling but not
reinforcement was found effective in increasing the target behavior.
When simulated clients were interviewed in a test of transfer, however,
the target behavior did not occur frequently enough to allow comparison.
Rank, Thoresen, and Smith (1972) compared modeling, videolecture,
and group discussion in increasing counselor trainees affective responses
in a group. The verbal interaction of each group member was measured
thirty minutes prior to and thirty minutes following the treatments.

Responses were coded using the Behavioral Interaction Description System

 

on the categories of delivery, content, process, and focus. Pre- and
posttest comparisons revealed significant differences for the modeling
treatment on all four dimensions and for the group treatment on the
focus dimension. No other comparisons were significant.

In a project designed to increase interpersonal group skills of
undergraduate volunteers, Rappaport, Gross, and Lepper (1973) studied
the relative effectiveness of modeling, sensitivity training, no train-
ing, and instructions. Each of three groups received modeling by
videotape, fourteen hours of sensitivity training, or no treatment.

For testing, half of the groups received general instructions and half
specific instructions about what to discuss in the group. Group skills
were measured using a technique called the Group Assessment of Inter-

personal Traits. Under general instructions, modeling and sensitivity

 

trained subjects performed better than controls, but with specific
instructions all groups performed equally well.

Kuna (1975) examined the effect of lecturing, reading, and
modeling on undergraduate students' restatement responses. Subjects
responded to audiotaped client statements and were rated on restatement

responses. Lecture and lecture-reading groups performed significantly

23

higher. The addition of modeling did not result in a significant
increase in restatement responses.

Using modeling and attentional variables, Eskedal (1975) studied
counseling students' acquisition of cognitive principles. Learning was
measured using a true-false test devised for the experiment. Analysis
of variance revealed differences between groups to be significant.
Scheffe post-hoc procedures showed the two modeling groups superior to
the two controls, the modeling-attentional group superior to the modeling-
only, and modeling-only superior to control-modeling. Modeling-only was
not significantly different from the attentional-control group.

Wallace, Horan, Baker, and Hudson (1975) compared lecture,
modeling, and feedback in teaching decision-making counseling to graduate
students. Treatments were administered cumulatively, so that group one
received a lecture, group two the lecture plus a videotaped model, and
group three both of these plus feedback. Group three performed
significantly better on an interview posttest than groups one or two.

The differences in the lecture and modeling groups were not significant.

In a two-by-two design, Chasnoff (1976) compared the effects of
videotaped modeling and ambiguity tolerance on interview behavior. Low
ambiguity tolerance has in the past been associated with ineffective
counseling behavior. Counseling students were selected for high or low
ambiguity tolerance and half of each group was assigned to a modeling
or no-modeling condition. As measured in an interview posttest by the
Counselor Interview Behavior Rating_Scale, there was a significant effect
for modeling but not for ambiguity tolerance. The author suggests that
rather than trying to select counselors on personality characteristics,

more effort be devoted to devising appropriate learning experiences.

24

In summary, some general observations can be made about these
eight studies. The modeling treatment was effective in most of them,
even though they employed a variety of dependent variables, including
both group and individual skills. Modeling was also effective as
measured by a variety of methods--written tests, responses to audio and

video client statements, and interviews with simulated clients.

Modeling and Empathy

 

Given that modeling appears to be effective in teaching various
interpersonal skills, the next issue is the efficacy of modeling in
teaching empathy. Again, there are a number of studies that examine this
question.

Using six treatment combinations, Payne, Winter, and Bell (1972)
studied the relative effectiveness of different supervisory styles and
modeling versus no modeling. Subjects were undergraduate psychology
students. The six treatment groups were techniques (didactic)
supervision, counseling (experiential) supervision, placebo-modeling,
placebo-no modeling, control-modeling, and control-no modeling. The
placebo treatment was a type of supervision in which the supervisor
discussed the client's psychodynamics. The modeling treatment was
provided on audiotape and consisted of examples of high and low empathy
with cues from a commentator. After listening to the modeling tape, each
subject to have supervision was assigned a supervisor and training
sessions were held. The experimental data consisted of subjects'
responses to taped client statements. Responses were judged by two
independent judges using a seven-point empathy scale based on those of
Traux and Barrett-Lennard. Analysis of the data showed significant

differences in the groups receiving modeling and those receiving

25

techniques-oriented supervision. No significant improvement was noted
in other groups.

In a follow-up of the.Payne, Winter, and Bell study, Payne, Weiss,
and Kapp (1975) assessed the relative contributions of modeling, didactic,
and experiential factors in training by examining each element separately
and in combination. They had six groups of undergraduate psychology
students receiving different combinations of treatments-~mode1ing-didactic,
no modeling-didactic, modeling-experiential, no modeling-experiential,
modeling-control, and no modeling-control. Subjects responded to clients'
tape-recorded statements and were rated on Carkhuff's Empathy Scale.
Results indicated a significant difference between the two methods of
supervision, with didactic groups being higher. The modeling groups were
also significantly higher than the no modeling. The modeling-didactic
group was significantly higher than the no modeling-didactic group. The
no modeling-experiential group was not significantly different from the no
modeling-control. The authors suggest that modeling and didactic super-
vision might be especially appr0priate for beginning students.

Dalton, Sundblad, and Hylbert (1973) investigated the effects
of a social learning program on the acquisition of empathy skills by
students in undergraduate rehabilitation education. The first group
received the learning program, which consisted of an advanced organizer,
an expert counselor model, and a covert practice session, all presented
via videotape. Group two, a treatment control, read about counseling
behavior. Group three was a no-treatment control. Subjects' written
responses to the Carkhuff Helpee Stimulus Expressions were rated using
the Carkhuff Empathy Scale. Performance was measured at the end of the
program and one month later. Analysis of variance on the first test

revealed significant differences among the groups. Post-hoe "t" tests

26

revealed significant differences between group one and each of the other
two groups, but no difference between groups two and three. The same
results were obtained on the one month follow-up, indicating not only
learning but also retention of empathic responses.

In examining the effects of modeling and instructions on
clergymen counselors' empathy, Perry (1975) combined high, low, and
no empathy audiotaped models with an instructions- no instructions
condition to form a 3x2 design. Empathy was measured in two phases using
Carkhuff's Empathy Scale. In Phase 1, subjects wrote responses to taped
client statements before and two times during treatment; in Phase II, they
interviewed a simulated client. In Phase I, the main effect of modeling
was significant; main effects of instructions and trials were not.
In an analysis of modeling and instructions only, modeling was signifi-
cant but the instructions condition was not. In Phase II, there were
no significant differences between groups. The researcher quotes the
subjects as saying they reverted to their usual methods of counseling
during the interviews, rather than using what they learned from the tapes.

Stone and Vance (1976) conducted a study to determine if there
are consistent factors, either singly or in combination, that facilitate
the learning of empathy by undergraduate volunteers. The training
components--modeling, instructions and overt rehearsal--were used to
form eight treatment combinations: instructions, modeling, and
rehearsal; instructions and rehearsal; modeling and rehearsal; instruc-
tions and modeling; rehearsal only; modeling only; instructions only;
and a no-treatment control. The training sequence, which was videotaped,
included a pretest, one-week delay, training and immediate posttest,
two-week delay, and interview measure. The written pre- and posttest

scores consisted of subjects' performance on the Communication Index as

 

27

measured by the Carkhuff Empathy Scale. Interviews were videotaped

and judged for empathy, also using the Carkhuff Scale. On the written
posttest, the scores of the seven treatment groups all increased
significantly, but interpretation of the four-way design is complicated
by some disordinal interactions. Instructions seemed the most
facilitative condition on posttest ratings. Scores on interviews showed
that combinations of training methods increased performance, while groups
receiving a signal training method tended to have scores comparable to
pretest scores. Combination groups were significantly higher than

single treatment groups in the Scheffe post-hoc analysis. Analysis of
interview data revealed that groups receiving modeling performed
significantly better than no-modeling groups. Instructions and rehearsal
also facilitated performance, but not significantly.

Dalton and Sundblad (1976) compared modeling to systematic
training as conceived by Carkhuff. Seventy-five resident-hall advisers
were assigned to one of three conditions, modeling-systematic training,
systematic training, and the resident-hall-adviser training program.

The ninety-minute video modeling treatment consisted of an advanced
organizer, modeled interviews, covert practice, and directions for the
posttest. Only group I saw the entire modeling tape. All subjects were

posttested at this point with Carkhuff's Helpee Stimulus Expression, a

 

written posttest that was rated using the Carkhuff Empathy Scale. Sig-
nificant differences were found only for group 1. Groups I and II

then received ten hours of systematic training over five days, while
group III received the resident-hall-adviser program. All groups were
then retested following the same procedure. Significant differences were
found for groups I and II. The authors comment that, while modeling is

effective in quickly raising the level of empathy, the levels were lower

28

than that portrayed by the model. They suggest individuals with
initial low levels of empathy need a longer time to reach acceptable
levels. They also note that systematic training is not superior to the
briefer modeling treatment.

In examining the effects of instructions and modeling on subjects
low in communication skills, Uhlemann, Lea, and Stone (1976) selected

fifty subjects with low scores on Carkhuff's Communication Index and

 

randomly assigned them by sex to one of five treatment groups: instruc-
tions, modeling, instructions plus modeling, modeling plus instructions,
and control. The instructions condition was a nine-minute videotape
describing reflection of feelings. The modeling was an eleven-minute
videotape of positive and negative examples of empathy. The control was
two minutes of videotaped instructions. After exposure to treatments,
subjects each interviewed a simulated client for fifteen minutes, then

again completed the Communication Index. Interviews were scored for

 

percentage of reflection of feeling statements and empathy (Carkhuff's
Scale). Analysis of reflection of feeling data revealed significant
differences for instructions and instructions plus modeling. Empathy
ratings were significant for both combined conditions. No significant
differences were found for modeling alone. The authors suggest that the
modeling treatment may have been too brief for learning to occur.
Comparing the effect of three supervisory styles--modeling,
feedback, and experiential--Ronnestad (1977) had forty graduate
counseling students respond to videotaped clients statements on three
occasions. Between the response sessions, students were assigned to one
of three supervisory conditions or a no-supervision control. For the
modeling condition, the supervisor responded to the taped client

statements. In the feedback condition, the Supervisor rated subjects'

29

responses using Carkhuff's Empathy Scale. While in the experiential
condition the supervisor explored the feelings of the subject as well

as those of the client. Subjects' responses were taped and were rated
using Carkhuff's Scale. Significant differences were noted for modeling
and feedback, with modeling being higher. There was an interaction of
treatment and sex; males scored higher on the feedback condition. The
author speculates that experiential supervision may take longer to
produce significant improvements.

Gulanick and Schmeck (1977) compared all possible combinations
of modeling, praise, and criticism to assess their single and combined
effects on empathy. The modeling treatment consisted of audiotaped
responses to client statements. Praise and criticism were given by a
supervisor for specific elements that contributed to or detracted from
empathic responses. Subjects (female undergraduate psychology students)
responded to three sets of taped client statements. Between the
responses they received the treatments. Responses were rated by two
pairs of raters using a modification of the Carkhuff Empathy Scale
(0.5 units). Significant results were found for modeling and stage of
training, with a modeling by stage interaction. No significant
differences were noted for praise or criticism. The authors' explanation
of their findings is interesting. They suggest that empathy is a concept
learning task; that is, that modeling facilitates subjects' learning a
general class of responses called empathy, which can then be used
appropriately depending on the situation. They feel the particular way
they administered praise and criticism prevented subjects establishing a
relationship among instances and acquiring a general concept. They also
point out that other experiments that did show significant effects for

praise and criticism were conducted differently, with the feedback given

30

all at once. Two studies of criticism, one significant and one not,
also fit their framework. The significant study gave 50 percent more
feedback. The authors state that concept-learning experiments support
the longer time needed for subjects to learn from negative instances.

Of the nine studies using modeling to teach empathy, all but
one yielded significant differences on immediate posttests, and this
one had only eleven minutes of modeling. This same study also chose
subjects low in communication skills (Uhlemann, Lea, & Stone, 1976).
Modeling seems, then, to be a powerful treatment, showing results with
from thirty to ninety minutes of exposure. In one case it was as
effective as ten hours of systematic training (Dalton & Sundblad, 1976).

Only two studies did follow-up posttests and the results are
conflicting. Dalton, Sundblad, and Hylbert (1973) found significant
differences for modeling four weeks after treatment, while Stone and
Vance (1976) noted significant differences for modeling and instructions
combined, but not for modeling alone after two weeks. The studies differ
in a number of ways. The Dalton, Sundblad, and Hylbert modeling condition
was sixty minutes in length, while Stone and Vance's was only twelve
minutes long. Also Stone and Vance's follow-up was an interview, not a
written posttest. Additional study of maintenance of behavior is needed
and will be included as a component in this research project. Maintenance
of behavior over time and not immediate performance is, after all, the
goal of teaching programs.

The dependent variable is usually tested by having subjects
respond to taped or written client statements. Only three studies
(Perry, 1975; Stone & Vance, 1976; Uhlemann, Lea, 8 Stone, 1976) also used
interviews as posttests, and none of these found significant differences

for modeling alone. Two of these used eleven- and twelve-minute

31

treatments, and the third did not state the time, but nevertheless
failure to demonstrate transfer to a realistic situation is a serious
problem. As with maintenance of behavior, transfer of learning is a
crucial goal of education. In an attempt to assess whether modeling
is effective in helping students demonstrate empathy during an actual
interview, an interview posttest using simulated clients will be
employed both immediately after treatment and for follow-up.

In summary, Chapter II has reviewed the literature of
the two areas relevant to the present research--empathy and modeling.
Under empathy, the issues included were the definition of empathy and
its importance in helping relationships, the construct validity of
tests purporting to measure empathy, and the status of research on empathy
in nursing. In relation to modeling, the first issue discussed was
the modeling paradigm, especially the importance of labeling and
rehearsal to retention. Studies of modeling and various interpersonal
behaviors were reviewed. Finally, studies using modeling to teach

empathy were discussed.

CHAPTER III

METHODS
This chapter includes a discussion of design, hypotheses,
measurement of the dependent variable, subjects, and procedures. Under
the procedures section, the aspects included are production of the
videotapes, subject recruitment, assigment, and attrition, training of
simulated clients, collecting the data, and rating the audiotaped

interviews.

Design

The experimental design combines three conditions--modeling (M),
labeling (L), and rehearsal (R)--to form four experimental groups.
There is also a fifth group, a no-treatment control (C). The conditions
are discussed in detail in the procedures section of this chapter, but
briefly they are as follows. The modeling condition includes segments
of videotaped interviews portraying empathy and lack of empathy;
labeling is verbal labeling by a narrator pointing out behaviors that
contribute to empathy or the lack thereof; and rehearsal is further
interviews with pauses between the client's and nurse's statements to
allow the viewers to construct their own responses. Since each student
was posttested twice, immediately after treatment and three weeks later,
a repeated measures design with time as a variable was employed.

Table 1 illustrates the design

32

33

Table 1

Experimental Design and Number of Subjects
in Each Group

 

 

Group Condition Time 1 Time 2

 

l C l3 l3
2 M 11 ll
3 M+L 11 ll
4 M+R 11 10
5 M+L+R 11 11

Total 57 56

 

Initially, there were fourteen subjects assigned to each group.
After some students dropped out, the distribution of subjects is given
in Table 1 under Time 1. Between Time 1 and Time 2, only one student

dropped out, leaving a total of a fifty-six subjects.

Hypotheses

 

The purpose of the study was to test the effectiveness of
modeling in teaching empathy to nursing students. Within this general
purpose, two hypotheses were investigated.

1. The treatment groups will be rated significantly
higher on empathic behavior than the control group.

a. The modeling-labeling and modeling-rehearsal groups
will each be rated higher than the control group.

b. The modeling-labeling-rehearsal group will be rated
higher than the other two treatment groups.

2. All groups will retain their level of empathic
behavior on follow-up tests three weeks later.

The predicted trend, then, for the five groups was that each component

added to the experimental treatment would improve the performance of the

34

group. This expectation is consistent with the research findings on
modeling reviewed in the last chapter.

In addition to the two hypotheses about modeling and empathy,
the following hypothesis about construct validity was investigated.

3. The three measures of the dependent variable will
be significantly correlated with each other.

This hypothesis, although not the primary focus of the research, was
considered important in relation to the construct validity of the
instruments used. In light of the discussion of construct validity

in the last chapter, it is a crucial question to address.

Measurement of the Dependent Variable
The dependent variable, empathy, was measured in three ways.

Interviews were audiotaped and rated using the Carkhuff Empathic Under-

 

standing in Interpersonal Processes Scale, hereafter referred to as the
Carkhuff Scale; simulated clients completed the Barrett-Lennard
Relationship Inventory; and students completed a written test, the

Empathy Test, constructed by the researcher. Parallel forms of the

 

Empathy Test were constructed, with Form I being administered at the

 

first testing and Form II at the second. (See Appendices A-D for
copies of the instruments used in the study.)

The Carkhuff Scale and the Barrett-Lennard were both discussed
in Chapter II. The training of raters and interrater reliability for
the Carkhuff Scale are included in the Procedures section of this chapter.

Reliability of the Barrett-Lennard and development of the Empathy Test

 

are discussed here.
Barrett-Lennard (1962) provides corrected split-half reliability
coefficients for each of the five subtests of the Relationship Inventory.

For the empathic understanding subtest, the realiability coefficient

35

15.86 based on a sample of forty-two therapy clients. Barrett-Lennard
also computed test-retest reliability for each subtest on a sample of
college students and obtained a reliability coefficient of .89 for the

empathic understanding subtest.

Empathy Test.

 

Two twenty-four item tests on empathy were constructed by the
researcher to be administered at the two posttest sessions. Each test
consisted of twelve true-false items and twelve application items.

In six of the application items, students were given a situation and
asked to choose the most empathic response of two possible responses
and in six they were to choose the least empathic response.

The true-false items were based on Rogers' definition of empathy
and on principles of empathy demonstrated in the videotapes seen by
the experimental groups. In the application items, students were given
both positive and negative examples of empathy, which was consistent
with the approach used in the videotapes of modeling both high and low
empathy.

It was not possible to pretest the Empathy Test on a group of

 

nursing students, since there was no appropriate group available to
the researcher other than those participating in the study. However,
several nursing faculty members were administered the tests, and
feedback was obtained regarding content and interpretation of the items.
Reliability coefficients (Kuder-Richardson 20), mean item
discrimination and difficulty indices, and standard error of measurement
for the two tests are reported in Table 2. The index of difficulty is
defined as the percentage of the total group marking a wrong answer or

omitting an item. The index of discrimination is the difference between

36

the percentage of the upper and lower groups each marking the correct

answer, using the upper and lower 27 percent to constitute the groups.

Table 2

Item Analysis Data for Form I and Form II
of the Empathy Test

 

 

 

Test Statistic Form I Form 11

Kuder-Richardson Reliability .24 .26
#20

Standard Error of Measurement 1.64 1.89

Mean Item Difficulty 27 33

Mean Item Discrimination 18 21

 

Since reliability and mean item discrimination indices were
both rather low, a second item analysis was carried out discarding items
with an item discrimination index of less than twenty in an attempt to
increase reliability. This reduced the number of items to eleven for
Form I and fourteen for Form II. This second item analysis is given
in Table 3.

As some improvement in reliability was obtained in the second
item analysis, students' scores from the second rather than the first

item analysis will be used when data are analyzed.

37

Table 3

Second Item Analysis on Form I and Form II of the
Empathy Test

 

 

 

 

Test Statistic Form I Form II
Kuder-Richardson Reliability .34 .27
#20
Standard Error of Measurement 1.39 1.70
Mean Item Difficulty 37 46
Mean Item Discrimination 38 32
Subjects

The participants in the study were junior and senior students
enrolled in the baccalaureate program of the School of Nursing at a large
midwestern university. The school enrolls students in the nursing major
in their SOphomore Year. As this study was carried out in the winter
term, junior students had had a minimum of four clinical nursing courses,
while senior students had at least seven such courses. Thus, all students
had some experience relating to patients, a factor of some importance
in the study. Since the interview posttest required students to interact
with a simulated client, the students' experience and comfort in doing
so might affect the outcome of the study.

The School of Nursing enrolls approximately one hundred students
to the nursing major each year, of whom at least 95 percent are women.

All the volunteers for this project were women students. Except for
being mostly female, the characteristics of students enrolled in the
nursing program resemble closely the characteristics of other undergraduate

students at the university; the majority are unmarried and are eighteen

38

years of age when admitted.

The School of Nursing admits as many qualified students to the
nursing major in the sophomore year as resources permit. There are,
however, three to four applicants for each available position. The
school reviews all applicants with a minimum grade-point average of 2.5
(on a 4.0 scale) and a minimum of 2.2 in science. The average grade-
point average of students actually admitted is 3.3 to 3.4 The competi-
tion for admission, it appears, results in admission of a group of
students who are high academic achievers.

It is difficult to say, for purposes of generalization of
results, what differentiates students who volunteered for this project
from those who did not. On a questionnaire completed at the end of
the study, most students indicated they participated because they hoped
to learn something and/or were interested in research. Although the
students were paid for their participation, this was a motivating factor
for some, but it was not for all. It seems, then, that students who
participated hoped to learn something about relating to patients by
doing 50. One cannot, however, conclude that those who did not
participate were not interested in learning. Why some students did

not volunteer is an unknown factor in this study.

Procedures

 

In this section, the topics discussed are videotape production,
subject recruitment, assignment, and attrition, training of simulated

clients, data collection, and rating of the audiotaped interviews.

Videotape Production
There were three components in the videotapes that, in various

combinations, formed the four experimental treatments discussed in the

39

design section of this chapter. The three components--modeling, labeling,
and rehearsal--will be discussed in detail, since they constitute the
Operational definition of the independent variable.

Modeling. The modeling component of the experimental treatment
consisted of three nurses interviewing simulated clients. The nurses
were faculty members of the School of Nursing at the university, all with
master's degrees in nursing, two in psychiatric-mental health and
one in public—health nursing. The nurses were selected for their ability
to convey empathy during an interview while being videotaped.

The three persons who played the role of clients were selected
with a view to varying sex, ethnic background, setting, and type of
problem. This was done to help students generalize, that is, to see
empathy as necessary and useful in relating to clients regardless of
the circumstances of the interaction.

The first client was an abusive mother who was visited in her
home by a public-health nurse following a referral from a local hospital.
The second client was an East Indian woman who played the role of the
wife of a graduate student. The client was hospitalized for various
somatic symptoms that her physician believed to be emotional in origin;
she was being visited by a mental health nurse for an assessment of her
situation and for possible outpatient counseling by the nurse. The
third client was a young man with a chronic kidney disease who was on
dialysis and who had been seeing the nurse as an outpatient for a year
to discuss adjustment to his life situation. The main problem he was
currently experiencing was depression and suicidal thoughts following
the death of a friend who had received a kidney transplant.

Scripts were written for each of the three situations giving

details about the client's problem and the general course the interview

40

should take. A line-by-line script was not used. Rather, nurses

and clients were instructed in the role they were to play and allowed

to develop some of the details themselves. One or two practice sessions
were held with each of the pairs prior to videotaping until the
participants felt comfortable and the researcher was assured they were
ready for taping. Since all the simulated clients had been in videotapes
previously, the training was relatively easy.

Each interview was videotaped twice, once with the nurse
portraying empathy and once with her portraying lack of empathy. These
two portrayals were called high empathy and low empathy, respectively.
The nurses were given general instructions about how to show high and low
empathy, but were allowed to use their natural style in carrying out the
instructions. This gave some variety to the ways of being empathic,

a factor considered important if students are to generalize. This
approach is consistent with Bandura's theory (1977) that what observers
learn from models is not specific behaviors, but rules that later guide
their behavior. The use of high and low empathy, or positive and negative
examples, was based on other studies that used this approach (Payne,
Winter, & Bell, 1972; Payne, Weiss, & Kapp, 1972; Stone & Vance, 1976;
Uhlemann, Lea, & Stone, 1976) and also on Gagné's (1977) suggestion

that concepts be taught by giving positive and negative examples of the
concept.

The guidelines to the nurses for high empathy included an
explanation of Rogers' definition of empathy, with instructions to be
empathic but to do so using their natural interview style. The guide-
lines for low empathy gave some examples of non-empathic behavior, such
as changing the subject and not giving one's total attention to the

client.

41

When the modeling segment of the videotape was edited, two
interviews from each nurse-client pair were used, one with high empathy
and one with low empathy. The modeling segment began with an introduction
and a definition of empathy. The interviews were given written labels
of high empathy or low empathy, with narration to give any necessary
details regarding the circumstances of the situation. Other than the
written labels, no explanation was given of why the empathy was high
or low. Such explanations were given, however, in the labeling component
of the videotape. (See Appendix E for the narration accompanying the
videotape.)

Labeling. The labeling component of the experimental treatment
consisted of re—showing short segments of the modeled interviews, with
a narrative explaining which specific behaviors of the nurses were high
or low in empathy. The exact narration is given in Appendix E.

Table 4 indicates the length of the modeling, labeling, and
rehearsal components of the videotape, as well as the length of
each section within each component.

Rehearsal. In the rehearsal component of the videotape, which
consisted of further interview segments, there was a ten-second pause
following each client statement, with instructions to the students to
mentally construct their own responses to the clients. (The complete
instructions to the students are contained in the narration in Appendix E.)
This rehearsal technique is based upon that used by Dalton, Sundblad,
and Hylbert (1973) and Dalton and Sundblad (1976). During this part
of the tape, the camera was focused only on the clients and the video
portion of the tape continued to roll during the ten-second pauses to
create a realistic effect as the students responded mentally to the

clients.

42

Table 4

Length of Videotape Components
and Length of Each Section Within Each Component

 

 

Component Length of Time
in Minutes and
Seconds*

 

1. Modeling 35:53

Interview I --High Empathy
Low Empathy

Interview II --High Empathy
Low Empathy

Interview III--High Empathy
Low Empathy

2. Labeling 18:12
Interview I --High Empathy Excerpts :16

0105 #01 0103
co
‘0

Low Empathy Excerpts :22

Interview 11 --High Empathy Excerpts :36

Low Empathy Excerpts ;15

Interview III--High Empathy Excerpts :32

Low Empathy Excerpts :13

3. Rehearsal 22:26

Interview I
Interview II
Interview III

4. Conclusion

--' \INON
no
N

:19

 

*The length of the sections within each component does not equal the
length of the component because the length of each section includes only
interview time, not narration.

‘43

Supject Recruitment, Assignment, and Attrition

The subjects were volunteer junior and senior students enrolled
in nursing at the university. They were recruited by the researcher, who
went to each of the three junior and three senior nursing courses to
explain the project and ask for volunteers.

Prior to recruitment, approval was obtained from the University
Committee on Research Involving Human Subjects to carry out this project.
The consent form, which was approved by this committee, was used to
explain the project to students. (See Appendix F for a copy of the
consent form.) Students who wished to participate signed the consent
form at the time the researcher visited class and were randomly assigned
to one of the five groups.

Random assignment was carried out by having students draw cards
numbered one through five. Assignment was randomized within each of
the six nursing courses rather than within the total group of subjects.
This was done to equalize systematic differences that might exist within
groups because of class level or learning experiences in a particular
course.

The researcher wanted to obtain seventy-five volunteers in order
to have fifteen in each of the five groups. Seventy students actually
volunteered. Between one-third and one-half of the students in each
course volunteered, except for maternity nursing, from which only two
volunteers were obtained. Students and faculty see this course as making
more demands on students' time, which probably explains the low response
rate.

Fourteen of the seventy volunteers withdrew from the project,
all but one before it began. In general, there were two reasons for

attrition-—illness and other demands on the students' time.

44

Training Simulated Clients

 

Most of the persons hired to play the role of simulated clients
had prior experience in playing client roles. Each person received about
two hours of training prior to participating in the project.

Simulated clients were male and female college students whose
ages ranged from twenty to twenty-seven. They were instructed to come
to the interview prepared to discuss a personal problem with a nurse.
The problem was to be one with which they were thoroughly familiar so
they could portray it realistically and give any necessary details. It
was suggested they use a problem experienced by a close acquaintance or
by themselves in the past, but not a current problem they might be having.
They were also instructed to respond to the nurse as if the problem were
real. For example, if they felt understood, they would probably talk
freely, while if they did not feel the nurse accepted them, it might be
difficult to talk.

Another aspect of training was discussion of the professional
atmosphere of the study. Simulated clients were asked to maintain their
roles at all times and not to have any contact or discussion with the
students that was outside this role. Following the rates set up by the
Clinical Simulation Laboratory, simulated clients were paid $4.00 per
hour for their participation in the study.

After the project had begun, the researcher monitored the audio-
taped interviews to ascertain whether the role playing was being carried

out appropriately. Further training was done when necessary.

Data Collection

 

Data were collected from each student at two different times. The

first time, the student viewed a videotape (unless in the control group),

45

took the written Empathy Test (Form I), and interviewed a simulated

 

client. The second time, three weeks later, the student took Form II

of the Empathy Test and interviewed a different simulated client.

Each interview was audiotaped for later rating using the Carkhuff Scale.
Immediately following each interview, the simulated client completed the

Barrett-Lennard Relationshipylnventory.

 

Data collection was carried out in a facility that has the
capability for videotape playback as well as for interviewing. Students
were scheduled during times when the facility was available, which was
three half-days each week. For each testing session, the time necessary
to accommodate fifty-six students was two-and-one-half weeks. Thus the
total data collection period was five to six weeks. Students viewed the
videotape in groups of from one to four. However, since there was no
discussion of the tape and students were taken to the interview room
immediately after completing the written test, each student is considered
to have received the experimental treatment independently. Students were
also asked not to discuss the project with their classmates.

Written directions were given to students at the beginning of
each session. (See Appendices G and H for specific directions.) The
researcher reviewed what was expected of the students at the beginning
of each step in the procedure. Questions pertaining to the directions
and the procedure were answered. Questions regarding the purpose and
nature of the experiment were not answered until the end of the
second testing.

Directions were designed to equalize conditions for all students.
Students were asked to learn as much as possible from the tape, but not
to take notes. As already mentioned, they were also asked not to discuss

the videotape or other aspects of the project.

46

The purpose of the directions regarding the interview was to give
the students a focus; that is, they were told the simulated client
wanted to discuss a personal problem with them, and that they were to
try to understand the client's problem and how he or she felt about it.

Students did not know, before coming to the first session, how
many treatment groups there were nor to which group they were assigned.
Students in the control group were told at the beginning of the session
that they were in the control group and would not be seeing a videotape.

Students were also asked to practice what they learned about
empathy between the first and second sessions. This instruction was
given to help counteract the effects of forgetting that might occur in
the three-week interval between sessions.

At the end of session two, students were asked to complete a
Feedback Sheet (see Appendix I) which was used to determine motivation
for participation in the study and whether they carried out the directions
as requested. Students were also paid $10.00 for their participation as
agreed by the researcher. In addition, questions about the project were

answered at this time.

Rating the Audiotaped Interviews

 

There were two aspects to rating the audiotaped interviews,
training of raters and establishing interrater reliability. Training
was carried out using the Carkhuff Scale (see Appendix A) plus some
special rules adapted from Steinitz (1976). (See Appendix J for Rg1g§_

for Rating Accurate Empathy in Special Cases.) The rules were guidelines

 

to deal with certain situations, such as rating questions of clarification

and expression of feelings by the therapist.

47

Since both raters had previous experience as therapists as well
as experience using the Carkhuff Scale, training was relatively easy.
The first rater had an M.A. in counseling, while the second had two
years of graduate work in clinical psychology.

Training began with a discussion of the Carkhuff Scale and
the meaning of the different points on the scale, starting with level
three or what Carkhuff calls the interchangeable response. The rules for
special cases were also examined. After a brief discussion of these
tools, a practice tape was rated. For purposes of computing reliability
during training, four three-minute segments were rated. The specific
procedure was to rate each response of the nurse and average responses
for each three-minute period. Reliability, which was computed using
Ebel's (1951) intraclass correlation coefficient, was .63 on the first
tape.

The raters then listened to the tape again, discussing it
response by response to reach better agreement on ratings. The
researcher listened to the discussion, but her participation was minimal.
Her role was to ascertain that one rater was not dominating the other
and to express her opinion if she did not agree with the raters.

Following the same procedure, the reliability on the next six
tapes was .57, .60, .88, .87, .99, and .99. Since a definite training
effect was noted after seven practice tapes, each rater was given several
tapes to rate independently. Both raters also rated some of the same
tapes so reliability could be checked. A severe decrease in reliability
to .06 was noted. While the reason for this loss of reliability cannot
be definitely determined, the researcher believes it was due to the
fact that the task of rating tapes is boring, and concentration is more

difficult working alone than working together. Also, the fact that the

48

raters were working in their homes probably made attention to the
task somewhat more difficult.

Because of these problems, the rating procedure was changed.

The original plan had been to have each person rate half the tapes and
do periodic reliability checks. The new procedure was to have both
raters rate all the tapes and use an average of the two ratings to
compute each subject's score on the Carkhuff Scale. The raters listened
to the tapes together, but each did her ratings independently.

Using the new procedure, reliability for the next six sessions
was .93, .80, .99, .99, .94, and .99. The researcher sat with the
raters for a few days to ensure that reliability was maintained. No
further training was necessary. Reliability appeared to be restored
immediately with the new procedure.

The procedure used in rating the interviews was somewhat
different from that used in training. The raters listened to the first
five minutes of the interview but did not rate this segment. The
purpose of doing this was to familiarize them with the simulated client's
problem. While some rating procedures use as little as three statements
by the therapist and client to rate therapist empathy, it seems to this
researcher that therapist empathy can best be determined if one under-
stands the client's situation. In other words, the context of both
client's and therapist's statements must be considered.

After listening to the first five minutes, each nurse statement
in the next ten minutes was rated. From all these ratings, an average
for each rater was determined, and from the two averages, a score for
the tape.

Raters listened to five tapes a day, which took about two hours

including breaks. Of these five tapes, four were randomly selected to

49

compute reliability. Reliability was checked every day for nine days,
and every other day for the remainder of the rating period. Average
reliability over all ratings was .97.

This concludes the discussion of methods used in the study.
To summarize, the aspects included were design, hypotheses, measurement
of the dependent variable, subjects, and procedures. Under the latter,
production of videotapes, subject recruitment, assignment, and attrition,
training of simulated clients, data collection, and the rating of

audiotapes were discussed.

CHAPTER IV

RESULTS AND DISCUSSION
This chapter first presents the results of the study, followed

by a discussion of these results.

Results
The results of the study are presented under three headings,
those related to the effect of modeling upon empathy (hypotheses one
and two), those related to construct validity (hypothesis three), and the

summary of the Feedback Sheet.

 

Modeling and Empathy_

The findings related to modeling and empathy for the participants
analyzed as a total group were not significant. Differences were found,
however, for junior students when juniors and seniors were analyzed
separately. This was an unpredicted but significant finding. Results
are, therefore, discussed in two parts--for the group as a whole and

for the subgroups of juniors and seniors separately.

Total Group Results. The experimental treatment had little if

 

any effect upon the group considered as a whole. Thus, the first
hypothesis, which stated that the treatment groups would be rated higher
on empathic behavior than the control group, was not borne out.
Hypothesis two, that the groups would retain their level of empathic

behavior on follow-up tests, was confirmed. However, considering that

50

51

the treatment effect was very weak, this finding is of questionable
value.

Table 5 displays each group's mean and standard deviation on
the three measures for two times, immediately following the experimental
treatment and three weeks later. (Scores for each subject on each
measure are given in Appendix K.)

Data analysis was carried out using the Finn program for multi-
variate analysis of variance (MANOVA). (See Scheifley and Schmidt, 1973.)
The first analysis of the data was a repeated measures anlaysis of
treatment-by-time; that is, it examined the difference between the two
posttestings. In the second analysis, the two posttest sessions were
analyzed separately.

The first analysis showed no treatment effect, no change over
time, and no interaction between treatment and time. Tables 6, 7, and 8
show the univariate and multivariate test results from the first data
analysis.

The second data analysis, in which the two posttest sessions
were examined separately, revealed a possible treatment effect (p<.lO)
on the multivariate test for time one but no effect was found in any of
the univariate tests nor in the multivariate or univariate tests for

time two. Tables 9 and 10 display the results for the two times.

52

 

 

.momogpcmcma cw mew meowpmv>mc uceecmum .cmme azocm on» m? co>wm amass: 9mg?» mzhm
Amo.~v ANm.¢_V Am~_.v Ako._v Aeo.¢2v Ae~.V
mm.“ mo.ee ~k.~ kw.m om.¢¢ Nh.~ cam: agate
Aam._v Ama.wv ASP.V Aam._v AN_.PFV Amm.v
_m.~ ao.om mk.~ mm.m mm.we No.~ Am+a+zv m
AoP.~V Amm.ev AGN.V Ame._v Awm.~Fv A¢~.V
oN.m om.Ne mm.~ oo.e oo.oe mw.~ A¢+zv a
Ase._v Amk.w_v Apm.v Aeo._v Amm.mpv Awm.v
co.“ em.ee Ne.~ Fm.~ mp.¢¢ m~.~ Aa+zv m
Amo._v Amm.e_v A_N.v Akm.PV ANN.~_V Am_.v
mm.k oo.¢¢ m~.~ Fm.o mm.~e o~.~ sz N
Aem.mv ANm.m~v Amm.v Am~.,v Aok.~FV ANN.V
we.“ Ne.mm mm.~ mm.e mm.ee aeo.~ on .
HH amok xcquEN ncmccoguppoccmm wwsgxgmu H amok aspmasm ucmccoguuuocgmm emacxcmu aaocw
N asap F wave

 

 

N wave new — meek toe ammppmoa comm :o mcompmw>oo ncmccmum use memo: aaocc

m wpnmh

53

Table 6

Multivariate and Univariate Analysis of Treatment Effects

 

 

 

 

Variable M.S. df F Ratio p less than

Multivariate 12,130 1.14 .33

Carkhuff .18 4.51 1.84 .14

Barrett-Lennard 202.27 4,51 .87 .49

Empathy Test 1.20 4,51 .37 .83
Table 7

Multivariate and Univariate Analysis of Repeated Measures Effects

 

 

 

 

Variable M.S. df F Ratio p less than

Multivariate 3,49 1.36 .27

Carkhuff .002 1,51 .05 .82

Barrett-Lennard 8.04 1,51 .05 .83

Empathy Test 11.57 1,51 3.20 .08
Table 8

Multivariate and Univariate Analysis
of Interaction of Treatment and Time

 

 

 

Variable M.S. df F Ratio p less than
Multivariate 12.130 1.33 .21
Carkhuff .07 4.51 1.72 .16
Barrett-Lennard 110.07 4,51 .62 .65

Empathy Test 6.73 4,51 1.86 .13

 

54

Table 9

Multivariate and Univariate Treatment Effects for Time 1

 

 

 

 

Variable M.S. df F Ratio p less than

Multivariate 12,130 1.60 .10

Carkhuff .11 4,51 1.83 .14

Barrett-Lennard 112.08 4,51 .57 .69

Empathy Test I 4.98 4,51 1.78 .15
Table 10

Multivariate and Univariate Treatment Effects for Time 2

 

 

Variable M.S. df F Ratio p less than
Multivariate 12,130 .82 .63
Carkhuff .14 4,51 1.79 .15
Barrett-Lennard 200.17 4,51 .95 .44
Empathy Test II 2.96 4,51 .72 .58

 

Results for Juniors and Seniors. Because the students participating

 

in the study were from two classes, juniors and seniors, data were
analyzed separately for these two subgroups. Tables 11 and 12 present means
and standard deviations for each of the five groups of juniors and
seniors.

Analysis of the data revealed differences on some of the posttests
for juniors, but not for seniors. On the repeated measures analysis,
junior students demonstrated treatment effects, especially on the

Carkhuff Scale. They also maintained their gain over time, as reflected

55

.mommcpcmcmq cw men meowpmw>wu ucmvcmum .cmoe azocm on» m? =o>wm conga: umgww mgpm

 

 

ANN._V ANN.¢PV ANN.V Aom.Fv AeN.¢_V ANN.V NNuz
Pe.N Pe.Ne mo.N mm.o om.N¢ oo.N cam: agate
ANN.NV ANN.mv ANF.V ANo.NV Ame.NPV ANP.V on:
NF.N om.me NN.N mm.o om.me mm.N AN+S+=V m
Amo.Nv AmN.mv AFN.V Aem.v A_m.m.v Aom.v a":
mN.m om.o¢ oo.m mN.m oo.mm ON.N AN+=V N
Amm._v Ame.m_v ANN.V ANF.FV Amm.mpv ASN.V Nu:
mN.m mN.N¢ om.N NN.N om.N¢ NG.N as+zv N
Amo.Nv Ace.m_v ANP.V AoN.NV Amm.mpv AN_.V mu:
oN.N oe.m¢ NN.N oe.o ON.N¢ mo.N sz N
ANN.~V AN_.NPV ANN.V Rom.Nv A_N.N_v AoN.v mu:
oe.m o¢.NN mN.N oe.m ow.me we.N ANN P
HH amok asumasm ucmccmguppoccmm meagxcmo amok xsumaEm ugmccopuuuogcmm mesgxemu macaw

N mEmh

p mark

 

 

N meek new _ meek Low ammuumoa comm :o meowumw>oo newccmam vcm memo: aaocw .mucocsum Lowcaw

pp mpamh

56

 

 

.mmmmgucocma cw mew meowumw>ou tgmucmum .cmos aaoem esp m? co>wm amass: umewe echo
ANN._V Apm.e_v AmN.v ANN.PV ANN.¢_V A_N.v NNuz
mm.N ao.ee mN.N GP.N Fe.me sN.N cam: agate
ANm._v ANe.mV A_F.v ANG._V Amm.mv Amm.v mu:
oe.N oe.mm GN.N oe.N ON.em NN.N AN+S+zV m
ANe.PV Aom.ev AmN.v ANN._V ANN.ONV Ao_.v on:
N_.N om.¢s NN.N NN.m Nm.oe NauN AN+zv e
ANm.NV ANo.NNV Am..v Ace..v Aom.m_v Amo.v Nu:
Ne.N oo.mm mm.N oo.N NG.N¢ NN.N As+zv N
ANN._V Ase.m_v AmN.V Amm.2v AeN.mv AN_.V on:
NN.N mm.Ne NN.N N_.N oo.N¢ NN.N sz N
ANe.NV Asm.mpv Amm.v ANm.v ANN.N_V ANN.V Nu:
oo.m om.Ne oN.N NN.N om.oe mN.N ANN F
HH amok mgpmasm ccmccmguuuoccmm wwzgxcmu “may aspwaEm cemccopupuocgmm emacxcmu azogo
N mace F wave

 

 

N wave ace F wave com ammuumom comm co meowpmw>mo cgmucmpm new memo: qzogw .mpcovzpm Lowcmm

NF epoch

57

in no significant differences for the repeated measures effect. In
addition, there was an interaction of treatment and time, probably
reflecting the lower scores of the control group and the higher scores
of several of the experimental groups on the Carkhuff Scale between time
one and time two. The treatment, repeated measures, and interaction

effects for junior students are presented in Tables 13, 14, and 15.

Table 13

Junior Students: Multivariate and Univariate
Analysis of Treatment Effects

 

 

 

 

Variable M.S. df F Ratio p less than

Multivariate 12,56 1.69 .09

Carkhuff .28 4,23 3.89 .01

Barrett-Lennard 92.64 4,23 .34 .85

Empathy Test 7.51 4,23 2.67 .06
Table 14

Junior Students: Multivariate and Univariate
Analysis of Repeated Measures Effects

 

 

Variable M.S. df F Ratio p less than
Multivariate 3,21 .61 .62
Carkhuff .02 1,23 .60 .45
Barrett-Lennard 1.14 1,23 .01 .93

Empathy Test 8.65 1,23 1.99 .17

 

Junior Students:

58

Table 15

Multivariate and Univariate
Analysis of Interaction of Treatment and Time

 

 

 

Variable M.S. df F Ratio p less than
Multivariate 12,56 1.65 .10
Carkhuff .16 4,23 4.48 .01
Barrett-Lennard 152.19 4,23 1.00 .43
Empathy Test 8.65 4,23 1.99 .13

 

Beside the repeated measures analysis of the data, separate

analyses of the two postteSt times were carried out for junior students.

These analyses are presented in Tables 16 and 17.

Junior Students:

Table 16

Multivariate and Univariate
Effects for Time 1

 

 

 

Variable M.S. df F Ratio p less than
Multivariate 12,56 .75 .69
Carkhuff .05 4,23 1.05 .41
Barrett-Lennard - 26.69 4,23 .12 .97
Empathy Test 4.69 4,23 1,29 .30

 

59

Table 17

Junior Students: Multivariate and Univariate
Effects for Time 2

 

 

 

Variable M.S. df F Ratio p less than
Multivariate 12,56 2.13 .03
Carkhuff .38 4,23 6.91 .001
Barrett-Lennard 218.08 4,23 1.07 .39
Empathy Test 11.47 4,23 3.25 .03

 

Analysis of senior students' performance revealed no significant
differences. Tables 18, 19, and 20 display the treatment, repeated

measures, and interaction effects for seniors.

Table 18

Senior Students: Multivariate and Univariate
Anlaysis of Treatment Effects

 

 

Variable M.S. df F Ratio p less than
Multivariate 12,56 1.32 .23
Carkhuff .04 4,23 .51 .73
Barrett-Lennard 298.04 4,23 1.45 .25

Empathy Test 3.05 4,23 1.13 .37

 

60

Table 19

Senior Students: Multivariate and Univariate
Analysis of Repeated Measures Effects

 

 

 

 

Variable M.S. df F Ratio p less than

Multivariate 3,21 1.01 .41

Carkhuff .04 1,23 1.64 .21

Barrett-Lennard 25.79 1,23 .12 .73

Empathy Test 3.50 1,23 1.11 .30
Table 20

Senior Students: Multivariate and Univariate
Analysis of Inteaction of Treatment and Time

 

 

 

Variable M.S. df F Ratio p less than
Multivariate 12,56 1.08 .40
Carkhuff .04 4,23 1.63 .20
Barrett-Lennard 92.39 4,23 .43 .79
Empathy Test .94 4,23 .30 .88

 

To determine which of the five groups among junior students differed
from each other, Scheffe post-hoc comparisons were carried out. Each
treatment group was compared to the control group on the measures for which
the univariate analysis was significant, that is, on the Carkhuff Scale
for the repeated measures analysis and on the Carkhuff Scale and Empathy
Test 11 for the single analysis of the second posttest session. The

contrasts are shown in Tables 21, 22, and 23. A contrast is significant

61

if the absolute value of the contrast (column 1) is larger than the

value used to obtain the confidence interval (column 3).

Table 21
Scheffe Post-Hoe Comparisons
for Junior Students on the Carkhuff Scale

 

 

 

Estimate of Contrast E ' 'ancea 7 (J-1)F4 23(.05) Est.Var.
MSw(l/n1+l/nj) ’

 

2,-22 (C-M)
3.36-3.78=-.42 .17 .56

Xl'XB (C-M+L)
3.36-3.66=-.30 .15 .50

71X4 (C-M+R)
3.36-4.03=-.67* .16 .59

XI-XS (C-M+L+R)
3.36-3.78=-.42 .18 .54

 

*
p<.05
a -
MS -.07
w

62

Table 22

Scheffe Post-Hoc Comparisons
for Junior Students on the Carkhuff Scale for Time 2

 

 

Estimate of Contrast Estimate of Variancea (J-1)F4 23(.05) Est.Var.

 

MSw(1/n]+1/nj)

2.29-2.72=-.43 .15 .50
71-73 (C-M+L)

2.29-2.50=-.21 .13 .44
71-24 (C-M+R)

2.29-3.00=-.71* .16 .54
71-25 (C-M+L+R)

2.29-2.82=-.53* .14 .47

 

*
p<.05

a _
MSW-.055

63

Table 23

Scheffe Post-Hoc Comparisons
for Junior Students on the Empathy Test for Time 2

 

 

Estimate of Contrast Estimate of Variancea (J-l)F (.05) Est.Var.

 

 

4,23
MSw(1/n]+1/nj)
1"X2 (C'M)
5.6-7.8=- 2.2 1.19 3.99
x]"-3 (C‘M+L)
5.6-6.75=-l.15 1.07 3.58
5.6-9.75=-4.15** 1.26 4.22
2,-75 (C-M+L+R)
5.6-8.17=-2.57 1.14 3.82
iComparison significant at p<.lO, but not at p<.05.
aM5w=3.53

Although the Empathy Test was significant at the second posttesting
(p<.O3), none of the Scheffe comparisons were significant at p<.OSO.
Possibly this is because the Scheffe, test gives a relatively large inter-
val for pair comparisons. (Tukey comparisons could not be done because
of unequal group sizes.) With p<.10, the Scheffe interval for the

modeling-rehearsal group was significant (4.15>3.74).

Construct Validity

 

Hypothesis three, that the three measures of the dependent
variable will be significantly correlated with each other, was confirmed
for some of the measures. Most importantly, the Carkhuff Scale and

the Barrett-Lennard Relationship Inventory were significantly correlated

 

64

correlated for both of the posttest sessions. There was also a
significant correlation between the Carkhuff Scale and Form II of the
Empathy Test and between the first and second testings using the

Carkhuff Scale. The correlation matrix is given in Table 24.

Table 24

Correlations of the Three Measures Used for Time 1 and Time 2

 

 

 

VlTl V2T1 V3T1 V1T2 V2T2
V2Tla .27*
V3Tl .13 -.01

V1T2 .43** .11 .04

 

V2T2 .ll .13 .Ol .31*

V3T2 -.003 .12 -.05 .46** .04
a *

V1 =Carkhuff Scale p .05

V2 =Barrett-Lennard **p .01

V3T1 =Empathy Test-Form I
V3T2 =Empathy Test-Form II
Tl =Time 1
T2 =Time 2

Feedback Sheet
To ascertain why students volunteered for the study and whether
they carried out directions as requested, each student was asked, upon

completion of the second testing, to fill out a Feedback Sheet. The

 

results of this survey are contained in Table 25.

65

Table 25

Summary of Students' Responses on the Feedback Sheet

 

 

Question Number Responding

 

1. Why did you sign up for this project?
(Circle as many as apply.)

1. Sounded interesting 39
2. Needed the money 29
3. Thought I might learn something 41
4. Interested in research 21

2. Did you try to carry out the following
directions for the project?

A. Not discussing it with classmates?

1. Yes 56
2. No 0

B. Practicing what you learned between
the two interviews?

1. Yes 42
2. No a 6
3. No mark 8

3. Do you feel you took the project seriously
and tried to do a good job?

1. Yes 56
2. No 0

 

aSome left this blank, indicating they were in the control group.

Following each question and at the end of the questionnaire,
students were asked to comment. A number of students stated they were
interested in learning and felt the project was helpful to them. For
instance, one student said:

It was really interesting. It was good to be put in a situation

where you had to help somebody with a problem, even though you

hardly knew them, because this happens so often in the hospital
setting.

66

Another student commented:

It takes more than watching a videotape to learn empathy and

the communication of it, but people like me need a few

examples and I think this study helped put ideas and methods

in my head so I'm constantly considering what I learned.

A number of students said and others commented in writing that
the videotape for group five (modeling-labeling-rehearsal) was too long
and that they became tired or bored. This reaction might help explain
the drop in performance on some of the measures for group five (see

Table 5, 11, and 12).

Discussion

 

Discussion of the results of the study follows the same plan
as their presentation. That is, the results pertaining to the effect
of modeling upon empathy are discussed first, followed by a discussion

of construct validity. Results of the Feedback Sheet are not discussed

 

separately, but are introduced where appropriate in the discussion of

other findings.

Modeling and Empathy

 

The main question to be addressed with regard to the effect
of modeling upon empathy is why the treatment was effective for junior
but not for senior students. Another finding to be explained is the
interaction of treatment and time for junior students. A third area to
be examined is the extent to which the hypotheses were confirmed. The
final issue is why change was observed on some but not all of the

measures used.

Results for Junior and Senior Students. The question to be
asked about any experimental treatment is not, "Is it effective?", but,

"Under what conditions is it effective?". In the present experiment,

67

this question takes the form of why modeling produced a change in one
but not in another group of students.

Several explanations are feasible. The seniors could be
expected to be functioning at a higher level even without treatment,
since they had had a year more of nursing education than had junior
students. And in fact, inspection of group means in Tables 11 and 12
reveals seniors' average performance on the Carkhuff Scale was better
than juniors, even considering the improvement of junior students at
the second testing. (Seniors X=2.84 for Time 1 and 2.79 for Time 2;
juniors X=2.6O for Time 1 and 2.66 for Time 2.) Perhaps, then, the
particular application of the modeling treatment used in this study is
more effective for beginning than for advanced students. Payne, Winter,
and Bell (1975) likewise suggest that modeling may be better for
beginning students.

A related explanation is that a more powerful or different
treatment may be needed once students reach a certain level of perform-
ance. Although the concept of a learning curve in which performance levels
off at the upper end of the curve is no longer popular among psychologists,
such a curve would explain the results observed in this study. A more
current explanation is to postulate an aptitude-treatment interaction
(cf. Cronbach and Snow, 1977) based on the principle that what is
effective for some students is not effective for all. The differences
(aptitudes) observed in students can be any number of traits or factors;
in this case, the difference is amount of education or experience.

Another possible reason why seniors failed to improve is that,
because of their greater experience, their interview styles were already
well established and therefore more impervious to change. Perry's (1975)

findings support this explanation. The clergymen counselors who were

68

her subjects showed differences on a written but not an interview
posttest. Their explanation of this was that during the interviews,
rather than use what they learned from the tapes, they reverted to their
usual method of counseling. This explanation supports the need for more
powerful or different approaches to change behavior that is already well
established.

Still another explanation is that the juniors, being less
experienced, felt more of a need to learn about communicating with
clients. Juniors are neophytes, relatively speaking, and it is likely
that they felt insecure about many of their nursing skills. This
insecurity might have served as a strong motivator, making them more
open to change and learning. A number of junior students expressed to
the researcher that they felt their curriculum was lacking in the area

of interpersonal skills. On the Feedback Sheet, too, many students

 

indicated desire to learn as the reason for their participation.

Unfortunately, the data collected on the Feedback Sheet did not allow

 

separation of responses for juniors and seniors.

Interaction of Treatment and Time. The univariate analysis for

 

the Carkhuff Scale revealed an interaction of treatment and time for
junior students (F=4.48, p<.Ol). The interaction of the control group
compared to the average of the four treatment groups is graphed in
Figure 2, while Figure 3 displays the change over time for the five
groups separately.

The prediction related to time was that all groups would maintain
their behavior on follow-up testing three weeks after treatment. An
interaction was not expected. Figure 2, however, indicates that the

performance of the treatment groups improved, while that of the control

69

5.00

‘1- Treatment groups .........
00 Control groups
.95
.90
.85
.80
.75 '_'.-o
.70 .,.—
.65 — '
.60
.55
.50
.45
.40
.35
.30
.25
.20
.15
.10
.05
.00

0T

Figure 2. Graphic representation of junior students' scores on the
Carkhuff Scale over Time 1 and Time 2 for the control and treatment
groups

Score on
Carkhuff
Scale

NNNNNNNNNNNNNNNNNNNNW

 

Time 1 Time 2

5.001’

a”

.00
95
.90
.85
.80
.75
.70 M-R
.65 M-L
.60 M
.55
.50 M-L-R
.45 C

.40
.35
.30
.25
.20
.15
.10
.05

.00
,L—

0T

Score on
Carkhuff
Scale

NNNNNNNNNNNNNNNNNNNNOO

 

Time 1 Time 2

Figure 3. Graphic representation of junior students' scores on the
Carkhuff Scale over Time 1 and Time 2 for the five groups.

71

group deteriorated over time.

One of the instructions given to students at the end of the
first testing period was to practice what they learned during the
three week interval between testings. This suggestion was made to
counteract the effect of memory loss that might be expected to occur
if the behavior were not practiced. Students' responses on the Feedback
Sheet indicated that forty-two tried to practice what they learned,
while only six said they did not. (Eight did not respond, some of them
indicating they were in the control group.) If the students did consci-
ously practice empathy in interacting with clients, this constitutes
an additional experimental condition-—overt rehearsal. The improvement,
then, is understandable and is consistent with research findings that
the highest level of learning results from the combination of covert
and overt rehearsal (Jeffery, 1976).

Explaining the lower performance of the control group at the
second testing is more difficult. It may be at least partly due to
error, as there were only five junior students in the control group.

It may also be from a loss of interest and motivation that was not
experienced by the experimental groups. As mentioned earlier, many
students volunteered for the study hoping to learn something. A number
of students in the control group expressed disappointment when they
arrived for the first session and were told they would not be seeing

a videotape. Possibly these students lost interest and did not try

as hard the second time.

Figure 3 is much more difficult to explain than Figure 2,
especially the disoridnal interactions among some of the experimental
groups. There is likely some error involved when the groups are looked

at separately because of the small number of subjects in some of the

72

groups. An individual high or low score affects the group average
more when the group is small.

The only treatment group to exhibit a lower level of performance
on follow-up testing was the modeling-labeling group (n=8). The two
groups receiving rehearsal showed the most gain from time one to time
two. If there is a theoretical explanation for the disordinal inter-
action observed, perhaps it has something to do with labeling and
rehearsal. Possibly the students' own symbolic coding or labeling was
superior to that provided in the videotape. The labels provided may
even have interfered somewhat with learning. Bandura, Jeffery, and
Bachicha (1974) state:

Evidence is now accumulating that memory codes differing

in characteristics (e.g., verbal, imaginal, reductive,

elaborative) can all enhance retention of observationally

learned responses. It is likely that a basic property--
meaningfulness--underlies the mnemonic utility of diverse
codes. Retention is improved by transforming the new into

the familiar, which already exists in long-term memory.

(p. 296)

If the labels provided in the videotape did not enhance coding, that is
were not meaningful, then neither performance nor retention would have
been aided. It is also feasible that the critical component enabling

students to benefit from practice between time one and time two was

rehearsal during treatment.

The Remaining Hypotheses. Some explanation still needs to be

 

made of the extent to which the hypotheses related to modeling and
empathy were confirmed. It has already been noted that juniors but
not seniors benefited from the experimental treatment, and that they
not only maintained but improved their behavior over time. It remains
to comment upon the extent to which the addition of labeling and

rehearsal improved performance. This question is specifically addressed

73

by the results of post-hoc procedures carried out on data for junior
students (see Tables 21, 22, and 23).

The only groups whose performance was statistically superior
to the control group were those receiving the rehearsal condition.
Neither modeling alone nor modeling with labeling were effective
conditions. It has already been suggested that the labeling used
may not have aided students in coding their learning in a meaningful
way. Possibly it would have been better to have students construct
their own labels or memory codes. Research studies vary in this respect,
with some providing labels (Friedrich & Stein, 1975; Bandura & Jeffery
1973; Bandura, Jeffery, & Bachicha, 1974) and some requiring subjects
to construct their own labels (Gerst, 1971). In general, the experiments
that provided labels were either laboratory studies using artificial
tasks as pottests or were carried out with young children. The study
that most closely resembles the present one is that by Gerst (1971) on
acquisition of hand signals for the deaf, and he had subjects construct
their own labels.

The covert rehearsal condition did significantly affect
performance, but not immediately. It was pointed out earlier that the
three-week interval may have constituted an overt rehearsal condition,
and that it was only after the three-week period that changes were
observed. If this conjecture is in fact true, then covert plus overt
rehearsal is the best combination for improving empathic behavior
through modeling.

The combination of modeling, labeling, and rehearsal did not,
as predicted, result in the best performance. In the repeated measures
analysis of the Carkhuff Scale (Table 21) it was not statistically

superior to the control group, although for the second testing (Table 22)

74

it was significant. It was mentioned earlier that students found this
videotape rather long; it is likely, then, that their attention

wandered. Also, the labeling condition may have interfered with students'
processing of information.

A final comment about the extent to which the hypotheses were
confirmed concerns the conditions under which the experiment was
conducted. Kerlinger (1973) contrasts laboratory and field studies in
relation to the degree of control that can be achieved in each. Because
laboratory studies control more variables, they have the potential to
detect smaller differences between groups. In field studies, differences
may be obscured because of the larger number of variables affecting
behavior. It requires a strong effect, in other words, to show through
in a field study situation.

The present study is more characteristic of field than laboratory
research. The experiment was conducted much like a class designed to
teach empathy. A particularly realistic aspect of this study, and one
that sets it apart from most studies of modeling and empathy, was the
use of an interview posttest. Although the interviews were with
simulated and not real clients, the interview situation required students
to interact and apply what they learned. This is a more rigorous and
less controlled test than one in which students respond to written
client statements. Thus, the fact that the present investigation did
find differences between groups attests to the power of the modeling

treatment, especially the rehearsal condition.

The Instruments in Retrospect. The final issue to be discussed
is why change was observed on some, but not all of the measures employed.

1>artly this is a question of the appropriateness or usefulness of the

75

instruments to the research that was carried out, and partly it is a
question of what the instruments measure. Whether all the instruments
measure the same construct is the issue of construct validity, and
this will be discussed in the next section. The appropriateness of
the instruments to the research that was done will be discussed here.
Despite significant correlations between the Carkhuff Scale

and the Barrett-Lennard Relationship Inventory and between the Carkhuff

 

Scale and the Empathy Test--Form II (see Table 24), only on the Carkhuff

 

Scale were significant differences noted for the repeated measures
analysis (see Table 14). In addition, when the two times were analyzed

separately, significant differences were found on the Empathy Test--

 

Form II (see Table 17).

The most puzzling finding regarding the instruments was the
lack of an observable pattern on the results of the Barrett-Lennard,
For the most part, the results on this instrument are not only insigni-
ficant, but uninterpretable. Except for the second testing of junior
students, there appears to be no pattern to the performance of the control
and the various treatment groups, and this despite its positive
correlation with the Carkhuff Scale (see Tables 11 and 12). This
finding is surprising and disappointing. It is surprising because
in the Kurtz and Grummon study (1972) the Barrett-Lennard emerged as
the best predictor of outcome. It was expected, then, that it would
be a sensitive measure of the empathic behavior of the participants in
this study. The Barrett-Lennard may still be the best predictor of
outcome; the present study did not address the question of client
outcomes. The point is, does the BarretteLennard measure empathy or
some other quality of the therapist that correlates with outcome? And,

given its correlation with the Carkhuff Scale, why was the

76

Barrett-Lennard not a better measure?

The failure of the Barrett-Lennard to measure learning is
disappointing as well as surprising. It is a simple and inexpensive
measurement tool, especially when compared to the Carkhuff Scale, which
takes considerably longer to score and requires trained judges. Were
it a good indicator of learning, it would certainly be the instrument
of choice for research on empathy. Its failure to show significant
differences leaves one little choice but to continue using a time-
consuming and expensive instrument.

There was one problem in using the Barrett-Lennard, which may
be an indication of other problems with this instrument. Two of the
items were inconsistently interpreted by the simulated clients. The
two items ("She understands what I say, from a detached, objective
point of view" and "She tries to understand me from her own point of
view") were intended by Barrett-Lennard to be negative qualities of
the interviewer, but they were interpreted by some of the simulated
clients as positive qualities. The items were deleted in scoring the
questionnaires, so they did not bias the results obtained. However,
since reliability was computed on sixteen items, removing two items
may alter the statistical qualities of the test.

This differing interpretation of items leaves one wondering,
too, about how often this has occurred in various uses of the test.
Questions of interpretation were raised because the isntrument was
used by simulated clients who completed the questionnaire a number of
times. Real clients, who would use the instrument only one or two
times, would probably not raise such questions.

Perhaps the Barrett-Lennard needs re—examination, especially

regarding the interpretation of the items. Although Barrett-Lennard (1962)

77

did extensive testing when he originally developed the Relationship

 

Inventory, it is possible that the process he went through did not
eliminate all ambiguity for the non-professional user of the instrument.

The last question about the Barrett-Lennard is whether it is
appropriate to use with simulated clients who are interviewed only one
time. The instrument was developed to be used by clients who see a
therapist over a period of time. What is being measured under these
conditions is likely the quality of a relationship developed over time
rather than the empathy conveyed during a single interview. The
measurements under these different conditions may not be tapping the
same quality.

Furthermore, the fact that simulated and not real clients are
involved in this study may make a difference in measurement. Simulated
clients are role playing a problem, and this may introduce unknown
factors into the measurement process. Even though they were instructed
to react as if they were real clients, it is not known to what extent
they were able to do so, nor whether this factor makes a difference in
measurement.

Finally it remains to make a few remarks about the appropriateness
of the other two instruments. Despite its low reliability,
gmpgthy Test II was correlated with the Carkhuff Scale (r=.46, p<.Ol).
It also showed significant differences fer junior students on the
second testing (p<.03), but not on the repeated measures analysis

(see Table 17). Empathy Test 1, however, was not correlated with

 

Empathy Test II, nor with any of the other measures used (see Table 24).

 

It appears, then, that the cognitive tests, which were intended to

be parallel, were not parallel. Possibly some of the parallel quality

of the tests was lost when certain items were eliminated.

78

Empathy Test II, though, served the purposes of the research
fairly well. The reason for using a test of cognitive understanding
of knowledge or principles of empathy was to test Bandura's theory of
mediation that was discussed in Chapter II. Bandura maintains that
what observers learn from models is principles, which are then used as

guides for behavior. Students' performance on Empathy Test II provides

 

some confirmation for a mediational theory. That is, it provides
some evidence that students learned principles of empathy from
observing the models.

Since this was the first use of Empathy Test II, statements

 

about it should be regarded as tentative. It may prove to be a useful
instrument, but further reliability and validity studies should precede
any definitive conclusions based upon it.

The Carkhuff Scale met the needs of this study. Although the
scale is expensive and time-consuming to use, it was possible to
achieve high reliability between independent judges and, apparently,
to make discriminations among different degrees of empathy. Thus,
although one might wish to use a simpler instrument such as the
Barrett-Lennard for research on empathy, it appears that for the present
it is still necessary to use several instruments, one of which employs

independent judges.

Construct Validity

 

The results of this study provide some support for the construct
validity of the instruments used. As seen in Table 24, the Carkhuff
Scale is correlated with itself for the two testings (r=.43. p<.Ol),
with the Barrett-Lennard for time one (r=.27, p<.05) and time two

(r=.31, p<.05), and with Empathy Test II (r=.46, p<.Ol). None of the

79

other correlations were statistically significant.

These correlations support the convergent validity of the
instruments used. Campbell and Fiske (1959), it will be recalled,
state that correlations of the same variable measured by different
methods must be statistically greater than zero to satisfy the
requirement for convergent validity. Discriminant validity was not
established because no variables other than empathy were measured.

The status of discriminant validity for the construct of
empathy is questionable. Traux (1972) cites evidence that empathy is
separate from genuineness and warmth. Rappaport and Chinsky (1972),
however, assert that the studies Traux uses as evidence of discriminant
validity are, in their judgment, evidence of a lack of discriminant
validity.

This study is consistent with some ways of examining construct
validity put forth by Cronbach and Meehl (1955). One method they
suggest is to measure changes brought about by an experimental study;
another is to compute correlations on tests designed to measure the
same variable. The present study fits both of these categories.
Changes in empathic behavior as influenced by the experimental treatment
of modeling were measured in nursing students. Changes were noted in
junior but not in senior students, for reasons discussed earlier in
the chapter. The fact that some of the instruments were sensitive to
change in the predicted direction lends support to their construct
validity. This applies to the Carkhuff Scale and to a lesser extent
to Empathy Test II.

The second method of Cronbach and Meehl is what Campbell and
Fiske call convergent validity, or the correlations of the same variable

by different methods. The extent of these correlations has already been

80

mentioned. One failure to find a correlation that should be discussed,
though, is that between the Barrett-Lennard and Empathy Test II (r=.12 for
Time 1 and r=.O4 for Time 2). The fact that both the Barrett-Lennard
and Empathy Test II are correlated with the Carkhuff Scale but not with
each other bears some explanation. One reason why tests correlate with
each other is that they measure the same underlying variable or factor.
Likewise, failure to find a correlation indicates that different
variables are being tapped. The extent of a correlation might be
thought of as the extent to which variance is accounted for, and this
is sometimes expressed as the squared correlation coefficient
(coefficient of determination). Unless correlations are quite high,
there is considerable variation unaccounted for. For instance, the

correlation of the Carkhuff Scale and Empathy Test II (r=.46), although

 

highly significant (p<.Ol), accounts for only 21 percent of the observed
variance. This leaves 79 percent of the variance unaccounted for.
Possibly, the lack of correlation between the Barrett-Lennard and

Empathy Test II is explained by the fact that they measure different

 

factors or account for different proportions of the variance tapped by
the Carkhuff Scale.

Of course, this is only one possible explanation, and it must
be considered tentative since it is based on a single study. Cronbach
(1971) mentions other reasons why one might not find high correlations,
such as restriction of range and the large number of variables affecting
performance in a given situation.

A final issue regarding construct validity is to compare the
results of this study to those of Kurtz (1970) and Langer (1972).
Kurtz as well as Rogers et al. (1967) noted positive and significant

correlations between tape-judged and client perceived empathy. Langer

81

found no relationship between measures. As has already been noted,
however, Langer's interrater reliability was rather low (r=.74).
Considering the results of the above studies together with the present
study, there is some support for the construct validity of the
Carkhuff Scale and the client fbrm of the Barrett-Lennard Relationship
Inventory. Definitive statements cannot as yet be made about the
construct validity of the instruments, though, for, as both Cronbach
(1971) and Hempel (1966) point Out, construct validation is part of

theory development, and this is a slow process indeed.

CHAPTER V

SUMMARY AND CONCLUSIONS
Chapter five presents a brief summary of the study and outlines

some conclusions.

Summar

The present study was undertaken to determine the effect of
modeling in teaching empathy to nursing students. The participants
were junior and senior students in a baccalaureate nursing program who
volunteered for the project. The seventy volunteers were randomly
assigned by class to one of five groups. Fifty-six students completed
the project.

The four treatment groups received variations of modeling,
labeling, and rehearsal. Specifically, the combinations were modeling-
only, modeling-labeling, modeling-rehearsal, and modeling-labeling-
rehearsal. All treatments were videotaped., The fifth group was a
no-treatment control.

The conceptual base for the study was Bandura's theory of modeling
which, simply explained, states that most behavior is learned through
modeling, that is, through observing others. The modeled behavior is
coded and remembered by the observer for later use in performance. Two
elements that have been found to enhance coding are labeling and
rehearsal. These elements were therefore included in the present study

and were varied in such a way that their single as well as combined

82

83

effects could be assessed.

A repeated measures design was employed with posttesting
immediately following the treatment and three weeks later. The follow-
up posttest was for memory and was deemed a crucial aspect of this
study, since most studies of modeling and empathy do not include
follow-up.

The dependent variable, empathy, was measured in three ways.
Immediately after treatment, students completed the Empathy_Test--Form I,
a cognitive test of principles of empathy constructed for this project.
Students then interviewed a simulated client. The interview was
evaluated in two ways. Clients completed the Barrett-Lennard

Relationship Inventory, and the interview was audiotaped for later

 

rating using Carkhuff's Empathic Understandingrin Interpersonal Processes
Scale. The second posttesting three weeks later was the same, except

that an alternate form of the Empathy Test (Form II) was used. The use

 

of an interview posttest was an important element in the study, since
most studies use only written posttests, and those that used interviews
showed no treatment effect for modeling alone.

Since most studies show modeling as an effective way to teach
empathy, it was hypothesized that the treatment groups would perform
better than the control group, and that the elements of labeling and
rehearsal would further improve performance. In relation to follow-up
testing, it was hypothesized that all groups would retain their level
of behavior three weeks later. Because the most serious problem in
doing research on empathy seems to be its measurement, a third
hypothesis related to construct validity was included, stating that

the measures used would be significantly correlated with each other.

84

Data were analyzed using multivariate analysis for repeated
measures, as well as for the two testing separately. There were no
treatment effects for the group as a whole. However, when data were
analyzed separately for junior and senior students, the results for junior
students showed a treatment effect on two of the instruments-- the

Carkhuff Scale and Empathyplest II. There was also an interaction of

 

treatment and time, with treatment groups improving on the second
posttesting.

To counteract the effect of memory loss, students were given
instructions to practice what they learned between the two testing
sessions. It was hypothesized that the students' practicing
constituted a condition of overt rehearsal and accounted for the
improvement seen at the second testing. Scheffe post-hoc pair
comparisons showed that only the two groups receiving covert rehearsal
performed significantly better than the control group. This finding
is consistent with other research on modeling, which indicates that covert
plus overt rehearsal produces the most change in behavior.

Several explanations are possible as to why juniors but not
seniors benefited from the modeling treatment. Since juniors' average
performance was lower than seniors', perhaps the treatment is better
for beginning students. Juniors may also have felt more of a need and
therefore been more motivated to learn. Seniors, because they have more
experience, may be set in their interview style and need different
learning experiences than those provided.

Some support was found for the construct validity of the
instruments used, as indicated by the significant correlations of the
Carkhuff Scale with the Barrett-Lennard Relationship Inventory and

also with Empathy Test II.

 

85

Conclusions

 

Conclusions from the study relate to three general areas--those
about the effectiveness of modeling in teaching empathy, some conclusions
pertaining to nursing education, and conclusions about the construct

validity of empathy.

Modeling and Empathy

 

One of the most serious criticisms of those who carry out and
use educational research is their tendency to over-generalize results.
More and more, researchers in education and other social sciences are
becoming aware of the specificity of many of their effects. Special
care is therefore taken not to generalize the results of this study to
all applications of modeling or to all populations of interest. The
conditions or operational definitions of modeling have been explained,
and the characteristics of the students who were participants in the
research have been described to the extent that it was possible to do so.
Given this information, readers can draw their own conclusions about
whether the results might apply to other groups.

With these cautions in mind, though, what can be concluded from
the study reported here? First, the combination of covert and overt
rehearsal appears to be the most effective application of modeling in
teaching empathy. As mentioned in Chapter IV, the fact that the
rehearsal condition produced a measurable effect in a field-like
situation is evidence that modeling is a powerful treatment. The use
of modeling alone or modeling with labeling did not produce significant
differences in behavior. Whether this was the result of the particular
application of these conditions or might be more generally true is a
question to be answered by future research. Another area that should

be investigated is the single and combined effects of covert and overt

86

rehearsal. The presence of an overt rehearsal condition in this study
is a conjecture, as has been explained, and it should be applied in a
controlled way.

Other suggestions for future research are for students to
construct their own labels or memory codes, and to ascertain whether
the use of positive and negative examples adds to student learning.
Various labeling conditions-~such as labels being provided versus labels
being constructed by the participants--should be compared. Likewise,
whether the use of negative examples enhances learning should be
studied. The approach of using positive and negative examples is based
on research on concept learning and was used in several other studies
reviewed as well as the present investigation. However, it has not
been tested empirically in teaching empathy.

Another comment about future studies on modeling and empathy
is that they should include an interview posttest. The studies
reviewed that require subjects to construct empathic responses to standard
client statements demonstrate fairly conclusively that modeling is
effective in producing this type of response. Future research should be
carried one step further and explore ways to teach empathy in more
realistic situations. Although laboratory or controlled research is
important in establishing basic relationships, most educators are
interested in practical applications of theory. Results of highly
controlled studies cannot be generalized to less controlled situations,
where many more variable are operating. Ways must be found to do more
research in applied settings. This study used interviews with simulated
clients, which is one type of applied test of learning. More tests
of this type should be used and, in addition, tests should eventually

be done using interactions with actual clients.

87

A related suggestion is that tests should be carried out
on whether behavior is maintained over time. As with the interview
posttest, few studies included the element of follow-up tests.
Maintenance of behavior is a primary goal of education. Most educators
are not interested in teaching knowledge and skills that will be learned
and soon forgotten. Whether or not they deliberately attend to memory,

few educators would say they do not think it important.

NursingiEducation

The research on modeling and empathy is not sufficiently
advanced to make many generalizations about its application in nursing
education. The suggestions made in the prior section about research on
modeling and empathy should be carried out using nursing students as well
as other groups.

One conclusion that seems to follow logically from the present
study is the necessity of introducing interpersonal skill learning early
in the students' education. The participants in this study expressed
the need for such learning. Furthermore, if it is true that seniors
did not benefit from this project because their interview styles were
already well established, then it is crucial to teach students early,
before habits become set. .

An interesting approach to teaching empathy to nursing students,
and one that could be studied empirically, would be to have a series of
modeling videotapes with a period of overt rehearsal between each tape.
This would allow a combination of covert and overt rehearsal and
perhaps other conditions, and would permit students to continue
learning finer details from models as they practiced their skills with

clients.

88

Finally, some outcome studies of the effect of empathy upon
clients should be done. Studies of this type have been done in counseling,
but none were located for nursing. Outcome studies address the question
of whether empathy makes a difference to clients, that is, of whether
clients benefit in some way. The particular outcomes selected depend
upon the research that is being done. One might, for example, measure
client satisfaction, clients' assumption of responsibility for their
own health, or maintenance of a treatment regimen for a chronic health

problem.

Construct Validity

 

As discussed in Chapter IV, this study gives some support to
the construct validity of the instruments used. More studies of both
convergent and discriminant validity should be done. The outcome studies
referred to in the prior section are also pertinent to construct
validity, because they connect the construct of empathy to a larger
theory of helping relationships.

As to the future measurement of empathy, it seems necessary to
continue using several instruments. Even though this process is
expensive and time-consuming, it appears to be the best approach until
the concept of empathy and its measurement are better understood.
Clearly, the most difficult problem in doing research on modeling and

empathy is the measurement of empathy.

APPENDIX A
Empathic Understanding In Interpersonal Processes

A Scale for Measurement

Appendix A

Empathic Understanding In Interpersonal Processes
A Scale For Measurement

Level 1

The verbal and behavioral expressions of the helper either do
not attend to or detract significantly from the verbal and behavioral
expressions of the helpee(s) in that they communicate significantly
less of the helpee's feelings and experiences than the helpee has
communicated himself.

EXAMPLE: The helper communicates no awareness of even the most
obvious expressed surface feelings of the helpee. The
helper may be bored or disinterested or simply operating
from a preconceived frame of reference which totally
excludes that of the helpee(s).

In summary, the helper does everything but express that he is listen-
ing, understanding, or being sensitive to even the most obvious
feelings of the helpee in such a way as to detract significantly
from the communications of the helpee.

Level 2

While the helper responds to the expressed feelings of the
helpee(s), he does so in such a way that he subtracts noticeable affect
from the communications of the helpee. '

EXAMPLE: The helper may communicate some awareness of obvious, surface
feelings of the helpee, but his communications drain off a
level of the affect and distort the level of meaning. The
helper may communicate his own ideas of what may be going on,
but these are not congruent with the expressions of the
helpee.

In summary, the helper tends to respond to other than what the helpee
is expressing or indicating.

89

90

Level 3

The expressions of the helper in response to the expressions of
the helpee(s) are essentially interchangeable with those of the helpee
in that they express essentially the same affect and meaning.

EXAMPLE: The helper responds with accurate understanding of the
surface feelings of the helpee but may not respond to or
may misinterpret the deeper feelings.

In summary, the helper is responding so as to neither subtract from nor
add to the expressions of the helpee. He does not respond accurately
to how that person really feels beneath the surface feelings; but he
indicates a willingness and openness to do so. Level 3 constitutes the
minimal level of facilitative interpersonal functioning.

Level 4

The responses of the helper add noticeably to the expressions
of the helpee(s) in such a way as to express feelings a level deeper
than the helpee was able to express himself.

EXAMPLE: The helper communicates his understanding of the expressions
of the helpee at a level deeper than they were expressed
and thus enables the helpee to experience and/or express
feelings he was unable to express previously.

In summary, the helper's responses add deeper feeling and meaning to
the expressions of the helpee.

Level 5

The helper's responses add significantly to the feeling and
meaning of the expressions of the helpee(s) in such a way as to
accurately express feelings levels below what the helpee himself was
able to express or, in the event of ongoing, deep self-exploration on
the helpee's part, to be fully with him in his deepest moments.

EXAMPLE: The helper responds with accuracy to all of the helpee's
deeper as well as surface feelings. He is "tuned in" on the
helpee's wave length. The helper and the helpee might proceed

together to explore previously unexplored areas of human
existence.

91

In summary, the helper is responding with a full awareness of who the
other person is and with a comprehensive and accurate empathic under-
standing of that individual's deepest feelings.

In the empathy discrimination training proper we concentrate
initially upon the question of the interchangeability of the communica-
tions of helper and helpee: Is the helper expressing essentially the
same affect and meaning, feeling and content that the helpee is
communicating? On the five—point scale for the assessment of empathy
level 3 is defined as follows: "The expressions of the helper in
response to the expressions of the helpee(s) are essentially inter-
changeable with those of the helpee in that they express essentially
the same affect and meaning." We have arbitrarily constituted level 3
as the minimal level of facilitative interpersonal functioning. At a
minimum we are suggesting that the helper must communicate an understand-
ing of at least as much material as the helpee has communicated in the
first place. If the helper can add significant affect and meaning in
that he extends the expressions of the helpee or enables the helpee to
understand himself at even deeper levels, then he receives higher
ratings.

The first determination, then, involves the question of inter-
changeability. This question can be answered by one helpee-helper
exchange in that order. That is, one response by the helper to a helpee
expression would conceivably be sufficient for rating purposes. Indeed,
during the initial stages of communication training we employ only one
exchange for rating. However, this is not to say that more exchange
between helpee and helper is not helpful. Rather, one helpee-helper
interaction is the minimum for making the determination of interchange-
ability.

If the helper's response is not interchangeable, it must be an
additive or subtractive nature. That is, it either adds noticeably
or subtracts noticeably from the expressions of the helpee. Such
helper expressions would seem to involve determinations more subjective
than the operational interchangeability. This difficulty can best be
handled by studying the effects upon the helpee. Does the helpee in
fact tend to employ the helper's response effectively? Does the helpee

92

in fact tend to explore himself meaningfully, searching out new meanings
and new understanding, in response to the helper's response? If so,
then we rate the helper above 3. If not, we rate him below 3. while
the question of interchangeability may be answered by a helpee—helper
interaction, then, the question of an additive or subtractive nature
requires a minimum of a helpee-helper-helpee interaction.

If there is consensus that the helper's response adds noticeably,
then, the trainee must assign the excerpt a rating above level 3. If there
is consensus that the helper's response subtracts noticeably, then the
trainee must assign the excerpt a rating below level. 3. The questions
of consensus are involved ones, however. Let us assume for the moment
that our trainees are well-selected people relatively free from neurotic
distortions. If, then, the helper in the excerpt continually does not
communicate to the great majority of people that he understands the
helpee, he will in all probability not communicate understanding to the
helpee. The greater the number of excerpts sampled, the greater the
probability that this will be so. In any event, over time and with
continuous exposure to the helper's taped response and implied perception
as well as to the perceptions of the trainer and the other trainees,
each trainee will increasingly move toward some kind of group conformity
in ratings. This conformity will be in large part a function of the
skillful efforts of the trainer.

The key to empathy ratings is the relation of the helper's
response to the helpee's expression, in terms of both the helpee's
expressed affect and content. In particular, the relation of the tonal
qualities as well as the verbal expressions of the helper to those of
the helpee are critical. Although there has been some controversy
concerning whether helper responses can be rated independently of helpee
expressions, it is not a meaningful argument, with the level of warmth
likely accounting for these independent empathy ratings. It is as
meaningless to rate helpee self-exploration in the absence of the helper's
response. Neither may be related to its counterpart. Indeed, each,
although apparently appropriate, when considered independently of the
other may be a defensive maneuver calculated to avoid relating with the
other.

93

Thus, if the helpee's initial expression relates to a deterior-
ating home situation and the attendant distressing affect, in order
to achieve a level 3 rating the helper's communication must incorporate
at least that much content and affect: "Right now things are just going
so poorly at home that you just don't know if you can make any sense
out of it any more." At higher levels the helper might not only reflect
the helpee's expressions but also tap in on the feelings of depression
and agitated hopelessness and extend the content of the helpee's express-
ions to all of his deteriorating relationships and the attendant area of
self worth, enabling the helpee to explore himself at even deeper
levels in the relevant areas: "It's really not just home but everything,
everywhere, is falling apart and it's got you feeling pretty low,
wondering about yourself." At lower levels the helper might not respond
to the affect or the content, even in some cases redirecting the helpee's
attention to those aspects of life, for example, his work situation,
for which he should be happy and greateful ("You should be grateful
for having such a fine job, though") or in more subtle ways simply
subtracting affect and meaning from the helpee's expressions ("Although
things aren't going well at home, there are other areas of your life that
must be rewarding").

Some confounding in the rating process, then, is unavoidable.
If available, the rater trainee may make his discrimination at least in part
on the basis of the helpee's activities both prior to and following the
helper's response. This is not to say that every empathic helper response
is reflected immediately in helpee process involvement. However, on the
average the utilitarian criterion of helpee response is supported by
both experiential and research evidence. Indeed, if the helpee cannot
understand the helper's responses, no matter how brilliant the helper's
insights or "understanding," diagnostic or otherwise, we cannot say that
the helper's understanding is immediately functional. If the helper
does not understand the helpee well enough to comnunicate to the helpee
in terms that the helpee can understand, then he does not understand the
helpee, at least not well enough to help him.

 

Source: Carkhuff, 1969(a), pp. 174-177.

APPENDIX B

Barrett-Lennard Relationship Inventory

fi‘s

PLEASE NOTE:

Copyrighted materials in this document
have not been filmed at the request of
the author. They are available for
consultation, however, in the author's
university library.

These consist of pages:

94-95

 

 

 

 

 

Universg'
Micr lms
International

300 N. ZEEB RD, ANN ARBOR,M148106(3131761-4700

I. C O ’1 Isl JIJIJIII’IJllllilJJl‘ll‘.lO-ln‘llj

Appendix B

Barrett-Lennard Relationship Inventory

Below are listed a variety of ways that one person could feel or behave
in relation to another person. Please consider each statement with
respect to whether you think it is true or not true about your inter-
view with the nurse. Mark each statement in the left margin according

to how

strongly you feel it is true or not true. Please mark every one.

Write in +1, +2, +3; or -1, -2, -3, to stand for the following answers:

+1: I
+2: I
+3: I
-l: I
-2: I
-3: I
1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

ll.

12.

13.

feel that it is probably true, or more true than untrue.
feel it is true.
strongly feel that it is true.
feel it is probably untrue, or more untrue than true.
feel it is not true.
strongly feel that it is not true.
She tries to see things through my eyes.
She understands my words but not the way I feel.
She is interested in knowing what my experiences mean to mg,
She nearly always knows exactly what I mean.

At times she jumps to the conclusion that I feel more strongly
or more concerned about something than I actually do.

Sometimes she thinks that I feel a certain way, because she
feels that way.

She understands me.

Her own attitudes toward some of the things I say, or do,
stop her from really understanding me.

She understands what I say, from a detached objective point
of view.**

She appreciates what my experiences feel like to me,

She does not realize how strongly I feel about some of the
things we discuss. ' '

She responds to me mechanically.

She usually understands all of what I say to her.

94

95
14. When I do not say what I mean at all clearly she still understands
me.
15. She tries to understand me from her own point of view.**

16. She can be deeply and fully aware of my most painful feelings
without being distressed or burdened by them herself.

Circle the number on items that were very difficult for you to answer
in these circumstances.*

 

*This statement was added to Barrett-Lennard's directions to
ascertain whether certain items were difficult to answer after only a
single interview.

**These two items were deleted in computing scores because there
was inconsistency among the simulated clients as to whether the items
described positive or negative qualities of the interviewer.

Source: Barrett-Lennard, 1962, p. 34.

APPENDIX C

Empathy Test - Form I

Appendix C
Empathy Test - Form I

Directions:

 

1) Put your name on the answer sheet and fill in the grid.
2) Mark answers on the answer sheet with a #2 pencil.
3) Please do not make any marks on the test.

Part I.

Mark each of the following twelve items true or false,

depending on whether you think it is generally true or generally false.
On the answer sheet, mark a "l“ for true and a "2" for false.

1.

d 10.

Understanding what a person means by what he says is the main
component of empathy. (F)

If a patient feels the nurse understands his thoughts and
feelings, the nurse is being empathic. (T)

Commenting on behavioral cues a person is giving detracts
from understanding what he or she is saying. (F)

A nurse can be empathic toward a patient without communicating
his or her understanding to the patient. (F)

Verbal expression of one's understanding of another's thoughts
and feelings is the most important part of empathy. (F)

One aspect of empathy is objectively understanding what a
person is experiencing. (F)

Empathy includes understanding another's experiences as well as
communicating this understanding to him or her. (T)

Empathy can be expressed verbally by reflecting mainly the
content of what a person says. (F)

Understanding a person's nonverbal communication is a requisite
part of empathy. (T)

Empathy implies understanding a person's experience from his
or her point of view. (T)

 

d

Deleted on second item analysis because of discrimination index below 20.

96

Part I.

d 11.

d 12.

Part II.

and two
believe

d 13.

d 14.

15.

d 16.

d 17.

18.

97

continued....

Empathy consists of accurately perceiving the feelings of
another person. (F)

It is necessary to express empathy nonverbally as well as
verbally. (T)

The following six items give a patient statement or behavior
possible responses of the nurse. Choose the response that you
to be most empathic.

 

A hospitalized patient is angry and asks the nurse to leave,
but he moves to one side so she can come in. The nurse:

1. leaves the patient alone.
*2. enters the patient's room.

A patient is explaining how someone she once considered a good
friend did not help her when she needed it. The nurse says:

*1. "You must have been disappointed."
2. "And what did you do then?"

A man who has had a heart attack tells the nurse in the clinic
that he is not following the prescribed activity restrictions.
The nurse says:

*1. "Do you feel that by over exerting yourself you can prove
the doctor is wrong?"

2. "Do you feel you would rather be dead than live with
restrictions?"

A patient edges her chair away from the nurse as they sit
talking. The nurse:

1. moves her chair closer.
*2. remains where she is.

A woman with three daughters has just given birth to her fourth.
She is crying because her husband wanted a son. The nurse says:

*1. "You feel you have disappointed your husband?"
2. "You feel hopeless about ever having a son?"

Five days after a radical mastectomy, a woman is sitting on the
edge of her bed staring at the floor. She does not notice the
nurse entering. The nurse puts her hand on the patient's
shoulder and says:

1. "You are really preoccupied today. Is something bothering
you?"
*2. "You seem down today. Do you want to talk about it?"

98

Part III. The following six items give a patient statement or behavior
and two possible responses of the nurse. Choose the response that you
believe to be least empathic. (Note that this section is opposite

from the previous one. Here you are to choose the response that shows
the least empathy. Note also that both responses may be low in empathy.)

19.

20.

21.

d 22.

d 23.

d 24.

A man confined to bed in his home turns his back toward the
visiting nurse when he starts to cry. The nurse:

1. walks around the bed so she can see the patient.
*2. leaves the patient, telling him she will visit again in a
couple of days.

A woman whose child is hospitalized does not want to leave when
visiting hours are over. The nurse says:

1. "I know it is hard for you to leave your son, but we will
take good care of him."
*2. "Hospital rules require that you leave now, but I'll
let you know tomorrow how he did."

A diabetic man is admitted with an ulcer on his leg. He had
some toes amputated four years ago and talks now about being
afraid of losing his leg. The nurse says:

*1. "This must be very frightening for you. You did the right
thing, though, by coming to the hospital."

2. "Do you want to share your thoughts about having your leg
amputated?"

A man whose wife is dying becomes angry at the nurse because
his wife still has pain after receiving pain medication. The
nurse says:

1. "Why don't you talk to the doctor about this when he makes
his rounds later? He can probably increase your wife's
medication."

*2. "I think your wife is very frightened, and some of her pain
is due to fear."

A woman in traction with a broken leg is crying because she has
to be in the hospital for Christmas. The nurse says:

*1. "Don't be upset. You will feel better in a couple days."
2. "But your family is coming to spend the afternoon with you,
aren't they?"

An alcoholic patient who did not eat his lunch asks the nurse to
order him a sandwich. The nurse asks the head nurse to do this.
The head nurse replies:

1. "No. Dinner trays will be here in an hour, and if he eats
now he will spoil his dinner."
*2. "He's just an old alcoholic. He doesn't need a sandwich."

Part I.

12.

Part II.
and two
believe

13.

14.

15.

16.

d 17.

18.

100

continued....

Empathy need not be expressed nonverbally if its verbal
expression is very clear. (F)

The following six items give a patient statement or behavior
possible responses of the nurse. Choose the response that you
to be most empathic.

A woman who has had two miscarriages and is now five months
pregnant is hospitalized for a threatened abortion. The nurse
who admits her says:

a. "Can you tell me what you are feeling now?"
*b. "This must be a very upsetting experience for you."

A fourteen year old boy who is admitted for plastic surgery on
burn scars about his face and arms looks away whenever the
nurse approaches to do something for him. The nurse:

*a. tries to talk to him about his turning away.
b. talks to him but does not bring up his turning away.

A man whose wife left him is hospitalized for a self-inflicted
gunshot wound. The nurse asks:

a. “You feel you cannot live without your wife?"
*6. "You feel you would be better off dead?"

A man whose wife is dying becomes angry at the nurse because his
wife still has pain after receiving pain medication. The nurse
says:

*a. "You are angry about your wife's condition?"
b. "You are angry with me because your wife still has pain?"

A woman in traction with a broken leg is crying because she has
to be in the hospital for Christmas. The nurse says:

*a. “It is hard for you to be separated from your family."
b. "Christmas is a difficult time for you."

A diabetic man is admitted with an ulcer on his leg. He had
some toes amputated four years ago and talks now about being
afraid of losing his leg. The nurse says:

a. "Can you tell me more about how you feel?"
*b. "So you are really afraid?"

PLEASE NOTE:

Page l0] is lacking in number
only. No text is missing.
Filmed as received.

UNIVERSITY MICROFILMS.

102

Part III. The following six items give a patient statement or behavior
and two possible responses of the nurse. Choose the response that you
believe to be least empathic. (Note that this section is opposite from
the previous one. ’Here you are to choose the response that shows the
least empathy. Note also that both responses may be low in empathy.)

 

 

19.

20.

d 21.

d 22.

23.

d 24.

A young married woman seeking abortion counseling says she
wants an abortion, but feels she has no good reason to have
one. The nurse says:

*a. "What do you feel would be a good reason to have an abortion?"
b. "Do you feel guilty about having an abortion?"

An eighty-five year old woman who cannot walk very well and is
confined to her home says she feels she has lived too long.
The visiting nurse says:

a. "Are you saying you want to die?"
*b. "Can you tell me why you feel this way?"

The parents of a child with leukemia express fears about their
child dying. The nurse says:

a. ”Do you have some reason for feeling this way?"
*b. "The outlook for recovery from leukemia is much better now
than it used to be."

A patient is explaining how someone she once considered a good
friend did not help her when she needed it. The nurse says:

*a. "That‘s too bad, but that is how things go sometimes."
b. "Your friend probably did not mean to hurt you."

A woman with three daughters has just given birth to her fourth.
She is crying because her husband wanted a son. The nurse says:

*a. "You should be grateful that the child is healthy."
b. "Perhaps next time you will have a boy."

A man who has had a heart attack tells the nurse in the clinic
that he is not following the prescribed activity restrictions.
The nurse says:

*a. "You know this could cause you to have another heart attack."
b. "You really ought to follow the doctor's orders, you know."

APPENDIX E
Narration for Videotape

"Empathy in Nursing"

II.

Appendix E

Narration for Videotape
"Empathy in Nursing"

Introduction

This is a film about empathy in nursing. Its purpose is to
show you some ways of being empathic in your interactions with
patients.

What, you might ask, do we mean by empathy? In a word, it
means understanding. Most writers identify two components of empathy,
the first being sensitivity to another person's feelings and the
second the communication of this understanding to him or her.

In this film you will see nurses demonstrating empathy in
their interactions with patients. The film has several segments,
each of which will be introduced as it begins.

Modeling

In the first segment, three nurses are interacting with
patients. Each nurse will do two interviews, one demonstrating
empathy and the second demonstrating lack of empathy. The purpose
of showing opposite behaviors is to help you learn what constitutes
empathy. Each interview will be identified at the beginning.

For purposes of this film, empathy will be labeled high
empathy and lack of empathy will be called low empathy.

The first set of interviews takes place in the patient's
home. The health department has received a referral from
Community Hospital to do follow-up on Mrs. H., a woman whose child
has been treated for a number of injuries in the past year, injuries

inflicted by Mrs. H.

103

III.

104

He open in the middle of the nurse's second visit.

The next home visit you will see, this one with low
empathy, shows the nurse doing an assessment of the family.

The second set of interviews takes place in the nurse's
office in the outpatient department. The patient has chronic
kidney disease for which he receives dialysis three times weekly.

He has been seeing the nurse for a year to discuss his treatment
program and his reactions to his illness. This interview is one
that takes place a few weeks after one of his friends, Rick, another
patient in the dialysis program, has died.

The next interview is the same situation, but this time the
nurse is nonempathic.

The third interview situation is the coffee shop in a
hospital, where the nurse and patient have gone for a private
conference. The nurse, a mental health consultant, has been asked
to see the patient, whose physician suspects her symptoms of insomnia
and gastronintestinal upset are emotionally based. We see them near
the beginning of their second talk.

The next interview is the same situation, but now with low
empathy from the nurse.

Labeling

In this next segment of the film, you will see short sections
from the previous interviews repeated, but this time with an
explanation of why the behavior is labeled high or low in empathy.

In the first interview, the one in the patient's home, the
nurse expresses empathy nonverbally by leaning toward the patient,
listening carefully, and maintaining eye contact while they are

speaking. Her facial expression is one of concern. Notice these

105

behaviors in the following brief segment.

In her verbal behavior, the nurse shows empathy in her
attempts to understand both the content and the feelings the
patient is expressing. In the following segment, she reflects content
as the patient talks about her relationship with her husband.

She also comments on the patient's feelings about what she
is experiencing.

In another segment, the nurse reflects the feelings the
patient seems to be experiencing and expresses her understanding.

During the interview in which the nurse shows low empathy,
she has a cool, business-like attitude toward the patient, and
rather than look at the patient, she writes during much of their
talk.

She focuses only on content and is more interested in getting
information than in understanding the patient.

She ignores the patient's feelings, for example, the
irritation she expresses in the following instance.

Not only does the nurse ignore the patient's feelings, she
has at times a rejecting and judgmental attitude toward her, as
shown in the next two sequences. _

In the second set of interviews as in the first, the nurse
conveys empathy through her total demeanor. She has an expression
of concern and leans toward the patient,listening carefully to what
he says.

She explores the content of the patient's conversation,
indicating her attempt to understand by pursuing what he says.

She expresses understanding by letting the patient know she

is aware of what he is feeling.

106

The same nurse, expressing a lack of empathy for the patient,
has an attitude of detachment, which is seen in her facial expressions
and in such behaviors as looking away from the patient.

Verbally, the nurse is not tuned in. She ignores the
patient's feelings and seems more concerned with her needs than with
his, as portrayed in the following sequences.

The nurse in the third set of interviews expresses empathy
nonverbally in the same way as the other two nurses, by attentive
listening and an expression of concern.

This nurse also combines recognition of the patient's
feelings with a nonverbal expression of reaching out toward her.

In the following sequence, the nurse hears the patient's
message, even though it is expressed indirectly and is disguised
with some humor.

Besides noting verbal content, the nurse acknowledges the
behavior cues the patient is giving.

The nurse also reflects the feeling implied in what the
patient says.

The same nurse, in showing lack of empathy, does so verbally
and nonverbally. Nonverbally, she indicates that her total
attention is not with the patient by such behavior as looking away
and drinking her juice.

The nurse changes the subject and cuts the patient off,
as in the fbllowing sequences.

Lack of empathy can also be expressed by attending only

to content and ignoring the patient's feelings.

IV.

107

Rehearsal

In the following section, you will have an opportunity to
think of your own empathic responses. You will see more interviews,
but this time there will be a pause after the patient's statement.

During this pause, think of an empathic response, that is,
one that expresses understanding of what the patient is saying and
feeling. Then, when you hear the response of the nurse in the
film, you can compare your response with hers. It is not expected
that your response will match hers. There is no single correct
response to any patient statement. The purpose of this section is not
for you to anticipate what the nurse will say, but to practice con-
structing your own empathic responses, so do not be concerned if
what you say is different from the nurse. You are to respond
mentally only. Please do not say your response aloud or write it
down.

To help you in responding, these interviews have been
filmed as if the patient were talking directly to you. Try to put
yourself into the situation and imagine that you are talking with
the patients in the film.

The first situation takes place in Mrs. H. home. We come
in in the middle of an interview.

The second situation is that of J. This interview takes
place several weeks after the last one we saw, in which he was
grieving over his friend who had died. He is discussing some
personal concerns.

The third and final situation is that of the nurse and Mrs. M.

in the hospital coffee shop.

108

Conclusion

This concludes the lesson on empathy in nursing.

You have seen different interviewers demonstrating empathy
and lack of empathy.

Empathy, remember, was defined as sensitivity to or under-
standing of another's feelings, and communication of this understand-
ing to him or her.

The purpose of this film has been to help you increase your

own empathy skills.

APPENDIX F

Empathy Teaching Project Consent Form

Appendix F

Empathy Teaching Project Consent Form

 

Name of Student

I have freely agreed to participate in a research project comparing
methods of teaching empathy which is being conducted by Janice Layton,
RN, MSN, Project Director, under the supervision of Walter Hapkiewicz,
Ph.D., Associate Professor of Educational Psychology.

I understand that my participation will involve viewing a videotape
(approx. 60 min.), completing a written test, and interviewing
persons hired to play the role of patients on two occasions about
three weeks apart (20-30 min, each time).

I understand that the interviews will be audiotaped and the tapes
rated for empathy, but that my identity will remain anonymous; i.e.,
tapes will be identified by code number, and only the project

director (Janice Layton) will know which numbers are assigned to

whom. I understand that the project director will not use the tapes
except for rating and will not disclose any information about indi-
viduals or their performance to anyone. It has also been explained

to me that the project director's interest is in group not individual
results, and that it is the methods and not individuals that are being
evaluated.

I understand that I may withdraw from the project at any time
without penalty.

I understand that I will be paid $10.00 upon completion of my
participation (viewing videotape, completing tests, and
conducting two interviews), but that if I withdraw before
completion I will receive no payment. '

I understand that, within the limits of anonymity stated above,
I can receive further explanation of the project and its results
from the director when it is completed.

 

 

Signature of Project Director Signature of Student

 

 

Date Date

109

APPENDIX G

Directions for Students
Empathy Teaching Project - Session I

Appendix G

Directions for Students
Empathy Teaching Project - Session I

Videotape

You will see a videotape designed to increase your knowledge
about empathy. Try to learn as much as you can from the tape, but do
not take any notes.

Please do not discuss the videotape or any other parts of the
project with your classmates. This is necessary to obtain an independent
measure of each student's learning. If you have any questions or
comments, the project director (Janice Layton) will be happy to talk to
you.

Written Test

 

Follow instructions on the test.
Interview

You are asked to conduct an interview with a simulated patient
(person hired to play the role of a patient). The interview
should last about 20 minutes, but may go as long as 30 minutes.
(Someone will knock on the door if you are going overtime.)

The interview situation is that of a patient in a clinic who
wants to talk to you, a nurse, about a personal problem (not a medical
problem).

Your goal is to try to understand the person's problem and how
he or she feels about it. As you talk to the "patient," try to apply
what you learned from the videotape.

The interview will be tape-recorded (audiotape, not videotape).
Second Interview

Please be sure you sign up for the second interview, which will
be three weeks from today, before you leave.

You are also asked to try to practice what you learned about

empathy in your interactions with patients between now and the second
interview.

*‘k*******

Please leave these instructions, the test, your answer sheet,
and the pencil here when you leave. '

110

APPENDIX I
Feedback Sheet

Appendix I

Feedback Sheet
DIRECTIONS:

Do not put your name on this sheet.

l. Why did you sign up for this project?
(Circle as many as apply)

Sounded interesting

Needed the money

Thought I might learn something

Interested in research

Other

th—I'

 

2. Did you try to carry out the following directions for the project?
A. Not discussing it with classmates.

. Yes
2. No
Comments

 

B. Practicing what you learned between the two interviews.

l. Yes
2. No
Comments

 

3. Do you feel you took the project seriously and tried to do a good
job?

l. Yes
2. No
Comments

 

4. Do you have any other comments or suggestions you would like to make?

112

APPENDIX J
Rules for Rating Accurate Empathy

In Special Cases*

10.

Appendix J

Rules for Rating Accurate Empathy
In Special Cases*

"Uh-hubs" and other single words expressing attention ("Yeah,"
"Right," etc.) were ggt_rated.

If a therapist response followed another therapist response after
a silence, twg_ratings were made.

Ratings segments always started and finished with client statements.

Appropriate questions of clarification by therapists were standardly
rated level 3 (i.e. questions which did not "take away" from client
affect .

Level 3 rating did ngt_have to include a feeling word if the
therapist response summarized the content and meaning of the client.

Raters could go back and listen to tape segments more than once if
they felt they could hear more data.

A therapist statement interrupted by a client "uh-huh" was also
rated once as long as the therapist was continuing a thought
(i.e. client saying "I'm listening.")

Therapist response without client reaction (at end of segment)
was not rated.

When the therapist expressed his feelings, generally a level 3
rating was given (unless significantly added or subtracted from
on-going material).

Inaudible therapist statements were not rated.

 

*Adapted from Steinitz (l976).

113

APPENDIX K

Scores for Each Subject on Each Variable by Group

Appendix K*

Scores for Each Subject on Each Variable by Group

Group I (Control)

Subject Class

09
13
26
31
35
20
21
23
32
33
37
44
46

MMMMMMMMLQLQQ

Carkhuff

NNNNNNNNNNNNN

.25
.56
.12
.70
.68
.97
.88
.83
.78
.79
.41
.92
.40

 

*Key to abbreviations
Subject - code numbers only given

Time l
B-L**

27
42
55
65
42
42
53
67
38
49
23
54
46

(k=11)

ET

3

\ioowwmxlxoooc-woooo

Time 2

Carkhuff B-L**

NwNNwNNwNNN-JN

.27
.86
.18
.57
.55
.06
.97
.48
.07
.50
.31
.15
.55

 

Class- J for junior and S for senior student
B-L- Barrett-Lennard Relationship'Inventory
ET- Empathy Test

56
16
30
43
17
53
49
50
60
11
27
48
42

ET
(k=14)

01050100

13

N05

10

**A constant (+27) was added to Barrett-Lennard scores to eliminate

negative numbers.

114

Group 11 (Modeling)

Subject

15
28
36
45
55
02
12
30
38
51
54

Class

C...

MMMMMMLQC—nb

Time l

Carkhuff B-L

N

NNNNOOOONNNN

.63
.79
.70
.55
.50
.02
.06
.91
.67
.88
.66

58
51
21
54
32
58
41
44
43
28
38

Group III (Modeling-Labeling)

01
03
11
29
39
40
47
53
04
19
43

C.-

MMMQLLLQQL

N

NNNNNNNNNOO

.46
.20
.71
.61
.65
.40
.70
.66
.88
.86
.92

50
45
54
58
29
12
54
38
29
49
68

115

Time 2
ET Carkhuff B-L

10 2.88 57
5 2.50 55
5 2.79 15
4 2.87 40
9 2.55 60
5 3.17 58
8 2.71 53
7 2.85 45
9 2.43 14
7 2.86 38
7 2.66 49
7 2.55 45
9 2.77 65
8 2.42 57
6 2.71 58
8 2.77 46
9 2.14 43
7 2.55 57
9 2.08 15
9 3.02 66
8 3.03 16
7 2.74 23

rm
.4

sowmwossoowdooosoo

mwmhmmooaiw

mc—l
o

Group IV (Modeling-Rehearsal)

116

 

Time l Time 2
Subject Class Carkhuff B-L ET Carkhuff B-L
08 J 2.63 4l 5 3.24 49
17 J 2.32 30 7 2.96 49
4l J 3.02 25 7 3.06 3l
48 J 2.82 60 6 2.74 33
06 S 2.88 48 6 2.23 44
l0 5 3.04 43 4 2.91 36
25 S 3.06 59 8 2.85 56
42 S 2.90 54 5 2.70 42
49 S 2.86 0l 8 2.83 45
50 S 3.09 39 4 2.82 44
Group V (Modeling-Labeling:Behearsal)
09 J 2.54 32 7 2.94 37
14 J 2.86 35 5 2.88 44
l6 J 2.45 39 7 2.79 50
l8 J 2.58 58 9 3.00 44
52 J 2.43 6l 7 2.79 52
56 J 2.30 36 3 2.50 46
05 S 3.l4 58 6 2.83 65
22 S 2.47 55 7 2.75 65
24 S 2.38 48 7 2.59 49
27 S 3.05 6l 10 2.79 44
34 S 2.57 49 8 2.86 55

E1

12
11

—l
WONONNOWCDOSN oomxowoxoooo

._o
\lo

LIST OF REFERENCES

LIST OF REFERENCES

Bandura, A. Principles of behavior modification. New York: Holt,
Rinehart, and Winston, Inc., 1969.

Bandura, A. Modeling theory: Some traditions, trends, and disputes.
In R. D. Parke (Ed.), Recent trends in social learning theory.
New York: Academic Press, 1972, 35-61.

 

Bandura, A. Social learningitheory. Englewood Cliffs, New Jersey:
Prentice-Hall, Inc., l977.

Bandura, A., Jeffery, R. W., and Bachicha, D. L. Analysis of memory
codes and cumulative rehearsal in observational learning.
Journal of Research in Personality, l974, 7, 295-305.

Campbell, 0. T., and Fiske, D. W. Convergent and discriminant validation
by the multitrait-multimethod matrix. Psychological Bulletin,
l959, 56, 81-l05.

 

Carkhuff, R. Helping and human relations. Volume I: Selection and
training. NewIYork: Holt, Rinéhart, & Winston, Inc., 1969 (a).

Carkhuff, R. Helping and human relations. Volume II: Practice and
research. New York: Holt, Rinehart, B Winston, Inc.:71969 (b).

Chasnoff, S. S. The effects of modeling and ambiguity tolerance on
interview behavior. Counselor Education and Supervision, l976,
6, 46-5l.

Cronbach, L. J. Test validation. In R. L. Thorndike (Ed.), Educational
measurement. Second edition. Washington, D.C.: American
Council on Education, l97l, 443-507.

 

Cronbach, L. J., and Meehl, P. E. Construct validity in psychological
tests. Psychological Bulletin, l955, 52, 28l-302.

Cronbach, L. J., and Snow, R. E. Aptitude and instructional methods:
A handbook for research on interactions. New York: IrVington,
1977.

Dalton, R. F., and Sundblad, L. M. Using principles of social learning
in training for communication of empathy. Journal of Counseling
Psycholggy, l976, 23, 454-457.

 

 

117

118

Dalton, R. F., Sundblad, L. M., and Hylbert, K. W. An application of
principles of social learning to training in communication of
empathy. Journal of Counseling Psychology, l972, 13, 425-429.

 

Ebel, R. L. Estimation of the reliability of ratings. Psychometrika,
l95l, lg, 407-424.

 

Eisenberg, S., and Delaney, D. J. Using video simulation of counseling
for training counselors. Journal of Counseling Psychology,
l970, 12, 15-l9.

Eskedal, G. A. Symbolic role modeling and cognitive learning in the
training of counselors. Journal of Counseling Psychology,
l975, 22, l52-155.

 

Eysenck, H. J. The effects of psychotherapy. IN H. J. Eysenck (Ed.),
Handbook of abnormal psychology. New York: Basic Books, 1960,
697-725.

 

Friedman, P. H. The effects of modeling, roleplaying, and participation
on behavior change. In A. Maher (Ed.), Progress in experimental
personality research. Vol. 6. Academic press, 1972, 4l-8l.

 

 

Friedrich, L. K., and Stein, A. H. Prosocial television and young
children: The effects of verbal labeling and role playing
on learning and behavior. Child Development, 1975, 46, 27-38.

 

Gagné, R. M. The conditions of learnipg. Third edition. New York:
Holt, Rinéhart, and WifiSton, l977.

 

Gerst, M. S. Symbolic coding processes in observational learning.
Journal of Personality and Social Psychology, l97l, lg, 7-l7.

Gulanick, N., and Schmeck, R. R. Modeling, praise, and criticism in
teaching empathic responding. Counselor Education and Supervision,
l977, 11, 284-29l.

 

Hempel, C. G. Philosophy of natural science. Englewood Cliffs, N.J.
Prentice-Hall, Inc., 1966.

 

Ivey, A. E., Normington, C. J., Miller, C. D., Morrill, W. H., Haase,
R. F. Microcounseling and attending behavior: An approach to

prepracticum counselor training. Journal of Counselin
Psychology: Monograph Supplement, 1968, 15, 1-12.
Jeffery, D. B. The influence of symbolic and motor rehearsal on obser-

vational learning. Journal of Research in Personality, l976, lg,
ll6-127.

 

 

Kalisch, B. J. An experiment in the development of empathy in nursing
students. Nursing Research, 1971, 29, 202-211.

 

Kerlinger, F. N. Foundations of behavioral research. 2nd ed. New York:
Holt, Rinehart, and Winston, l973.

 

119

Kuna, D. J. Lecturing, reading, and modeling in counselor restatement
training. Journal of Counseling Psycholpgy, 1975, 22, 542-546.

 

Kurtz, R. R. A comparison of different approaches to the measurement
of counselor empathy in personal counseling. Unpublished
doctoral dissertation, Michigan State University, 1970.

Kurtz, R. R., and Grummon, D. L. Different approaches to the measurement
of therapist empathy and their relationship to therapy outcomes.
Journal of Consulting and Clinical Psychology, 1972, 55, 106-115.

 

LaMonica, E. L., Carew, D. K., Winder, A. E., Haase, A. M. B., and
Blanchard, K. H. Empathy training as the major thrust of a
staff development program. NursingyResearch, 1976, 25, 447-451.

 

Langer, L. The implications for construct validity and Rogerian theory
in three measures of empathy. Unpublished master's thesis,
Michigan State University, 1972.

Magnusson, D. Test theory. Reading, Mass.: Addison-Wesley Publishing
Company, 1966.

 

Mansfield, E. Empathy: Concept and identified psychiatric nursing
behavior. Nursing Research, 1973, 22, 525-530.

Payne, P. P., Weiss, S. D., and Kapp. R. A. Didactic, experiential,
and modeling factors in the learning of empathy. Journal of
Counseling Psychology, 1972, 15, 425-429.

 

 

Payne, P. P., Winter, 0. E., and Bell, G. E. Effects of supervisor
style on the learning of empathy in a supervision analogue.
Counselor Education and Supervision, 1972, 11, 262-269.

 

Peitchinis, J. A. Therapeutic effectiveness of counseling by nursing
personnel. Nursing Research, 1972, 21, 138-148.

 

Perry, M. A. Modeling and instructions in training for counselor
empathy. Journal of Counseling Psycholpgy, 1975, 22, 173-179.

 

Rank, R. C., Thoresen, C. E., and Smith, R. M. Encouraging counselor
trainee affective group behavior by social modeling. Counselor
Education and Supervision, 1972, 11, 270-278.

 

Rappaport, J., and Chinsky, J. M. Accurate empathy: Confusion of a
construct. Psychological Bulletin, 1972, 12, 400-404.

 

Rappaport, J., Gross, T., and Lepper, C. Modeling, sensitivity training,
and instructions: Implications for the training of college
student volunteers and for outcome research. Journal of Consult-

 

ing and Clinical Psychology, 1973, 49, 99-107.

Rogers, C. R. The necessary and sufficient conditions of therapeutic
change. Journal of Consulting-Psychology, 1957, 21, 95-103.

 

120

Rogers, C. R. A theory of therapy, personality and interpersonal
relationships as developed in the client-centered framework.
In S. Koch (Ed.), Psycholpgy: A study_of science. Vol. III.
Formulations of the person and the social context. New York:
McGraw-Hill, 1959, 184-256.

 

Rogers, C. R. Empathic: An unappreciated way of being. The Counseling
Psychologist, 1975, 5, 2-10.

 

 

Rogers, C. R., Gendlin, E., Kiesler, D., and Traux, C. B. (Eds.)
The therppeutic relationship_and its impact. Madison: The
University of Wisconsin Press, 1967.

 

Ronnestad, M. H. The effects of modeling, feedback, and experiential
methods on counselor empathy. Counselor Education and
Supervision, 1977, 15, 194-201.

 

 

Rosendahl, P. L. Effectiveness of empathy, nonpossessive warmth, and
genuineness on self-actualization of nursing students. Nursing
Research, 1973, 22, 253-257.

Scheifley, V. M., and Schmidt, W. H. Jeremy D. Finn's Multivariance--
Univariate and multivariate analysis Of variance, covariance,
and regression: Modified and adapted for use on the CDC 6500.
Occasional paper No. 22, Office of Research Consultation,
School for Advanced Studies, College of Education, Michigan
State University, October, 1973.

 

 

 

Standards for educational and psychological tests. Washington, D.C.:
TAmeriEan—Psychological Association, Inc., 1974.

 

Steinitz, H. S. Perspectives on psychotherapy process and empathy.
Unpublished doctoral dissertation, Michigan State University,
1976.

Stetler, C. B. Relationship of perceived empathy to nurses communication.
Nursing Research, 1977, 25, 432-438.

 

Stone, G. L., and Vance, A. Instructions, modeling, and rehearsal:
Implications for training. Journal of Counseling_Psychology,
1976, 25, 272-279.

 

Traux, C. B. The meaning and reliability of accurate empathy ratings:
A rejoinder. Psychological Bulletin, 1972, 12, 397-399.

 

Traux, C. B., and Carkhuff, R. Toward effective counselino and
psychotherapy. Chicago: Aldine Publishing Company, 1967.

 

Uhlemann, M. R., Lea, G. W., and Stone, G. L. Effect of instructions
and modeling on trainees low in interpersonal-communication
skills. Journal of Counseling Psychology, 1976, 25, 509-513.

 

121

Wallace, W. G., Horan, J. J., Baker, S. B., and Hudson, G. R.
Incremental effects of modeling and performance feedback in
teaching decision-making counseling. Journal of Counseling
Psychology, 1975, 22, 570-572.

 

GENERAL REFERENCES

GENERAL REFERENCES

Aspy, D. N. Empathy: Let's get the hell on with it. The Counseling
Psychologist, 1975, 5, 10-14.

 

Avery, A. W., D'Augelli, A. R., and Danish, S. J. An empirical
investigation of the construct validity of empathic understanding
ratings. Counselor Education and Supervision, 1976, 15, 177-183.

 

Bandura, A., and Jeffery, R. W. Role of symbolic coding and rehearsal
processes in observational learning. Journal of Personality
and Social Psychology, 1973, 25, 122-130.

 

 

Beutler, L. E., Johnson, D. W., Neville, C. W., and Workman, S. N.
Some sources of variance in "Accurate empathy" ratings.
Journal of Consultingyand Clinical Psychology, 1973, 39,
167-169.

 

Blaas, C. D., and Heck, E. J. Accuracy of accurate empathy ratings.
Journal of Counseling Psychology, 1975, 22, 243-246.

 

Bozarth, J. D., and Krauft, C. C. Accurate empathy ratings: Some
methodological considerations. Journal of Clinical Psychology,
1972, 25, 408-410.

Brammer, L. Re:Views: Training packages. Personnel and Guidance
Journal, 1977, 55, 612-618.

 

Breisinger, G. 0. Sex and empathy, reexamined. Journal of Counseling
Psychology, 1976, 25, 289-290.

 

 

Carkhuff, R. R. The development of systematic human resource development
models. The Counseling Psychologist, 1972, 5, 4-11.

 

Carkhuff, R. R. New directions in training for the helping professions:
Toward a technology for human and community resource development.
The Counseling Psychologist, 1972, 5, 12-30.

 

Carkhuff, R. R. What's it all about anyway? Some reflections on help-
ing and human resource development models. The Counseling
Psychologist, 1972, 5, 79-87.

 

Chalmers, D., and Rosenbaum, M. Learning by observing vs. learning by
doing. Journal of Educational Psychology, 1974, 55, 216-224.

Chinsky, J. M., and Rappaport, J. Brief critique of the meaning and
reliability of "Accurate empathy ratings." Psychological
Bulletin, 1970, 25, 379-382.

 

122

123

Conklin, R. C., and Hunt, A. S. An investigation of the validity of
empathy measures. Counselor Education and Supervision, 1975,
15, 119-127.

Cronbach, L. J. Processes affecting scores on "Understanding of others"
and "Assumed similarity." Psychological Bulletin, 1955, 52,
177-193.

 

Deutsch, F., and Madle, R. A. Empathy: Historic and current
conceptualizations, measurement, and a cognitive theoretical
perspective. Human Development, 1975, 15, 267-287.

Doster, J. A. Effects of instructions, modeling, and role rehearsal
on interview verbal behavior. Journal of Consulting and Clinical

 

Psychology, 1972, 55, 202-209.

 

Dowling, T. H., and Frantz, T. T. The influence of facilitative
relationships on imitative learning. Journal of Counselipg
Psychology, 1975, 22, 259-263.

 

 

Feshbach, N. D. Empathy in children: Some theoretical and empirical
considerations. The Counseling Psychologist, 1975, 5, 25-30.

Gazda, G. M., Asburg, F., Balzer, F., Childers, W. C., and Walters, R. P.
Human relations development: A manual for educators. 2nd ed.
Boston: Allyh and Bacon, Inc., 1977.

Gazda, G. M., Walters, R. P., and Childers, W. C. Human relations
development: A manual for health sciences. Boston: Allyn and
Bacon, Inc., 1975.

 

Glass, G. V., and Stanley, J. C. Statistical methods in education and
psychology. Englewood Clifs, N.J.: Prentice-Ha1l, Inc. 1970.

Gormally, J., and Hill, C. E. Guidelines for research on Carkhuff's
training model. Journal of Counselipchsychology, 1974, 21,
539-547.

Green, A. H., and Marlatt, G. A. Effects of instructions and modeling
upon affective and descriptive verbalizations. Journal of
Abnormal Psychology, 1972, 59, 189-196.

 

Greif, E. B., and Hogan, R. The theory and measurement of empathy.
Journal of Counseling Psychology, 1973, 29, 280-284.

Grummon, D. L. Client-centered theory. In H. M. Burks (Ed.), Theories
of counselipg, New York: Mc-Graw Hill Book Company, 1978.

Hefele, T. J., and Hurst, M. W. Interpersonal skill measurement:

Precision, validity, and utility. The Counseling Psychologist,
1972, 5, 62-70.

Heller, K. Effects of modeling procedures in helping relationships.
Journal of Consulting and Clinical Psychology, 1969, 55, 522-526.

124

Hill, C. E., and King, J. Perceptions of empathy as a function of the
measuring instrument. Journal of Counselinchsychology, 1976,
23, 155-157.

 

Hogan, R. Empathy: A conceptual and psychometric anlaysis. 155
Counseling Psychologist, 1975, 5, 14-18.

 

Horwitz, M. B. A comment on "An empirical investigation of the construct
validity of empathic understanding ratings." Counselor Education
and Supervision, 1977, 11, 292-295.

 

Iannotti, R. J. The nature and measurement of empathy in children.
The Counselinchsychologist, 1975, 5, 21-25.

 

Ivey, A. E. A time for action: Can counseling psychology respond to
systematic training programs? The Counseling Psychologist, 1972,
3, 70-74.

 

Kagan, N. Observations and suggestions. The Counseling Psychologist,
1972, 5, 42-45.

 

Kagan, N. Influencing human interaction: Fifteen years with IPR.
Unpublished manuscript, Michigan State University, 1977.

Kalisch, B. J. Strategies for developing nurse empathy. Nursing
Outlook, 1971, 15, 714-718.

Kalisch, B. J. What is empathy? American Journal of Nursing, 1973, 25,
1548-1552.

 

Kanfer, F., and Phillips, J. Learning foundations of behavior therapy.
New York: John Wiley & Sons, Inc., 1970.

 

Kratochvil, D. W. Changes in values and in interpersonal functioning of
counselor candidates. Counselor Education and Supervision, 1969,
8, 104-107.

 

Labouvie, E. W. The dialectical nature of measurement activities in
the behavioral sciences. Human Development, 1975, 15, 396-403.

 

Lebow, J. A., and Doty, D: W. An evaluation of empathic listening in
telephone counseling. Journal of Counselinchsychology, 1976, 25,
532-537.

Lincoln, R., Layton, J., and Holdman, H. Using simulated patients to
teach assessment. Nursipg Outlook, 1978, 25, 316-320.

 

McGuire, D., Thelen, M. H., and Amolsch, T. Interview self-disclosure
as a function of length of modeling and descriptive instructions.
Journal of Consulting and Clinical Psychology, 1975, 55, 356-362.

PopP, J. A. What is the problem of construct validity? Paper presented
at the—American Educational Research’Association, Washington, D.C.,
April, 1975.

125

Resnikoff, A. Critique of the human resource development model from
the viewpoint of rigor. The Counselinchsychologist, 1972, 5,
46-55.

 

Saltmarsh, R. E. Development of empathic interview skills through
programmed instruction. Journal of Counseling Psychology, 1973, 29,
375-377.

Shantz, C. U. Empathy in relation to social cognitive development.
The Counseling Psychologist, 1975, 5, 18-21.

Shapiro, J., and Gust, T. Counselor training for facilitative human
relationships. Counselor Education and Supervision, 1974, 15,
198-206.

 

Stone, G. L., and Gotlib, I. Effect of instructions and modeling on
self-disclosure. Journal of Counseling Psychology, 1975, 22,
288-293.

 

Stotland, E. Exploratory investigations of empathy. In L. Berkowitz
(Ed.), Advances in experimental social psychology. Vol. 4.
New York: Academic Press:’1969, 271-314.

 

Straham, C., and Zytowski, D. G. Impact of visual, vocal, and lexical
cues on judgments of counselor qualities. Journal of Counseling
Psychology, 1976, 25, 387-393.

 

 

Thoresen, C. E. Constructs don't speak for themselves. Counselor
Education and Supervision, 1977, 12, 296-303.

 

Traux, C. B. Therapist empathy, warmth, and genuineness and patient
personality change in group psychotherapy: A comparison between
interaction unit measures, time sample measures, patient
perception measures. Journal of Clinical Psychology, 1966, 22,
225-229.

 

Traux, C. B., and Wargo, D. G. Psychotherapeutic encounters that change
behavior: For better or for worse. American Journal of
Psychotherapy, 1966, 25, 499-520.

 

Wallston, K. a., and Weitz, L. J. Measurement of the core dimensions
of helping. Journal of Counseling Psychology, 1975, 22, 567-569.

 

Whalen, C. Effects of a model and instructions on group verbal behaviors.
Journal of Consulting and Clinical Psychology, 1969, 55, 509-524.

 

MICHIGAN STATE UNIV. LIBRARIES
1|HI1111111111111111111111111111111111111
31293104097989