Impact of counselor race and power base on black and white college students


Material Information

Impact of counselor race and power base on black and white college students
Physical Description:
ix, 152 leaves : ; 28 cm.
Peddle, Mark J
Publication Date:


Subjects / Keywords:
Student counselors   ( lcsh )
Race relations   ( lcsh )
Interpersonal attraction   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1983.
Includes bibliographical references (leaves 140-151).
Statement of Responsibility:
by Mark J. Peddle, III.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000466429
notis - ACN0725
oclc - 11557235
System ID:

Full Text








Copyright 1983


Mark J. Peddle, III


I would like to thank all of the members of my committee for

their thoughtful consideration throughout the progress of this disser-

tation. Dr. Rod McDavis, chairperson, was constantly available with

positive support. He showed significant sensitivity to my personal


In addition, I would like to thank all of my counselors: Louise

Ousley, Janice Guinyard, Jan Stretch, and Lynn Solomon. They patiently

gave me much of their valuable time. Donna Dolittle deserves thanks

for her tireless energy and cheerful assistance. Barbara Smerage

gave me a professional product with a touch of kindness. Finally,

I reserve my deepest appreciation, love, and respect for my wife,

Elovia. Her devotion and confidence sustained me through the most

extreme stress.


ACKNOWLEDGMENTS. . . ... .... iii

LIST OF TABLES . . ... ...... .vi

ABSTRACT . . .. . viii



Problem Statement . .... 1
Purpose . . ... 3
Need..................... 4
Significance of the Study ........... 5
Definition of Terms . 7
Organization of the Study . 8


Black Attitudes Toward Counseling . 9
Black Responses to Counseling ... .15
Suggested Approaches and Theory ... .22
The Interpersonal Influence Model ... .32
Expertness . . .... 38
Attractiveness. ........ .. 45
Expertness and Attractiveness ........ 52
Summary . . ... 58

THREE METHODOLOGY . . ... .. .60

Research Design . ... 60
Rationale for Design . 61
Hypotheses . . .. 62
Subjects . . .. 62
Instrumentation ........ . 64
The Counselor Rating Form ......... 64
Problem Solving Attitude Scale. .. 66
Behavioral Measure . ... 67
Treatment . . 68
Interview Roles . 68
Interviewers and Interviewer Training 70
Procedures . . 71
Data Analysis . . 73

Methodological Assumptions. ..
Limitations . .


Results . .
Hypothesis One. .
Hypothesis Two. .
Hypothesis Three. .
Hypothesis Four .
Quality Control .
Summary . .
Discussion. . .


Conclusions . .
Implications. . .
Summary . .
Recommendations .


















. 107

. 111

* 116

. 118

. 122

S. 125

. 128

. 130

. 135

. 137

S. 139

S. 140

. 152

. t .
. e .


S . 76
S . 76

. . 78

S 82
S 83
S 86

S 98

S 98
S 105



1 Treatment Conditions . .

2 Counselor Expertness-Attractiveness Rating--
Preexperiment . . .

3 Three Factor Analysis . .

4 Treatment x Client Race . .

5 ANOVA of Counselor Race by Role and Client Race
for Expert Variable . .

6 ANOVA of Counselor Race by Role and Client Race
for Attractiveness Variable . .

7 Means and Standard Deviations for Counseling
Variables by Client Race . .

8 Means and Standard Deviations for Counseling
Variables by Counselor Role . .

9 Rank Order of Means of Perceived Expertness by
Race and Treatment Condition . .

10 Rank Order of Means of Perceived Attractiveness
by Race and Treatment Condition . .

11 ANOVA of Treatment by Race for Problem Solving
Attitude Variable . .

12 ANOVA of Counselor Race by Role and Client Race
for Problem Solving Attitude Variable .

13 Means and Standard Deviations for Problem Solving
Attitudes by Client Race . .

14 Rank Order of Means of Problem Solving Attitude by
Race and Treatment Condition . .

15 Loglinear Analysis of Treatment by Client Race
for Return of Questionnaire . .

S 71

S 74

S 74

S 79

S 80

S 81

S 81

S 81

S 82

S 83

S 84

S 84

S 85

S 85


16 Questionnaire Response Frequencies by
Treatment and Race. . ... 87

17 Rank Order of Questionnaire Return by
Treatment and Race. . ... 87

18 T-Test Results on Role Maintenance by Counselors. .. 89


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



Mark J. Peddle, III

August 1983
Chairman: Rod McDavis
Major Department: Counselor Education

The study was designed to assess the effects of black and white

counselors' expert and referent power bases on the perceptions, at-

titudes, and behavior of black and white clients. The study makes use

of the interpersonal influence counseling model which specifically

identifies the components of different counselor roles. The expert and

referent roles were individually defined by behavior and reputational


In a factorial design black and white college students were

randomly assigned to either a treatment or control condition. Treat-

ment consisted of seeing either a black or white counselor who was

trained to be either expert or referent. Those subjects in the con-

trol group merely read the same decision making literature discussed

in the counseling interview. Subjects who saw counselors rated their

perception of the counselor's expertness and attractiveness. All

clients' attitudes towards problem solving were assessed. There was

also a behavioral measure of rate of return of a follow-up


questionnaire. The independent variables were counselor race, coun-

selor role, and client race. The dependent variables were client

perception of the counselor, client attitude toward problem solving,

and the frequency of return of a mailed questionnaire.

It was hypothesized that there would be no differences among

black and white perceptions of counselors, no differences among black

and white attitudes towards problem solving, no differences among

black and white rates of return of a questionnaire, and no interac-

tions between any of the independent variables.

Analyses of data determined the black clients perceived all the

counselors in general to be more expert than the white clients per-

ceived them. There was a significant main effect for counselor role

such that all clients perceived the referent counselors as being sig-

nificantly more attractive than the expert counselors. Black subjects

had a significantly more positive attitude toward problem solving

than white subjects. There were no significant differences in rate of

return of a questionnaire. There were no significant interactions.

Implications for practice, training, and research are discussed.


Problem Statement

Griffith (1977) stated that race is unquestionably an influence

in both racially similar and dissimilar therapeutic relationships. For

Griffith the critical question is not whether race influences the re-

lationship between client and therapist, but under what conditions and

in what ways. During the past 15 years there has been a significant

increase in the amount of research devoted to the investigation of

race as a variable in counselor client relationships. Unfortunately,

a large body of the research has avoided the significant aspects of the

questions posed by Griffith (1977). Instead, most of the research has

consisted of essays and discussions about the nature and needs of the

interracial therapeutic dyads.

Making theoretical assumptions about the nature of black-white

relationships in therapy is fundamentally sound and expected at the

beginning of an investigation of most fields of study. Essays by

Vontress (1969, 1970) and Williams and Kirkland (1971), which pointed

to cultural differences and educational and experimental needs, pro-

vided the subject matter for experimentation in the field. The past

15 years, however, have provided a lot of conflicting material, much

of which is merely anecdotal and uncontrolled (Griffith, 1977).

Reviews of literature on the racial issue in therapy between

blacks and whites have mirrored the conflicting results of the


information available (Griffith, 1977; Harrison, 1975; Sattler, 1977).

According to Griffith (1977) the research does support the conclusion

that "social differences have a somewhat negated effect upon psycho-

therapy" (p. 33). Sattler (1977) takes the opposite view that race

"is not a significant variable in affecting the performance and reac-

tions in these types of nonclinical initial interviews" (p. 39).

Harrison (1975) acknowledges the mixed results of the available lit-

erature but holds for a general trend counselees tend to prefer

counselors of the same race, particularly if they are black counselors"

(p. 131). All three reviewers call for research on the counseling

process and on behavioral outcomes of interracial counseling. They

suggest the need for the study of interactions. Bryson and Bardo

(1975) more explicitly conclude that counselor race is in itself insuf-

ficient to predict effectiveness of counseling but must be considered

with other variables.

The results of reviews of literature on racial dyads in counseling

lead to such different conclusions that looking at racial differences

alone comes into question as being too simplistic. There are too few

experimental studies with control or comparison groups and too few

that look at interactions between race and counseling variables

(Parloff, Waskow, & Wolfe, 1978). If the essential questions for

counselors remain--what kind of treatment is more effective with what

kinds of client in what kind of situation--then those questions should

be addressed by research.

Some attempts have been made to investigate direct versus indirect

therapist styles with different racial populations. Even in these

studies, however, the results have been mixed or the control lacking.

Tucker (1973) showed black counselees benefitted more from his action

model than from an insight oriented approach, but his study lacked

white clients for comparison. Peoples and Dell (1975) used such a

comparison group but found both black and white students preferred

active versus passive counselors. Atkinson, Maruyama, and Matsui

(1978) found a directive approach made the counselor much more

credible and approachable to Asian-American students. The only

comparable study using black clients has been a recent doctoral dis-

sertation by Boger (1981). In her analogestudy, no significant

differences in perceptions existed for black versus white subjects

who viewed tapes of directive versus an insight oriented approach.


The purpose of this study was to include critical elements that

have been lacking in earlier research on interracial counseling dyads.

The study made use of a counseling model (Strong & Matross, 1973)

which specifically delineates the components of different counselor

styles or roles. Some components are in terms of behavior, some in

terms of reputation, and some in terms of objective evidence such as

decor. The specific counselor roles were "expert" and "referent."

Other critical elements that were included were randomization, con-

trol groups, multiple independent variables, a factorial design, and

a variety of dependent variables including perceptual, attitudinal,

and a behavioral measure.

The purpose was to assess the effects of black and white coun-

selors' expert and referent power bases on the perceptions, attitudes,

and behavior of black and white clients. Black and white counselors

from each role taught a problem solving strategy to student inter-

viewees. Immediately following the interview students took a Problem

Solving Attitude Scale (Carey, 1958) and the Counselor Rating Form

(Barak & LaCrosse, 1975) to measure their perceptions of counselor

expertness and attractiveness (referent power). A behavioral measure,

of inferred therapist influence, was the rate of return of a mailed

questionnaire. Each dependent measure allowed the investigation of

possible separate and combined effects of counselor race, client race,

and counselor rule.


One difficulty which may have plagued earlier studies of thera-

pist style may have been the complexity of the therapist behavior.

The complexity of a very general style of behavior may mask the very

details that need to be identified as effective or detrimental. In-

stead of using untested proposals it may be more advantageous to make

use of a specific counseling model that has increasing empirical

support. If a style can be specifically identified, described, and

supported by research as having specific effects with one segment of

the treatment population at large, then that style can be logically

investigated for its effectiveness with minority groups in our

society. If a "style" or "styles" are part of a theoretical model

and are specifically defined, then benefits can be derived not only

for the target minority population but also for the counseling model's

development. The basis for such a model has existed in social

psychology. The model has received extensive explanation and inves-

tigation since first introduced by Strong in 1968. The revised model


of Strong and Matross's (1973) interpersonal influence model of counsel-

ing may be the means for advancing our knowledge of the process of

counseling in general and of counseling blacks in particular.

Even within the proscribed area of social influence research there

exists a need to further define the counselor and client characteris-

tics that alter its effectiveness. Merluzzi, Merluzzi, and Kaul

(1977) conducted the only study of the Strong and Matross (1973) model

that systematically varied race, and in that study only counselor race

was varied. All the clients were white. A short review of this study

states .. subjects were influenced more by racially similar re-

ferent and racially dissimilar expert counselors (Corrigan,

Dell, Lewis, & Schmidt, 1980, p. 425). A real need in this research

area has been to ascertain the possible differential responses of

black clients to different power bases used by different race thera-


Expert and referent power bases have evidential and behavioral

cues that have been specifically delineated in the literature. If

these specific variables were studied in interaction with client race

and counselor race, then specific conclusions might be drawn for

counseling blacks in similar situations. Additionally the social

influence model might be expanded to include the client characteristic

of race as a determinant of power base effectiveness.

Significance of the Study

Establishing a relationship between counselor race, counseling

style, and client race could yield benefits for a counseling model,

research, training, and practice. The Strong and Matross (1973)

social influence model of counseling has concentrated on explaining

the characteristics of the counselor role that determine the effec-

tiveness of an influence attempt. A study by Merluzzi et al. (1977)

has introduced the counselor characteristic of race as a determinant

of counselor effectiveness. This study proposed to investigate the

need to include the client characteristic of race as a variable that

interacts with counselor role and race to affect counselor influence.

The effective demonstration of a relationship between counselor

race, counseling style, and client race would enable researchers to

focus on the relative strength of each specific behavior subsumed

under the counselor roles of expertness and attractiveness. It may

be that just one element of any given role is singularly effective in

influencing different race clients. Differential role effectiveness

would also encourage an expanded exploration of counselor roles to

include other racial and ethnic groups within our society.

Empirical evidence of the differential effectiveness of specific

roles with black clients would provide impetus for changes in the

academic and experiential training of future counselors. Counselor

education programs are increasingly addressing the need to consider

the different value systems of minority clients. This study should

provide some specific guidelines for counselors to adopt, at least in

the beginning stages of working with black clients.

Practicing counselors would hopefully see the demonstration of

techniques that are effective in working with black clients. Clinical

application of counseling behaviors that are empirically demonstrated

as effective with blacks would enhance the overall effectiveness of

a given agency's service delivery. Black clients would directly

benefit by the elimination of inappropriate counselor behaviors and

by the encouragement of effective behaviors.

Definition of Terms

Attractiveness--perceived similarity between communicator and re-

cipient. Similarity is in terms of basic values and attitudes

(Strong, 1968). Attractiveness is also a dimension measured by

the CRF (Barak & LaCrosse, 1975).

Career Planning Problems--indecision regarding alternative occupations

or alternative steps leading to one occupation; also ignorance

of logical decision making steps that can be taken to aid in

occupational choice or attainment.

Expertness--extent to which communicator is perceived as source of

valid assertions. Pertinent knowledge and skills as perceived

by the client (Strong, 1968). Expertness is also a dimension

measured by the CRF (Barak&LaCrosse, 1975).

Expert Power Base--represents client's need to reduce cost of a

particular goal and counselor's ability to do so. It is

impersonal and structured.

Power Bases--distinct sources of counselor power that consist of

different pairs of client needs and counselor resources.

Problem Solving--a behavioral process, whether overt or cognitive

in nature, which (a) makes available a variety of effective

response alternatives to a situation and (b) increases the

probability of selecting the most effective response

(D'zurilla & Goldfried, 1971).

Referent Power Base--predicated on interpersonal attraction. Rep-

resents the client's self-perceived inconsistency between

behavior and values encountering a perceived similarity with

the counselor's world views. The client needs greater

consistency between values and behavior; the counselor is

presumed to have that consistency.

Organization of the Study

The remainder of this study is organized into four chapters.

The second chapter includes a review of the related literature. Topics

covered in this section include a review of research on black atti-

tudes toward counselors, black responses in counseling, suggested

approaches for counseling blacks, the interpersonal influence model

(Strong & Matross, 1973), expertness, attractiveness, and both dimen-

sions together in counseling. The third chapter represents the

research methodology. The fourth chapter contains the results of

this study. The final chapter summarizes the study, discusses its

implications, and makes suggestions for further research.


The literature review presents a discussion of the research

which forms the background for this study. There are seven sections

to this chapter. The first of these deals with studies of black

clients' attitudes and preferences for counselors. The second sec-

tion concentrates more on the process of counseling and the black

client's responses to it. The third section presents the approaches

and considerations that theorists have suggested for counseling the

black client. The fourth section is a description of the bases for

the interpersonal model for counseling. The fifth section concen-

trates on research of counselor expertness, while the sixth section

concentrates on counselor attractiveness. The final section presents

research into the differential effectiveness of expertness versus


Black Attitudes Toward Counseling

In 1967 Banks, Berenson, and Carkhuff conducted a study focusing

on the interaction of counselor race and training in counseling black

clients. The results from this study prompted much of the research into

the demographic variable of race that was to follow. They used one

black and three white counselors in initial interviews. Taped excerpts

from these interviews were rated on relationship skills such as empathy,

genuineness, rapport, depth of exploration, and concreteness. All of


these values had been shown to be related to constructive client change

(Carkhuff, 1966). The results of the study showed that the black and

two inexperienced white counselors trained in the relationship oriented

approach were significantly more effective than the traditionally

trained white therapist. The researchers acknowledged, however, that

their most important result was probably the discovery that all black

clients would return to see the black counselor, but none would return

to see the traditional white counselor. "Further, if the data on the

Negro counselor were not considered, 16 of 24 or two-thirds of the

clients would not return to see a white counselor for a second ses-

sion" (Banks et al., 1967, p. 72).

In a study specifically geared to assessing black attitudes

toward counseling, Cimbolic (1972) randomly selected black freshmen

from the University of Missouri student body. Subjects were randomly

assigned to either experienced or inexperienced counselors who in turn

were either black or white. The selection and design procedures

were a deliberate attempt to improve on the Banks et al. (1967) study.

The results of Cimbolic's study were a surprising rebuttal to

the earlier findings of Banks et al. (1967). The subjects did not

seem to favor the counselors on the basis of race but rather on the

basis of experience and skill level in facilitative conditions. In

addition, all of the black counselees were willing to return to at

least one of the white counselors.

A similar positive result for the white therapist-black client

pairing was found by Ewing (1974). His study strongly agreed with the

implications of Cimbolic's (1972) study. The results did not favor

the hypothesis that black counselors were more effective with black

clients. All interview sessions consisted of one time precounseling

interviews. College students were randomly assigned to black and

white counselors, all of whom were highly experienced. Not only did

the black clients rate white counselors as favorably as they did black

counselors, but the black clients tended to rate counseling more fav-

orable overall. Ewing concluded that it is not the similarity of

racial background that is important but the training and experience

of the counselor.

In his discussion Cimbolic (1972) made some suggestions which may

help explain the differences in results between the Banks et al. (1967)

study and Cimbolic's (1972) and Ewing's (1974) studies. Cimbolic's

clients participated in a four interview session study. Cimbolic sug-

gested that a future study look at precounseling attitudes of the

black clients. He suggested some process could have taken place over

time to cancel out the effects of different races.

A later study by Wright (1975) tracked the change in attitudes of

black clients over five sessions. Initially, no preconceived notions

were reported. As the number of sessions increased, there was an

apparent increase in the favorable perception of counselors of the

opposite race. Using the Barrett-Leonard Inventory (1962) Wright

managed to show there was a relationship that existed between counselor

race and student perceptions of trust, and that this relationship

changed over time. This finding offers a reasonable explanation for

the different results between the Banks et al. (1967) study and that

of Cimbolic. Unfortunately the issue is not clear because Ewing used

only one interview session to obtain positive results for the inter-

-racial dyad.

The vast majority of research depicts a black client population

that not only prefers black counselors before entering counseling, but

is more satisfied with them afterwards. Heffernon and Breuhl (1971)

trained college males in Rogerian principles of counseling. The four

black and four white counselors were each assigned to three black

eighth graders. The eighth graders were matched on intelligence plus

several other variables. Significant results were obtained on a

behavioral measure, when the boys were given a choice to go to the

library or return to visit the counselor. All of the black counsel-

ors' boys returned, but less than one half of the white counselors'

boys did.

Race, education, and experience were the variables investigated

by Gardner (1972) to assess their effects on black clients' perceptions

of counselor effectiveness. All three variables were found to be

significant for effectiveness with black clients. Race was the most

significant. Gardner had used 24 male and 24 female black students

with four white and four black counselors in a format similar to the

Banks et al. (1967) study. Counselors were rated not only by the

students but by two Ph.D level judges on their interpersonal


The Gardner study was positive in two of its conclusions. First,

the results indicated that individuals with backgrounds similar to the

client's could be recruited and trained as counselors. Secondly, the

author felt individuals who were not similar in background, but who

could generate facilitative conditions, would be appropriate for use

with black clients.

Linking possible interaction of the racial issue in counseling to

experience of the counselor seemed to be the beginning of a decade-

long attempt to study possible interactions with the demographic

variable of race. Jackson and Kirschner (1973) found that there was

a significant difference for counselor preference based upon the

black's self-designation as a black, Afro-American, or Negro. Those

who referred to themselves as blacks or Afro-Americans preferred a

counselor of African descent. No significant differences were found

relating to sex, age, or socioeconomic background.

Focusing entirely on the relationships of race and social class,

Wolkon, Moriwaki and Williams (1973) had findings that support a

social class factor operating in conjunction with black attitudes.

Social class was the significant variable related to orientation to

therapy, but race was crucial with self-disclosure and with treatment

satisfaction. Regardless of class, blacks disclosed less even to

other blacks and were not satisfied with the treatment they received.

Gordon and Grantham (1979) had similar results in studying the

helper preference of disadvantaged students. They discovered a slight

preference for a helper of the same sex, age, and race. There was a

definite preference for a helper of the same social class background.

Studying counselor preference among a sample of delinquent girls,

Gamboa, Tosi, and Riccio (1976) looked at three possible problem

areas: personal-social, educational, and vocational. In general,

the girls were most willing to see a counselor to discuss educational

and vocational issues. Black girls had a significant negative attitude

toward personal-social counseling. This information is similar to what

Woldon et al. (1973) noted about self-disclosure.

Most studies of black attitudes and responses to counseling have

been in a university setting. In that environment Thompson and

Cimbolic (1978) looked at black student preferences and attitudes

toward counseling center use in a predominantly white university.

When forced to choose between a black and white counselor, the students

tended to choose black counselors. The black students also indicated

that their likelihood of using the counseling center increased with

the likelihood that their preference would be satisfied. Similar

results were obtained in dissertations by Peoples (1978) and Albert

(1978). Albert's study compared predominantly white and black

universities. On both campuses race was a significant issue in the

black client's preference for a counselor.

In a study of the three racial groups, Schneider, Laury, and

Hughes (1980) found main effects for race, sex, and provider groups

for Chicano, black and white college students. Blacks and Chicanos

reported they would more likely take their personal problems to

professional mental health workers. On the surface this result seems

counter to that reported by Gamboa et al. (1976) and Wolkon et al.

(1973), but it may actually reflect blacks and Chicanos would go

to the "more prestigious" professionals for their personal problems,

if they had to go. The therapists were broken down into classifi-

cations: psychiatrist, clinical psychologist, counseling psycholo-

gist, college counselor, high school counselor, and advisor.

Preference for the professions was in the order listed, implying a

hierarchy of perceived expertise.

Harrison's (1975) review of the literature on race as an issue

in counseling comes to the same conclusion as this author. Although

the results of the research on black preference for counselors some-

times offers mixed results, the preponderance of evidence is indicative

of a definite trend to prefer same race counselors.

Black Responses to Counseling

There exists no consensus of evidence on one side or the other

regarding the issue of race in counseling process and client response.

On the contrary, the results are mixed and frequently contradictory.

An early review by Banks (1971) of the process and outcome research on

interracial therapy cites relatively few studies of any merit. Banks

notes that much research in counseling and psychotherapy that deals

with the racial issue is anecdotal. Two strong early studies cited

by Banks are Banks, Berenson, and Carkhuff (1967) and Carkhuff and

Pierce (1967). The first study has already been described as a

prompting of much research that was to follow.

Carkhuff and Pierce (1967) designed a study to look at the effects

of the counselor's race and social class upon the patient's depth of

self-exploration. They used black and white counselors of both upper

and lower class with black and white patients of both classes. A

significant interaction was found between counselor and client vari-

ables. Race and social class of both patient and counselor had sig-

nificant effects upon depth of that self-exploration. Patients most

similar to the race and social class of the therapist explored

themselves the most.

In a study of his own Banks (1972) looked at the effects of race,

social class, and level of empathy in the counseling process. He

hypothesized that racial similarity had an effect on the initial

interview relationship. Using the same design as Banks, Berenson,

and Carkhuff (1967), he found that racially similar pairings resulted

in greater self-disclosure. Surprisingly, social class was not found

to have an effect, but no lower class counselors were used as in the

Carkhuff and Pierce (1967) study. The evidence suggests that empathy

had a greater effect on rapport than race.

Looking closer at the nature of the counselor-client interaction,

Bryson and Cody (1973) studied the relationship of counselor and

client race to the level of understanding between the two of them.

They discovered that race is related to understanding in the counsel-

ing relationship. Thirty-two undergraduates, evenly divided according

to race and sex, were assigned to graduate level counselors in such a

way that each counselor saw a client of each race and sex. Indepen-

dently trained judges evaluated the levels of agreement on specific

statements recorded in the interviews. Results suggested that the

race of the counselor was significant to understanding in the counsel-

ing process, but the race of the client was not related to understand-

ing. White counselors understood white clients best, and black

counselors understood black clients best.

In another study on depth of self-disclosure and racial inter-

action, Grantham (1973) varied counselor sex, race, and language

style. Subjects were a special population of disadvantaged black

students. Grantham was interested in demonstrating that there was a

direct relationship between client-counselor similarity and progress

made in counseling. As in earlier studies cited in this section,

counselor-client racial similarity had a significant effect on client

satisfaction. Post hoc analysis, however, resulted in controversial


In Grantham's study (1973) satisfaction did not appear to be

influenced by level of facilitative conditions. The greater satisfac-

tion was related to race, not to the level of facilitative conditions

used by the counselor. The level of facilitative functioning of the

black counselor was not as important as that counselor's race. If all

counselors had the same level of functioning, black counselors were

preferred. There was even a reverse effect of lower self-exploration

when black counselors were highly facilitative. Grantham suggested

the possibility of facilitative conditions operating differently across


In two studies of black client responses in the mental health

setting (Sue, McKinney, Allen, & Hall, 1974; Sue, 1977a), it was found

that blacks were not even receiving treatment equal to the quality

given to white clients. They were placed in inferior treatment pro-

grams and assigned to paraprofessionals rather than to professional

personnel. The response of the black clients in over 50 percent of

the cases was to terminate treatment early (Sue et al., 1974). Sue

proposes that at least in some agencies the differences in outcome may

be more indicative of misapplication of current techniques than in

differences in clientele. Responsive services would include trained

specialists, independent services in the minority community, and

development of new therapeutic approaches (Sue, 1977a).

The literature so far cited in this section has been one-sided

in rating significant racial effects on the counseling process and

outcome. The next grouping of studies offers mixed and in some cases

contradictory evidence. One particular study serves as an ideal

transition reference between the arrays of evidence in that it can

be interpreted both ways.

In 1974 Williams found no differences in trust and self-disclosure

ratings of black students who were seen by white professional counsel-

ors or by black lay counselors who had several hours of training in

facilitative skills. Williams concluded that his results supported the

contention that the race of the therapist is not a significant variable

in counseling. A review by Parloff, Waskow, and Wolfe (1978) aptly

pointed out that the results can be interpreted in just the opposite

direction. They ask the reader to consider how impressive it is for

slightly trained blacks to do as well as white professionals with the

black clients.

Another study with mixed results was conducted by Jones (1978).

Fourteen female clients (seven white and seven black) were seen by

therapists in all four possible matches. Clients were seen for 10

sessions. Outcome was measured by both therapist and client responses

to questionnaires. Process rates and recordings were subjected to

specially designed Q-sort. There were no differences in outcome as a

result of racial matching. There were no differences in attitudes,

impressions, or dropout rate. Race, however, did influence process.

If the client was white, it mattered little if the therapist was

white or black, there was no difference in process. If the client was

black, regardless of the race of the therapist, a discussion of race-

related concerns evolved as an important aspect of therapy. If the

client was black, there was a significant erotic transference and a

greater focus on issues of concern if the counselor was also black.

Jones concluded his discussion with a cautionary warning about the

generalizability of those studies which are based on one interview.

He discovered such dramatic changes in his 10-session study that the

first interview was drastically different in terms of counseling

relationship than later sessions.

Other studies have consistently found results in the direction

suggesting no difference in counseling effects that can be attributed

to racial pairing. Woods and Zimmer (1976) made use of interracial

dyads and an attitude measure based on the semantic differential. He

found no differences in client perceptions of attraction in the

analogue study. Vail (1978) studied early termination from therapy

in an inner-city mental health setting. Black male and female clients

were randomly assigned to male and female, black and white therapists.

Variables considered for influence included counselor and client sex,

race, racial attitudes, and perception of level of understanding in

therapy. The only significant correlate of termination was the sex

pairing of client and therapist. Although the race of the therapist

did not matter statistically, most of the interaction occurred between

male therapists and black male clients. Female therapists also were

less effective with their own sex client.

Proctor and Rosen (1981) studied possible effects of expectancies

and preferences for therapy on intermediate outcomes of treatment.

They used both white and black outpatient veterans in their study.

All counselors were white. Both white and black clients expected to

be treated by a white counselor, but preferences were different. One-

half of each racial group had no preference; the other half had

preferences significantly in favor of a counselor of their own race.

Drop-out rate and satisfaction with treatment, however, showed no

differences between the races. Not having black counselors incor-

porated in the design prevented inferring that blacks do as well with

white counselors as they do with blacks.

The studies addressed so far have focused either on the simple

issue of racial differences, attitudes, or on the interaction of essen-

tially demographic variables when counseling black clients. The as-

sumptions and claims of theorists recommending specific approaches to

be taken with the black clients began to be more actively addressed in

the mid-1970's. The hope apparently has been that by identifying

specific components of counselor style, the confusion over conflicting

results can be dissipated to the benefit of client, practitioners, and

researcher. As early as 1967 Bancroft, then later Russell (1970),

suggested that counselor style rather than race might be the reason

that counselors were reporting differential success with minority

clients. At that time Bancroft argued that blacks might not be accus-

tomed to a reflective style of counseling. Russell (1970) argued

that blacks might not be desirous of it.

The first significant study to look at different racial reactions

was Roll, Schmidt, and Kaul (1972). They made use of Strong's (1968)

interpersonal influence model of counseling as well as information

from Kaul and Schmidt's (1971) research which provided definitions of

verbal and nonverbal cues of trustworthiness. Roll et al. used

videotaped segments of combinations of content and behavior. They

discovered that there was a significant content-manner combination

which was perceived as trustworthy by both black and white convicts.

Trustworthy cues included eye contact, forward body lean, hand

gestures, head nods, and an upright but relaxed posture. Untrust-

worthy behavior included distracting movements, an overly relaxed

posture, leaning away from the client, and absence of hand movement.

There was no difference between the perceptions of whites and blacks.

In 1975 Peoples and Dell conducted an analogue study to measure

black and white student reaction to other types of black and white

counselors. Both counselors were female but one was white, the other

was black, and they played both an active and a passive role. Peoples

and Dell hoped to account for the varied findings reported in other

studies and considered counselor style as a strong possible interacting

variable. Results demonstrated a significant preference for the active

counselor by both races. The authors reasoned that the amount and type

of activity in the active role made that counselor appear much more

competent. Dreeman (1977) reports that university counseling center

clients in general prefer active counselors who worked hard toward

system relief of both cognitive and behavioral problems.

In a related study of the possible differential effects of coun-

selor race and approach on subject's perceptions, Atkinson, Maruyama,

and Matsui (1978) compared directive versus nondirective approaches

with Asian-American students. They were able to provide strong evidence

supporting the use of a direct, logical, structured approach for use

with Asian-American students. More importantly, their results imply

that specific counseling techniques are not necessarily generalizable

to all cultures.

Fry, Kropf, and Coe (1980) looked at the effects of client and

counselor racial similarity from a different perspective. They inves-

tigated the effects of racial similarity on the counselor's response

patterns in therapy. Similarity in race in the counseling dyad re-

sulted in a smoothly flowing communication interaction. Dissimilarity

led to disproportionate expressiveness towards black clients and

increased attending toward white clients. The interaction of client

and counselor race were shown to be prime mediators of counselor


Suggested Approaches and Theory

Rogers' (1957) article on the necessary and sufficient conditions

of change in therapy provided the impetus and gave the direction for

countless studies over the past two decades. The successful applica-

tion of a theoretical model or technique to one population, however,

is no guarantee of its application to a culturally different popula-

tion. As Harper (1973) points out "counselors must be careful in

applying various concepts and propositions of psychological theory to

the behavior of black counselees, especially poor blacks" (Harper,

1973, p. 113).

The theorists in the field of cross cultural or minority coun-

seling have at various times called into question the ability, or

at a minimum the difficulty, of the white counselor to achieve some

of the core conditions with black clients. Is the therapist truly

congruent in terms of feeling and behavior toward the black? Does

the therapist experience unconditional positive regard? Can the white

therapist have an empathic understanding of the client's iiernal

frame of reference? Is this communicated? These questions are posed

in terms of Rogers' theory (1957), but significant and thought-

provoking issues have come from other sources.

Vontress published several articles in the late 60's and early

70's which helped establish a new foundation of interest in working

with the black client. His 1969 article cited significant, but not

insurmountable, barriers to the establishment of a successful black-

white counseling relationship. Prejudicial racial attitudes, black

reservations about self-disclosure, different language styles, the sex

and race taboo, and mutual ignorance about each other were all put

forth by Vontress as significant parts of the counseling barrier. The

greatest blockage was attributed to the counselor's lack of understand-

ing of the client's background.

This led Vontress to make specific suggestions about counselor

training programs (Vontress, 1970). He stressed the need for inculcat-

ing greater counselor sensitivity. This was to be achieved by increas-

ing service training in the black environment, not in the middle-class

service setting. Counselor curriculums were to be revamped (Vontress,

1971) to take into account unique aspects of racial differences:

easier rapport with women than with men, urban-rural differences,

lower rates of self-disclosure. The occurrences of affective experi-

ences with black clients in their settings would humanize counselors

and help remove barriers to the establishment of rapport (Vontress,


The result of Vontress's speculation on the differences between

blacks and whites was an optimistic approach which was aimed at capi-

talizing on the similarities of the human condition. Vontress (1979)

stated that humans are fundamentally more alike than they are differ-

ent, and that these similarities should be used by counselors to com-

municate with and understand those from different cultures. Vontress

argued for an existential approach which puts aside analyses of dif-

ferences and instead concentrates on the commonalities of people.

Borrowing terms from existential philosophy, Vontress pointed out that

all humans had to deal with the same Umwelt; that is, we confront the

same basic survival problems. We have the same social needs and

longing for others (Mitwelt). There is a part of us that is difficult

to share with anyone no matter what our culture (Eigenwelt). There is

a constant interaction of this private world with the Umwelt and

Mitwelt and this is the experience counselor and client can share.

Existentialism for Vontress offers great promise for bridging cultural

and racial differences. As a philosophy it has great merit, but as a

practical approach to counseling too much is missing.

Harper and Stone (1974) presented a theory for counseling blacks

that was highly specific in its recommendations. It grew out of Harper's

(1973) idea that the social needs of blacks need to be understood be-

fore theories are applied to them. Assessment and action are the two

key processes for their approach. Based on their assumptions that black

clients do not have the patience and time for slow passive therapies,

they recommended action and responsibility oriented action approaches.

In a rotating sequence from assessment to action to assessment, the

black clients' basic needs, responsibility, behavior, and personal

growth are dealt with. The goal is to help the black client transcend

his environment.

Tucker (1973) also recommended a very action-oriented approach.

Tucker stated that preparing counselors to work with black clients

took special efforts because of the widespread poverty and high visi-

bility of black Americans. His action model was based on specific

goal setting and aggressive commitment to change. Like Vontress

(1974) and later McDavis (1978) the counselors were to become actively

involved in the client's environment, but he was not just referring to

student experiences but rather an aggressive outreach program. Verbal

and nonverbal cues were emphasized to build up basic trust. From that

point on specific short-term and long-term goals were prioritized,

strategies were assessed, and behaviors implemented. His program has

elements highly similar to Harper and Stone's (1974) assessment and

action process, but each step is much more specifically defined.

Tucker assessed the merits of his model with 40 black students. Not

only were the action counselors perceived as more helpful, but action

counselors differed significantly regarding degree of satisfactory


Williams and Kirkland (1971) claimed that white counselors would

hamper black client growth patterns, unless they received special

training from a black perspective. This training had several key

components including a commitment to an advocate, a bicultural role,

and certification by black psychologists. Williams and Kirkland

emphasized the necessity for the white counselor to reject the deficit

model of the black experience and the need to accept the difference

model. The difference model posits that differences between blacks

and whites are not due to pathology.

Smith (1977), in a highly emotional article, attacked the founda-

tion of many approaches to counseling blacks which are based on

stereotyped concepts. She questioned the validity and assumptions of

testing results, the social consequences of research, the arbitrary

use of intervention techniques, and the role of stereotyping in keeping

counselor and client apart. Discerning the truth about the black

client may not be possible for a counselor believing one set view

which pictures the black client as being merely opposite everything

that is white middle class. Smith believes counselors have actually

contributed to the stereotyping of blacks and that we must increase

our awareness of their real needs.

Derald Sue has been a prolific writer in the field of counseling

minorities in general and Asian-Americans in particular. His 1977

article on counseling the culturally different recommended a sys-

tematic approach that took into account four variables that char-

acterize the culturally different client as distinct. Those vari-

ables are different cultural values, different class values, language

factors, and the unique experience of oppression. General counseling

approaches may not always be appropriate with such variables, and

even if they are, the goals selected may be inappropriate. Sue

believes it is critical to not only have an awareness of the minority

group's culture but of the value assumptions inherent in psycho-

logical approaches. Choices between abstract versus concrete goals,

action versus introspection, and privacy versus disclosure are

some of the critical valued issues involved in counselor-client


Sue and Sue, in a 1977 article on barriers to counseling the

culturally different, described in detail third world variables of

language, class, and culture and made recommendations for dealing

with them for several racial groups. For blacks, Sue recommends a

concrete, structured approach that is action-oriented. Importance is

placed on nonverbal behavior in communication. Instead of letting

the client's ethnicity serve as a barrier to the counselor-client

relationship, it can be used to therapeutic advantage.

Sue (1978a, 1978b) provided a general theory of how race and

culture interacted to provide people with specific world views.

Those world views were characterized by two concepts--locus of con-

trol and locus of responsibility. Locus of control is one's orien-

tation on whether reinforcements are contingent on one's behavior

or occur as a result of chance. Locus of responsibility is one's

blame system. Success or failure is attributed to either one's

skills or to the environment. Sue proposes that the Western World

counseling approach is based on a world view of internal control

and internal responsibility, which frequently clashes with minority

world views of passivism, oppression, and even militancy. The

minority world views vary depending on whether an individual's

orientation is internal or external on the two loci, control and


A client's particular world view has implications for treatment.

If internal on both loci, then the client would benefit from current

approaches. If the client's orientation is external on either or

both loci, then a counselor with standard therapeutic approaches

could well be imposing his world view. Sue calls this cultural

oppression. A more amenable solution is to take the client's world

view into account. For example, new coping skills could be taught

to those who give control and responsibility to the environment.

Research is lacking exploring the appropriate approaches to these

different world views.

Kinzie (1978) sought to identify those aspects of cross cul-

tural psychotherapy that were technically effective in application

of the medical model of therapy. He notes that the "sick role" is

an accepted role in most cultures and has a pretty clear relation-

ship with the "healer" role. The medical figure appears to be use-

ful in using his moral authority to promote patients' health.

Sensitivity to nonverbal communication and the subjective phenom-

enological world of the client is important. He also cites awareness

of value systems inherent in therapy as being important before they

are arbitrarily applied.

A comprehensive approach that is truly eclectic has been advo-

cated by McDavis (1978). This approach makes use of inputs from six

existing counseling approaches. The counselor does not have to

learn new techniques, but how to combine them in a new way, when

working with a black client. The techniques chosen take into ac-

count the research and theories that have been influential over

the past decade. Specific concepts are drawn from three approaches

as a philosophical foundation. The orientation is existential.

Concepts from the client-centered approach are acceptance, under-

standing, and congruence. Reality therapy concepts involve commit-

ment of counselor and client and a nonjudgemental attitude.

Application of the model is relatively straightforward. The

philosophical orientation is evidenced by counselor outreach and

involvement in the living and social life of the black community.

This is reminiscent of the suggestions made by Vontress (1970,

1971). Establishing a relationship and making a commitment could

be facilitated by counselor self-disclosure. The client is encouraged

to be the sole judge of his own behavior and to make a commitment to

the counseling process.

Techniques are borrowed from Gestalt, Behavioral and Rational-

Emotive approaches. An emphasis on a transition to self-support

comes from the Gestalt approach. The use of out of session tasks

and changing irrational belief systems comes from rational-emotive

therapy. Finally, goal setting, role playing, and modeling are

drawn from behavioral approaches. It is up to the counselor to

decide the appropriate timing of each technique. This is to be

based on the nature of the client's needs and progress. The focus

of individual client needs and active counselor involvement has

been a relatively common thread running through the literature.

In response to criticisms in the literature of traditional ap-

proaches to counseling blacks (Tucker, 1973; Vontress, 1970, 1971)

Schauble, Parker, Probert, and Altmaier (1979) attempted a novel

strategy in introducing minority students to a positive experience

with counseling. The researchers implemented a three step program

which introduced the psychologist, group experience, and individual

counseling in a nonthreatening manner. The psychologists initially

came into the classroom in an instructive role--to teach academic

skills and some basic psychological principles. In the second step

students became involved in small group interactions to explore

personal interaction issues. Finally, the students were given the

opportunity to voluntarily seek individual counseling with their

group leader. Assessment of the program was informal, but the

higher percentage of student voluntary use of the university counsel-

ing facilities plus their referral of peers suggested this out-

reach program had significant merit for use of predominantly white


Reviews of the literature on race as a counselor client variable

appear to have as much disagreement as the individual research

articles. If there is any consensus at all, it appears to be that

the results are mixed on many issues and that a large portion of the

research accomplished has been anecdotal and uncontrolled (Griffith,

1977). Higgins and Warner (1975) cited the paucity of quality research

in the area of counseling blacks, the mixed results, and the incon-

sistency of the literature. They did conclude that counselors should

emphasize empathy in working with minorities and should use an action

approach. They also held that black counselors were more effective

not as a function of race but as a function of cultural understanding

and language.

Bryson and Bardo (1975) reported three types of literature related

to counseling blacks--client characteristics, counselor characteris-

tics, and the counseling process itself. Lower class blacks were

described as submissive and suspicious, while upper class blacks were

described as possible and uncommunicative in therapy. Counselor race

was viewed as potentially obstructive in good therapy. Inappropriate

techniques were found to be used in therapy with blacks. "The general

conclusion seems to be that although counselor race as a simple vari-

able is insufficient qualification for predicting effectiveness, it is

a factor that must be considered (Bryson & Bardo, 1975, p. 75).

Sattler (1977) is one reviewer who does not conclude there is a

trend for race as a significant variable in therapy. He reached this

decision because his review included a number of studies, already

cited in this text, that demonstrated no difference, or even a dif-

ference in the reverse direction expected.

Harrison (1975) admits the results are mixed but claims there is

enough evidence to note a trend for clients to prefer same race

counselors. Black clients in particular prefer black counselors. The

crucial variable, regardless of race, appears to be the counselor's

ability to communicate empathic understanding. During the counseling

process blacks tend to respond negatively to white counselors in terms

of preference, self-exploration, and language. Harrison believes two

theoretical viewpoints are the bases for the different findings. The

first is that black counselors assumed similarity of experience with

black counselors. The other theory is one of role expectations gener-

ated by black and white social history.

As late as 1978, in their review of therapist variables in

processes and outcome, Parloff et al. concluded that there is not much

definitive evidence on the effects of race per se. They reason that

attitudes of the therapist may be more important than his or her sex

or race. It is highly likely that the very manner, or style, of the

therapist communicates a great deal to the black client about that

therapist's attitudes towards a host of issues. A great deal of re-

search on stylistic behavior and its bearing on a therapist's per-

ceived expertness, attractiveness, and trustworthiness has been con-

duced with white clients. The model for such research has been the

interpersonal model of counseling (Strong, 1968). This model has not

been systematically examined for applicability to the black client.

The Interpersonal Influence Model

The origin of the thought processes that evolved into the Strong

and Matross (1973) model of interpersonal counseling can logically be

traced back to the notion of systematic causality as espoused by Kurt

Lewin (1935). This dynamic view of change assumes that all behavior

is caused by forces acting on an individual at the time a given behavior

is emitted. "Influence is understood to be a product of the changing

relationship between two individuals rather than a function of the

static and discrete characteristics of the influence and influence.

The basis of influence is interdependence" (Johnson & Matross, 1977,

p. 397). The counseling relationship can be viewed as such an inter-

dependent relationship.

Another major assumption comes from the dissonance theory of

Festinger (1957). From this theory is derived the motivating force for

the client. It assumes there is a tendency for consistency in our

thoughts about ourselves and about our environment. The existence of

inconsistent cognitive elements produces dissonance. The discomfort

of dissonance motivates the individual to seek relief.

There are two other assumptions made for the model which stem

logically from the first two. The first is that the change process in

counseling is a direct result of the interaction of the counselor and

the client. The second is that counseling is specifically defined as

an interaction process in which the counselor's task is to influence

the client to change, and the client's task is to be influenced by the

counselor. The task of influence should not be confused with manipula-

tion for the counselor's purposes. As Johnson and Matross (1977) cau-

tion, influence is focused on the achievement of the client's chosen

and agreed goal. "The therapist influences a client to the extent

that the therapist furnishes resources needed by the client for the

accomplishment of highly valued goals and to the extent that the client

cannot obtain these resources at a lower cost from other relationships"

(Johnson & Matross, 1977, p. 405).

How does the counselor influence the client? Different pairs of

client needs and counselor resources are considered to be sources of

counselor power. Each pair is considered a power base. The original

classification of power bases came from French and Raven (1959). They

also described influence strategies in social psychology in terms of

social power; the basis of power is the nature of the relationship

between two people. French and Raven (1959) considered five sources

of power: reward, coercive, legitimate, referent, and expert. The

last three were considered to be individual sources of influence.

The typology used by Strong (1968) comes more from the revised

typology and definitions of power bases in Raven (1965). In this

typology of Raven's, expert power "stems from the attribution of

superior knowledge or ability to the influencing agent" (Raven, 1965,

p. 374). Referent power exists "when a person uses another person as

a 'frame of reference,' as a background, or as a yardstick against

which he evaluates some aspect of himself" (Raven, 1965, p. 374).

Legitimate power is based on "the influence's acceptance of a rela-

tionship in the power structure such that the agent is permitted or

obliged to prescribe behaviors for him ." (Raven, 1975, p. 375).

Strong (1968) labeled his power bases in terms of communicator

characteristics which prevent the client from derogating the counselor,

when there is perceived dissonance. Because of these positive charac-

teristics, the counselor is a highly credible communicator and dis-

sonance is likely to be reduced by change in the direction of influence.

These source characteristics, drawing from Raven (1965) and French and

Raven (1959), are expertness, attractiveness, and trustworthiness.

Attractiveness as a source characteristic corresponds to Raven's (1965)

referent power base. Sometimes the term "referent" is used in current

research. Trustworthiness is defined as perception of the counselor's

interest to communicate the most valid assertions.

A fourth version of the power bases appeared in Strong and Matross

(1973). They listed five power bases: expert, referent, legitimate,

informational, and ecological. The first three are the main sources

of power for the therapist in the interview. Counselor expert sources

are described as knowledge and skills as perceived by the client.

Counselors can be referents depending upon the similarity of values

and attitudes with their clients. Counselors have legitimate roles

as help givers in personal problems (Strong & Matross, 1973).

The power of a counselor to influence change is dependent upon the

counselor-client relationship. The client's dependence on the counselor

is the measure to which he perceives the counselor's resources cor-

respond to his needs. The key element is client perception of his

needs and of counselor attributes not actual needs nor actual counselor


Development of the interpersonal influence model was influenced

and supported by other attitude change studies. The works of Hovland,

Janis, and Kelly (1953) led directly to definitions of the source

characteristics used by Strong (1968). Zimbardo (1960) conducted a

study on opinion change in which it was discovered that the magnitude

of the opinion change was dependent on the involvement of the subjects,

and the attractiveness and credibility of the communicator. Patton

(1969) also conceived of counseling and therapy as a social influence

process. He tested client responses to interpersonal attraction of the

communicator and the congruence of expectations. He found client

responses were contingent upon both factors.

McGuire (1969) presented a review of influential source charac-

teristics which included credibility, attractiveness, and power. His

descriptions closely corresponded to the three power bases used by

Strong (1968). Simmons, Berkowitz, and Moyer (1970) reviewed the

literature on attitude change and were able to distinguish between

cognitive and affective components of change. The cognitive component

was affected by expertise and prestige, while the affective component

was influenced by liking and friendliness.

Strupp (1973) also sees the therapeutic situation as a power base

for psychological influence. He argues that there is no such thing as

nondirective psychotherapy. Influence is inherent in the helper role.

Client attraction to the counselor accounts for some of the variance

in counselor ability to create an effective relationship.

An entire series of research studies went into the development of

the Counselor Rating Form (CRF) (Barak & LaCrosse, 1975) which provide

ample evidence for the existence of the three dimensions of perceived

counselor behavior: expertness, attractiveness, and trust. In the

original investigation of the three dimensions predicted by Strong

(1968), Barak and LaCrosse (1975) showed over 200 subjects filmed

interviews of Rogers, Ellis, and Perls. Subjects were then asked to

rate the counselors based on a 36 adjective list. The list itself was

constructed from a much larger one. All the items on it had been

selected by a panel of four expert judges who were familiar with re-

search in the area. After rating forms were collected and separated

according to each counselor, they were factor analyzed. According to

test results, the three factor approach seemed to be the most appropri-

ate. Across the ratings of all three counselors, the dimensions of

expertness, attractiveness, and trust appeared to be distinct from

each other. The average percentage of total variance accounted for by

the factors was 52 percent. Any attempt to extract another factor

resulted in meaningless additional factors of only one or two items.

In 1976 LaCrosse and Barak tested the reliability of the CRF and

its ability to differentiate between counselors on the perceived

dimensions. The reliability coefficients across counselors were .874

for expertness, .850 for attractiveness, and .908 for trustworthiness.

The data also indicated that the counselors (taped interviews of

Rogers, Perls, and Ellis were used) were differentially perceived

with p < .01 (LaCrosse & Barak, 1976).

A replication and extension of the previous study was made by

Barak and Dell (1977). Their data showed the CRF to be sensitive to

the perceived differences among counselors on the three predicted

dimensions. The CRF was sensitive to those differences even among

counselors of low or moderate experience. Barak and Dell also

reported a significant positive relationship between perceptions of

expertness, attractiveness, and trustworthiness and willingness to

see a counselor.

In a comparative study of client perceptions of the three relevant

dimensions, LaCrosse (1977) used actual clients in a mental health

setting. Following one of their interviews, clients were asked to

fill out the CRF and the Barrett-Lennard Relationshios Inventory

(BLRI) (Barrett-Lennard, 1962). The interview occurrence ranged

from the first to the twenty-fourth for any particular client. There

were no differences in perception due to nonfirst session (p < .05).

Observers ratings of counselors were consistently more conservative

than client ratings. Clients rated the counselors highest, followed

by the counselors self ratings, followed by observer ratings. Inter-

correlations obtained for the CRF with the BLRI were high, ranging from

.51 to .80.

The LaCrosse (1977) study supported the results of Barak and

LaCrosse (1977) in which both counselors and supervisors provided

lower perception of a counselor's performance than the clients did.

Even the novice counselor in the Barak and LaCrosse (1977) study

rated his performance lower than the client. Both studies strongly

suggest the utility of the instrument and the use of the three dimen-

sions with different population groups: students, mental health

patients, counselor trainees, supervisors, and practitioners.

In 1980 LaCrosse conducted a study into the predictive validity

of the CRF. He hypothesized that a positive relationship would occur

between perceptions and outcomes. Total CRF ratings correlated mod-

erately highly with outcomes (r = .53, p < .001). Expertness ratings

correlated most highly, followed by attractiveness and trustworthiness.

The three CRF variables accounted for 35 percent of the variance on

outcome. Perceived expertness accounted for 31 percent of the vari-

ance by itself. The study generally supported the validity of the CRF

and the usefulness of the social influence model. Additionally, the

client's initial "perception of helper expertness was the single most

powerful predictor of counseling outcomes of those predictors studied"

(LaCrosse, 1980, p. 395).


A counselor's expertness, or the perception of a counselor as a

valid source of assertion, can be affected by reputational cues, be-

havioral cues, and objective evidence (Strong, 1968). Expertness is

one major component of a counselor's credibility, the other being

trustworthiness (Hartley, 1969). In an attempt to begin providing

empirical support for the definition of expertness, Schmidt and Strong

(1970) conducted a study in which counselor behavior was videotaped

and then rated by 37 male college students. In addition they asked

the students to list characteristics which they believed to be expert

or inexpert. The order of expertness on the ratings was nearly the

reverse of the order of the counselor's actual level of training. Two

factors seemed to characterize those behaviors rated expert:

(a) greater responsiveness to the client and (b) logic of the ques-

Strong and Schmidt (1970) then took the results of their earlier

study and varied interviewer introduction and counselor performance

with 49 college male subjects. Subjects were assessed for change in

need for achievement. Those subjects treated to both "expert" conditions

changed their ratings significantly more than those subjects with

"inexpert" introduction and behavior. Neither the introduction nor

the behavior by themselves was significant. The introduction was

either to a Ph.D or to a peer counselor. Expert behavior consisted

of facial and behavioral responsiveness, alert posture, and hand

gestures. Expert verbal behavior was structured and organized. In-

expert behavior was nonreactive, stiff and formal, confused, and lacked


Schwartz (1971) tookanother look at the inverse relationship

reported by Schmidt and Strong (1970) between counselor experience

and rated expertness. Using a least squares analysis he was able to

demonstrate that expertness is not correlated in any simple way with

experience, but rather the subjects apparently liked one counselor,

disliked another, and thought the others were equal.

A later study by Dell and Schmidt (1976) helped clear up the

issue. They showed videotapes of counselor interviews with 120 male

and female undergraduates. The male and female counselors were at

three levels of experience. All subjects rated their willingness to

refer close friends to the counselors and gave their reasons. The

researchers found that only individual counselor performance is related

to perceived counselor expertness in the initial interview situation.

Specific behavioral cues of expertness were identified including

preparedness, gesticulation, use of client's first name, relaxed

posture, and nonmonotonic voice. There were no sex differences in

perceived expertness.

Several studies have identified a reputational component of

expertness. Guttman and Haase (1972) randomly assigned first year

male college students to conditions of high and low induced expertness.

Although the information recall of those under the expert condition

was higher, the clients appeared to respond more favorably to the

nonexpert in terms of expressing their feelings about the relation-

ship's success. The authors concluded that their findings supported

the earlier Schmidt and Strong (1970) study which showed an inverse

relationship between experience and expertness. The Schwartz (1971)

study plus the other research cited here indicates that information

can reliably affect the counselor's perceived expertness as long as it

is at least congruent with the counselor's behavior. Of the two vari-

ables, reputation versus behavior, it may verywell be that behavior

is the more influential variable.

An early study by Price and Iverson (1969) provided initial en-

couragement for believing in the effects of reputational cues of expert-

ness. Tape recordings of initial interviews were heard by 120 students.

The sessions were described as being done either by a "head counselor"

or by a "trainee." Reflective or directive styles were not distinguished

in the ratings. High commitment was the most significant variable for

being rated capable. Those who were the "head counselors" and who

conformed to role expectations made the most positive impressions.

Atkinson and Carskadden (1975) made use of a prestigious introduc-

tion and psychological jargon to identify their expert. Videotaped

counselors were rated on their knowledge of psychology. Greater

knowledge was attributed to the counselors' using jargon. The sub-

jects were more willing to see those counselors with the prestigious

introduction. Suggestions of counselor expertness also led to higher

evaluations of counselors heard on audio tape.

Spiegel (1976) used four levels of similarity/expertness sug-

gestions each of which was paired with an academic and a friendship

problem. Although the manipulation of the similarity variable was too

weak, the expertness was perceived as intended for both problem areas.

Scher (1975) planned his study of counselor experience and its

effects on success in counseling in such a way that "having experience"

was defined as having a Ph.D. Counselor experience was found to be a

relevant variable in influencing outcome, while sex and verbal activity

were not. Because of the way experience was defined, however, and

because some of the non-Ph.D counselors had as much counseling ex-

perience as Ph.D counselors, Scher had to admit the possible role of

reputation. He concluded "it may be something as obvious as the

counselor's title of Doctor that is the important factor in the

influence of experience on counseling" (Scher, 1975, p. 100).

The strength of presession information on the counselor's per-

ceived competence is strongly supported by Scheid (1976). Counselors

were introduced as high status, low status, or not given status at all.

Male and female undergraduates viewed videotaped segments of one of

two staged interviews. One type was arranged to be level three of

the case conditions and the other type was level one (Carkhuff, 1969).

The most impressive results of the study were that even when counsel-

ors were viewed in the nonfacilitative behavior, they were rated as

competent, if they had a high status introduction.

External devices, accoutrements, and decor have been investigated

for their effects on client perceptions of counselors. Bloom, Weigel,

and Trautt (1977) investigated the effects of office decor and sug-

gested counselor gender information on male and female undergraduates.

A "traditional" office was decorated with a file cabinet, professional

texts and manuals, diplomas and certificates, and a desk situated

between the client's and therapist's chair. The "humanistic" office

contained a desk placed against a wall at the back of the room. Chairs

were closer together. Modern posters were on the wall with philosophi-

cal sayings. The room also had a beanbag chair with several throw

pillows. Sex pairing had no effect on the perception of credibility.

There was, however, an interaction effect for decor and therapist

gender on credibility. Female therapists in the "traditional" office

and male therapists in the "humanistic" office were viewed as more

credible, safer, and more dynamic. Environmental effects were sig-

nificantly related to degree of perceived credibility.

Heppner and Pew (1977) used a 2 x 2 factorial design to evaluate

the effects of objective evidence of competence and counselor sex on

perceived expertness. Male and female counselors, with and without

certificates and diplomas on display in their offices, interviewed

undergraduate university students in 30-minute sessions. Counselor

behavior was standardized by training. There was no interaction effect

between counselor sex and treatment condition. Initial perception of

counselor credibility was significantly enhanced by the presence of

visual, objective evidence of competence such as diplomas, awards, and


In a pair of studies (Siegel & Sell, 1978; Siegel, 1980) objec-

tive evidence and nonverbal counselor behavior were manipulated to test

their effect on the client's perception of counselor expertness. It

was hypothesized that such perceptions could be made without overt

introductory cues. The net result of both studies is that male and

female clients do not differ in their perception of expertness cues.

Results led the authors to further conclude that the failure of thera-

pists to display objective evidence of their status and professional

standing and the failure to use specific nonverbal behavior may make

the establishment of expertness more difficult. Examples of specific

nonverbal expert behavior include smiles, eye contact, angle of

shoulder orientation toward the client, and gestures. Both studies

indicate that the nonverbal behavior is consistently more effective

in determining the client's perception.

In a multifactor fixed effects design Heppner and Dixon (1978)

attempted to extend the knowledge about counselor influence in an

expert role to actual behavior beyond the interview. As they noted,

most research prior to theirs had been limited to assessing expert

influence on perceptions and attitudes. Ninety female undergraduates

took part in the experiment with the overt purpose of discussing their

level of problem solving skills. Subject self ratings following inter-

view sessions were significantly more influenced by expert than by

inexpert counselors.

In analysis of past interview behavior, Heppner and Dixon dis-

covered that interviewers playing the expert role were significantly

more influential in getting students to seek problem-solving handouts

than interviewers playing an inexpert role (p < .0001). Very few

subjects in any category, however, were influenced to attend a work-

shop. Need was not a significant variable in either measure. The

results not only continue the support of the differential influence of

expert versus inexpert roles on perceptions and attitudes but provide

some clues about measured influence on behavior. Students were

influenced to perform a behavior requiring a relatively minimal

commitment but not a more significant behavior. It must be noted

that even though the behavior influenced was a modest one, it was

done following a short counseling analogue.

Two studies exist which suggest or imply an issue to be resolved

about the origin of the perception of expertness. It is possible that

just being a counselor or mental health professional may be sufficient

to have some degree of expertness. In evaluating attitudes towards

seeking professional help, Cash, Kehr, and Salzbach (1978) discovered

that help seeking attitudes were significantly tied to perceptions of

a counselor's expertise, trustworthiness, and genuineness. Although

the subjects who reported more favorable attitudes were those who

reported prior professional contact, the implication is that for some

part of the prospective client population, expertness is an expected

part of the counselor's role. This does confuse expert power and

legitimate power.

Corrigan (1978) had over 200 undergraduate students rate the

relative importance of expertness, attractiveness, and trustworthi-

ness for a friend and a mental health professional. Results indicated

that expertness was a salient characteristic for a mental health

professional but not for a friend. The suggestion for the social

influence model (Strong & Matross, 1973) is that client expectation of

expertness may affect the perception of expertness at least in the

initial stages of counseling.


Festinger's (1957) hypothesis that people who are similar to us

with regard to attitudes or opinions will serve as referents regarding

that opinion is the basis for Strong's (1968) and the Strong and

Matross (1973) concept of referent power base. Strong (1968) claimed

that perceptions of a counselor's attractiveness would be based on

"perceived similarity to, compatibility with, and liking for "

(p. 216) the counselor. As with expertness, the attractiveness

dimension has been shown to possess behavioral and reputational cues.

In an early study in 1971 Schmidt and Strong designed an analog

interview experiment in which college males were interviewed by

psychology graduate students playing either attractive or unattractive

roles. In the attractive role, the interviewer greeted the subject

warmly, looked and smiled at the subject, and responded warmly

throughout the interview. Shared experiences and attitudes were also

disclosed by the counselor. The unattractive interviewers did not

greet the client, did not smile, and revealed dissimilar experiences.

Subjects achievement motivation were assessed before, immediately

after, and one week after the interview. In the interview the coun-

selors had attempted to influence the subjects' self-ratings by

telling them their score was two stanines higher than in reality on

the preinterview measure. In spite of very strong feelings about the

interviewers, the subjects were influenced equally in both cases. There

was an apparent confounding of roles, however, in that both counselors

were presented as experts. While not providing independent evidence

of the role of attractiveness, the study did provide evidence of its

relative importance.

The Schmidt and Strong (1971) results were hardly expected,

particularly since Patton in 1969 had been able to show distinct dif-

ferences in client responses to counselors to whom they were personally

attracted. Client responses also appeared dependent upon congruence

of expectations. Strong and Dixon (1971) studied the relationship

between expertness and attractiveness hoping to unravel the confusion

the earlier results generated for their model. Their study had two

hypotheses. First, they wanted to check out if the two dimensions

combined additively, and secondly they wanted to determine if expert-

ness masked attractiveness.

Their results supported the second hypothesis and helped explain

the results of the earlier Schmidt and Strong (1971) study. It ap-

peared that expertness and attractiveness may combine additively to

affect perceptions but not for influence. Expert interviewer's at-

tractiveness will not affect their influence role. Inexpert inter-

viewer's attractiveness will be affected only by attractiveness in

terms of influence. The masking effect of expertness occurs only in

the unattractive role. Over time it is possible that this mildly

positive influential effect will deteriorate.

Murphy and Strong (1972) then went on to further help define the

nature of referent power (attractiveness). They concentrated on study-

ing the effect of the nature and frequency of self-disclosures on

students' reactions in an interview. The number of disclosures ranged

in even numbers from zero to eight in the 20-minute interview. The

self-disclosing interviewers were seen as warm and friendly. Positive

self-disclosures that were similar to students' experiences enhanced

students' belief that they were being understood. Since the

self-disclosures were cued by an outside observer, some disclosures

were made at inappropriate times. When this occurred, the students

apparently noticed the nonsequiter and felt they were being tested

for a new theory. Timing, therefore, is a critical function of self-

disclosure to enhance interviewer influence.

Giannandrea and Murphy (1973) extended the Murphy and Strong

(1972) research by investigating the effect of frequency and timing of

self-disclosures on attitudes and behavior following an interview.

College students were engaged in a 20-minute interview to discuss

decision-making situations. Ten subjects were randomly assigned to

each of five treatment conditions. Students received 0, 2, 4, 8, or

12 disclosures from the interviewer. Results indicated that an inter-

mediate number of self-disclosures led to a significantly greater rate

of return for a second interview. The number of disclosures differen-

tially affects interviewee behavior. The use of a moderate number of

self-disclosures was cited as an effective technique for establishing

a good counselor/client relationship. The results were supported by

the similar later research of Davis and Skinner (1974) and Mann and

Murphy (1975).

In addition to similarity in levels of disclosure, similarity in

content has been shown to be effective in influencing interpersonal

attraction. Daher and Banikotes (1976) studied the effects of levels

and content and their possible interaction. Subjects were 87 male

undergraduates. On the basis of an inventory subjects were divided

into high and low disclosers. Each group then responded to four bogus

inventories which were manipulated for content and frequency of self-

disclosure. The amount of disclosure and the similarity of the content

were shown to be significant in attracting the subjects. Interaction

effects indicate that the amount of disclosure is attractive only when

content is similar.

Content similarity as an important issue in attractiveness is also

supported by early attitude change and influence studies. Berscheid

(1966) conducted two experiments in which attractiveness was controlled

and separated from the attempt at opinion change. In the first experi-

ment freshmen were divided into two groups. One group was told the

communicator held values similar to theirs in education, while the

second group was told the communicator had similar values to theirs in

international relations. The communicators then tried to influence

each group on first, a topic that was relevant to the similarity and

then, a topic that was not relevant to the similarity. Results sup-

ported the hypothesis that similarity of attitudes,uninfluenced by

attractiveness, led a person who shared a relevant opinion to view the

communicator as a referent for that opinion. When two share a

similarity that is not relevant to the opinion being influenced, then

the communicator does not serve as a referent.

The second experiment (Berscheid, 1966) provided further elabora-

tion and support. In this experiment it was discovered that if people

perceive that they have views on an issue that are dissimilar to the

views of a communicator, some reverse effects are likely. If the

opinion communicated is relevant to the shared dissimilarity, then

the receiver's opinion is likely to move away from that of the

communicator. Thus in this instance, the communicator acts as a

negative referent. "When attractiveness is controlled, an important

criterion for determining who will or will not constitute a referent

is often simply knowledge of whether or not the person possesses

attributes relevant to the issue in question" (Berscheid, 1966,

p. 679).

Nonverbal behavior also plays a role in the referent power

base--a very strong role. Haase and Tepper (1972) used the counselors

as subjects to explore the empathy perceived in a filmed counselor's

behavior. Nonverbal measures including eye contact, body lean, and

interpersonal distance interacted with each other and with verbal

expressions to affect ratings of empathy. More significantly was the

discovery that nonverbal effects accounted for twice the amount of

variance accounted for by verbal expression.

LaCrosse (1975) drew heavily on the work of Mehrabian (1969,

1970a, 1970b, 1971) in studying nonverbal behavior and counselor

attractiveness. Summarizing Mehrabian's work LaCrosse stated "In

studies of affiliative and persuasive behavior, the nonverbal cate-

gories found to be important are communicator smiles, positive head

nods, gesticulations, eye contact, angle of forward body lean, and

angle of shoulder orientation" (p. 563). LaCrosse used all of the

above-mentioned variables in his 1975 study. Eye contact was 80 per-

cent, body lean, 20 percent, and shoulder orientation zero degrees

for the affiliative behavior. For unaffiliative behavior eye contact

was 40 percent, body lean was 20 percent reclining, and angle of

shoulders was 30 degrees away. The camera lens served as the eye of

the observer. The ratings by the male and female undergraduates sup-

ported the hypothesis about the attractiveness and persuasiveness of

the affiliative nonverbal behavior.

Evidential cues such as attire and decor appear to be much more

limited for attractiveness than they are for expertness. Stillman and

Resnick (1972) devised a Counselor Attractiveness Rating Scale to

determine the degree to which subjects found a counselor confederate

attractive. Fifty male undergraduates saw either a "professionally"

attired counselor who wore a sport coat and tie, or a "casually"

attired counselor who was neat, but wore a sport shirt and slacks.

In addition to the Attractiveness Rating scale, students also filled

out a Disclosure scale. Analysis yielded no significant differences.

Counselor attire had little effect on client disclosure and client

perception of counselor attractiveness. Unfortunately, one of the

terms used to define attractiveness in the study was "an expert who

could help the subject and others if they were seeking help" (p. 347).

This indicates a confounding of expert and attractive roles. The

data of Bloom et al. (1977) showed the "traditional" role, as evi-

denced by decor, as being expert, not attractive.

Amira and Abramowitz (1979) noted the confusion in the literature

on the function of attire and decor and investigated those variables

with regard to therapeutic attraction. Male and female undergraduates

viewed one of four five-minute videotapes in which attire and decor

varied. A main effect for room formality was obtained for therapist

competence. The only other significant finding was a dress formality

by room formality interaction for favorable attitude toward the thera-

pist. Further analyses revealed that subjects preferred a combination

of formality and informality in the therapeutic setting. "Pure" com-

binations were not favored. The results tend to converge with the

predominantly negative and mixed data (Bloom et al., 1977; Stillman &

Resnick, 1972). There is a consistent absence of main effects for

professionalism of the therapeutic setting and for therapist attire.

Physical attractiveness was not directly included in Strong's

(1968) description of his counseling model, but several studies have

shown that physical attractiveness does have an effect on interpersonal

influence, although a limited one. In 1975 Cash, Begley, McGown, and

Weise had subjects first view a videotaped introduction by a counselor

then listen to an audio tape of a session. The male counselor was

either in a "physically attractive" or "physically unattractive" con-

dition. Subjects' overall impressions significantly favored the

"physically attractive" counselor and placed more confidence in that

counselor's effectiveness.

Including female counselors and clients in their study Lewis and

Walsh (1978) attempted to replicate and extend the Cash et al. (1975)

findings. The counselors were ranked on 12 traits and 15 personal

problems. The subjects perceived the counselors as intended; that is,

there was a main effect for physical attractiveness. Further analyses

revealed a shortcoming, however, in that the attractiveness manipulation

was not successful for the male groups. A univariate analysis of the

female groups was still carried out and yielded significant differ-

ences in perception. Female subjects perceived the attractive coun-

selor as more competent, professional, interesting, relaxed, and

assertive than the unattractive counselor.

In another attempt to replicate Cash et al. (1975), Carter (1978)

used photographs of male and female counselors. She did not get sig-

nificant differences between her attractive and unattractive conditions.

She did not have a genuine unattractive condition. Her results led

her to speculate that Cash et al.'s (1975) results may be due more

to the negative effects of the unattractive condition than to a posi-

tive effect for the attractive condition. Carter also reported an

interaction between sex and attractiveness such that female counselors

receive higher ratings on impression variables, particularly in the

attractive condition.

In attempting to extend the knowledge of the effect of attrac-

tiveness beyond an introductory period, Cash and Kehr (1978) found

results which supported Carter's (1978) speculation. Cash and Kehr

used attractive, unattractive, and anonymous conditions in their study.

Attractiveness effects were found on all dependent variables without

regard to sex of counselor. There were no differences, however, be-

tween the attractive and anonymous conditions. They therefore con-

cluded that "the comparisons of attractive and unattractive counselors

with the physically anonymous control conditions revealed a debilita-

tive influence of unattractiveness as opposed to a facilitative influ-

ence of attractiveness" (Cash & Kehr, 1978, p. 341). A more recent

study by Cash and Salzbach (1978) provides some support that this bias

towards the unattractive conditions can be ameliorated somewhat by a

moderate number of self-disclosures. Behavior again appears to

outweigh external cues in influence.

Expertness and Attractiveness

Several studies have included variables in such a manner so as

to examine simultaneously the effects of expertness and attractiveness.

Such studies have allowed the delineation of each dimension's cues to

be improved while others have allowed contrasts of the relative effec-

tiveness of each dimension.

In an early study that previewed the comparison of the two dimen-

sions, Greenberg (1969) randomly assigned 112 undergraduates to one of

four treatment conditions. Each group listened to a 15-minute tape of

a simulated therapy session. Prior to hearing the tape the therapist

was described as being warm or cold, and experienced versus inexperi-

enced. As discussed in earlier sections of this review, the labeling

information can affect perceptions significantly. In this study sub-

jects who were told the therapist was warm or inexperienced were more

attracted to the therapist, more receptive to his influence, and

evaluated him more positively. Subjects told the therapist was warm

were more willing to meet with the therapist and were more influenced

by his communications. This study incorporated early components of

what later were more systematically defined as expert and referent

power bases. The significant effects were all due to reputation

attributed to the counselors.

Claiborn and Schmidt (1977) studied the effects of presession

written information on the subjects' perceptions of counselor expert-

ness, attractiveness, and powerfulness. Subjects viewed a videotape

of a counseling interview in which the counselor attempted to influence

the client to take some specific action. Significant results were ob-

tained for higher ratings of expertness for the expert versus the

referent power base and higher ratings on expertness for the low

versus high status within the expert power base. The results for low

status expert could have been the result of a uniquely described peer

with some highly expert accomplishments.

Thus, both "high" and "low" status experts could in reality have

been "high" status experts. No differences were found on perceived

attractiveness. The authors concede that they may have described

similarities which were appropriate to the taped counselor-client but

not for the real subject. This failure to achieve the attractiveness

results anticipated points out the subtlety of such manipulations.

Fortunately the importance of nonverbal behavior in differential per-

ception has helped researchers in other studies (LaCrosse, 1975).

The possibility that the type of satisfaction required by the

client, the nature of the problem, might determine the relative ef-

fectiveness of an influence style led Tessler (1975) to look at the

relationship of several types of variables. "Experienced" versus

"inexperienced" introduction labels, similar versus dissimilar

counselor-client values, and formality versus informality were all

manipulated. Dependent measures were client problem-centered and

relationship-centered satisfaction. Relationship-centered satisfac-

tion was greater with value similarity which corresponds to the

referrent power base. Problem-centered satisfaction was greater with

greater experience which corresponds to the expert power base.

Spiegel (1976) attempted a similar study in which she manipulated

biographical sketches of counselors, presented a tape, then assessed

reactions. In her study suggestions of expertness led to higher

evaluations of counselors for both affiliative and academic problems.

Although Spiegel did not note this reason in her discussion, the fact

that she used merely the demographic variables of age and class as

her criterion of similarity may serve as an explanation for no results

differentiating similar versus dissimilar sketches.

An earlier study (Strong & Dixon, 1971), discussed in the section

on attractiveness, concluded that expertness masks attractiveness,

particularly in the unattractive roles. In another study Sell (1974)

successfully induced conditions of attractiveness, but results ob-

served were not attributed to that manipulation but rather to experi-

mental demands. Sell claimed this further supported the masking


A much more successful differentiation of expert and referent

role was achieved by Kerr and Dell (1976). They varied interviewer

role, setting, and attire with 80 undergraduates in an interview

situation. Using CRF they determined that perceptions of attractiveness

seemed to have exclusively been a function of behavior. For the expert

role, attire interacted with behavior in determining perceived expert-

ness. In every within-cell comparison the relative ratings of expert-

ness and attractiveness were dependent on the interviewer's role.

Dell (1973) studied the effect of counselor characteristics and

the generation of resistance using the Strong (1968) social influence

model. He not only successfully manipulated expert and referent roles

but managed to combine them in his experimental conditions such that

incongruence or illegitimate influence attempts were made. He combined

this novel approach with a measure of behavior outside the interview.

Expert and referent power bases were perceived as intended. "Pure"

expert and "pure" referent counselors were equally effective. Coun-

selors with congruent roles induced more behavioral compliance than

those with incongruent roles.

Dell (1973) suggests that the ability of the expert to have such

influence so early in a relationship-building situation is contrary to

what one would expect from general counseling approaches (Truax &

Carkhuff, 1967). He does suggest, however, that an expert's lack of

attractiveness later in the relationship could erode the counselor's


Merluzzi, Merluzzi, and Kaul (1977) continued Dell's (1973) ex-

tension of works on the social influence model. They continued the

emphasis on behavioral outcome but added an investigation into the

counselor characteristic of race. They used expert and referent power

bases with white and black counselors. The clients, all white, were

assessed for perceptions, attitude change, behavior change, and cog-

nitive retention. A significant race by role interaction was dis-

covered. White subjects responded more favorably to black experts

versus black referents. Taking into account the results of the

Atkinson et al. (1978) study for the effect of expertness on race, one

can infer particular approaches by particular race counselors might be

more effective with a particular race client. But an inference is

all that can be drawn, because in social influence literature,

counselor race has been systematically varied only in the Merluzzi

et al. (1977) study, and client race for that study was white.

One social influence study that did vary client race was done by

McKay, Dowd, and Rollin (1982). They investigated the effect of client

race on the perceptions of counselor social influence. Black and

white subjects viewed videotapes of counselors trained in both high and

low social influence roles. Black subjects and subjects of lower

education rated the low influence counselor significantly higher on

all measures of the Counselor Rating Form. The results suggested

the importance of the client characteristic of race and the possibility

that white middle class values might be inherent in the counseling

model. Unfortunately, counselor race was not varied and the ratings

were of a videotape rather than of a counseling experience.

Recent reviews of social influence in counseling have aptly

described its limitations and its potential. Corrigan, Dell, Lewis,

and Schmidt (1980) note that most studies have used only one interview,

have concentrated on attitude change primarily, and therefore counselor

influence is difficult to generalize. Perceived expertness and trust-

worthiness appear to be legitimate aspects of counselor power. The

role of attractiveness, at least for white clients, may come more into

play beyond the first session.

Blocher (1980) sees the work of Strong (1968) and later Johnson

and Matross (1975) as being part of a process of building a more

active empirical investigation of the counseling process. He believes

it will lead to a more systematic application of models to particular

settings, problems, and populations.

Heppner and Dixon (1981) summarized what is known about the

events affecting counselor power. They cited objective evidence,

counselor behavior, and prestigious cues as sources of counselor

expert power. Several responsive nonverbal behaviors and self-

disclosures were consistently associated with counselor attractiveness.

They note that although race as a counselor characteristic has been

investigated, such research involving the interpersonal influence

process has produced mixed results. They did not address the lack of

a systematic exploration of the client characteristic of race.


Investigation of the literature dealing with the black client

reveals a collection of ideas, propositions, and theories that are

largely in essay form. Empirical research is either lacking or has

provided conflicting information in many respects. Yet there are

some general conclusions which can be drawn from the available litera-

ture. The majority of the research indicates that black clients not

only prefer to work with black counselors but apparently are more

satisfied with them. Racial differences by themselves do not explain

the mixed results in outcome research. Interactions appear to be

present. Finally, there are some indications that different counsel-

ing approaches or models may be key components in the interaction


The social influence model (Strong, 1968) has a decade of research

behind it defining specific behaviors and manipulations that can

readily be applied in an empirical setting. A counselor's expertness

can be defined by reputation, behavior, and objective evidence. This

expertness role has been repeatedly shown to be effective in influ-

encing others. A counselor's attractiveness, or referent power base,

has also been defined by reputation and behavior. In their "pure"

forms both expertness and attractiveness appear to be equally effective

with white clients.

Counselor race has been systematically varied in only one study

in this area resulting in a significant race by role interaction for

white clients. All that is lacking is the application of a similar


study to an exploration of race by role interactions with the black

client. With the identification of specific behaviors for each role,

their relative effectiveness with the black client can be assessed.


The purpose of this study was to assess the effects of black and

white counselors' expert and referent power bases on the perceptions,

attitudes, and behavior of black and white clients. A secondary

purpose was to further the development of the Strong (1968) and

Strong and Matross (1973) social influence model of the counseling

process. It may very well be that expert and referent power bases may

interact with the race of the client and be differentially effective

as a result of the client's race. This chapter contains the methodol-

ogy that was used in the study. The research design, rationale for

the design, hypotheses, subjects, instrumentation, treatment, pro-

cedures, data analyses, methodological assumptions, and limitations

are described in this chapter.

Research Design

A factorial design with volunteer client subjects from each race,

random assignment to groups by race, and random assignment to treatment

was used. All measures were posttest. The four treatment groups were

compared with a control group. The nature of the study lent itself

to a factorial design which allowed multiple comparisons and analyses

of possible interactions.

The three independent variables were counselor race, counselor

role, and client race. Treatment cells consisted of counselor race

and counselor role combinations. There were three dependent variables.

The subjects' perceptions of the counselors were the first dependent

variable and were measured by the Counselor Rating Form (Barak &

LaCrosse, 1975). The second dependent variable was subject's attitude

toward problem solving. This was measured by an attitude scale

developed by Carey (1958). The third dependent variable was a be-

havioral measure of the strength of the therapist's influence on the

client. This measure is similar to the procedure used in the

Merluzzi et al. (1977) study. The third dependent variable was a

frequency count of those subjects who return a mailed questionnaire

within a prescribed time limit.

Rationale for Design

Although analogue research has been severely criticized as being

not generalizable at all (Goldman, 1977), it has been supported as an

analytic method of investigating questions that might be impractical

to evaluate in clinical situations (Kazdin, 1978). The nature of

analogue research itself is becoming a subject of research, and Helms

(1978) reported results that were surprisingly similar between an

analogue study and naturalistic research.

The very nature of analogue research contains both its strengths

and its weaknesses. It gives the experimenter control over the in-

dependent variables and subject assignment, thus enhancing internal

validity. On the other hand, analogue research threatens external

validity because of the artificiality that control produces.

According to Kazdin (1978), all treatment research is an analogue

of a clinical situation. Johnson and Matross (1977) defend experimental

counseling research strongly by writing, "internal validity is the more

important, for if the experimental effect does not occur, there is no

need to worry about its generality" (Johnson & Matross, 1977, p. 425).

They further state, "External validity of research studies is always

an empirical question which may be answered only by systematic repli-

cations of an experiment in a variety of settings with procedures

which adequately operationalize the conceptual variables in each

setting" (Johnson & Matross, 1977, p. 425). This study attempted a

systematic replication and extension of Merluzzi et al. (1977) based

on roles delineated in Dell (1973).


The following hypotheses were tested in this study:

1. There were no differences among black and white

clients' perceptions of counselors.

2. There were no differences among black and white

clients' attitudes towards problem solving.

3. There were no differences among black and white

clients in being influenced to return a question-


4. There was no interaction among the independent



Subjects for this study were undergraduate students registered for

the Spring 1982 semester at the University of Florida in Gainesville,

Florida. Gainesville is located in North Central Florida with a city

population of 83,000 and an urban area population of 118,000. The

University of Florida is the largest state supported institution of

higher learning in Florida and has a total student body of approxi-

mately 31,000.

In order to obtain the desired number of subjects for the study,

an offer of career counseling was made in several ways. In the

first offering the registrar's office provided separate alphabetized

lists of black and white undergraduate students and their addresses.

All students were assigned a number from one to the nth individual.

Using a table of random numbers, 250 students were randomly selected

from each list. These individuals were mailed an announcement (see

Appendix J) offering a limited number of students an opportunity for

an interview with a counselor, if they were interested in discussing

problems of problem solving and career planning (Merluzzi et al., 1977;

Dell, 1973). The announcement noted that while all participants

would not receive an interview, all would receive an information and

referral packet. The interview and packet were described as part of

a study in problem solving and career planning. If interested, stu-

dents would then phone the experimenter. Students were asked if they

had ever been to the Career Resource Center. If the students said

"yes," they were thanked for responding but were not included in

the study. This method resulted in only 10 subjects, 5 black and

5 white.

The same information as contained in the mailed announcement was

presented in several different university settings. It was posted on

the psychology department experimental sign-up board. Psychology

students are required to participate in a minimum number of studies.

This added 26 more subjects, 25 white and 1 black. Oral presentations

of the announcement notice were made at the following settings result-

ing in the participation of the indicated number of subjects: a voca-

tional exploration class--lO white students; the Black Student Union--

11 black students; a black fraternity--8 black students; a black

sorority--lO black students; the athletic study hall--5 black students.

Once a minimum n for each group (40 blacks and 40 whites) was

obtained, the separate racial groups were randomly assigned to groups

of 8 and then randomly assigned to one of five treatment conditions.

The groups were as follows:

Treatment Conditions

Black Black White White
Counselor Counselor Counselor Counselor
Expert Role Referent Role Expert Role Referent Role Control

Subjects 8 8 8 8 8

Subjects 8 8 8 8 8

Students did not know whether they had an interview scheduled or an

information and referral packet until they arrived.


The Counselor Rating Form

The Counselor Rating Form (CRF) (Barak & LaCrosse, 1975) was

developed to measure the dimensions of expertness, attractiveness, and

trustworthiness. These were the key dimensions in Strong's (1968)

social influence counseling model. This was systematically defined

by Strong and Schmidt (1970), Strong and Dixon (1971), and Kaul and

Schmidt (1971). The CRF, modified by LaCrosse and Barak (1976),

consists of 36 bipolar seven-point scales. From an original list of

83 adjectives suggested by communication research, four judges selected

and assigned 36 to appropriate influence dimensions. All judges agreed

on 22 of the assignments, while three judges agreed on the remaining

14. Students then rated filmed interviews using the new scale.

Across the ratings of three different counselors the dimensions of

expertness, attractiveness, and trust appeared to be distinct from

each other. Attempts to extract another factor produced meaningless

results. Each of the three dimensions is measured by 12 items. The

range of scores is from 12 to 84.

The CRF can discriminate between and within counselors on the

three dimensions (Barak & Dell, 1977; LaCrosse, 1977; Kerr & Dell,

1976; LaCrosse & Barak, 1976). The CRF (see Appendix A) also has

very high intercorrelations with the more established Barrett Rela-

tionship Inventory (Barrett-Lennard, 1962) for both client and ob-

server ratings. The predictive validity of the CRF was tested

(LaCrosse, 1980) with strong indications of a positive relationship

between measured perceptions and outcome. LaCrosse and Barak (1976)

report split half reliabilities of .87, .85, and .90 for expertness,

attractiveness, and trustworthiness, respectively.

Problem Solving Attitude Scale

The problem solving attitude scale (see Appendix B) is an 18 item

Likert type scale that was developed by Carey in 1958. She was inter-

ested in investigating the differences between men and women in problem

solving ability and hypothesized that the difference in problem solving

performance was actually a reflection of a difference in attitude

toward problem solving. According to Carey, an attitude is liking or

not liking to solve problems or "to the valence which problem solving

has for the individual" (p. 256). Originally she and a group of re-

searchers selected 63 items which they believed were related to prob-

lem solving. She then conducted an item analysis using 32 males and

100 females from introductory psychology classes at Stanford University.

From the original items the investigators were able to select 36 which

differentiated between high and low scores at the .05 level or better.

The 36 items were then split up on the basis of content to form two

tests, Form A and Form B. Many of the items refer directly to an

interest in or frequency of problem-solving activity. These were the

most highly differentiating items in the analysis.

The instrument appears to be relatively reliable. Intercorrelation

of the two forms on the first sample was .94. With a second sample of

59 men and 50 women equivalent forms reliability was reported as .83.

The functional characteristics of the test were successfully assessed

upon a third sample of 48 women and 48 men also from Stanford Univer-

sity. The only estimates lacking are those for test-retest reliability.

According to Shaw and Wright (1967) the instrument is average

or above average in validity. "Some internal evidence of validity was

provided by the fact that items that referred directly to interest in

or frequency of problem solving activity were among the most highly

differentiating items in the items analysis" (Carey, 1958, p. 257).

Evidence for predictive validity also exists. Carey's results (1958)

showed a positive relationship between performance scores and scores

on her attitude scale. Further results reported by Carey are that men

were found to have significantly higher scores than women and that

women showed a change to a more positive attitude after an influence


Scoring for the scale can range from 18 to 90 with the high score
being the most favorable. Individual items are scored from 1 (strongly

disagree or almost never) to 5 (strongly agree or almost always).

Weightings for negative items are reversed. High scores indicate a

positive attitude toward problem solving while low scores indicate a

negative attitude.

Behavioral Measure

The behavioral measure was a duplication of the approach used by

Merluzzi et al. (1977). It was an indirect assessment of therapist

influence, since no specific suggestion was made that it be accom-

plished. The behavioral measure consisted of an accounting of those

individuals who returned a follow-up questionnaire within two weeks

of its mailing. This follow-up questionnaire consisted of the alter-

nate form of the problem solving attitude scale and open-ended

questions allowing student feedback.


Subjects were randomly assigned to one of five possible conditions:

1) Black expert counselor, 2) White expert counselor, 3) Black referent

counselor, 4) White referent counselor, or 5) Control group. Inter-

view roles were patterned after the portrayal of the expert and re-

ferent power base roles used by Merluzzi et al. (1977) and Dell

(1973). Dell (1973) designed the format for the role portrayal based

upon the definitions of the expert and attractive power bases in

Schmidt and Strong (1971) and Strong and Schmidt (1970). Additional

material for the role portrayed came from the work of Strong and

Matross (1973) on the impelling forces of counselor power. Both the

expert and referent roles provided evidential, reputational, and

behavioral cues to the subjects (Corrigan et al., 1980).

Interview Roles

The expert interviewer was described to the subject as

Ms. who is in a counseling doctoral program at the

University of Florida. She was described as being aware of a lot of

recent research in career counseling and decision making. Following

Dell's (1973) expert role development, the expert interviewer did

not rise from her chair but directed the subject to the seat.

The expert's opening remarks structured the sequence of events

to follow. This summary structured the roles of the client and the

interviewer. Items addressed included expert questioning, student

responses, expert commentary and literature citation, expert sugges-

tions and description of problem-solving strategy, and finally a

recommendation and referral to the Career Resource Center.

All expert questioning was explicit and came from a formal proto-

col. Any comments made by the expert were solely in regard to other

students she had counseled or to research. No comments were made

citing any similarity between the counselor and the subject nor was

disclosure of personal experiences made. A minimum of four references

were made by the counselor to establish their expertnesss."

The expert was attentive and professionally interested in the

subject. Eye contact was maintained, and the counselor was responsive

to the subject's behavior. Hand gestures were used to emphasize

critical points. The expert performed with an air of confidence in

her ability. Expert attire was not casual; a pants suit or dress were

worn (Kerr & Dell, 1976).

The referent interviewer was introduced as a graduate student

counselor who has had to deal with career planning problems herself

and who was now interested in helping others with similar concerns.

The referent interviewers attempted to convey an open friendly attitude

and liking for the client. She greeted the client at the door, used

her first name, and made comments to put the client at ease.

The referent counselor was verbally responsive to the client

throughout the interview and made use of reflective statements to dem-

onstrate understanding. No written protocol was used for the inter-

view. Four self-disclosures were made by the counselor to stress

similarity of attitude and experience with the client (Merluzzi et al.,

1977). No reference was made to the counselor's previous counseling

experiences or psychological knowledge.

Interviewers and Interviewer Training

The interviewers for this study were two white female and two

black female graduate students in the Counseling Psychology Program at

the University of Florida. The interviewers were paid at the rate of

five dollars per hour of training and counseling. Each interviewer

worked with 16 subjects. Training consisted of three phases. In the

first phase all interviewers were brought together at the Counseling

Laboratory. Each interviewer was provided with general instructions

(see Appendix C), a written description of the appropriate counseling

role (see Appendices D and E), plus a copy of the problem solving

strategy (see Appendix F) to be given to all students. Expert inter-

viewers additionally received a copy of a protocol (see Appendix G)

to follow in interviewing the students, plus a special version of the

problem solving strategy (see Appendix H). This special version con-

tained research references and results. In this first phase all

interviewers studied their procedures and were given the opportunity

to clarify their understanding of their roles.

In the second phase of training expert interviewers were trained

separately from referent interviewers. The second phase activities

consisted of modeling and rehearsal. Interviewers viewed an enactment

of their role as portrayed by the experimenter. Respective roles were

then rehearsed for two hours.

The third phase consisted of evaluation of role standards. Each

interviewer was videotaped and judged by a three-member panel of

advanced counseling psychology students using the Counselor Rating

Form. The expert interviewers did not see subjects until their

expertness rating was significantly above their attractiveness rating

for all three judges. The referent interviewers did not see subjects

until their attractiveness rating was significantly above their

expertness rating for all three judges (p < .10). Results of this

judging are in Table 2. To ensure some measure of control over the

way the roles were actually enacted with the subjects, a random

selection of eight interviews (10 percent) were videotaped and rated

by the panel of three advanced graduate students.

Counselor Expertness-Attractiveness Rating--Preexperiment

Treatment Mean T PR > T

White Referent -15.67 -7.73 .02

White Expert 20.34 6.86 .02

Black Expert 22 9.53 .01

Black Referent -22 -3.99 .06


All subjects had specific appointment times for their interviews

in the Counseling Laboratory of the Counselor Education Department.

Interviews were scheduled to last 20 to 25 minutes with a 5-10 minute

break between sessions. Thus, one interview was completed every 30

minutes. The career development laboratory room was used as a recep-

tion area. Subjects were greeted by a receptionist who gave them

informed consent forms to be signed (see Appendix I).

Treatment subjects were then brought to one of two treatment

rooms reserved for the experiment. En route to the interview room,

the receptionist gave each subject a short description of the inter-

viewer. The description was appropriate to the interviewer's role,

expert or referent, and was given to enhance the effectiveness of that

role. Control subjects were given their information and referral

packets and given 10 minutes to read them.

All interviews lasted 20 to 25 minutes. During the last 10 min-

utes of the interview, an influence attempt was made that was con-

sistent with the respective roles. The interviewer introduced a

problem-solving strategy and applied it to the specific situation

being discussed. The problem-solving strategy used the five stages

identified by D'zurilla and Goldfield (1971) and extended by Heppner

(1978). This same information was contained in the handout given to

the control subjects upon arrival and to interviewees after testing.

Finally, a suggestion was made at the end of the interview that the

subject visit the Career Resource Center very soon. A referral card

was given to each student. The suggestion was incorporated as part

of the problem-solving task.

Immediately after the interview, subjects completed the attitude

scale and the Counselor Rating Form. The control subjects completed

the attitude scale but not the Counselor Rating Form. All participants

received information and referral packets after this testing. Each

individual had a referral slip to turn in to the receptionist at the

Career Resource Center.

At the Career Resource Center, the students were given the

opportunity to sign up for minicourses in 1) center services,

2) cooperative education, 3) job hunting tactics, and 4) career

planning. Center services also include 1) a career information library

with a special section for minority students, 2) audio-visual resources,

3) a computer aided decision-making activity, 4) job interview sign-up

lists, and 5) personal vocation counseling.

Seven days after each individual's interview, the follow-up

questionnaire was mailed out. Each mailing contained Form B of the

problem-solving attitude scale (Carey, 1958), several open-ended

questions, a cover letter, and a stamped self-addressed envelope.

Subjects had two weeks to return the questionnaire. Those returned

after two weeks did not meet the behavioral criteria for the variable.

Data Analysis

In order to assess the effects of expert and referent power bases

on client perceptions, a three-factor analysis of variance was used

as depicted in Table 3. In order to take into account the counselor

and client race variables, a three-factor analysis of variance was

conducted on perceptions of expertness and then on perceptions of

attractiveness. This procedure allowed the determination that the

roles were differentially perceived. It also allowed the study of

the separate and combined effects of counselor role, counselor race,

and client race on perceptions.

In order to assess the effect of treatment (counselor race x

role or control), client race, or interactions on client attitudes, a

two-factor analysis of variance design was used. This allowed the

incorporation of data from the control groups (Table 4).

Three Factor


Counselor Race A

Treatment x Client Race




9 .9.




Client 8 8 8 8 8

Client 8 8 8 8 8



to gain a more detailed understanding of the nature of the interac-

tions, a more refined analysis was possible by taking the additional

step of dropping control groups and performing a three-factor analysis

of variance as depicted in Table 3. The Tukey multiple comparison

procedure was used for all anovas in order to make all possible pair-

wise comparisons. A multiple comparison error rate of .05 was used

to provide a 95 percent simultaneous confidence interval.

The response to the behavioral measure was a dichotomous variable

which did not lend itself to analysis of various procedures. In the

past decade, however, statistical theory has led to advances in the

analysis of multidimensional cross classified categorical data that

are analogous to analysis of variance procedures. These loglinear

models as described in Bishop et al. (1975), Reynolds (1977), and

Fienberg (1977) are able to "describe the structured relationship

among the variable corresponding to the dimensions of the table"

(Fienberg, 1977, p. 3). Interaction is defined as being based on the

cross product ratios of expected cell values. "The models are linear

in the logarithms of the expected value scale" (Fienberg, 1977, p. 3).

From this is derived the term loglinear. The coefficients of loglinear

models are based on maximum likelihood estimation.

In order to assess the affects of counselor role, counselor race,

and client race on return of the follow-up questionnaire, a loglinear

analysis was used. The loglinear analysis made use of a 2 x 5 table

as depicted in Table 4. The analysis made use of the X2 statistic.

The results were a rank order of counselor race x role treatment

conditions according to client race, determination at the .05 level

of difference, and an assessment of possible interactions.

Methodological Assumptions

Use of the F ratio assumes independent random samples, normal
population distribution, and equal population variances. Moderate

departures from normality of populations can be tolerated. In the
case where sample sizes are equal, the F test is robust to violations
of the assumption of equal variances.

The X2 statistic can be used only with frequency data composed of
independent measures. The X2 tests are approximate in the sense that

the true sample distribution of the X2 statistic approaches the X2

distribution as n increases (Agresti & Agresti, 1979). The tests
should not be used if several of the fe's are close to zero. When fe
sets close to zero, the statistic tends to blow up. If these condi-

tions are not met, there are analogs of Fisher's exact test for the

r x c table (Agresti & Wackerly, 1977).


One dimension which posed a particular limitation for this study

was the population. Taken together, the target problems of problem
solving and career planning were highly relevant areas of counseling
for a university client population. Results of this study would not
necessarily be appropriate for a nonuniversity population. A second

limitation was that the sample used in the study was not selected
randomly and is not truly representative of the university as a whole.

The sample is, however, more nearly representative of those students

likely to take advantage of career oriented counseling as offered by

a university agency. A third limitation was that only female

counselors were used. Further research will be needed to assess the

applicability of findings for male counselors. The fourth and fifth


limitations were the duration of the interview and its one-time

occurrence. These two dimensions were at the heart of the analogue

question. Effects attributed to treatment during such a short session

would be most appropriately generalizable to the early stages of a

counseling relationship or to a one-visit session.


The purpose of this study was to assess the effects of black and

white counselors' expert and referent power bases on the perceptions,

attitudes, and behavior of black and white clients. The results of the

data analyses for the four hypotheses are presented in this chapter,

followed by a discussion of the results.


Hypothesis One

The first hypothesis was that there would be no differences

among black and white clients' perceptions of counselors. A three

factor analysis of variance was conducted, first on client perceptions

of expertness, and then on client perceptions of attractiveness. A

3x2 ANOVA procedure was used to allow the study of the separate and

combined effects of counselor role, counselor race, and client race

on perceptions.

In the analysis of the expert means a significant main effect was

found for the race of the subject. Table 5 presents the results of the

analysis of variance for the dependent variable of expertness. The

p value for race was significant at the .05 level. There were no

other significant effects or interactions. This result means that the

black clients in this study perceived the counselors in general to be


ANOVA of Counselor Race by Role and Client Race
for Expert Variable

Source df MS F

CRACE 1 4.52 .06

ROLE 1 0.02 0.00

CRACE*ROLE 1 97.52 1.28

RACE 1 328.52 4.30*

CRACE*RACE 1 92.61 1.21

ROLE*RACE 1 83.27 1.09

CRACE*ROLE*RACE 1 5.65 .07

ERROR 56 76.43

*p < .05.

more expert than the white clients perceived them. Therefore, the

first hypothesis was rejected for the variable expertness.

In testing the first hypothesis a three factor analysis of vari-

ance was also conducted on the clients' perception of counselor

attractiveness. The results depicted in Table 6 indicate a significant

main effect for role. The p value for role was significant at the

.05 level. This result means that all clients perceived the referent

counselors as being significantly more attractive than the expert

counselors. The hypothesis that no differences existed among black

and white clients' perceptions of attractiveness was not rejected.

ANOVA of Counselor Race by Role and Client Race
for Attractiveness Variable

Source df MS F

CRACE 1 13.14 0.19

ROLE 1 425.39 6.21*

CRACE*ROLE 1 0.39 0.01

RACE 1 19.14 0.28

CRACE*RACE 1 192.52 2.81

ROLE*RACE 1 206.64 3.02

CRACE*ROLE*RACE 1 1.26 0.02

ERROR 56 68.52

*p < .05.

Further analyses were made of the data generated for the first

hypothesis. Means and standard deviations for counseling variables by

client race are contained in Table 7. Means and standard deviations

for counseling variables by counselor role are contained in Table 8.

Tukey's studentized range test was used in the further analysis of

both the expertness and attractiveness variables. The test provided

rank orders of counselor race x role treatment conditions for all

groups according to client race. The rank order for expertness means

is in Table 9. There were no significant differences. The rank order

for attractiveness means is in Table 10.

The results in Table 10 indicate that there was a significant

difference at the .05 level in the perception of attractiveness by

Means and Standard Deviations for Counseling Variables
by Client Race

Race--White Race--Black


Expertness 70.25 10.49 74.78 6.09

Attractiveness 72.75 10.29 73.84 6.71

Means and Standard Deviations for Counseling Variables
by Counselor Role

Role--Expert Role--Attractive


Expertness 72.53 9.26 72.52 8.48

Attractiveness 70.72 9.46 75.88 6.94

Rank Order of Means of Perceived Expertness by Race
and Treatment Condition

Race--White Race--Black

Treatment X Treatment X

Black Referent 71.38 Black Referent 76.625
White Referent 71.38 White Referent 76.000
White Expert 71.00 Black Expert 75.875
Black Expert 67.25 White Referent 70.625

Rank Order of Means of Perceived Attractiveness
by Race and Treatment Condition

Race--White Race--Black

Treatment X Treatment X

White Referent 79.38* Black Referent 76.125
Black Referent 74.88 Black Expert 74.125
White Expert 70.50 White Referent 73.125
Black Expert 66.25* White Expert 72.000

*p < .05.

white clients. This means that white clients perceived white referent

counselors to be significantly more attractive than black expert


Hypothesis Two

The second hypothesis was that there would be no differences

among black and white clients' attitudes towards problem solving. It

was desirable to be able to compare client attitudes towards problem

solving in all treatment conditions. Each treatment condition con-

sisted of either a control group or a counselor's race and role. A

two factor ANOVA assessed the effect of treatment and client race on

problem solving attitude. The data presented in Table 11 show a strong

effect for client race that is significant at the .01 level. There

was no significant effect for treatment and no interaction between

treatment and client race. The net result was that black clients

had a significantly more positive attitude toward problem solving

than the white clients. The second hypothesis was therefore rejected.

ANOVA of Treatment by Race for Problem Solving
Attitude Variable

df MS F

Treatment 4 28.58 0.36

Race 1 696.20 8.84*

Treatment*Race 4 105.20 1.34

ERROR 78.75

*p < .01.

For further analyses the control groups were dropped and a three

factor ANOVA was conducted. Again the only significant factor was a

main effect for client race at the .01 level. These data are in

Table 12. Table 13 contains the means and standard deviations for the

black and white problem solving attitude scores.

From the 2x5 ANOVA design which includes all treatment and

control conditions by client race, a Tukey multiple range test was

conducted. This yielded a rank order of counselor race and role

conditions for each race of client which is depicted in Table 14.

The results show that when each client race group was considered

separately, there was no significant difference in problem solving

attitude between any two treatment levels.

Hypothesis Three

The third hypothesis was that there would be no differences

among black and white clients in being influenced to return a

ANOVA of Counselor Race by Role and Client Race
for Problem Solving Attitude Variable

df MS F

CRACE 1 0.76 0.01

ROLE 1 28.89 0.33

CRACE*ROLE 1 83.26 0.96

RACE 1 708.89 8.14**

CRACE*RACE 1 199.52 2.29

ROLE*RACE 1 141.01 1.62

CRACE*ROLE*RACE 1 34.52 0.40

ERROR 87.09

**p < .01.

Means and Standard Deviations for Problem Solving
Attitudes by Client Race

Race--White Race--Black


53.53 9.33 60.19 9.28

Rank Order of Means of Problem Solving Attitude
by Race and Treatment Condition

Race--White Race--Black

Treatment X Treatment X

Black Referent 57.75 White Expert 64.75

Control 55.75 Black Referent 59.38

White Referent 53.63 White Referent 59.38

Black Expert 52.63 Control 58.62

White Expert 50.12 Black Expert 57.25

questionnaire. A loglinear analysis of the 2 x 5 tables using the

chi-square statistic was conducted. Results are depicted in Table 15.

Neither treatment, nor race, nor an interaction of the two was sig-

nificant. This means that no difference could be claimed for black

and white clients in being influenced to return a questionnaire. It

Loglinear Analysis of Treatment by Client Race
for Return of Questionnaire

Source df Chi-Square Probability

Treatment 4 1.65 .80

Race 1 3.52 .06

Treatment*Race 4 4.80 .31

also means that neither client race nor treatment (counselor race and

role) influenced the rate of questionnaire return. The third hypothe-

sis was not rejected. It was noted that the client race factor did

approach significance. Response frequencies by treatment are in

Table 16. Table 17 contains the treatment rank order by race.

Hypothesis Four

The fourth hypothesis was that there would be no interaction

among the independent variables. The independent variables were

counselor race, client race, and counselor role. The dependent

variables were client perception of the counselor, client attitude

toward problem solving, and client return of a mailed questionnaire

within a time limit.

Possible interactions among the independent variables were as-

sessed at each step of the process of testing the first three hypothe-

ses. Tables 5 and 6 contain results of three-factor ANOVA's of

client perceptions of expertness and attractiveness, respectively.

No significant interactions were found with any independent variables.

Table 11 contains results of a two-factor ANOVA of treatment by race

for problem solving attitude. No interaction was noted. A three-

factor ANOVA was also conducted for the problem solving attitude

variable. Results in Table 12 show no significant interactions.

Finally, the loglinear analysis results for the return of the ques-

tionnaire, which are in Table 15, show no interaction.

No interaction effects were noted as significant in any of the

analyses. Therefore, the hypothesis of no interaction among in-

dependent variables was not rejected. This means that no

Questionnaire Response
Treatment and

Frequencies by

Treatment Race--White Race--Black Total

Control 4 6 10

White Referent 6 5 11

Black Referent 5 5 10

White Expert 2 6 8

Black Expert 4 7 11

TOTAL 21 29 50

Rank Order of Questionnaire Return
by Treatment and Race

Race--White Race--Black

Treatment Frequency Treatment Frequency

White Referent 6 Black Expert 7

Black Referent 5 White Expert 6

Black Expert 4 Control 6

Control 4 Black Referent 5

White Expert 2 White Referent 5

interaction between or among counselor race, counselor role, or client

race could be claimed as a factor in influencing client perception of

expertness or attractiveness, attitude toward problem solving, or re-

turn of the questionnaire.

Quality Control

A procedure was instituted to monitor the actual enactment of the

counselor roles during the study. A random selection of two inter-

views for each of the four counselors was videotaped. Interviewers

and clients never knew which sessions were being videotaped. Each

videotaped interview was rated by the same panel of intern level doc-

toral counseling students using the Counselor Rating Form (Barak &

LaCrosse, 1975). For an expert interviewer to be considered signifi-

cantly within her designated role, she would have to be perceived by

the judges as significantly more expert than attractive. For a

referent interviewer to be considered significantly within her desig-

nated role, she would have to be perceived by the judges as signifi-

cantly more attractive than expert. The criterion for significance

was a p value of .10.

T-test results for each treatment condition are shown in

Table 18. The white referent counselor was judged significantly more

attractive than expert on her first tape. There was no significant

difference on her second tape. The white expert counselor was sig-

nificantly more expert than attractive on her first tape, but there

was no significant difference on her second tape. The black expert

counselor was significantly more expert than attractive on both of

her tapes. The black referent counselor was not significantly more

T-Test Results

on Role Maintenance by Counselors

Treatment Tape # T PR > T

White Referent 1 7.95 .02*

White Referent 2 1.79 .22

Black Referent 1 2.56 .12

Black Referent 2 2.29 .15

White Expert 1 -4.62 .04*

White Expert 2 -1.38 .30

Black Expert 1 -6.55 .02*

Black Expert 2 -4.99 .04*

*p > .05.

attractive on either of her tapes. Although the results consistently

show an attractiveness-expertness difference in the desired direction

for every tape, they also suggest a lack of consistency in the degree

to which the roles were enacted.


The first analysis dealt with the separate and combined effects

of counselor role, counselor race, and client race on client percep-

tions. Results were different from what was expected in the case of

perceived expertness. Counselor roles were experimentally manipulated

to be either expert or referent, yet in the analysis of the expert

scores there was no main effect for role. There was, however, a main

effect for race. Black clients perceived the counselors in general

to be more expert than the white clients perceived them. There was no

racial difference between the clients in their perception of the

counselor's attractiveness. All clients perceived the referent

counselors as being significantly more attractive than the expert

counselors. In this case there was a main effect for role. A within

race analysis for clients indicated that for white clients, white

referent counselors were significantly more attractive than black

expert counselors.

Significant results also led to the rejection of the second

hypothesis. A significant main effect for client race was again

noted, this time in regards to attitude toward problem solving. Black

clients had a significantly more positive attitude toward problem

solving than white students. Testing of differences in the rate of

return of a questionnaire yielded no significant results. Finally,

an assessment of interaction influence on any of the dependent vari-

ables yielded no significant results.


As has been described in the preceding chapter, counselors were

trained to use specific behaviors and attire which research has shown

delineates either the expert or referent role, respectively. Dif-

ferent specific introductions were also used for each role. It was

therefore expected that regardless of other possible results, the

counselor's role would be a significant factor in the client's per-

ception of the counselor's expertness or attractiveness. Results

indicate that the expert counselor role, as defined by introduction,

attire, and behavior, had no effect on client perception of counselor

expertness. This is contrary to all expectations derived from the

results of extensive early research as reviewed by Corrigan et al.

(1980). On the other hand, the referent counselor role was a sig-

nificant factor in the client perception of counselor attractiveness.

The results raise the question of whether there was successful

incorporation of expert and referent role criteria in the training

and/or performance of the counselors. Results located in Table 2,

Chapter Three, indicate that all intern judges rated both expert

counselors as significantly more expert than attractive on their

training tapes. Results from Table 18 indicate the intern judges

rated the white expert counselor as more expert than attractive on one

of two quality control tapes. The intern judges rated the black

expert counselor as more expert than attractive on both quality control

tapes. The data from videotaped sessions indicate that the expert

criteria were incorporated in training but not necessarily consistently

in practice.

The data on referent training and practice are even more il-

luminating. Both referent counselors were judged by the interns as

being significantly more attractive than expert on their training tapes.

On the quality control tapes, however, the interns judged the white

referent counselor as being more attractive than expert on only one

of two tapes. The black referent counselor was not judged as being

more attractive than expert on either quality control tape. Yet for

the clients as a whole there was a significant main effect for the

attractiveness role. It was differentially perceived by the clients.