The Relationship of repression-sensitization to aspects of marital dyad functioning

Material Information

The Relationship of repression-sensitization to aspects of marital dyad functioning
Day, Dennis Alvin, 1944-
Place of Publication:
Gainesville, Fla.
Dennis Alvin Day
Publication Date:
Copyright Date:
Physical Description:
xi, 115 leaves. : ; 28 cm.


Subjects / Keywords:
Dyadic relations ( jstor )
Human aggression ( jstor )
Husbands ( jstor )
Philosophical psychology ( jstor )
Psychological stress ( jstor )
Ratings ( jstor )
Self ( jstor )
Social interaction ( jstor )
Spouses ( jstor )
Stress ratio ( jstor )
Adjustment (Psychology) ( lcsh )
Dissertations, Academic -- Psychology -- UF ( lcsh )
Married students ( lcsh )
Psychology thesis Ph. D ( lcsh )
Repression (Pshchology) ( lcsh )
bibliography ( marcgt )
non-fiction ( marcgt )


Thesis -- University of Florida.
Bibliography: leaves 112-114.
Additional Physical Form:
Also available on World Wide Web
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
000577263 ( AlephBibNum )
13944038 ( OCLC )
ADA4958 ( NOTIS )


This item has the following downloads:

Full Text

The Relationship of Rcprcssioni-Sensitization
to Aspects of llarital Dyad Functioning





EEEEEEEE"11111111111 i

To my wife,



The writer wishes to gratefully acknowledge the

assistance he received from the members of his supervisory

committee: Drs. Benjamin Barger, Chairman; Carl Clarke,

Marvin Shaw, and Richard MicGee of the Department of Psy-

chology; and Dr. Gerald Leslie of the Department of Sociology.

He also wishes to thank Mlr. Larry Carstensen for assisting

with the statistical analysis and the computer programming.

He wishes to express a special word of appreciation to

Dr. Carl Clarke, who provided invaluable professional com-

petence and personal encouragement in all phases of this









. . .

* .

* .

. . *

S. iii

. viii


. .

Theoretical Background . .
Self-Perception . . .
Personal Adjustment . .
Group Process . . .
Implications for Present Research .
Study of R-S in the Marital Relationship

II MrETHOD . . . . .

Subjects . . . .
Instruments . . . .
Procedure . . .
Analysis . . . .
Hypotheses . . .




. 1

S. 28

. 41

. 52

. 70





Health and Opinion Survey

Locke-Wallace .

Marriage and College
Environment Inventory .

Bales' Rating Scales .

* *

* *

* S *

* .





Inventory of Ibarital Conflict

Additional Statistical Tables

. * * .

Pa ge

. 94

. 109

. 112

S. .115


Table Par

1 Analysis of Variance of Locke-Wallace Ratings
for Repressors, Neutrals, and Sensitizers 42

2 Ranked Means of Locke-Wallace Ratings for
Repressors, Neutrals, and Sensitizers . 42

3 Ranked Eleans of Locke-Wallace Ratings for
All Neutral Dyads . . . 43

4 Analysis of Variance of IACE Stress/
Satisfaction Ratio for Repressors, Neutrals,
and Sensitizers . . . . 44

5 Ranked Means of MACE Stress/Satisfaction
Ratio for Repressors, Neutrals, and
Sensitizers . . . . 44

6 Ranked Means of iACE Stress/Satisfaction
Ratio for Neutrals in NR, NN, and NS Dyads 45

7 Ranked Means of INC "Win index" for
Both iates in RS Dyad . . . 46

8 Ranked Means of Bales'"Difference Score" for
Repressors, Neutrals, and Sensitizers . 47

9 Analysis of Variance of Bales'"Difference
Score" Repressors, Neutrals, and
Sensitizers . . . . 48

10 Analysis of Variance of Bales'(4-8) Self-Ratings
for Repressors, Neutrals, and Sensitizers 49

11 Ranked Means of Bales'(4-8) for Repressors,
Neutrals, and Sensitizers . . 49

12 Analysis of Variance of Bales' (4-8)
Rating-of-Spouse Score for Repressors,
Neutrals, and Sensitizers . . 49

13 Ranked Means of Bales' (4-8) Rating-of-Spouse
Score for Repressors, Neutrals,
and Sensitizers . . . ... 50


Table Pai

14 Analysis of Variance of IfLACE "Agreement
Score" for Represeors, Neutrals, and
Sensitizers . . . . 50

15 Ranked i.,eans of 1-ACE "Agreement Score" for
Repre;ssors, Ileutrals, and Sensitizers . 51


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



Dennis Alvin Day

June, 1972

Chairman: Dr. Benjamin Barger
Co-Chairman: Dr. Carl T. Clarke
Ilajor Department: Department of Psychology

The present investigation represents an attempt to

examine and extend findings from previous studies of the

repression-sensitization, R-S, personality construct into

the context of the marital dyad. In the R-S dimension,

repression has been represented as an orientation of avoid-

ance of threatening stimuli, while sensitization has been

represented as an orientation of approach to threatening

stimuli. Relevant literature was reviewed with focus being

made upon those studies which related R-S to aspects of small

group interaction. It was pointed out after this review

that research efforts were needed to relate the R-S variable

to the process and outcome features of natural groups. Also,

it was presented that information about the neutral on the

R-S dimension could serve to answer important issues in R-S

theory. Finally, the need to compare self-rating data with

data collected by other means was emphasized. The marital

dyad was chosen as the crucible for examining issues from

R-S research because of several reasons: (1) the marital


dyad is a substantive area in need of research from many

perspectives, (2) it is naturally-found, (3) it is self-

selected, and (4) it is of a long-term nature. The specific

areas of marital dyad functioning that were assessed by

research measurements were as follows: (1) marital adjust-

ment, (2) the perception of satisfaction and stress of

marriage in a college setting, (3) the resolution of marital

conflicts, (4) the perception of own aggressiveness, (5) the

perception of spouse's aggressiveness, and (6) the congruency

between self-perception and perception by one's mate.

The participants in this investigation were randomly

chosen and individually contacted married college couples.

Couples were successively contacted until 60 dads, repre-

senting all combinations of repressor, neutral, and sensitizer

mates, were identified and were willing to participate in

the study. Data relevant to this research's interests were

collected during both the initial identification session of

the study and a second measurement session.

The main findings of this study, as related to each of

the separate hypotheses, are as follows:

(1) Repressors reported significantly better marital

adjustment than did the sensitizers and neutrals.

(2) For all three R-S categories, repressors, neutrals,

and sensitizers, the lower their spouse's R-S score, the

better w'as their marital adjustment.

(3) Sensitizers reported significantly greater stress
from their life situation, relative to their reporting of

satisfaction, than did repressors and neutrals.

(4) For all R-S categories, repressors, neutrals, and

sensitizers, the lower their spouse's R-S score, the lower

was their assessment of stress from their life situation.

(5) The sensitizer, vhcn married to the repressor,
prevailed in having his judgment endorsed by his repressor

mate, when both were confronted with ambiguous choice situ-


(6) The sensitizer perceived himself to be more aggres-

sive in social situations than his mate perceived him to be.

There was no such discrepancy in the case of the repressor.

(7) The sensitizer rated himself as behaving at a

significantly more aggressive level than the repressor rated

himself to be.

(8) The feelings that the repressor held about aspects

of his life situation were more congruently perceived by his

mate than was the case for the sensitizer.

It was argued that these findings support the following

inferences: (1) that R-S is linearly related to measures of

personal and marital adjustment, (2) that R-S denotes a con-

struct that is more general and comprehensive than previously

considered, (3) that R-S represents more than just a "response

bias," and (4) that the expression and reaction to "aggres-

siveness" seems an important differentiating behavior relative

to R-S.

It was posited that R-S differentiates a generalized

style of reactivity to both personal and interpersonal

events, which in turn relates to experiencing both the

impact and evaluation of those events in characteristic

terms of stress-satisfaction and negativity-positivity.

Finally, possible implications of this research for marital

counseling were discussed.



The personality construct Repression-Sensitization (R-S)

has stimulated several research efforts to assess its value

and to determine its relationship to individual functioning

in situations of threat, ambiguity, and conflict. More

recently, additional interesting findings have been reported

relating R-S to various aspects of group functioning. The

present research investigation represents an attempt to exam-

ine and extend findings from previous studies of the R-S con-

struct into the context of the marital dyad.

Theoretical Eackground

Before presenting an explanation and rationale for the

present study, the historical development and scope of this

research area will be outlined.

A point of clarification should be offered about the
relationship of R-S to the instrument used to assess R-S
(Health and Opinion Survey). On this self-rating instru-
ment (Derived from M~IP scales), an individual who achieves
a low HOS score falls at the repressing end of the R-S con-
tinuum, while an individual who achieves a high HOS score
falls at the sensitizing end. Thus, a direct linear relation-
ship between R-S and a dependent variable would indicate that
the sensitizer ranks high on that particular factor, while an
inverse linear relationship would indicate that the repressor
ranks high on that factor.

The R-S construct evolved from studies in the late

1940's and the 1950's that have since been referred to as

the "new look" in perception. The essential feature of this

"new look" orientation was its approach to perception as

being an active process, ,.'herein features of the perceiver

figured perhaps as importantly in the process of perception

as features of the perceived stimuli. This orientation to

perception as an active process was described by Bruner and

Postman (1947),

Perception is a form of adaptive behavior. Its
operation reflects not only the characteristics
of sensorineural processes, but also the dominant
needs, attitudes, and values of the organism.
(p. 69)

John C. Raven (1951) also describes this orientation as

follows, ,

Our perception of events is intentional in the
sense that we always respond to and perceive
.events with intent, as if reaching out with desire
or effort directed toward the aoorehension of
some objective, although at the tiie not neces-
sarily realized and not necessarily pursued with
deliberation. (p. 15)

A newly identified phenomenon was reported in this early

research one which stimulated many diverse studies. The

experimental situation for these studies involved the measure

of recognition thresholds for tachistoscopically presented

words of different affective value. It was found that two

distinct patterns of differential recognition thresholds

existed for some subjects when visually perceiving these

"neutral" and "threatening" stimuli. One of these response

patterns was identified by subjects who exhibited a signifi-

cantly higher threshold for the recognition of "threatening"

stimuli than for "neutral stimuli." This response style was

characterized as an avoidance of threat or fear and was termed

"perceptual defense." While this finding is congruent with

long-established concepts about the avoiding of unwelcome or

fear-provoking stimuli, a perceptual pattern was discovered

that was not of such a predictable nature. This latter re-

sponse pattern was identified by subjects who revealed a

significantly lower threshold for the recognition of threat-

ening stimuli than for neutral stimuli. This response style

was characterized as an "approach" to threat or fear and it

was termed "perceptual vigilance."

Since the most important of the early investigations of

these perceptual findings were done by Jerome Bruner and Leo

Postman, a review of their experimental method and procedure

will be presented. In this study (Bruner and Postman, 1947)

subject's associative reaction times were elicited to a vari-

ety of stimuli words. These stimuli included words of both

neutral and potentially "threatening" connotation. Examples

of the latter, threatening words, include rape, penis, and

death. From the associative reaction time data, an individ-

ualized list of three groups of words was arranged for each

subject. These three groups consisted of those words that

had the longest associative reaction times, the shortest

associative reaction times, and the midmost reaction times.

At a later experimental session each subject was asked to

recognize this latter selection of words as they again were

presented tachistoscopically. The exposure time for each


stimulus word was progressively increased until the word w'as

correctly recognized. Two different recognition patterns

were found to exist for some of these subjects. One of the

patterns was defined by a longer exposure time for the recog-

nition of the threatening words as compared to the exposure

time necessary for the recognition of the neutral words.

While the other pattern involved exactly the opposite rela-

tionship, i.e. a relatively shorter exposure time neces-

sary for the threatening words compared to the neutral words.

The first of these patterns, or perceptual styles, exemplifies

perceptual defense and the second exemplifies perceptual


The term perceptual defense has subsequently evolved into

the term "repression," and the term perceptual vigilance has

evolved into the term "sensitization." Since the sensitizer,

perceptual vigilant, recognizesthe threatening stimuli

earlier than the neutral stimuli, this perceptual style was

also characterized as one of "approach', while the style of

the repressor, perceptual defender, which revealed the oppo-

site relationship was characterized as one of "avoidance."

Both of these perceptual modes share the fact that they in-

volve differential awareness to threatening stimuli. Conse-

quently, both have been thought to represent polar opposites

in the handling of the anxiety or fear that is aroused by

aversive stimuli. Hence, the perceptual styles have also

been thought of as defense mechanisms, while their polarity


has been considered to anchor an approach-avoidance dimension

of individual reactivity to aversive stimuli.

The majority of research subsequent to the investigation

by Bruner and Postman can be divided into two successive con-

ceptual endeavors. The earlier efforts consisted of diverse

attempts to verify, explain, and extend the nature of these

identified perceptual phenomena, while the second group of

investigations proceeded from the established fact of indi-

vidual differences in perceptual style, in an effort to de-

termine how this relates to self-perception, person percep-

tion, personal adjustment, and interpersonal interaction. It

is with this latter group of investigations that the present

research derives its conceptual ground.

A crucial feature of this second development in R-S

research w'as the establishment of a psychometric assessment

instrument of the Repression-Sensitization dimension. This

psychometric identification of perceptual style contrasts

with the original determination of R-S accomplished by psycho-

physiological methods. One of the first attempts to utilize

selected Minnesota i.iultiphasic Personality Inventory (MiPI)

scales as a measure of R-S was made by Jesse Gordon (1957).

He also originally coined the term sensitization. His effort

was followed by several attempts to utilize single scales of

the MrPI to assess R-S. The next development was by Altrocchi,

Parsons, and Dickoff (1960) who utilized multiple 1%MPI scales

to assess R-S style. However, it was Byrne (1961) who


accomplished certain advances in the usage of multiple I.1'PI

scales that established a measure that was accepted by others

in this research area. His instrument, the Health and Opinion

Survey (HOS) (1963), has been shown to have split-half reli-

ability of .94 and test-retest reliability of .83 (three


The R-S construct has thus become the generally accepted

term for the dimension of perceptual defense-perceptual vigi-

lance: Repression being the counterpart of perceptual de-

fense and sensitization being the counterpart of perceptual

vigilance. Repression represents a generalized marde of avoid-

ance to potentially threatening stimuli and sensitization

represents a generalized mode of approach to potentially

threatening stimuli. Inferences have been made in regards to

the characteristic defense mechanisms subserved by both ex-

tremes of the R-S dimension. It is hypothesized that the

repressor utilizes defense behaviors such as denial and avoid-

ance, while the sensitizer utilizes the defensive behaviors

of manifest worrying, obsession-compulsion, and intellectual-


Since the present investigation is concerned with the

relationship of R-S to features of interpersonal functioning

in the marital dyad, selected research in the areas of self-

perception, person perception, and personal adjustment in

group interaction as related to R-S will be reviewed.


Altrocchi et al. (1960) reported that R-S was positively

correlated with self-ratings for being rebellious, aggressive,

and self-effacing. This pattern of self-description resulted

in a general portrait of the sensitizer as one who manifested

a poorer self-image than was true for the repressor. In

fact, the related finding that the sensitizer had a greater

discrepancy between his self-image and his ideal-image than

did the repressor, was also accounted for by the sensitizer's

lower self-image. That is, both the sensitizer and the re-

pressor had a similar ideal-image, but the sensitizer's lower

self-image resulted in a greater self-ideal discrepancy.

In an investigation by Byrne, Barry, and Nelson (1963),

the previous finding of there being a greater self-ideal dis-

crepancy for sensitizers than repressors was again examined.

This time, however, the newly revised R-S scale developed by

Byrne (the Health and Opinion Survey, HOS) was utilized to

identify repressors and sensitizers. The previous findings

were replicated, in that the sensitizer revealed a greater

self-ideal discrepancy than the repressor. Also, as before

this discrepancy was shown to be a function of the sensi-

tizer's lower self-image. A second aspect of this study in-

volved testing the hypothesis that R-S would be related to

differences in the ways repressors and sensitizers handle

hostility. To assess this hypothesis, the Hostility Incon-

gruency Test (Byrne, 1961) was correlated with R-S. The

former measure represents a further development of an instru-

nent devised by IMcReynolds (1958). This instrument assesses

the extent to which one's feelings and values are incon-

sistent for specific areas of behavior. The procedure for

this latter assessment consisted of having the subject eval-

uate various statements, all involving some degree of hos-

tility, along three identified dim.n*nsions: Like-Dislike,

Good-Bad, and Pleasant-Unpleasant. Incongruency w'as defined

and measured by totaling items that were placed at con-

flicting poles across the three dimensions. For instance,

with the Like-Dislike and Good-B*ad dimensions, a score of in-

congruency was counted whenever a person rated a particular

behavior as occupying both the Like and Bad poles simulta-

neously, or whenever a behavior was rated at both the Dislike

and Good poles simultaneously. It was found that incongru-

ency was correlated significantly and positively with R-S

across all of the incongruency measures. And, more impor-

tantly, it vwas found that only one of the two possible sub

varieties of incongruency accounted for this correlation.

This is, the Like-Bad, Like-Unpleasant, and Bad-Pleasant sub

scores were significantly correlated with R-S, while the

Dislike-Good, Dislike-Pleasant, and Bad-Pleasant scores were

shonm to be uncorrelated with R-S. Thus, it was found by

this investigation that sensitizers tend to report conflicted

feelings and attitudes related to their liking and enjoying

hostile behavior that they also consider to be bad or morally

wrong, while this was not the case for repressors.

Altrocchi, Shrauger, and IcLeod (1964) employed two

separate measures of hostility, the Rosenweig Picture Frus-

tration Test and a self-rating scale, to further examine the

relationship of R-S and hostility. They obtained results

th.-t were consistent with those of the previous investigations.

That is, sensitizers revealed more hostility than did reprcs-

sort on both of these measures.

Byrne and Sheffield (1965) explored the hypothesis that

sc-nnitizers v:ould react with greater verbalized anxiety than

would repressors in a situation involving sexually arousing

stimuli. A factorial design was utilized wherein different

groups of repressors and sensitizers read either sexually

explicit passages or sexually neutral passages that were

taken from the same book. Self-ratings of arousal were then

obtained and significant differences were four to exist be-

tween the experimental groups. Significantly greater arousal

was found for both the repressor and sensitizer groups v:ho

read the sexually explicit passages. More interesting, how-

ever, was the difference between repressors and sensitizers

that emerged when a correlation was made between the scale

for sexual arousal and the remaining rating scales. It was

found that distinctly different patterns of feelings associ-

ated with sexual arousal emerged for the sensitizers and the

repressors. Whereas the sensitizers reported feelings of

(1) being entertained, (2) lack of boredom,and (3) being

anxious -- the repressors revealed feelings of (1) being

disgusted, and (2) being angry. These differences between

sensitizers and repressors in their self-rated feelings

associated with sexual arousal are open to several possible

interpretations. One interpretation that would appear to fit

especially well would involve the idea that the repressor

feels less anxiety in potentially threatening situations due

to directing his feelings outward, while the sensitizer feels

more anxiety due to directing of his feelings inward.

A study by Altrocchi (1961) provides information of how

the repressor's and the sensitizer's evaluation of their o.wn

characteristics compares to their evaluation of others. The

subjects in this study rated themselves and three classmates

on the Leary Interpersonal Check List. An analysis was made

of the difference between the subject's self-perception and

his perception of others. This analysis was expressed in

terms of a measure of the assumed dissimilarity between self

and others. It was found by this index that the sensitizer

assumes a significantly greater disparity between himself

and others than is true for the repressor.

The several previously described investigations reveal

consistent findings that the sensitizer evaluates both him-

self and others differently than does the repressor. It was

also established that the general direction of this difference

was in terms of the sensitizer manifesting a less positive

evaluation of himself than he has for others. Other investi-

gations will now be cited which attempt to ascertain whether

R-S might also be related to differences in the area of

personal adjustment.

Personal Adjustr.ent

Byrne, Golightly, and Sheffield (1965) tested the hy-

pothesis that R-S would be related to personal adjustFmenet in

a curvilinear fashion. This hypothesis was based on the

argument that "Neither obsessional concern with conflicts

nor selective forgetting of them should result in opti;ial

adjustment" (p. 586). Here, the "obsessional concern" with

conflicts would characterize the sensitizer, and the "selec-

tive forgetting" of conflicts would characterize the re-

pressor. 'he California Psychological Inventory (CI) \was

administered as the measure of adjustment, and each of the

eighteen CPI scales was correlated with R-S. It was shown

that significant correlations existed between R-S and seven

of the CPI scales. The direction of this correlation, with-

out exception, indicated that sensitizers were more malad-

justed than were repressors. The scores of persons falling

in the "Neutral" range of the R-S scale were also exn.ained

in this study. They were found to fall between those of the

repressors and the sensitizers. Thus, a significant linear

relationship, rather than a curvilinear relationship, was

found to exist between R-S and this measure of personal ad-


Thelen (1969) also investigated the relationship between

R-S and adjustment for college students who sought psycho-

therapy, compared to a control group who did not. He, too,

employed the CPI as the measure of personal adjustment.

There were tw.o important findings from this study. First,

those who sought psychotherapy had significantly higher

scores on the R-S scale than those :who did not. And, second,

when the therapy--neeking group and the non-therapy-seeking

group '..ere equated for adjustment, the therapy-seeking group

weroe found to be significantly higher in sensitization. In

regards to this latter finding, Thelen stated that "Perhaps

the R-S scale measure[-i sic] 'adjustment' as well as the tend-

ency to approach or avoid stress. Such a relationship does

not make the terms interchangeable and certainly does not

reduce the value of the R-S concept" (p. 164).

Byrne, Blaylock, and Goldberg (1966) tested the hypoth-

esis that the personality dynamics of the repressor, which

are charact'r-rized repressing and denying defense mechan-

isms, v'ould fit the personality pattern associated with dog-

matism. R-S was correlated with dogmatism, measured by

Rokeach's Dogmatism and Opinionation Scales, in two inde-

pendent samples. The relationship between dogmatism and R-S

was found to be both significant and positive in the inde-

pendent samples. Thus, dogmatism was found to relate more

strongly to sensitizing defenses than to repressing defenses.

The authors presented tw:o possible interpretations to explain

this unpredicted relationship. First, that while dogmatism

might serve as a defense against anxiety, the specific de-

fenses employed are sensitizing rather than repressing.

Second, that the close-minded or dogmatic person has basic

beliefs that man is alone, isolated, and helpless -- and that

this there is also reflected by the sensitizing person. The

authors conclude, "Thus, the dogmatic, sensitizing, person-

ally unhappy individual tends to express negative feelings

toward self and toward others" (p. 741).

Gayton and Bernstein (1969) investigated two issues

raised by R-S theory and previous R-S research. They also

assessed Byrne's hypothesis that R-S would be related in a

curvilinear fashion to personal adjustment. They investi-

gated specific areas of personal conflict via Trehub's ego-

disjunction measure of incompatible needs. Briefly, Trehub's

measure is an index arrived from responses to the Edwards

Personal Preference Schedule (EPPS). Eight of the EPPS's

scales are grouped into four pairs of conflicting needs,

"deference-aggression, autonomy-abasement, succorance-nur-

turance, and order-change" (p. 192). The joint magnitude of

each of these pairs is used as a measure of incompatible need

strength. The results of this investigation supported the

previously established linear relationship between R-S and

personal adjustment. Additionally, it identified only two

of the four need-pairs as contributors to this incompatibility

in need strength: succorance-nurturance and autonomy-abase-

ment. The remaining need-pairs showed no such need discrep-

ancy. Thus, the need areas of succorance-nurturance and

autonomy-abasement might be tentatively identified as areas

of particular conflict for those at the sensitizing end of

the R-S dimension.


In an attem-t to assess the extent that R-S might relate

to a differential incidence of physical illness, Byrne,

Steinberg, and Schwartz (1963) examined both a self-rated

incidence of illness and an incidence of illness based upon

visits made to a student health center. The prediction was

made that both repressors and sensitizers would have a

greater incidence of illness than would neutrals. Instead,

a linear relationship was found between R-S and both measures,

at least for the male student sample. Thus, in this study

both the self-assessment and the behavioral assessment of

physical illness were found to be related linearly to R-S.

Up to this point I have reviewed several investigations

that dealt with the relationship of R-S to measures of (1)

self-perception,(2) person perception:and (3) personal ad-

justment. Now, this review will focus upon studies that are

essentially concerned with R-S as related to group process

and group outcome variables.

Ground Process

One of the earliest investigations that attempted to

study the relationship between R-S and group functioning w'as

by Joy (1963). Subsequent to establishing that R-S corre-

lated -.87 with leadership (as measured by an 1.1I.PI scale)

Joy arranged experimental problem-solving groups. These

groups were each composed of three members: a repressor, a

sensitizer, and a neutral. Each group was assigned a human

relations problem as a group discussion task. Following this

group interaction, each of the group members rated himself

and the other two members on a number of variables. There

were two important findings. First, on the ratings of

others, the sensitizers were significantly loss often chosen

to be work partners. Second, the repressors rated themselves

to be more concerned for maintaining friendly group relations

than did the sensitizers.

Perhaps the first study that related R-S to a short-

term, non-experimentally created, group was that by Turk

(1963). In this investigation he examined the congruence of

self-rating and rating of others in student nurse and student

physician teams. The student nurses were the only dyad mem-

bers that were identified on the R-S dimension. These phy-

sician-nurse teams worked together in a clinic for a period

of three weeks. The dependent variable under study was the

perceived "enjoyment" by each of the dyad members during this

work experience. Each of the subjects rated himself and the

other dyad member on a multi-faceted questionnaire. A sig-

nificant correlation was found to exist between R-S and the

rating of other for task enjoyment, only when the nurse being

rated was a sensitizer. That is, the self-rating of enjoy-

ment for repressor nursing students was not accurately per-

ceived by the doctors with whom they worked, while it was so

for sensitizer nursing students. A second important finding

was the existence of a significantly greater assumed simi-

larity of other rating for the repressor nurse group than for


the sensitizer group. That is, the sensitizer nurse assumed

a greater difference between herself and her dyad partner

than did the repressor nurse. Parsons and Fulgenzi (1968)

followed up early findings in R-S research to examine the

developing conclusion that "hostility is of special impor-

tance in this personality dimension" (p. 537). Subjects in

this study were identified via the HOS into three groups:

repressors, sensitizers and neutrals. Two separate proce-

dures were employed to determine if these groups differed in

their expression of hostility. The first procedure involved

the administration of the group Rorschach. The second pro-

cedure involved the formation of five-person groups of heter-

ogeneous R-S composition that subsequently engaged in a group

story construction task. During this latter task, experi-

enced judges rated each group member's interpersonal behavior

on several behavior rating scales. Finally, each of tie group

members rated all other group members on a twenty-five-item

scale. The findings showed no significant differences among

the three R-S groups on the hostility score derived from

their Rorschach protocal. However, on the behavior ratings

by the judges, repressors were assessed to be significantly

more aggressive and hostile than were sensitizers. Supporting

this finding were the ratings from each of the group members

as they too consistently rated the repressor as more aggres-

sive and hostile than the sensitizer.

Parsons, Fulgenzi, and Edelburg (1969) report a study

which investigates this problem area by comparing the self-

assessment, behavioral-assessment, and psychophysiological

assessment of a specific interpersonal interaction. Again,

they utilized five-person discussion groups. The group ma-

jority was structured so that an equal number of groups

would have repressors or sensitizers as predominate. Two

independent investigations were then carried out. The first

was essentially a replication of the previously reported

experiment, Parsons and Fulgenzi (1968). The results of this

part of the rudy were also in agreement with the earlier

study; i.e. (1) repressors rated themselves to be less ag-

gressive than sensitizers rated themselves,and (2) there was

greater discrepancy between the judges' rating of aggression

and the self-ratings of aggression for the repressors. The

authors concluded this part of the study with the observation

that these repeatedly established findings "provide an experi-

mental analogue to the psychotheropeutic encounter (p. 239).

That is, like some therapy clients, the repressor reveals

behavioral signs of hostility and aggressiveness but does not

report a parallel state of emotional arousal, while the sensi-

tizer, who does not reveal such behavioral assertiveness,

reports that he is in fact quite emotionally aroused.

This repeatedly observed discrepancy between overt

behavior and self-assessed emotional experience, resulted

in the subsequent study that made a psychophysiological meas-

ure of the emotional state of repressors and sensitizers.

Five-person groups, with either repressors or sensitizers in


majority, engaged in a half-hour discussion while (1) their

skin conductance responses (SCRs) were recorded, (2) their

interaction was rated for aggressiveness of behavior, and (3)

their verbal behavior was recorded. The results show the

repressors to have a higher level of aggressive behavior,

that ',was in turn accompanied by a higher level of SCHs, while

the sensitizers had a lower level of a'ggressivener.s that

was accompanied by a lower level of SCRs. The authors follow

by stating,

Repressors, then, appear to be highly involved,
both by behavioral and psychophysiological cri-
teria, in the group discussion, goal oriented
(getting on with the task), and perceived by
others as aggressive but not reporting themselves
as aggressive. The sensitizers are less aggres-
sive and not as effectively aroused at least as
measured by the SCR. On The other hand, they
rate themselves higher than repressors on aggres-
siveness and hostility but are not so rated by others.
(p. 242)

When these empirical relationships were evaluated in the terns

of Lacey's (1959) "transactional" interpretation of autonomic

responses, the inference was presented that repressors were

more task and goal oriented, while the sensitizers were

more oriented to the emotional quality of the interpersonal


A study by Wilkins and Epting (1971) provides findings

that relate to the issue of differential interpersonal ori-

entation between sensitizers and repressors. They employed

Bieri's Interpersonal Cognitive Complexity (ICC) measure to

differentiate the nature of interpersonal orientation related

to R-S. According to the authors,

Interpersonal Cognitive Complexity is defined in
terms of the degree of differentiation bctwoen the
dimensions of a construct system. A construct
system \which highly differentiates persons
in the social environment is considered to be
cognitively complex in structure. On the other
hand, a construct system which poorly differ-
entiates among persons is considered to be cogni-
tively simple. (p. 1)

The authors predicted that the sensitizer, with obsessive

traits, would manifest interpersonal complexity, and that the

repressor, with traits of avoidance, would manifest inter-

personal simplicity. The results supported these predictions.

These findings were interpreted to reveal that repressors are

less discriminating in their interpersonal orientation than

are sensitizers. These findings also support the inference

from the previous study (Parsons et al.) -*- that sensitizers

are particularly attuned to the interpersonal quality of

social events, while repressors are not.

Cohen and Foerst (1968) present evidence that supports

the second inference from Parsons et al. that repressors are

more goal directed in problem-solving situations. It has

previously been shown that some personality variables are

related to differences in group performance. Thus, it is

appropriate to assess whether R-S too has such a relationship.

In this study, experimental groups were homogeneously com-

posed of either five sensitizers or five repressors. They

were situated in a structured communication network and

asked to participate in group problem-solving tasks. The

iiajn findings were summarized by the authors as follows,

"R groups formed appropriate problem-solving systems earlier

than S groups, had faster times in performing tasks, and

exhibited greater continuity of leadership (p. 214).

These results seemed consistent with the inference by Parsons

et al. that the repressor may, be more goal oriented and less

vulnerable than the sensitizer to interpersonal distraction.

Two studies by Cohen and Carrera (196'?) and Carrera and

Cohen (1968) also relate R-S to aspects of a problem-solving

interaction. They utilized five-person experimental groups

homogeneously composed of repressors or sensitizers. Both

investigations involved group interaction that led to experi-

ences of group success or group failure. Both studies had

similar procedures and results, thus the second study might

be essentially thought of as a replication of the first.

Although there was a tendency, for "extreme" sensitizers to

verbalize more hostility subsequent to a failure experience,

there were no significant differences in affect, nor in judg-

ments concerning their performance, for either repressors or

sensitizers subsequent to group failure or group success.

The general conclusions inferred from this absence of meas-

ured difference in verbal behavior and in evaluative judgments

were as follows:

It was suggested that group factors may consti-
tute a set of mitigating conditions that inter-
vene between the induction of stimuli and their
effects on the manifest productions of personality
(p. 221),


In both studies the repression-sennitization
variable proved to be of relatively minor impor-
tance for interpersonal behavior within the con-
text of small groups. (p. 13)

In contrast to studies that related individual response

tendencies to R-S, the studies by Carrera and Cohen indicate

crucial limitations in the value of R-S as a relevant vari-

able in small group interaction. The following section will

discuss this crucial implication as related to the present

study as well as several general findings presented in this

selected review of R-S research.

Implications for Present Research

Several conclusions and implications can be distin-

guished in the previous R-S investigations. First, and most

comprehensively, the R-S construct has shown both theoreti-

cally and empnirically congruent relationships with several

other psychological constructs and measures. Second, the

R-S instrument, the Health and Opinion Survey, has evidenced

good reliability and construct validity. Third, the R-S

concept has been found useful in delineating features of

individual response tendency and in delineating some features

of small group interaction.

In addition to these very general positive features of

R-S research, several limitations and areas of neglect can

also be identified. First, because of the very nature of the

R-S variable, self-rated assessments need to be related to

more objective modes assessing the same behavior. Second,

few R-S studies have directed attention to the behavior of

the intermediate scorers (Neutrals) as related to extreme

scorers (Repressors and Sensitizers). Third, while there

have been only a fe' efforts to study the relationship of

R-S to aspects of small group or dyadic process, the infer-

ence has been reported that the R-S variable is of little

significance in group interaction.

These general findings and limitations in R-S research

provide grist for the present investigation. The observa-

tion is presented that previous studies of small group inter-

action have been based almost exclusively on artificially

organized "experimental" groups. These groups did not have

the characteristics of being (1) naturally found,(2) self-

selected, nor (3) were they of a long-term nature. It is

argued that if R-S were a valuable dimension in group pro-

cess and outcome, that the study of natural groups, embodying

the just enumerated characteristics, would be very desirable.

As stated by Byrne,

Any pervasive personality variable such as
repression-sensitization, is potentially an
important determiner of some aspects of inter-
personal behavior. An individual's socially
relevant motives, his perceptions of others,
his response to the demands of group situations,
and his effect on others are likely to be in
part a function of his characteristic defense
modes. (1964, p. 203)

The formulation of defense modes has been of long-standing

interest and value in personality theory. However, there

seems to have been little attempt to measure and study these

defense modes as they relate to aspects of interpersonal


relationships. The general purpose of the present study is

thus to investigate various R-S findings and inferences as

related to aspects of a particular interpersonal relation-

ship the marital dyad.

Study of R-S in the Marital Relationship

The current study utilized the marital dyad in the

examination of several factors relevant to R-S; those factors

being (1) marital adjustment, (2) the perception of satis-

faction and stress in marriage in a college setting, (3)

the resolution of marital conflicts, (4) self-perception,

and (5) spouse-perception. The marital dyad was chosen as

the relationship of focus for several critical reasons

which will not be discussed. Previous R-S research has been

concerned, among other variables, with the relationship of

R-S to the perception of aggression in self and other group

members and with the effectiveness of group performance.

The groups utilized in these studies were experimentally

created, and as such they differ considerably from naturally

formed groups. It is argued that if individual differences,

measured in terms of R-S, do reveal differential effects in

these experimental groups, then such differences should be

even more manifest in long-term naturally formed groups.

Also, it seems evident that the resolution of conflict is a

feature of all but perhaps the most casual or formalized

relationship, and that probably few relationships could

involve the extent of conflict resolution that would occur in

the marital dyad. Thus, the marital dyad, where an indi-

vidual functions intensively and extensively in a self-

chosen relationship, appears to provide a most excellent

natural group for the investigation of hypotheses drawn from

R-S research.

It has been argued that the marital dyad represents a

particularly good setting for the investigation of issues

drawn from the R-S literature. The following section pre-

sents the reasoning involved in the several areas that will

later be delineated by specific hypotheses.

It has been reported (Byrne, 1964) that social desira-

bility scores are highly correlated with R-S scores. This

relationship is one in which those with repressor scores also

have high social desirability scores, while those with sensi-

tizer scores have very low social desirability scores. The

picture is presented that the repressor, v:ith a high social

desirability set, is one who would like to think of himself

in very positive terms while the sensitizer operates with

the very opposite orientation. It would then be expected

that the repressor would paint a very positive picture of

himself and his marital relationship on a self-rated instru-

ment while the sensitizer would be expected to do the


Earlier studies have also shown that the repressor rated

himself low in aggressive behavior in a group interaction,

but that he is rated by both co-acting peers and observing

judges as high in aggressive behavior. The sensitizer

presents quite the opposite picture -- he rates himself to

be high in aggressive behavior, but he is rated by peers and

judges as manifesting little aggression. An important fact

here is that comparable ratings have been given by both the

interacting peers and the judges. In a marital conflict-

resolution situation it would be expected that the greater

aggressiveness of the repressor would be revealed by his

judgment being accepted more frequently than the sensitizer's

judgment. In this situation the sensitizer would also be

expected to rate himself as being more aggressive than he is

perceived to be by his mate, and the repressor would be

expected to rate himself as less aggressive than he is per-

ceived to be by his nate.

In summary up to this point, the R-S construct refers

to a dimension of individual differences in reactivity to

emotionally arousing stimuli especially those of an aversive

nature. There are two most central features of the poles of

this dimension. The first is characterized as a bias to

operate toward (repressor), or away from (sensitizer), so-

cially desirable responses, while the second is character-

ized as the lack of correspondence between how the repressor

and sensitizer perceive aspects of their behavior and how

they are perceived by others. It is argued that the exami-

nation of responses made by neutrals (those who have inter-

mediate socres on the R-S scale) may provide a new perspective


when related to these and to other aspects of the R-S picture.

The neutral can be viewed as one who does not operate with

the degree of positive or negative response bias as the sensi-

tizer or repressor. Consequently the ratings made by the

neutral, of his marital relationship, should be less dis-

torted than those made by the repressor and the sensitizer.

Therefore, the most objective evaluation of the quality of

a marriage between a repressor and a neutral, or between a

sensitizer and a neutral, would be expected to be that made

by the neutral spouse.

The marital ratings received from the neutral should

thus establish a clear-cut test of Byrne's cornerstone

assumption that R-S is related curvilinearly to adjustment

and well-being. That is, the excessive avoidance or approach

modes of defensive behavior of the repressor or sensitizer

spouse is predicted to limit the adjustment and satisfaction

partners in the marital dyad could achieve. Therefore,

neutrals would be expected to report more satisfaction in a

marriage with another neutral than in a marriage with either

a repressor or a sensitizer spouse.

Also, from previous R-S research, it was shown that the

perception of enjoyment in a dyadic working relationship is

much more accurate when the perceptual target is a sensitizer

than when the target is a repressor. It would be expected

that the sensitizer spouse will be more congruently evaluated

by his mate, than will the repressor spouse, on a measure of

environmental stress and satisfaction.


Finally, the inference reported by Carrera and Cohen,

that R-S is a variable of relatively minor importance in

small group interaction, could be evaluated on the basis of

whether there are significant relationships established

between R-S and the several dependent measures in the present





The program of this experimental research began with

the identification of marital dyad partners in terms of

their R-S scores on the Health and Opinion Survey (HOS).

Sixty couples who had been married for at least one year, and

in which at least one member was a University of Florida stu-

dent, were needed for this study. Prospective couples were

personally contacted and asked to respond to the HOS. This

procedure was followed until the six experimental cells,

presented below, were represented by ten couples each. The

experimental cells consisted of dyads with all possible com-

binations of repressor (R), sensitizer (S), and neutral (N)

mates. According to the most common practice in R-S research,

subjects were identified as repressors if their score was

below 39, as sensitizers if it was above 55, and as neutrals

if their score fell between 39 and 55. Thus the following

cells were represented:

Mate 1 Mate 2


The Health and Opinion Survey (Byrne, 1963) was used to

identify sensitizers, repressors, and neutrals. This instru-

ment is composed of 127 out of 182 possible items from six

Minnesota Multiphasic Personality Inventory scales (D,L,K,

Hy, Pt, and Welsch Anxiety). The deletion of some of these

182 items was initiated by Byrne to eliminate the item over-

lap of earlier R-S scales. This revised R-S instrument, HOS,

has shown a Brown-Spearmen corrected split-half reliability of

.94 and a three month test-retest correlation of .82 (Byrne,


The Inventory of IMarital Conflict (I.'iC) by Olson

(1969) is comprised of 18 vignettes designed to induce con-

flict in problem resolution between the husband and wife on

12/18 stories, and to induce agreement in spousal judgments

on the remaining 6/18. Each spouse is asked to individually

read the vignettes, which describe relatively common conflict

experiences that a couple may have, and assign a judgment of

responsibility for the described problem. After the spouses

individually assess problem responsibility, they must jointly

reach a single judgment as a couple. Twelve of the eighteen

vignettes are written such that spouses are led to believe

that different story characters are in the wrong. Six of

the 18 vignettes are written so that both spouses will

judge the same story character as being in the wrong.

A "Win" index was derived based upon which spouse's

opinion was accepted during the joint resolution of the 12/18

conflict vignettes. Winning was determined by whose decision

was jointly endorsed when the husband and wife differed in

their individual judgments. The win index was computed by

awarding each spouse one point when his individual judgment

was later endorsed by his mate in the joint resolution. This

sum was then divided by the total number of stories that

they differed upon. It is inferred froir this index which

spouse is the more aggressive and ,which thie more submissive

in the conflict resolution interaction.

Further information will now be presented regarding

aspects of the construction and validation of the IiCC. The

conflict situation in each vignette is presented with slight

differences in slant regarding the information that each

spouse of the experimental dyad receives. Thus a delicate

balance must be attained so that a difference in the assign-

ment of responsibility for the described conflict situation

could be achieved without having the couple realize that

they had in fact been misled. The judgment that this desired

balance has been achieved can be substantiated by data pro-

vided by the IMC author (Olson, 1969). For example, on the

story situations designed to produce a conflict in judg-

ments, 86 percent of the stories designed to produce con-

flicting judgments did so. In general, the II.IC represents

a significant improvement in assessing interaction in a

conflict resolution situation over its more widely known

predecessor Strodtbeck's Revealed Difference Technique


Behavior RPatinc Scales. Following the 1-IC interaction,

each spouse filled out an eight scale assessment of inter-

personal boeiavior. They rated both themselves and their

spouse on each of the scales. These 8 scales represent

categories defined by Bales (1950) in the Interaction Process

Analysis, and modified by Parsons et al. (1969). Five of the

scales assess aggressive behavior (scales 4-8), and three

assess non-aggrrcssive, but interaction-relevant behavior

(scales 1-3).

The Locke-'.allace "Short maritalal Adjustment Test" (1959)

is a fifteen-item inventory composed of selected items from

marital inventories which proved to have the highest dis-

criminative value and which did not seem to overlap other

items in content. The authors have presented evidence that

their shortened inventory has essentially the same reliability

as the longer form. The range of scores possible to achieve

is from 2 to 158 points.

The .i-arriage and College Environment Inventory (LIACE)

by Clarke (1969) is a recently devised instrument for married

couples wherein either one or both of the mates are college

students. It is designed to assess the extent to which various

personal and interpersonal aspects of the marriage experience

and features of the college environment are viewed as sources


of stress or satisfaction. The I:ACE is composed of 175 items,

has both a husband and wife form, and is designed for the

spouses to complete alone without consultation with each

other. There are four possible response categories available

for each of the 175 items: (1) Non-applicable, (2) Generally

satisfactory, (3) Generally stressful, and (4) Neither.

The i'iACE was administered on two separate occasions to

each subject. The first administration requested each sub-

ject to respond to the MACE statements on the basis of his

own feelings. A proportion score was generated from these

responses by dividing the number of items that he rated as

"Generally stressful" by the number of items that he rated as

"Generally satisfactory." This proportion score was termed

the IMACE stress/satisfaction ratio. The higher this ratio

is, the greater the amount of stress the subject is experi-


On the second administration of the iiACE, the subject

was requested to respond on the basis of how he perceived his

spouse to feel. A second score was thus derived by comparing

each subject's own responses to the MACE with the responses

ascribed to him by his spouse. This resulting score was

termed the MACE agreement score. This score was arithmetically

computed by summing one point for each L.LACE statement that

was responded to similarly by the subject's self-rating and

by his spouse's rating of him. This sum was then divided by

the total number of responses that fell in the "Generally

stressful" and "Generally satisfactory" categories for that

person. The MACE agreement score could thus serve as a

statistic for comparing the degree of congruency between a

person's self-rated feelings and his spouse's perception of

his feelings.


The 60 couples, 10 couples in each cell, who participated

in this study, were selected from names provided by the

Registrar at the University of Florida. This list included

all persons who were then currently enrolled married students.

Couples were randomly selected from this list and asked to

participate in a research project. This project was described

as a study of aspects of college married life.

Initial contact with each prospective subject couple

was made by the investigator at the subjects' home. It was

felt that such personal contact would maximally encourage a

positive attitude toward participation in the study. If the

couple were willing to participate in the investigation, a

date and time were agreed upon for the investigator to return

with the initial questionnaires, i. e. HOS, MACE, and Locke-

Wallace. To insure non-collaborative responses on these

measures, the couples were asked to fill out the question-

naires with the investigator present. At the end of this

first phase of testing, the couples were informed that they

might be contacted in the near future for the purposes of

an unrelated investigation that would provide a nominal re-

muneration ($4.00). This reference to a future investigation

was actually an attempt to promote participation in the second

phase of the study.


Coupleos who were identified as having the desired co.-

binatiLcn of R-S scores were then contacted by telephone to

particirate- in the second phase of data collection, i. e.

If;C, L:1-s' rating scales, and ,MACE. This final phase of

da.a co.e].';.ion was again carried out in the subjects' home,

while .t.o investigator was present. The subjects were first

girver.- thL- 13 v*inettes of the IMC and asked individually to

assj-r. -r3sor'.ioiility for the described problem situations.

Upon cc:.eri ion cf this task, the coupe was brought to-

gether_ in order to make a mutually endorsed assignment of

res pon~;i: ility for the same 18 vignet-es. They were next

a2::ed t: rate themccc.lve and their mate on the 8 Eales'

scales i.n regards to their behavior during conflict resolution.

L?.tly, thcy :~a~-.n responded to t.e V[J.CE; however, this time

they responded the ;.'ay that they thought their spouse would


A. a cor;:ente on the apparent effectiveness of these pro-

cedur-- to enlist cooperation, it can be noted that it only

took contact \:.ith 62 coul].es, having the desired combination

of R- -cores, to fill the 6 experimental cells of 10 couples

each. The total number of couples who were tested in order

to identify those having the desired R-S combinations was 197.

The statistics appropriate for testing the hypotheses

of this study were derived from analyses of variance with

planned comparisons of cell means. Further statistics were

computed with multiple compariscns made among the R-S dyads