THE RELATIONSHIP AMONG
STATUS CHARACTERISTICS,
PERFORMANCE EXPECTATIONS.
PROCEDURAL EXPECTATIONS.
AND VIOLATIONS OF
PROCEDURAL EXPECTATIONS .

Dissertation far the Degree of Ph. D.
MICHIGAN STATE UNIVERSITY
LYNN M. BRODY
1975

 
   
   

  

LIBRARY

Michigm State
University

0"" 5w

  

f“

This is to certify that the

thesis entitled

THE RELATIONSHIP AMONG STATUS CHARACTERISTICS,
PERFORMANCE EXPECTATIONS, PROCEDURAL
EXPECTATIONS. AND VIOLATIONS OF
PROCEDURAL EXPECTATIONS

presented by

Lynn M. Brody

has been accepted towards fulfillment
of the requirements for

ED.D.. _.__degree in Jhilnsnphy

Eféfm

Major professor

DateWZ 5"

0-7639

 
  

 
  

some “333'- \
aoox smom me.

LIBRARY SINGERS
SPRIIGPORT. moment, 1

’2‘ 5:" ‘/V\
[- L" t
. \ f
15‘ / ‘4
LI" x‘ I
‘A\. 1' "

\ " ABSTRACT
I

L,“

THE RELATIONSHIP AMONG STATUS CHARACTERISTICS,
PERFORMANCE EXPECTATIONS, PROCEDURAL
EXPECTATIONS, AND VIOLATIONS OF
PROCEDURAL EXPECTATIONS

By

Lynn M. Brody

The work described here is heavily influenced by the
Expectation Theory of Joseph Berger and associates and by Harold
Garfinkel's and Peter McHugh's ethnomethodological research on
expectation violations.

The influence of the Berger group (1966) revolves around
their conceptions of performance expectations, external statu§_

characteristic, status characteristic, and diffuse status characters

 

jstjg, In a task situation where person (P) and other (0) are com-
mitted to "successful" completion of the task, it is assumed that the
actors will feel pressured to determine their relative abilities at
the task. These beliefs concerning relative task ability are labeled

performance expectations. If P and 0 differ in an external status

 

characteristic (such as profession or sex), and if these criteria are
not directly relevant to the task, but if there are no other available
criteria for determining their differential expertise at the task,
then beliefs associated with the status characteristic will become the

basis for assignment of performance expectations.

Lynn M. Brody

A status characteristic stands for an evaluation by an
actor that he or she is "better" or "worse" than another actor. How-
ever, there are apparently two kinds of evaluations: specific and
general or diffuse. First, judgments of specific abilities are con-
cerned with evaluations that are made with respect to specific traits
associated with the status characteristic; for example, professors
are considered good writers. Second, these specific evaluations be-
come generalized or diffuse; for example, a professor is considered
not only a good writer (a specific evaluation), but also to be capable
generally. The specific evaluations that become generalized or diffuse
represent the abstract definition of a diffuse status characteristic.
Because it is assumed that status characteristics determine the
assignment of performance expectations, and because it is assumed that
performance expectations determine the observable power and prestige
order of the group, the initial status differences between P and 0
should come to correspond to the differences in power and prestige
between them.

In addition to extending the theory of Berger and associates,
this work investigates a new aspect, namely, assumptions about vio-

lations of procedural or normative expectations. The latter are

 

defined as the beliefs that P and 0 have for self and each other in
terms of their following the task rules. The theory posits that
performance and normative expectations are related, and are related
in degree: The higher the performance expectations that P holds for

0, the higher the normative expectations that P holds for 0.

Lynn M. Brody

The theory also predicts that when a high status 0 breaks
the normative order, a lower status P will respond later, register-
ing fewer occurrences, and less positively to the violation than will
a P of higher status interacting under the same conditions with a
lower status 0.

The other major influence on this work derives primarily
from the research of Garfinkel and McHugh on the topic of violations
of procedural expectations. However, their efforts lack precision
of conceptualization and testable derivations, and an attempt to
avoid these shortcomings was made here.

In order to test the derivations of the theory, a laboratory
test was designed and executed. The two independent variables were
the high or low performance expectations that the subjects held for
their opponent, and the procedural expectations that were violated.
The two dependent variables were the subjects' responses to violations,
and their feelings toward their opponent. Fifty subjects were run in
the experiment, 27 assigned randomly to one condition and 23 assigned
randomly to the other. Ten subjects were excluded from the analysis
since they did not meet the scope conditions of the theory, leaving
an n of 20 in each condition. At the beginning of the experiment,
subjects were introduced either to a high (H-L) or low (L-H) status
confederate. Their task was to play a word game with their opponent
which included seven critical trials. These latter were defined as
those in which a procedural violation occurred. The proportion of

times the subjects in each condition detected and/or scored a violation

Lynn M. Brody

was considered a measure of their acceptance or rejection of the
other's performance outputs.

After the experiment, subjects were administered an
adjective test. The proportion of times that subjects responded
favorably or negatively toward the opponent, in each condition,
was considered a measure of their feelings toward their opponent.

The data demonstrate that the derived predictions tend
in the directions expected. Subjects in the L-H condition responded
to violations later during the sequence of trials and fewer times,
and they felt less positively toward the other than did subjects in
the H-L condition.

In addition to the data relating to the major tests of
the theory, also discussed are serendipitous findings, interpreta-
tions of some unexpected data, evaluation of the experimental
setting, and suggestions for future changes in the experimental
design and extensions of the theory.

The data appear to support the theoretical assertions
made here. In general, the derived predictions basically were
confirmed in terms of the expected differences between conditions.
The following conclusions seem justified:

(l) The differential performance and procedural expectations
that P holds for P' (self) and 0 seem to produce an
observable behaVioral difference; and

(2) the differences between the conditions in terms of re-
actions to violations of normative expectations are

substantial and in the predicted direction.

THE RELATIONSHIP AMONG STATUS CHARACTERISTICS,
PERFORMANCE EXPECTATIONS, PROCEDURAL
EXPECTATIONS, AND VIOLATIONS OF
PROCEDURAL EXPECTATIONS

BY

A to
Lynn M: Brody

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Sociology

1975

Copyright by
Lynn M. Brody
1975

ACKNOWLEDGMENTS

I first would like to extend my appreciation to my
committee. Although I had chosen this topic before Dr. Bo
Anderson became my thesis director, my conceptualizations were
not precise. He offered many insights into the theoretical
formulation and the experimental design. He also provided
thorough critiques of my written material. Dr. Santo F. Camilleri
directed much of my graduate training. He was especially helpful
in assisting me to obtain the formal structure of my theory and
with the data analysis. Dr. Hans Lee provided both task and a
tremendous amount of socio-emotional support. His excitement
concerning my experimental design was an inspiration. Dr. Barrie
Thorne, both a committee member and dear friend, offered constant
support.

The Department of Sociology, Michigan State University,
merits my appreciation for providing me with a stimulating and
enjoyable learning process and the opportunity to gain teaching
experience. My thanks also to the National Institute of Mental
Health for a traineeship in Formal Theory that was awarded me.

My main research assistants, David Williams and Linda
McMillan, warrant my gratitude and respect for their excellent
work. I also am grateful to all those anonymous subjects who made
the research possible.

ii

Dr. James McKee deserves my appreciation for simply being
who he is, and my very special thanks to to Dr. Thomas Ely Lasswell
and Professor Marcia Lasswell. The former's inspirational teaching
prompted me to enter the field of sociology, and both have extended
their faith, support, and highly valued friendship over the years.

To Marilyn Lovall and Betty Barron for their contribu-
tions, to Margaret Cook for her excellent assistance in typing
drafts of this thesis, and to Elayne Ballance for typing the final
draft, my heartfelt thanks. My dear friend, Elizabeth Johnston,
provided constant faith in this project and its completion, even
when I had my doubts, and valuable editorial assistance.

My parents deserve much appreciation. Their attitudes
toward education and science and lack of prejudice toward other
people contributed tremendously toward my choice of profession.

Dr. Howard Kabakow, who held the same attitudes, also was a signi-
ficant influence.

Last, but most important, Nicholas B. Waite, my dearest
friend, made the completion of this project possible. Without his

support and sacrifices, the task would have been much more difficult.

iii

TABLE OF CONTENTS

LIST OF TABLES ........................
LIST OF FIGURES .......................

INTRODUCTION TO THE PROBLEM AND OVERVIEW
OF CHAPTER ......................

INTRODUCTION ....................
PLAN OF STUDY ...................

ETHNOMETHODOLOGICAL PERSPECTIVE, EXPECTATION
THEORY, AND SELECTED LITERATURE ON LATITUDE
ALLOWED TO HIGH STATUS DEVIANTS ............

THE ETHNOMETHODOLOGICAL DISCOURSE .........

The Garfinkel Perspective ...........
Ethnomethodological Research ..........
Limitations of the Ethnomethodological

Perspective .................
General Theoretical and Methodological

Weaknesses ..................
Specific Design Problems ............
Summary .....................

EXPECTATION THEORY .................

Contributions of Expectation Theory and
the Ethnomethodological Discourse to
the Present Work ...............

SELECTED LITERATURE ON HIGH STATUS DEVIANCE . . . .
SUMMARY ......................

A THEORY OF REACTIONS TO VIOLATIONS OF
PROCEDURAL EXPECTATIONS ................

INTRODUCTION ....................
SCOPE CONDITIONS AND TERMS OF THE THEORY ......

iv

Page

\JU'IU'IUT

IO

l3
l3

l7
l9

23
23
24

Chapter

EXTENSION OF EXPECTATION THEORY ..........
ASSUMPTIONS AND DEFINITIONS ............
SUMMARY ......................

EMPIRICAL TEST OF DERIVATIONS FROM THE THEORY . . .
OPERATIONALIZATION OF VARIABLES ..........
REJECTED ALTERNATIVES ...............

Measures of Perceptions and Reactions
to Violations ................
Timing Devices .................

CHOSEN EXPERIMENTAL SITUATION ...........
PHYSICAL SETTING .................
SAMPLE ......................
SEQUENCE OF THE EXPERIMENT ............

The Word Cognitive Ability Task ........
Posttrials Phase ................

ANALYSIS OF DATA ...................
INCLUSION OF DATA .................

General Criteria for Exclusion .........
Specific Cases Excluded ............

RESULTS ......................

Responses to Violations of Expectations . . .
Benefit of the Doubt ..............
Feelings toward the Confederate ........
Performance Expectations ............

DISCUSSION AND INTERPRETATION OF THE RESULTS

Responses to Violations of Procedural
Expectations: Critical Trial Data .....
Responses to Violations of Procedural
Expectations: The Postexperimental
Interview ..................
Benefit of the Doubt ..............
Feelings toward the Confederate ........
Performance Expectations ............

Page

27
28
37

39
39
4O

74
75

78

Chapter Page

VI. SUMMARY, CONCLUSIONS, AND SUGGESTIONS

FOR FURTHER RESEARCH ................. 8l
SUMMARY ...................... Bl
GENERAL CONCLUSIONS ................ 84
Responses to Violations of
Procedural Expectations ........... 84
Recommendations for Revisions
of Future Research ............. 85
Recommendations for Future Extensions
of the Theory ................ 87
APPENDICES
A. Complete Experimental Procedures ........... 90
B. Procedure for Recruitment of Subjects ........ lO3
LIST OF REFERENCES ..................... lO7

vi

LIST OF TABLES

Table Page
5.l Total Frequencies of Violations Detected
on All Critical Trials, by Condition ....... 64
5.2 Frequency of Detected Violations, by Condition . . . 64

5.3 Percentage of Violations Detected on
Critical Trials, by Condition .......... 66

5.4 Proportion of Total Noticed Violations
to Benefit of the Doubt ............. 67

5.5 Proportion of Subjects with Positive, Neutral,
and Negative Feelings toward Opponent ...... 68

vii

LIST OF FIGURES

Figure Page
4.l Activity Room .................... 48
4.2 Observation Room .................. 50

viii

CHAPTER I
INTRODUCTION TO THE PROBLEM
AND OVERVIEW OF CHAPTER
INTRODUCTION

In many relationships between people there will be a
system of expectations defining what constitutes appropriate and/or
inappropriate behavior for each person. Behavior most likely is
defined by the interactants as either conforming to or deviating
from these expectations. This study attempts to formulate a theory
and test empirically some of the derivations from the theory con-
cerning reactions to deviations from expectations. More specifi-
cally, the theory centers on reactions to violations of procedural
expectations.

Procedural expectations are beliefs that an actor holds
for another actor concerning "proper" behavior in a situation. If,
for example, two people are playing a game, neither will expect the
other nor himself to "cheat." However, if "cheating" or some viola-
tion of a procedural expectation does occur, the crucial questions
are: Given a violation of a procedural expectation, what reactions
will be engendered, and will the reactions be dependent upon specific
characteristics of the relationship between the actors? With respect
to the latter, one may ask whether the relative status of the actors

is related to their reactions to violations of procedural expectations.1

 

1The activation or situational dependence of expectations
being defined as appropriate or inappropriate would be an interest-
ing theoretical issue to pursue in the future.

I

The basis of an answer to these questions can be developed
from two relatively recent lines of thought. The first of these is
the Ethnomethodological School, in particular the work of Harold
Garfinkel and Peter McHugh on violations of procedural or constitu-
tive expectations. Their basic premise is that there are procedural
rules which define proper situational behavior; although universal
and unstated, these rules are recognized by each person, and he
expects himself and those with whom he interacts to abide by them.
When expectations are violated, confusion and the necessity to re-
define the situation will result. Although the Ethnomethodological
School encompasses the work of several people as well as Garfinkel
and McHugh, throughout this study the term will be used solely in
reference to the work of these two writers.

The second source from which this study derives, Expecta-
tion Theory, deals with status relations and the interaction between
individuals. Formally developed by Joseph Berger and his associates,
this theory makes several assumptions which are relevant here.

(l) When individuals of differing statuses are in a task situation,

a need arises for them to evaluate their differential ability at

the task. (2) The abilities usually are assigned according to
beliefs that the individuals have about their differential statuses.
Under certain circumstances, people will consider someone of higher
status than themselves to be more competent at a task, although the
other's status or expertise may be unrelated to task ability. (3)
These initial evaluations will affect all future task interaction be-

tween the individuals.

Despite its significant contributions, Expectation Theory
has not included conceptions of procedunalexpectations or violations
of procedural expectations. The major argument of this work is that
people form not only task ability but also procedural expectations
of others, and these are related. The higher the task ability
evaluation, the higher the procedural expectation; when the latter
are violated, individuals will react differentially depending upon
their initial ability evaluations.2

In summary, this work extends Expectation Theory by add-
ing assumptions about reactions to violations of procedural
expectations and adjoining them to conceptions of status character-

istics. From this extension it will be possible to construct

testable derivations.

PLAN OF STUDY

Chapter II presents in detail both the Ethnomethodological
School's conceptions of violations of procedural expectations and
Expectation Theory's theoretical structure. The chapter also
briefly describes other relevant work. Suggestions gleaned from
the literature analysis then are incorporated into the theory, which

is presented in Chapter III.

 

2In the following work, procedural expectations are denoted
as "high" or "low" and refer to evaluations made by actors concerning
how demanding or exacting their procedural expectations are for
another. For example, people of low status will have more demanding
("high") expectations concerning conformity to procedural rules for
those of higher status than people of high status will have for
people of lower status.

Chapter 111 states the formal assumptions and scope
conditions of the theory,which is an extension of Expectation
Theory. The theory presented here adds and defines the concepts
of procedural expectations and reactions to violations thereof.
Also included in Chapter III are the theoretical derivations which
are to be tested in Chapter IV.

Chapter IV describes the empirical situation in which
the derivations from the theory were tested. The initial design
problems, the eventual design chosen, the physical setting, and
the mechanical equipment used are discussed.

Chapter V reports the results of the study and analyzes
the data. A discussion and interpretation of the empirical find-
ings is included. The chapter also presents some suggestions for
refinements of the experimental design as well as extensions of the
theory.

Chapter VI contains a summary of the research, conclusions,
and detailed recommendations for future changes in the experimental

design and extensions of the theory.

CHAPTER II

ETHNOMETHODOLOGICAL PERSPECTIVE, EXPECTATION
THEORY, AND SELECTED LITERATURE ON LATITUDE
ALLOWED TO HIGH STATUS DEVIANTS

THE ETHNOMETHODOLOGICAL
DISCOURSE

Some of the most heuristically valuable discussions of
expectation violations come from Harold Garfinkel and Peter McHugh.
Although the discourse is lacking in rigor and precision and the
empirical research is not of a systematic nature, it offers sub-

stantial insights.

The Garfinkel Perspective

Garfinkel suggests that to understand order, one must
produce disorder. His example of a stable situation (order), a
game, is a very useful tool for recognition by self and others of
the required behaviors for that game.

A game is selected because the basic rules of play serve

each player as a scheme for recognizing and interpreting

the other players "as well as his own behavioral displays

as events of game conduct." The basic rules of a game de-
fine the situations and normal events of play for persons

who seek to act in compliance with them (a player) (Garfinkel,
l963, p. l90).

According to Garfinkel, the rules of any game have three

properties.

l. From the standpoint of a player, out of alternative
territories of play, numbers of players, sequences of

moves, and the like, they frame a set that the player
expects to choose regardless of his desires, circum-
stances, plans, interests, or consequences of choice
either to himself or to others.

2. The player expects that the same set of required
alternatives are binding upon the other player as
are binding upon him.

3. The player expects that, as he expects the above of
the other person, the other person expects it of him
(Garfinkel, 1963, p. l90).

These three properties are labeled collectively as constitutive

 

expectancies. Garfinkel suggests that there is a basic order in
human behavior based on the "unstated terms of contract" which
essentially results in consensus among people to follow the rules
of the contract. Constitutive expectancies are present for each
player, and it is only when the rules of the game are broken that
people must reassess both their expectations for themselves and
others and the rules by which they supposedly are playing. Garfinkel
is trying to establish that there are empirical regularities re-
specting certain normative orders that are independent of a person's
personal biography. The constitutive rules are so structured and
so defined that any breach of them is called to the attention of
the players immediately.

Being invariant to actual courses of game play, these

expected uniformities serve as standards, i.e., as de-

finitions of correct play. Thereby they serve as the

basis for recognizing the strange move, the move that

is "outside" of the game (Garfinkel, l963, p. 194).

Garfinkel suggests that the rules of the game comprise the

"constitutive order of events," and trust represents compliance to

the order. For trust to occur, "the player takes for granted the

basic rules of the game as a definition of his situation, and that
means of course as a definition of his relationships to others"
(Garfinkel, 1963, p. 194). Players must believe and expect that
the basic rules of the game hold for themselves as well as the people
with whom they are playing. The basic rules provide a player with
a means of interpreting the appropriateness of a move or a behavior,
both his own and the other player's. When a person has committed
himself to following the basic rules and these are breached by
another, the action becomes "senseless, i.e., it acquires the peg;
ggjygg_properties of unpredictability, arbitrary occurrence,
indeterminateness, lacking causal texture, means character, and
moral necessity“ (Garfinkel, 1963, p. 196). The player will become
confused and therefore must reassess and redefine what constitutes
social reality for this interaction, because he supposedly no longer
has a frame of reference within which he can interpret the breach.
In summary, Garfinkel claims that there are basic expecta-
tions that people have of one another concerning "proper behavior";
although these are "unstated," they always are present. When these
expectancies are broken, people become "confused" and search for a

redefinition of the situation.

Ethnomethodologipal Research

Garfinkel has conducted various research projects on breaches
of constitutive expectancies. Although these studies are interesting,

one project in particular is most pertinent to the topic under

discussion. It was conducted both by Garfinkel (1962; 1967) and,
later, McHugh (1968), and the "experiments" were so similar that
they will be described as if they were one, although relevant
differences will be noted.1 Subjects were recruited ostensibly to
help determine whether there is a means of giving psychotherapy
more simply than "the present time-consuming method." When S
arrives he is told to think about a problem, give some background
to it, and subsequently ask ten questions of a psychiatrist (in

the Garfinkel study, a student counselor) that can be answered only
by a "yes" or "no." The "psychiatrist" (E), who is in a separate
room from S, gives his answers according to a predetermined sequence.
After each answer from E, S shuts off the intercom so that E cannot
hear his comments, and 5 records what he understands E to be tell-
ing him. After all ten questions are asked and answered, 5
summarizes what he has learned from the experience. In order to
create disturbance, in the Garfinkel study the answers were divided
equally between yes and no answers. In the McHugh study the pro-
portion of yes and no answers varied between 100 and 50 percent.
The situation then was constructed so that E, reading answers from
a predetermined sequence, was not in fact answering S's questions.
McHugh predicted two outcomes: (1) The greater the split, the more
likely that disruption will occur because S will be surprised by

receiving the same answer to so many questions; and (2) disruption

 

1In this discussion S stands for subject and E for
experimenter.

will occur in those splits when S asks the same question more than
once and receives a different answer the second time.

Each of these techniques should make S entertain the

possibility that the answers he gets are not motivated

by the questions he asks. This in turn should make him

question whether his original definition of the situa-

tion, psychotherapy, is actually occurring, and so

create the possibility of anomie (McHugh, 1968, pp. 66-67).

Garfinkel and McHugh report basically the same results,

and in both cases the outcome was not as strong as anticipated. The
latter anticipated possible "anomie," and the former expected "be-
wilderment, uncertainty, internal conflict, massive incongruity,
psychosocial isolation, acute general anxiety, loss of identity,
and various symptoms of depersonalization" (Garfinkel, 1963, p. 189).
In response to incongruous answers, subjects, instead of treating
them as illegitimate, sought to legitimate them by doubting their
own perceptions and imputing expertise and reasons for the dis-

crepancy to the therapist.

Limitations of the Ethno-
methodological Perspective

Although the work of Garfinkel and McHugh is heuristically
insightful, it lacks an explicit theoretical structure and testable
derivations, is imprecise, and is therefore untestable. In the dis-
cussion that follows, some of the serious theoretical and methodo-
logical weaknesses of these studies will be dealt with very briefly,
some of the specific design problems will be noted, and improvements

in their formulation will be recommended.

10

General Theoretical and
MethodoTogical Weaknesses

Concerning theoretical and methodological weaknesses,
Garfinkel and McHugh state no scope conditions to indicate the
boundaries of their formulation's application; consequently, its
usefulness as a predictive tool is seriously affected. Further-
more, theoretical structure is missing, and there are no testable
derivations from a theory. All of the above make it difficult to
determine empirical confirmation or disconfirmation, thus seriously
limiting the scientific usefulness of the formulation. Also, the
McHugh and Garfinkel research does not reflect control of extraneous
variables, hence their conclusions may be spurious.

Due to the lack of a guiding theory and systematic re-
search, we only can deduce that their conclusions derive from ad_
hgg interpretations of the data. To be fair to the Ethnomethodologi-
cal School in general, it does not require a theory to guide
structured research, or even structured research. They label their
methodology descriptive and claim they are interested in hpw_questions,
not why questions or explanation. That is, the purpose is to discover
hpw the actor is defining the situation, not from the scientist's
point of view, but from the actor's. Given their perspective, it is
difficult to evaluate the validity of their conclusions. There is
nothing inherently wrong with description, but the jump from description

to generalization cannot be justified.

ll
Specific Design Problems

In addition to the theoretical and methodological short-
comings mentioned above, the Garfinkel and McHugh work contains two
specific design problems. The first revolves around the assumption
that their research situation involved an explicit violation. In
the psychotherapy experiment, the psychiatrist or student counselor
was established as a figure who would aid the student. Therefore,
that expectation was stated explicitly as part of the “experimental"
situation. Rosenthal (1966) and others have demonstrated the
qualities of a demand characteristic. In the case under discussion,
the fact that the psychiatrist (counselor) was presented in this
light may have influenced the subject to such a degree that it
would be difficult for him to perceive that the therapist was not
helping. The experiment would have been far sounder had the sub-
ject's expectations not been part of the experimental instructions.

The second, and even more important, design problem con-
cerns the assumption that subjects have expectations for a thera-
peutic setting. It is more than likely that most college freshmen
and sophomores (the sample) have few such expectations because they
probably have not participated in therapy. It is entirely possible,
then, that the subjects' unfamiliarity with the situation would con-
found their reactions. If they did not have well-defined expectations
at the outset, and since they were told by the experimenters that they
would be helped (strongly influencing and possibly even determining
their expectations), then it is not unreasonable to suggest they

might be unable to recognize or perceive a violation. If one were

12

to run the same "experiment,“ but substitute another freshman or
sophomore for the therapist, would the same results occur? In the
situation constructed by Garfinkel and McHugh, the participants

were very different in status. A therapist or psychiatrist generally
is considered to have high status, and a student usually would con-
sider himself to be of lower status.

The sociological literature (Wahrman, 1970; 1972)
generally accepts the premise that people of high status are con-
sidered competent by people of lower status, and that indications
of incompetence are likely to be ignored or rationalized away by
those of lower status. When a person of low status perceives a
high status person violating his expectations, he is much more
likely to assume that his perceptions are incorrect than that the
other person is incompetent.

The literature also mentions a "halo“ effect surrounding
people of high status. This effect refers to characteristics
attributed to a person as a consequence of his status, but not
directly related to it. For example, a psychiatrist probably
would be considered not only competent at therapy, but also honest
and kind. As was stated above, the Garfinkel and McHugh situation
involved a high status confederate and a low status subject. The
explanation for the lack of disturbance on their subjects' part well
might be due to the subjects' reluctance to perceive the therapist
as a violator. They would doubt that a violation had occurred and

try to make their initial expectations of the therapist (competency,

13

honesty, and so forth) consistent with their observations. In
fact, this is what Garfinkel and McHugh found. Subjects did tend
to question themselves rather than the therapist, giving him the

benefit of the doubt.

Summar

In summary, the ethnomethodological perspective has pro-
vided this work with interesting heuristic insights concerning
conceptualizations of violations of procedural expectations. It
lacks, however, a precise theory, testable derivations, and an
important substantive assertion relating status conceptions to
reactions to violations of procedural expectations.' Some of these
gaps have been filled by the work of Berger and associates (1967;

1974), to whose research we now turn our attention.

EXPECTATION THEORY

The contribution of the work of Berger and associates
(1966; 1974) to the theory developed in Chapter III is that it pro-
vides a precise and testable theory of the relationship between
status conceptions and their effect on interaction.

Although much has been written about Expectation Theory,
the work most relevant to this study is "Status Characteristics and
Expectation States" (Berger et al., 1966). Because the entire theory
is not pertinent to the formulation developed in Chapter III, the

discussion will be confined to the relevant assumptions.

14

Berger and his associates are investigating how power and
prestige orders emerge and are allocated in groups initially differ-
entiated with respect to status and whose members are task-oriented.
In such groups, members will feel constrained to evaluate their
respective abilities in order to determine their probability of
successful task completion. Because there are no prior task
expectations, history of interaction among group members, and
commitment to succeed at the task, the available basis for evalua-
tion is the criterion upon which they are differentiated initially.

In Expectation Theory, the differentiation is based upon
a status characteristic, that is, a dimension of an evaluation by
an individual that he is better or worse than another individual
at doing something. These evaluations constitute beliefs concern-
ing positive or negative characteristics that can be attributed to
the person. There are two components to this evaluation. First,
an individual will be assessed as better or worse in regard to
specific traits associated with the status characteristic. For
example, professors are thought to be verbal and able writers, or
men are thought to be more mechanical than women. Second, these
specific generalizations become generalized and diffuse; a halo
effect develops, and people attribute their positive or negative
evaluations to the person as a whole. For example, instead of say-
ing that professors are verbal and able writers, one would say that
professors are capable in general.

Associated with perceptions of differences in status

characteristics are the beliefs individuals develop concerning

15

various abilities or other attributes of an actor in a specific

situation. In other words, actors form expectations about per-

 

formance of specific tasks, as expressed in the statement: "Women
are more capable at cooking than men." People also will form
generalized beliefs, such as, "Women are warmer than men.“ The
actor believes that, despite the situation, the person will do

better. These generalized beliefs are referred to as generalized

 

expectations.

The conceptions discussed above constitute an abstract
definition of a diffuse status characteristic. In essence, an
individual's perception of a status attribute (how people signify
to themselves and others that they have a defined status) will
cause him to assess very generalized meanings or beliefs that are
associated with the status attribute.

Expectation Theory deals partially with external status

 

characteristics, those unrelated to the situation. For example, if
an M.D. were fixing his car, his status characteristic would be
external to the task.

In a task-oriented situation it is likely that an actor
will want to assess the relative ability of himself and others at
the task in order to estimate his probabilities of succeeding. Be-

liefs about who will and will not do well are defined as performance

 

expectations. If the actor has prior expectations about the relative
abilities of himself and others at the task, he does not need to use

status conceptions to determine ability. When task ability has not

16

been specifically associated with or disassociated from the diffuse
status characteristic, and if there is no other basis upon which an
actor can infer his own and others' task abilities, then the diffuse
status characteristic will be the basis for evaluation.

To summarize, in a task-oriented situation where actors
are committed to the successful completion of the task, there will
be a need for them to assign performance expectations to self and
other. If actors P and O initially are differentiated on the basis
of an external status characteristic, and if there are no other
criteria for assigning performance expectations, they will be
assigned on the basis of stereotyped beliefs that people have for
individuals who possess similar states of the status characteristic.
Moreover, if the task requires an ability that is not relevant to
these stereotypical conceptions, the basis for assigning performance
expectations still will be dependent upon the status characteristic
because of the halo effect. As mentioned above, the theory indicates
that a halo effect represents the process whereby specific generali-
zations become generalizations attributed to the person as a whole.

Once expectations are formed concerning task ability, they
affect all future interaction in the task situation. If an actor
has assigned higher performance expectations to another, then he will
have less influence and prestige in the task situation than an actor
who has high performance expectations of himself and low expectations

of another. In the former situation the actor probably will be more

17

deferential, make fewer attempts to influence the group, and evaluate

his own performance less favorably than those who have high perform-

ance expectations for themselves and low expectations of others.
Contributions of Expectation Theory

and the Ethnomethodological
Discourse to the PresentTWork

Having presented the portion of Expectation Theory relevant
to this work, there now is a basis for discussing that theory and the
ethnomethodological discourse in relation to this research. In
evaluating the contributions of each, it becomes apparent that both
have contributed significantly to the theory developed in this study.

The McHugh and Garfinkel work did not consider at least
one important variable--status--that might have determined their
subjects' responses. As was mentioned, the experimental condition
consisted of a high status confederate and a low status subject.
Their results then are not explainable solely on the basis of inter-
action between a subject and a confederate, because Expectation
Theory has demonstrated empirically that perceptions of status in
initially differentiated task groups determine the group's future
interaction in terms of allocations of power and prestige for the
members. Garfinkel and McHugh have not accounted systematically for
how or why the process operates, simply that it seems to. It would
be impossible to make any responsible generalizations from their re-
search. However, Expectation Theory provides the tools to facilitate
a comprehension of this process. It assumes that once the power and

prestige order is stabilized, it maintains itself. This assumption

18

might facilitate an understanding of the Garfinkel and McHugh re-
sults. With a high status-low status situation and the subjects'
ambiguous expectations for the therapeutic process, it is possible
that they granted the therapist the credibility that their stereo-
typical beliefs about a person of high status would lead them to do,
despite contrary evidence. Subjects therefore questioned their
perceptions more than the actions of the therapist and gave the
therapist the “benefit of the doubt."

Since the major theoretical assertion of this work is
that the perceived relative statuses of interactants differentially
will affect their reactions to violations of expectations, two
goals become clear. First, a clear characterization as to what
constitutes a violation of a procedural expectation must be developed.
The Garfinkel and McHugh work comes the closest of any other in the
literature to meeting that goal, but it is insufficient. Second,
an explicit, precise, and testable theory concerning the effect of
status on interaction is needed. Expectation Theory meets this
goal, but its present formulation does not include conceptualizations
of the effect of status on reactions to violations of procedural
expectations. Therefore, the extension of Expectation Theory that
is presented in Chapter III will incorporate the Ethnomethodological
School's ideas concerning violations of procedural expectations.

This work, thus far, has dealt with the most relevant litera-
ture. There is, of course, other literature relating high status

deviance and reactions to it. Although these sources are less germane

19

to the present argument, they will be presented below as indications

of gaps in the literature that need examination and refinement.

SELECTED LITERATURE ON
HIGH SIRIUS DEVIANCE

Ralph Wahrman (1970) has made an extensive and definitive
review of the literature on violations of expectations. He notes,
and correctly so, that authors are quite divided in their opinions
about high status violators. For example, Homans (1950), Sabath
(1964), Hollander (1958; 1960), Hughes (1946), Blau (1960), and Sherif
and Sherif (1956) observed that higher status deviants are allowed
less leeway than lower status deviants. Homans (1961), Hollander
(1958), Verba (1961), and Sherif and Sherif (1964; 1965) have argued
that high status deviants are protected from sanctions and therefore
allowed more leeway. These studies deserve some mention here, but
they contain rather serious theoretical and methodological weaknesses,
some of which will be discussed below. It is not the intention of
this author to present an extensive review of this literature; it is
not directly applicable to the theory formulated in Chapter III, and
Wahrman's work provides ample treatment. Instead, the major focus
will be to point out some of the shortcomings contained in the litera-
ture and to suggest improvements, some of which will be incorporated
into the theory in the next chapter.

The literature on high status deviance contains various
shortcomings: Conceptualizations or definitions of deviance are
not precise; reactions to, or sanctions of, deviance are not clearly
defined; and there is a general neglect of definitions concerning
intensity or magnitude of a violation. Discussion of these problems
and methods for dealing with them in the present work are discussed

below.

20

It is difficult to integrate the studies done on high
status deviance due to the nature of their conceptualizations.

There are no specifications for kipg_pj_deviance. Some describe
but do not define procedural deviance, while others do not state
or define what kind of deviance they are investigating. It is
difficult, then, to determine exactly what is being measured.
Therefore, it is important to the present formulation to define
precisely and operationalize what kind of deviance is being in-
vestigated.

Treatments of reactions to, or sanctions of, high status
violators as described in the literature suffer from the same lack
of definition. Thus, it is difficult to discern what constitutes
a sanction. The theory presented in the next chapter is less con-
cerned with sanctions than with the process by which people react
to violations and why. The intent is to understand the process of
reactions to violations, not simply the end result. The emphasis
is on why reactions occur as they do, not simply that they occur.

It is not sufficient to argue that people will react
positively or negatively (sanction) to high status violators. A
thematically more interesting argument involves why these differential
reactions exist. The theory formulated in Chapter III attempts to
clarify this issue.

It also seems important to specify the intensity or magni-
tude of a violation. For example, "cheating," "not helping,‘l "murder,"

and "infidelity" obviously are not violations of the same intensity,

21

and, therefore, reactions to them should differ. The literature is
seriously lacking in such distinctions. The present theory will
attempt to determine on basically intuitive grounds, because no other
bases are available, what would constitute a "moderate" violation,
one that is not inconsequential or severe, but one that hopefully
would fall somewhere in the middle. In the future it will be very
important to determine empirically how violations vary in intensity.

Most conceptions of violations refer to negative breaches.
An interesting perspective indicating that this is not always a
valid conclusion has been offered by S. R. Maddi (1961). He has
proposed a "variety" position in contrast to what he calls consistency
assumptions. He argues that the unexpected can, indeed, be satisfy-
ing due to the fact that novelty, unexpectedness, change, and com-
plexity may be inherently satisfying. He argues that consistency
assumptions trivialize life and underestimate man, and, in contrast,
the unexpected relieves boredom and sometimes provides delight. How-
ever, it is reasonable to expect that there is most probably a
continuum of violations ranging from fairly inconsequential, to
satisfying, to extremely upsetting. Because this work deals with
negative violations of procedural expectations, it is important to
conceptualize a violation of a procedural expectation that will be
of sufficient magnitude to evoke negative responses.

In conclusion, the analysis of the literature on reactions
to high status deviants indicates a need for precise definitions,

precise formulations, and testable derivations which are amenable

22

to empirical confirmation or disconfirmation. Where relevant, the
discussion has suggested how the theory in Chapter III will improve

these shortcomings.

SUMMARY

The chapter has presented the two lines of thought most
pertinent to this study, the works of Harold Garfinkel and Peter
McHugh and the writings of Berger and associates on Expectation
Theory. The former contributed conceptualization about violations
of procedural expectations, and the latter offered assumptions
about the relationship between status conceptions and interaction
in initially differentiated task-oriented groups. Because the
theory of the present work predicts that status differentially will
affect perceptions of and reactions to violations of procedural
expectations, the importance and limitations of both these works
for the theory developed in Chapter III has been discussed.

Other relevant literature on reactions to high status
deviance has been presented. Suggestions for improvements of these
formulations and incorporation of these suggestions into the theory

presented in the next chapter have been developed.

CHAPTER III

A THEORY OF REACTIONS TO VIOLATIONS
OF PROCEDURAL EXPECTATIONS

INTRODUCTION

The theory presented here argues that reactions to vio-
1ations of procedural expectations are dependent upon the relative
statuses of the interactants. Before a reaction can occur, a
person first must assess the relative status of both himself and the
other. The justification for this assertion derives from our
assumption that expectations about a person's procedural conformity
or deviance are associated with their prestige. Thus, if an other
(0) is considered to be higher status than person (P), P's expecta-
tions for 0's conformity to procedural rules will be more exacting
(demanding) than those of a P who considers himself to be higher
status than 0. In the latter instance, P's expectations for 0's
conformity to procedural rules will be less demanding (exacting).

In this presentation, the scope conditions and terms of
the theory will be stated; where they differ from those of Berger
and associates (1966; 1974), an indication will be given. The
extension of Expectation Theory then will be discussed, and the
assumptions, definitions, and derivations of the theory will be

offered.

23

24

SCOPE CONDITIONS AND
TERMS OF THE THEORY

The interaction situation consists of members of a group,

P and 0, who are required to perform a valued, competitive task, T,
which has only two outcomes--success or failure.1 The task is con-
sidered to be valhed because it is assumed that P perceives the
possible outcomes of the task as either success or failure. The
assumption is that P is committed to success. The task is defined
as competitive, both because only one of the actors can win and be-
cause of the ease of operationalizing a competitive structure in a
game situation.

The theory applies only to task behavior; behavior related
exclusively to social and/or emotional relationships is excluded.

The theory is formulated from the point of view of an
actor, P, who is oriented to P' (the actor as an object to himself)
and one other, 0. It is P's reaction to the task situation that
will be described. From this point on, the notation P' will refer
to the actor's expectations for himself, and 0 will refer to the
actor's expectations for the other.

The status characteristic of the culture of P is a given in
the theory. For example, P perceives that a professor is more highly
valued than a student, or that it is better to be honest than dishonest.

Thus, it is assumed that these are universal evaluations.

 

1Expectation Theory uses collective, not competitive, tasks.

25

An external status characteristic (female, male, president,
professor, student, and so forth) refers to an evaluation by an actor
that he is, for example, "better" or "worse" than another actor on
some state of the status characteristic, such as verbal ability.

The theory requires that there be only one status characteristic
which discriminates between P and 0 in situation S, and that the
only information they have about each other concerns this character-
istic.

Evaluations of competence involve two factors: specific
and general evaluations. The former refer to judgments of competence
concerning specific traits associated with the status characteristic.
For example, professors are believed to be verbal and able writers.
Specific evaluations are generalized so that they become general
evaluations. Instead of saying that professors are able writers
and verbal, a halo effect develops, and professors become evaluated
as being generally able. This generalization process constitutes an
abstract definition of a diffuse status characteristic.

In the present theory, it is assumed that prestige is a
diffuse status characteristic, and it is given only two distinct
states. In other words, in reality there is a continuum of prestige,
but for the sake of simplicity, prestige will be designated here as
either high or low.

Performance expectations are defined as P's beliefs con-
cerning P' and 0's relative abilities at the task. Since status

characteristics have been shown to affect performance expectations

26

(Berger et al., 1966), and since it is assumed that initially P

will have to evaluate the relative capabilities of P' and 0 at the
task (performance expectations) in order to assess his probability
of success, it is essential that P have no basis for assignment of
expectations to P' and 0 other than the state of the diffuse status
characteristic. If there were preexisting knowledge concerning

task ability, there would be no need for P to use status conceptions
in order to determine the relative task ability of P' and 0. There-
fore, the theory requires that members have pg_prior expectations
for their relative ability at_tfle_ta§k,

Let us say that P is a student and O, a professor. If
there are no other status characteristics that differentiate P from
0, then P will use the diffuse status characteristic to assign per-
formance expectations for P' and 0, even though the diffuse status
characteristic may not be relevant to the task. In such a situation,
P usually will assign concomitant states of the status and the
diffuse status characteristic, that is, P would perceive the status
characteristic for 0 as "professor" and the diffuse status character-
istic as "prestige." If P perceives a professor as high status,
then P probably will perceive the professor as able because a halo
effect usually develops such that the professor is considered not
only prestigious but also able, moral, and honest. This generalized
evaluation of a person's ability will be referred to as a generalized
expectation state. Although this construct is theoretical and not

directly observable or measurable, it has importance for the theory.

27

The expectation state usually carries with it expectations for
future performance and is stable throughout time. Thus, P will
expect 0 to be competent not only at the present moment, but also
in the future and in a variety of areas, even though they may not
be relevant to the task.

The presentation above completes the discussion of

Expectation Theory relevant to the present theory.

EXTENSION OF EXPECTATION THEORY

The theory presented here, in addition to extending
Expectation Theory, will deal with a new dimension, assumptions
concerning procedural expectations and their relation to perform-
ance expectations. .The theory requires that P initially forms
procedural expectations for the task.2 These are defined as be-
liefs formed by P concerning P' and O's following of the task rules.
It will be predicted that actors form both performance and procedural
expectations for others as well as themselves and that they are re-

lated in degree. It is also asserted that performance expectations

 

2Although procedural expectations involve various components,
for example, manners, rules of the game, moral norms, and rules deal-
ing with aesthetics and cleanliness, the theory presented here restricts
the definition of procedural expectations to "honesty," which could be
classified as a moral norm. No paradigm of procedural expectations
exists, and little, if any, research has been done on the topic. Con-
sequently, the present theory has no previously established foundation
upon which to build. Since it is only logical to assume that viola-
tions of different procedural norms would evoke different responses,
it was necessary to choose only one type of procedural norm in order
to delimit the field of study. Furthermore, restricting study to
only one type of norm ensures that what one seeks to measure is, in-
deed, being measured.

28

precede procedural expectations. If P has higher performance
expectations for 0 than for P', then that P will have higher pro-
cedural expectations for 0 than a P who has lower performance

expectations for 0 than for P'.3

To be more specific, the theory
is concerned with the relationship between performance expectations
and the effect on P of violations of procedural expectations by 0.
The theory only applies to negatively evaluated violations pf p39;
cedural expectations. Such a violation occurs when P perceives 0
as violating the task rules in a way that threatens P's possibility
of obtaining the rewards that P expected. A positively evaluated
violation of a procedural expectation would occur if P perceived 0

as violating the task rules in order to give P more reward than

P expected.

ASSUMPTIONS AND DEFINITIONS

Situation S will consist of two actors, P and 0, who are

performing a task. The theory makes no predictions for 0's behavior.

 

3The usage of "higher" or "lower" in reference to procedural
expectations may be an imprecise semantic distinction. These words
normally are used in a quantifiable sense, and procedural expecta-
tions are not strictly quantifiable. However, in this work, the
usage of "high" and "low" as qualifiers for both performance and
procedural expectations seems stylistically preferable. "High“ and
"low" procedural expectations refer to evaluations made by actors
concerning how demanding or exacting their procedural expectations
are for an other. It is asserted here that actors of low status will
have more demanding procedural expectations (concerning conformity to
procedural rules) for those of higher status than will actors who have
a higher status than an other.

29

Definition 1: An interaction is a task situation, S, if and only
if

a. there are at least two actors, P and 0;
b. P and 0 are task oriented;
c. P and O are performing a task;

d. P and 0 are competitively oriented to task 1
in situation S;

e. P and 0 are motivated to successful completion
of the task;

f. P and 0 have no prior expectations for their
own or each other's performance at the task;

9. P and 0 initially are differentiated with
respect to status; and

h. there is only one characteristic which dis-
criminates P from 0.

This completes the definition of situation S. The necessity

of this definition to the theory already has been discussed.

Assumption 1:

Given the task condition in situation S, if P and O possess
different states of a single external status characteristic, and if
these states are not directly relevant to T, but if there is no other
basis for evaluation, then the status characteristic will become the
basis for discrimination.

Because P and 0 will feel a need to assign performance expecta-
tions concerning the task for self and other, and if no other criterion
exists for such an assignment other than the external status character-
istic (such as sex or profession), the status characteristic will be-
come the basis for the assignment of performance expectations due to

the halo effect.

30

The theory presented here deals primarily with the relation-
ship between performance and procedural expectations; therefore, it
is necessary to assert that P and 0 do form performance expectations
and pply_a§_a_§e§plt_gf situation S, and that performance expecta-
tions are formed only on the basis of the diffuse status characteris-
tic. If, for example, prior performance expectations did exist, it
would be difficult to determine whether P's reaction to situation S
was due to the experimental manipulation or some other variable or
variables.

Berger and associates (1966) assume that once performance
expectations have been assigned by P to P' and 0, these will deter-
mine the future power and prestige order of the members of the group.
For example, if P holds a low performance expectation for P' and a
high performance expectation for O with respect to ability at the
task, if P is committed to success at the task, and if P believes
that O is more likely to do well, "P will . . . make fewer attempts
to influence 0 than will a second P who holds a high performance
expectation for self and low for other" (Berger et al., 1966, p. 40).

It is assumed that this statement also applies to the
present theory. Once the performance expectations are assigned by
P for P' and 0, they will determine the power and prestige order in
such a way that the probabilities of distributions of influence,
disagreements, deference, and so forth, will be different for members
who have a higher expectation for themselves at task ability than for

those who have a lower expectation.

31

The theory asserts that P will have not only performance
expectations for P' and 0, but also procedural expectations for P'
and 0, and that they are related in degree.

Definition 2: In S, P will have two kinds of expectations for P'
and 0, and they are related in degree:

a. performance expectations, and
b. procedural (normative) expectations.

Berger and associates (1966) also posit that the external
status characteristic carries a halo effect such that a high status
person will be considered not only able, but also kind, moral, honest,
and so forth. In this theory the halo effect is reformulated to
assert that P has not only performance expectations for P' and 0,
but also procedural expectations. In other words, in a task situa-
tion P has beliefs concerning P's and O's normative behaviors, and
these beliefs are related to P's performance expectations for P' and
0. We expect, then, that a P who has higher performance expectations
for 0 than for P' also will have higher procedural expectations
than will a P who has higher performance expectations for P' than
0.

To summarize, when people are involved in a situation, 5,
they not only form performance expectations concerning the relative
task abilities of the group members, but also have normative
expectations concerning how members should behave in the situation,
that is, what constitutes the proper behavior. These unspecified,
informal rules are brought by members into a situation and define

what constitutes proper (normative) behavior, for example, cheating

32

is not proper. If an actor violates one of the task rules, one
certainly would expect some reaction from the other participant.
It is proposed that the effect that performance expectations have
on the power and prestige of a group is identical to the effect
procedural expectations will have on a group, and that performance
expectations precede and are related in degree to the formation of
procedural expectations. That is, a P who has higher performance
expectations for 0 than for P' and who is the recipient of a vio-
lation of a procedural expectation by 0 will be more deferential
and disagree less with 0 than a P who has higher performance
expectations for P' than 0.

It is posited that performance expectations precede pro-
cedural ones. P first will assess performance expectations for P'
and 0 and then assign procedural expectations to P' and O and to
the same degree. In other words, a person first will be defined as
high status and more generally capable at a task, and this evalua-
tion then will be generalized so that the person also is seen as
being more honest, and so forth.

Some notation needs to be introduced.4

From this point on,
P's who have higher performance expectations for 0 than P' will be
referred to as L-H, and P's who have higher performance expectations

for P' than 0 will be referred to as H-L.

 

4The notation in this research will follow that of Berger
and associates.

33

The issue of reactions to violations of procedural
expectations is more complex than what has been presented thus
far. Although L-H recipients of a violation will be more defer-
ential and disagree less with an 0 than will H-L recipients, their
perceptions of and reactions to the violation will be a function
91 the subjective probability they hold for the likelihood 91 ED:
occurrence pf a_violation. In other words, L-H P's not only will
be more deferential once a violation has occurred, but also will
be less likely to identify the violation and will react differently

to 0, depending upon O's status relative to P.

Assumption 2:

The higher the performance expectations P holds for 0
relative to P', the less P will expect 0 to emit a violation of ‘
a procedural expectation.

Some research indicates that people will refuse to per-
ceive a high status person as a violator. The Balance and Consistency
Theorists (Heider, 1946, 1958; Festinger, 1957; Newcomb, 1953, 1961;
Osgood and Tannenbaum, 1955; and Rosenberg and Abelson, 1960) suggest
that people's cognitive processes tend toward balance and consistency
because they are uncomfortable with the opposite; when presented with
inconsistent or imbalanced information, people try to make the cogni-
tions consistent by denying them, changing them as little as possible,
or by reevaluation. Particularly with regard to high status non-
conformity, Osgood (1960), Osgood and Tannenbaum (1955), and Krech,
Crutchfield, and Ballachey (1962) have indicated that the more extreme

34

the attitude, the harder it is to change. Here we assume that P's
expectations (attitudes) about cheating are stronger for a status
superior and weaker for a status inferior. Thus, L-H P's will be
less likely to notice a violation than will H-L P's, and, because
they have stronger beliefs about the honesty of high status people,
it will be much more difficult for them to change their opinion.

If people actually have higher procedural expectations
for a high status person, for example, that they are honest, then
it is reasonable to assume that a status superior is expected to
"cheat“ much less often than a status inferior. The Cognitive
Dissonance Theories point out the dissatisfaction, tension, and
discomfort that people experience when their world is not "balanced,‘I
that is, when the unexpected occurs. It is also reasonable to
assume that it takes longer for someone to perceive and act on the
unexpected, especially in an unfamiliar situation. We assert that
L-H P's will be less likely to perceive (and it will take them
longer to do so) that a violation is occurring than H-L P's. L-H
P's will be much more likely to see themselves as "seeing things"
or misperceiving the situation.

.QLL If P has higher performance expectations for 0 than
for P', P is less likely to perceive violations of procedural ex-
pectations emitted by 0.

g;g_ If P has higher performance expectations for P' than
for O, P is more likely to perceive violations of procedural ex-

pectations emitted by O.

35

g;§_ The higher P's performance expectations for 0 than
for P', the longer it will take P to notice and/or react to a vio-
lation of a procedural expectation emitted by 0.

g;g_ The higher P's performance expectations for P' than
for 0, the more quickly P will notice and/or react to a violation
of a procedural expectation emitted by 0.

If, for example, P's in the L-H condition have higher
procedural expectations for 0 than P's in the H-L condition, a
violation by O in the L-H condition will be less expected than a
violation by O in the H-L condition. Because the violation is less
expected in the L-H than in the H-L condition, it is assumed that

the L-H P will feel less positively toward 0 than will the H-L P.

Assumption 3:

The more unexpected a behavior, the stronger will be the
reaction.

g;§_ If P has higher performance expectations for 0 than for
P', and if 0 violates a procedural expectation, once P notices the
violation, P will feel less positive toward 0 than would a P with
higher performance expectations for P' than 0.

It is asserted here that unexpected behavior causes strong
negative reactions and that the higher the status of an actor, the
higher are the expectations that he will conform to procedural rules.
Conversely, the lower the status of an actor, the less he will be
expected to conform to those rules. Because nonconformity to pro-

cedural rules is less expected by an L-H P than by an H-L P, reactions

36

to violators in the L-H situation will be less positive than in the
H-L condition.

An entirely separate issue arises when P's overt responses
to a violation are considered. It is assumed in this theory that
overt responses to a violation always will be less likely in L-H
than in H-L situations, but this most definitely will be the case
when 0 has sanctioning power over P. If P would lose some rewards
by sanctioning 0 (there is a higher probability of this in the L-H
than in the H-L situation), it is assumed that overt and covert
reactions might be very different. If a manager strikes an employee,
the latter is less likely to strike back than if his assailant were
someone of lower status. Although in this instance the behavior
is less expected and therefore should evoke a stronger response, the
potential loss of rewards should mitigate an intense overt reaction.
However, a stronger covert response is anticipated because such be-

havior is less expected from a status superior.

Assumption 4:

The higher the relative performance expectations that P
holds for 0 relative to P', the less the likelihood that negative
behavior will be overtly manifested when 0 violates P's expectations,
but especially if 0 controls sanctioning power over P.

g;§_ If P holds higher performance expectations for 0 than
for P', and if 0 violates P's procedural expectations, P's reactions

will be more covert than overt.

37

gpz, If P holds higher performance expectations for P'
than for O, and if 0 violates P's procedural expectations, P's
reactions will be fairly equally covert and overt.

The reasoning for d.7 is that in this situation 0 has
little sanctioning power over P. Therefore, P has nothing to lose
by making the reactions congruous.

Finally, there is the issue of resolution of the disturb-
ance and how that resolution comes about over time. For example,
can one violation qualitatively change a relationship between two
people temporarily or forever? Although this question is extremely

interesting, it will be deferred to future research.

SUMMARY

The presentation of the theory is completed. It differs
from previous Expectation Theory formulations in three ways. First,
it adds and defines the concept of procedural expectations, in
particular, violations of those expectations. Second, it posits a
relationship between performance and procedural expectations.
Third, it makes perceptions of and reactions to violations of pro-
cedural expectations probabilistically dependent upon performance
expectations. In particular, the higher P's performance expectations
for 0 relative to P', the higher the probability that P will take
longer to notice and/or react to a violation emitted by O, and the
less positive will be P's response to O, in general, as opposed to

a P who holds higher performance expectations for P' than 0. Also,

38

in the L-H situation, P's overt and covert responses to 0'5 viola-
tion of procedural expectations will be much more discrepant than
P's responses in the H-L situation.

Through the testable derivations of the theory, the
author will seek to determine its confirmation status in Chapter IV.
Because, as reported in Chapter II, many of the derivations of
Expectation Theory already have been tested, the focus of this
endeavor will be to test the predictions that are specific tg EELS
m-

The derivations to be tested are:

g;§_ The probability that P will perceive a violation is
less in the L-H situation than in the H-L situation.

g;g_ In the L-H situation, the probabilities are greater
that P will take longer to notice a violation and/or react to a
violation than in the H-L situation.

ELJQ_ The probability of P feeling negative toward 0 in
the L-H situation is greater than in the H-L situation.

These predictions will constitute the empirical tests of
the theory, presented in Chapter IV.

Chapter V will describe the experimental setting, the

developmental task problems, and the physical equipment used.

CHAPTER IV
EMPIRICAL TEST OF DERIVATIONS FROM THE THEORY

This chapter first will deal with the major objective of
a task situation, namely, operationalization of the derivations
from the theory. It then will discuss the problems encountered in
meeting this objective. Finally, the solutions to the various

problems will be presented.

OPERATIONALIZATION
OF VARIABLES

In order to operationalize the states of the diffuse status
characteristic (independent variable), the subject and confederate
were introduced to each other and told their states of the diffuse
status characteristic. In the L-H condition, subjects were told
that the confederate was Dr. Gordon, a professor from the University
of Michigan. In the H-L condition, the confederate was introduced
as a high school student. The subjects were undergraduate students
at Michigan State University.

A violation of a procedural expectation, the second inde-
pendent variable, was operationalized by having the confederate not
follow the rules of the game. The trials during which this occurred
will be referred to as the critical trials.

The dependent variable concerning subject's feelings to-

ward the confederate was operationalized by having the subjects

39

40

'complete, after the experiment, an adjective test that revealed
positive or negative feelings.

Differential perceptions and reactions to violations, the
second dependent variable, were much more difficult to operational-
ize. It was necessary to devise a measure for these factors on the
critical trials. Asking subjects at the end of the experiment if
they had noticed cheating and on what trials would be unsatisfactory
for three reasons: (1) recall may be biased; (2) subjects might be
reluctant to verbalize their perceptions to an experimenter; and
(3) this method would not directly test the prediction. Various
alternatives for operationalizing the second dependent variable

were explored, and many were rejected as inappropriate.

REJECTED ALTERNATIVES

In order to understand why so many attempts to operation-
alize the dependent variables were rejected, it must be noted that
one of the major goals in developing the research design was an
effort to construct a standardized experimental situation. The
principal feature of such a setting is that the experimenter gains
a large degree of control over the situation, the advantage being
to minimize the effect of extraneous variables. The researcher can
feel more confident that what initially was to be measured was, in-

deed, measured.

41

Measures of Perceptions and
Reactions to Violations

Alternatives for operationalizing perceptions and reactions
to violations were explored. One possibility was a face-toaface
interaction between subject and confederate with the experimenters
behind a one-way mirror coding various facial and verbal responses.
This technique was rejected because face-to-face interaction, re-
gardless of the skill of the confederate, would involve the confed-
erate giving cues to the subject, such as nonverbal cues, and because
verbal interaction could occur. For example, it would be very easy
for the subject to question the confederate on the critical trials,
and it would be extremely difficult to standardize the confederate's
reaction to ensure that each subject received the same response;
different subjects might ask different questions to which a standard-
ized response would not be appropriate.

Another alternative considered was the use of physiological
measures. A Grass Recorder could have provided readings of skin
conductance, gut motility, EKG, and perspiration rates. However, a
review of the literature showed contradictory opinions about what
these devices actually measure. Pupil dilation, another physiological
measure, seemed to have more support in the literature in terms of
reliability, but this technique was not practical in terms of the
equipment and expertise available to the researcher.

Also explored was the method of having the subjects record

on paper, trial by trial, their reactions to the task. This

42

alternative had two drawbacks: (1) possible increase in suspicion
rates of subjects; and (2) more complicated instructions for the
subject to remember.

Another alternative was to give the subjects 30-second
rest periods during the experiment. They simply could press a
button labeled "rest," which would indicate to the other that they
also had 30 seconds to rest. The assumption here is that the more
disturbed the subjects are, the more rest periods they will take.
This method was discarded for pragmatic reasons. Assuming subjects
did take rest periods, additional time would be added to the experi-
ment, and the number of subjects that could be run on a given day
would be decreased. In addition, there were standardization
problems. It is impossible to predict the effect that 30 seconds
of extra time would have on the subjects; all subjects probably
would not be taking rest periods, and if a difference occurred be-
tween the two conditions, it would be difficult to determine whether
the rest period was in some manner responsible.

Another trial-by-trial approach considered was some sort
of challenging mechanism. Participants could challenge their
opponent's word, write "challenge'' on a board, and consult a
dictionary. If their challenge was correct, they would add points
to their score; if incorrect, their opponent would deduct points
from their score. The assumption here would be that L-H subjects
would be less likely than H-L subjects to challenge their opponent.

The fact that they did or did not challenge would constitute a

43

measure of their perceptions of and reactions to a violation. The
challenge technique presented standardization problems; the
challenge might affect subjects' responses in some manner that

could not be measured. This alternative also was rejected because
(rf the extra time the experiment would entail. Experiment length

is an important factor because subjects can become bored or apathetic
if the experiment is too time consuming.

The last alternative to be discussed involved having the
subjects in separate rooms, but allowing them to send message slips
to each other. It was conjectured that subjects who caught a vio-
lation would send more message slips to the confederate than subjects
who did not. This procedure seemed unnecessarily complicated. It
would be necessary to have someone act as messenger, and the messen-
ger might give off cues. This procedure also would create a lengthy
experiment. A content analysis of the notes would be cumbersome,
and it would be difficult to predict what the subject's messages

would contain and therefore difficult to standardize the confederate's

replies.

Timing Devices

Problems were encountered with attempts to find an appro-
priate timing mechanism to demarcate the beginning and ending of a
trial. When the experiment was conceived as face-to-face interaction,
the possibility of using a regular sand timer in conjunction with a

buzzer was explored. Subjects alternately would play and time each

44

other, then press a buzzer to let each other know when to begin or
end a phase. However, the sand timer proved undesirable when tested
by the researcher. It was found that (1) some people became so in-
volved in their opponent's progress that they forgot to notice when
time had expired, and (2) in continual usage some of the sand in

the timers did not return to its original cavity. Both difficulties
certainly would cause a standardization problem because trial times
would vary. Moreover, it was possible that subjects who accidentally
received more time might impute meanings to this behavior which
would confound the experiment results.

Two other timing alternatives were considered: (1) piping
the experimenter's voice into rooms to announce the beginning and
the end of the trial, and (2) having an experimenter in the room
designated as the timer to call out "start" and "finish." These
options were thought to be unsatisfactory because any source other
than the confederate might bias the subject's reaction, for example,
the subject might expect the experimenter to interfere when cheat-
ing occurred. Since the theory does not include statements con-
cerning reactions to sources other than the opponent, an experimenter's
presence might confound the results.

Many possible ways to violate expectations were considered.
If the sand timer were employed, the confederate deliberately and
consistently could allocate more or less time to the subject by mani-
pulating the timing mechanism. But the necessity for the subject to

observe such manipulation obviously involved face-to-face interaction,

45

a possibility that already had been discounted; furthermore, use
of the sand timer also had been rejected. Other alternatives were
to have the confederate continue to work after time was up, or to
create violations through the way in which game rules were stated.
The choice of violation and why it was selected will be discussed
later in this chapter.

CHOSEN EXPERIMENTAL
SITUATION

It was determined that in order to control variance the
subject and confederate must be in separate rooms with no means of
communication other than that necessary for the experiment. In a
face-to-face interaction, for example, subjects could receive
unnecessary cues concerning status or task performance. Since the
theory requires that the subject and confederate be differentiated
on only one status characteristic, it was important to ensure that
subjects did not receive additional cues. The standardized design,
then, would keep the subject and confederate completely separated.
Although the final procedure chosen involved rather complicated
manipulation of the laboratory machinery, it proved to be the most
satisfactory alternative by far. In fact, it simplified the
experiment tremendously.

To increase standardization and optimize time efficiency,
the entire task situation was recorded on videotape and transmitted
to the subject via television. This assured that all subjects would
receive the same stimuli under all experimental conditions. (See

Kruse [1972] for a discussion of the benefits of videotape usage in

46

experimental situations.) Under all experimental conditions, all
cues, innuendos, and messages are uniform. If the experimenter
were to have direct contact with subjects, biasing cues could be
projected. In this particular experiment the subject was exposed
only to the experimenter's voice in order to diminish any possible
visual cues. Moreover, a speech specialist read the audio portion
of the taping and was requested to be as neutral as possible (not
give off unnecessary vocal cues). The subjects found the entire
procedure very credible.

Another benefit of the videotape method is the effective
communication barrier it constructs between subject and confederate.
The confederates also were taped and were introduced visually (not
verbally) at the beginning, which prevented any other cues being
emitted during the rest of the experiment. The technique proved
to be a time and money saving device; there was no need to arrange
schedules around the availability of the confederate or to pay her
to be present during the experiment.

Other measures were taken to increase standardization. The
same experimenter always was present, and she dressed and looked as
similarly as possible each time; the same recruiter obtained all the
subjects; the same person telephoned subjects to schedule and remind
them of their appointments; and the same message was used each time.
Subjects responding to ads that had been placed in the student news-
paper or in their dormitories spoke to an automatic answerphone device

and received the same message. This mechanism was very useful not

47

only because it maximized standardization, but also because it was
unnecessary for someone to be in the laboratory to answer the
telephone.

Further discussion of the solutions to the general task
problem and the reasons for choosing them will be presented during

the description of the sequence of the experiment.

PHYSICAL SETTING

The activity room was completely separated from the
observation room in this experiment (the one-way mirror between
the rooms was completely covered up). Figure4~l depicts the
structural layout of the room and the necessary apparatus. The
subject enters and is seated at a table. At the end of the desk
is a television, from which all instructions and experimental
stimuli will be received. On the desk are the materials (see
Appendix A) required for participation in the experiment. There is
a dummy camera in the ceiling which subjects are told is the camera
that televises their faces and introduces them to the other. (Of
course, a confederate has been pretaped for the purpose of intro-
duction.)1 A camera to the left of the table is focused on the spot
where subjects will put the materials to be transmitted to the opponent.
There is a buzzer on the table for the subjects to use in the event

they have any questions.

 

1Except for references to the introduction of the confederate
on the tape, all further statements about other, confederate, or
opponent refer to the experimenter.

48

 

 

 

ONE WAY MIRROR

 

 

 

 

 

a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

c4

 

 

 

 

 

 

 

 

 

 

 

 

A - television
B - word and communications box
C - communications pad

D - buzzer

E - list of rules of game

Figure l.

r—XLHIGD‘TI

Activity Room

word sheets
pencil

scoring sheets
camera

table

subject

door

 

 

49

Figure 4.2 depicts the observation room. On the table on
the left wall are three monitor television sets. One picks up the
output from the videotape machine, the second transmits the sub-
ject's material, and the third transmits the confederate's material.
The first set is to ensure that the visual portion of the videotape
is functioning; the second set enables the experimenter to see and
score the subject's material; and the third set transmits the con-
federate's responses to the subject. On top of the third television
is the audio monitor for the videotape.

0n the desk along the other wall is the videotape machine,
a camera to focus on the confederate's material, a box containing
the confederate's standardized material, and a mechanism that
allows the confederate to switch back and forth from videotape to
camera input into the subject's room. This is necessary because
the experiment involves switching from the videotape (of the con-
federate's pretaped introduction) to live action and responses

(when the experimenter scores individual subjects).

SAMPLE

Subjects were recruited from the undergraduate population
of Michigan State University for "a study involving 'Word Cognitive
Ability.'“ Three recruiting methods were used. (1) Requests were
made during classes according to a modification of a standardized
procedure used in previous experimental research on Expectation

Theory (see Appendix B), and (2) ads were placed in the school

 

50

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:) |\
“ A I "/
3 * 0.44
a C ~
‘ E
I /\ \\
°o '—
i
g F
k'
S o
1§ o
>§ O
‘V
S e
“i
2
<3 0
0
i
.T
Fl
A - Videotape machine F - Monitor TV
B - Switching box G - Monitor TV
C - Confederate materials H - Monitor TV
0 - Word and communications box I - Experimenter
E - Camera J - Door

Figure 2. Observation Room

 

51

newspaper and (3) on bulletin boards in dormitories and classroom
buildings (see Appendix B). Those who expressed interest in class
filled out forms asking their name, sex, and telephone number.
Those answering ads, which listed the laboratory number, spoke to
an answerphone which asked them to leave their name and telephone
number (see Appendix B).

From the subject pool thus accumulated, it was decided
that only female subjects would be chosen for this experiment. The
reasons were twofold: (l) The experimenter was female, and it was
felt that a same sex experimenter and subject would not introduce
unnecessary bias (Rosenthal, 1966), and (2) females at Michigan
State University participate in experiments three times more often
than males, which means they are easier to obtain.

Subjects were scheduled for experiments by telephone
according to a standardized procedure (see Appendix B). They were
contacted again the night before to remind them of their appoint-
ment. (This one technique reduced the no-show rate to about l per-
cent, even during exam week.) If the subject did not plan to appear,
this reminder call gave the experimenter time to schedule someone
else. Subjects were excluded from the sample if the recruiter deter—
mined that they previously had participated in a social-psychological
experiment involving deception. Although the question of race and
its differential impact on subject responses still is not clearly
understood, there is enough evidence to indicate that it may have some

biasing effect (Cantril, 1944; Williams, 1964; Williams and Cantril,

52

1945; Katy, Robinson, Epps, and Waly, 1964); therefore, members of
racial minority groups were excluded. A total of 50 subjects were
run, 27 assigned randomly to one condition and 23 to the other.

The study had two experimental conditions. As stated in
Chapter II, the L-H condition consisted of a high status confederate
and a low status subject, and the H-L condition involved a high
status subject and a low status confederate. The status variable
was manipulated simply by visual videotape introduction of the con-

federate to the subject.

SEQUENCE OF THE EXPERIMENT

The subject arrived at the laboratory and was greeted by
the experimenter. As the two walked down the hall they passed a
closed door bearing the sign "Participant #2." The experimenter
pointed to the door and said: "You will be playing with partici-
pant number two." The subject was led to her room, which had a
sign reading "Participant #1." She was asked to sit down and look
and listen to the television set in front of her. The experimenter
then left and closed the door.

Immediately after the door was closed, the experimenter's
voice came into the activity room through the subject's television.
She was thanked for joining the study, which was explained as an
attempt to judge the effectiveness of groups working together using

new kinds of communication equipment rather than face to face.

53

The Word Cognitive Ability Task

In order to ensure that performance expectations were
based solely on the diffuse status characteristic, subjects were
told (Appendix A) that this task was a ppjgge_one, having nothing
at all to do with verbal, artistic, mechanical, or mathematical
ability. They then were told that they would be introduced to each
other, and, depending upon the condition (L-H, a professor, or H-L,
a high school student), the confederate was flashed on the screen.
This manipulation constituted an operationalization of the diffuse
status characteristic.

After the introduction, the task situation was explained.
Subjects were told that some letters would be flashed on their
screen, and from these letters they were to make as many allowable
(according to the rules of the game) words as possible. They were
instructed to copy the letters from the screen onto their word
sheets (see Appendix A) and then construct as many words as possible
in a forty-second time period. Subjects were told that a buzzer
would sound to denote that time was up, at which point they were to
put their sheets in a box.2 A camera would transmit their words to

their opponent. As soon as they saw their opponent's words on the

 

2The box was necessary because the camera could focus
accurately on small writing only in a very limited area. This pro-
cedure had other positive effects: (1) Subjects could not see the
confederate, and (2) the confederate's words could be written out
in advance. Both of these factors increased standardization.

54

screen, they were to copy the letters for that trial onto their
scoring sheet (see Appendix A), copy their opponent's words, score
them, put the score on a separate sheet, and place the sheet in the
box so that the scores could be transmitted to each other. They
then were to copy their own score onto the scoring sheet below the
score they had entered for their opponent. Subjects were given
sixty seconds in which to complete the scoring process, at which
time a buzzer sounded. Five seconds later a buzzer designated the
beginning of a new trial.

The word sheets were used for two reasons. First, be-
cause the sheets were programmed (see Appendix A), they condensed
the instructions and gave the subject a step-by-step guide for how
to play the game. It was felt that this would strongly diminish
any possible misunderstanding of the task instructions. Second,
the sheets would provide evidence of whether or not subjects under-
stood the rules (a scope condition of the theory). The scoring
sheets also were used for these two reasons and because they provided
empirical data about the subject's reactions to the critical trials.
These latter involved the confederate making up words that vio-
lated the rules of the game. How subjects scored these critical
trials, that is, whether or not points were deducted for a violation,
is an operationalization of perceptions of and reactions to viola-
tions, the dependent variable.

It was thought that if the subject had to recopy the

letters on the scoring sheet (already having copied them on their

55

word sheet) apg_copy their opponent's words, then any violation
should become much more apparent to them than if they just scored
off the screen. The assumption was that copying the letters twice
would imprint the allowable letters for that trial on the subject
agg_make any violations more obvious. Furthermore, having to write
their opponent's score apg_their own score just beneath it should
give the subjects a trial-by-trial comparison of their respective

abilities at the task3

and also should provide an indirect measure
of the status manipulation when they reported who they thought won
the game (as opposed to who actually won).4 It is assumed that
the subject's assessment of who won the game would be more a function
of the initial performance expectations for P' and O and less a
function of reality.

The timing problem was solved by the decision to use the
videotape equipment. Because the entire experiment was recorded
on videotape, the buzzes that denoted the beginning and ending of
each trial phase when the experiment first was recorded were also
on the tape. Therefore, all timing intervals were exactly the same

for every subject. This seemed a measurable improvement over the

other possibilities considered and also avoided any problems the

 

3Most Expectation Theory experiments do not include this
trial-by-trial feedback.

4Notice that there is no ability manipulation in this
experiment, and no predetermined "winner." It is equally likely
that the subject or the confederate could win.

56

experimenter might have had with stop watches (they become faulty)
or other devices that could require varying times to switch on or
off during the experiment.

The experiment involved switching from the video portion
of the tape to live action. At the stage when word sheets were to
be shown to the subject, the experimenter switched to a camera
focused on the word and scoring box in the experimenter's room.
Live action continued until the scoring process was completed.

Each subject was given two practice trials. The experi-
menter then knocked on their door, entered the room, and asked if
there were any questions. Subjects' sheets were checked to make
sure they had understood the instructions.

The total number of trials was twenty, seven of which
were critical: trials 4, 6, 9, 13, 14, l7,and 18. To decrease
suspicion, it was decided that critical trials should not be spaced
at regular intervals, that is, two noncritical, two critical, and
so forth. It also was deemed important that no critical trials
occur in the beginning of the experiment; an early violation might
increase suspicion.

The violations eventually chosen were from a list of many
possibilities submitted to a panel of academics, who were asked to
rate them on scales of severity and blatancy. The trials progressed
from nonsevere and nonblatant to severe and blatant violations. In
addition to decreasing suspicion, as noted above, this progression
procedure was chosen for an even more important reason._ On nonsevere

trials it was expected that subjects in the L-H condition definitely

57

would be more likely to give the confederate the "benefit of the
doubt" than would subjects in the H-L condition, which operational-
izes the prediction that L-H subjects would be slower in perceiving
and reacting to a violation than would H-L subjects. That is, L-H
subjects would question themselves before they would question the
confederate, the opposite of what is expected in the H-L condition.
However, if a violation were severe, it would be difficult for
either H-L or L-H subjects to ignore it.

All of the confederate's materials, derived from thirty
test cases submitted to a panel, were standardized. The words
selected had appeared with the most frequency on each test case
trial, and they were averaged for inclusion in the material.

Originally, instead of letters, words were going to be
flashed. However, in the process of trying to develop "neutral"
words and testing them on the panel, it was found that no word
really had neutral, nonaffective meanings. Thus, ten lettered dice
were thrown twenty times. The letters from each throw were used

for each trial.

Posttrials Phase

After the trials were completed the experimenter knocked
on the door, entered the subject's room, collected the materials,
and asked the subject to fill out an adjective test (see Appendix A)
rating her opponent. This constituted an operationalization of

subjects' feelings toward the opponent. When finished, they were

58

led into the debriefing room, where a postexperimental interview

was administered (see Appendix A). This interview sought to measure
suspicion rates and subjects' initial purported relative performance
expectations. In other words, subjects were debriefed. The study,
the reasons for it, and the manipulations involved were explained to
them, and any questions were answered. Their cooperation in not
discussing the experiment with other students was solicited.

After the debriefing, subjects were asked a final question:
“At any time during the experiment did you feel that your opponent
was cheating?" This question was posed after debriefing to see if
answers differed according to H-L and L—H conditions. Such a
blatant question entailed no risk of suspicion at this point in the
experiment, and it was felt that the replies might provide some
heuristic insights for the future. It should be emphasized that
this last question was included mainly for heuristic purposes.

The subjects were thanked and paid $2.00 for their parti-
cipation. They again were asked not to repeat the specifics of the
experiment to their friends.

The analysis of the data from this experiment and the
confirmation status of the derived predictions from the theory will

be presented in Chapter V.

CHAPTER V

ANALYSIS OF DATA

This chapter first describes the inclusion and exclusion
criteria that were used to obtain data for analysis. Second, it
presents and discusses the results and their implications. Third,

some suggestions are made for future improvements of the research.

INCLUSION OF DATA

General Criteria for Exclusion

In executing an experiment it is impossible to ensure
that all subjects will meet the fixed initial scope conditions of
the theory; if they do not, their data must be excluded from the
analysis. For example, the theory requires that all subjects have
no prior expectations for the task, but a subject may have partici-
pated in or read about deception experiments; therefore, he or she
would not meet one of the fixed initial conditions as stated in the
definition of situation S. When a subject does not meet one or
more of the theory's scope conditions, his or her data must be ex-
cluded because the theory does not make any predictions about his
or her behavior.

Specific criteria for exclusion were established before
the experiment began. Thus, the experimenter could not exclude a

subject who might have had an unusual pattern of responses on the

59

6O

postexperimental data. This precaution was, and always is, necessary
since it prevents the inclusion or removal, on intuitive grounds,

of a subject who might bias the results for or against the theory's
predictions.

There were five specific exclusion criteria. First, data
were to be removed if subjects showed suspicion about some part of
the experiment. In this category were placed subjects who did not
believe they really were playing against the stated, or any,opponent,
or subjects who believed that the confederate's words were pre-
determined or that the game was "rigged" by the experimenter.
Second, those subjects who did not remember who they played with
were to be removed from the analysis. For example, some in the L-H
condition said they played against a college student instead of a
professor. Third, lack of task orientation would warrant exclusion,
such as subjects who did not complete the task or did not follow
instructions. Fourth, also to be removed were subjects who failed
to have a competitive orientation. As it turned out, only one
subject fell in this category. Fifth, screening problems also were
to be a basis for exclusion. In this category eventually were in-
cluded some subjects who should have been omitted in the recruiting
process and a handicapped subject whose inability to perform the
task properly did not become apparent until actual trials were run.

Exclusion criteria for all but the fifth category were
determined from the following information. The postexperimental

interview contained standard questions designed to determine whether

61

the scope conditions of the theory were met. For example, subjects
were asked how they felt about their opponent and the task, and how
they expected their opponent to do before the task began and why.
They also were asked if they ever had read about or participated in
an experiment "like this one," probing for deception experience or
other disqualifying criteria. The subjects' scoring sheets then
were examined to make sure that they had followed the experimental
instructions. If in either of these stages a reason for exclusion
was found, the data were removed from the analysis.

The researcher hoped to increase standardization of the
interview process and reliability of inclusion-exclusion criteria
by conducting all interviews herself and by checking all decisions
for inclusion or exclusion with two others involved in the experi-
mental process. All interviews were listened to and checked for

reliability against the written responses.

Specific Cases Excluded

Of the 50 subjects run, 27 in the L-H condition and 23 in
the H-L condition, 4 were excluded because they were suspicious.
One had participated previously in a deception experiment, one was
a senior psychology major, one had seen the film "The Milgram
Experiment," and one simply did not believe that she was playing
against a professor. These 4 subjects constituted .08 of the original
sample. Considering the amount of deception in this experiment and

the fact that suspicion rates normally run between 20-30 percent,

62

this rate was extremely low. Moreover, 3 of these 4 subjects never
should have been included in the sample, but should have been
screened out by the scheduler initially.

Also excluded from the analysis were 3 subjects who could
not remember with whom they were playing. One in the L-H condition
simply did not recall, but after probing said: "maybe a college
student." Another said: "someone own age," and the third also could
not remember who she played with. In future experiments, it may be
advisable to give subjects a longer exposure to the confederate in
order to decrease this kind of problem.

Of course, one subject who did not follow instructions
properly was excluded.

One subject had to be excluded for lacking a competitive
orientation because she could not recall that she even had an
opponent.

The fifth type of exclusion case, screening problems,
already has been discussed above. Three of the four subjects
omitted because of their suspicion also could be included in this
category, as was the one physically handicapped subject.

Once these specific cases were excluded, the data analysis
then was based on results obtained from the remaining 40 subjects,

which yielded an n of 20 in each condition.

63

RESULTS

Responses to Violations
of Expectations

The data from each subject consisted of the subject's
reaction to the critical trials, that is, those that contained the
violations. For this phase of the data the subject's method of
scoring the confederate constituted the empirical measure of her
reactions. It was expected that reactions to violations would vary
according to the experimental condition; the effectiveness of the
status manipulation,and the consequent performance expectations
that are formed,differentially should affect the subject's responses
to violations in each of the experimental conditions. The main
statistic for this portion of the data analysis is the total pro-
portion of subjects in each condition that detected violations and
the proportion of subjects in each condition that registered the
violations on the critical trials.

The major tests of the theory involved the prediction that
the subjects' responses will vary according to their initial per—
formance expectations for self and other.

egg, The probability that P will perceive a violation is
less in the L-H situation than in the H-L situation.

Subjects in the L-H condition are expected to score fewer
violations than those in the H-L condition. Table 5.1 presents the

data for total violations detected per condition, in frequency form.

64

Table 5.1. Total Frequencies of Violations Detected on All
Critical Trials, by Condition.

 

 

Violations
Condition Detected
L-H
Percentage . .32
n (45)
H-L
Percentage .41
n (58)

 

Table 5.1 indicates that the data are in the direction
predicted. Of the subjects in the H-L condition, 41 percent
detected violations, as opposed to 32 percent in the L-H condition.
A more specific kind of examination breaks down, by condi-
tion, the percentage of subjects that responded to the seven possible

violations. Table 5.2 presents these data.

Table 5.2. Frequency of Detected Violations, by Condition.

 

Violations Detected

 

Condition 0 l 2 3 4 5 6 7
L-H
Percentage 0 .20 .40 .35 .05 O O 0
n (0) (4) (8) (7) (l) (0) (0) (0)
H-L
Percentage .10 O .30 .30 .10 .20 O O

n (2) (0) (6) (6) (2) (4) (0) (0)

 

65

As stated above, the theory predicted that subjects in
the H-L condition would register more violations than subjects in
the L-H condition. By comparing rows 1 and 2, the prediction seems
to hold. Table 5.2 indicates that, in terms of percentages, sub-
jects in the L-H condition responded to far fewer violations than
did subjects in the H-L condition. Probably the most interesting
aspect of this table is that 20 percent of the subjects in the
H—L condition noted five violations, whereas no subjects in the
L-H condition registered more than four. It is also interesting
that only 5 percent. of the subjects in the L-H condition detected
more than three violations, whereas 30 percent. in the H-L condition
registered more than three. The data, then, seem to support the
prediction that subjects in the L-H condition would register fewer
violations than subjects in the H-L condition.

The theory also states that the differential performance
expectations should cause the subjects in the L-H condition to take
longer to recognize and/or score a violation than those in the H-L
condition.

g;g_ In the L-H situation, the probabilities are greater
that P will take longer to notice and/or react to a violation than
in the H-L situation.

Thus, an examination of the data should demonstrate that
the trials with more "subtle" violations will contain higher dis-
crepancies between the percentage of subjects detecting violations

than trials with more "severe" violations. If this assumption is

66

Table 5.3. Percentage of Violations Detected on Critical Trials,
by Condition.

 

Detected Violations on Critical Trials

 

Condition 4 6 9 13 l4 17 18
L—H
Percentage .10 .05 .05 .25 .75 .05 1.00
n (2) (1) (l) (5) (15) (l) (20)
H-L
Percentage .30 .25 .10 .50 .90 O .85
n (6) (5) (2) (10) (18) (0) (l7)

 

accurate, a comparison of rows 1 and 2 of Table 5.3 should show
rather different percentages. When rows 1 and 2 are compared, the
discrepancy between registering violations is, for the most part,
much more pronounced during the early critical trials, with H-L sub-
jects detecting, in some cases, markedly more violations than L-H
subjects.1
Benefit of the Doubt

An implicit assumption of the theory is that subjects in

the L-H condition will give more benefit of the doubt2 to the con-

federate than will subjects in the H-L condition.

 

1Trials 9 and 17 were apparently not as severe as the design
asserted. Examination of the table indicates that they were the least
scored critical trials. When these are removed from the analysis the
prediction for trials 4 through 14 is completely in the expected direc-
tion. Trial 18, however, indicates an unexpected finding: L-H sub-
jects scored more violations than H-L subjects. Even though this
finding will be reserved for later discussion, note that there is a
minimal discrepancy between the conditions on these trials.

2That is, subjects will tend to doubt themselves and their
ability to judge a violation rather than doubt the confederate's
ability at the task.

67

Table 5.4 shows the percentage of subjects in each condi-_
tion that gave the confederate the benefit of the doubt on all
noticed violations. "Benefit of the doubt" here means that the
subject, on her scoring sheet, indicates that she has questioned
the confederate's word, but deducts nothing, giving the confederate
her full score. For example, a subject might have written beside
a word on her scoring sheet, "I don't think this is a word," but

might have given her opponent a full score despite her doubt.

Table 5.4. Proportion of Total Noticed Violations to Benefit of

 

 

the Doubt.
Proportion of
Total Noticed
Total Violations to
Violations Benefit of Benefit of
Condition Scored the Doubt the Doubt
L-H 45 4 .08
H-L 58 2 .03

 

Table 5.4 indicates that more than twice as many benefit
of the doubt situations occurred in the L-H condition than in the
H-L condition. However, caution should be applied when interpret-
ing the significance of these results due to the n in each cell being
so small. Although this assumption was not an explicit prediction of

the theory, the results do tend in the direction anticipated.

68
Feelings toward the Confederate

The theory also predicts that subjects in the L-H condi-
tion will have less positive feelings toward the confederate than
will subjects in the H-L condition.

ngQ_ The probability of P feeling negative toward 0 in

the L-H situation is greater than in the H-L situation.

Table 5.5. Proportion of Subjects with Positive, Neutral, and
Negative Feelings toward Opponent.

 

Feelings toward Opponent

 

Condition Positive Neutral Negative
L-H .62 .20 .18
H-L .78 .19 .03

 

Table 5.5 presents the data related to this prediction.
The table indicates that subjects in the H-L condition felt more
positive toward their opponent than did subjects in the L-H condi-
tion. The table also demonstrates that subjects in the L-H condi-
tion felt much more negative toward their opponent than did subjects
in the H-L condition. Hence, the data seem to support the prediction.
Since much of the literature (such as Berger et al., 1966) indicates
that high status actors will be more favorably evaluated than low
status actors, the reversed direction of the evaluations in Table 5.5

makes this finding even more significant.
Performance Expectations

Although performance expectations are not directly observable,
the questions from the experimental interview permit some indirect

measures of whether the subjects did, in fact, initially assign

69
differential performance expectations to themselves and their
opponents. As the theory predicts that performance and procedural
expectations are related, it is necessary to determine whether the
status manipulation was operative.

In the L-H condition, 18 out of 20 subjects reported that
they initially expected their opponent to do better at the task than
they would do. One subject said she expected to do equally well be-
cause the instructions told her that ability did not make any differ-
ence for this test. Two subjects reported "they just didn't know."

It would appear from these data that, in the L-H condition, subjects,
in general, initially formed higher performance expectations for
their opponent than for themselves.

Added support for this acceptance of higher initial perform-
ance expectations for the opponent by L-H subjects comes from a seren-
dipitous finding of the study. Although most subjects in the L-H
condition won the game by a large majority (they beat the confederate
by 489 total points, or an average 24.45 points per subject), only
60 percent of them reported accurately who they thought won the game.
This figure is in marked contrast to H-L subjects, who lost the game by
a total of -15 to the confederate (or an average of -.75 points per
subject); 95 percent of these subjects responded correctly to the

question of who they thought won, and only 5 percent were inaccurate.3

 

3The fact that L-H actors under some experimental conditions
develop high performance outputs has been demonstrated. The theoretical
and empirical status of this effect is indeterminate (see, for example,
the replication of the Adams and Jacobsen [(1964)] experiment reported
to Anderson and Shelly [(1970)] and references cited in their biblio-
graphy.) However, the marked difference in performance outputs here
seem to indicate that future theoretical work is needed.

70
Perhaps the reluctance of the subjects in the L-H condition to report
they won the game was a function of their initial performance
expectations.

The data for the H-L condition concerning initial per-
formance expectations are rather sketchy and difficult to interpret.
Only three subjects reported feeling initially that the confederate
was going to do worse than they would; the majority of the others
were noncommittal. Possibly the question concerning initial per—
formance expectations is not a good measure in this case because
H-L subjects' severe losses to someone of lower status, probably
unexpected, might have clouded their remembrance of initial feel-
ings. Postexperimental interviews of H-L subjects contain such
statements as, "I've never been good at word games," or "high
school students are really learning better and faster these days."
It is also possible that a college student does not feel she is
that much higher in status than a high school student.

Another serendipitous finding may lend support to the
contention that subjects felt some connection between performance
and procedural expectations. Although of a much more qualitative
nature than any of the results reported above, this finding might
be fruitful to pursue heuristically. When the subjects were asked
if they felt that the confederate cheated at any time during the
experiment, the majority in both conditions (17 in the H-L, 18 in
the L-H) reported "no.“ However, of those reporting "no" in the
H-L conditions, few subjects elaborated on their reply, while in
the L-H condition, many of the subjects gave fuller explanations.

The latter also seemed to mention more of the violations, although

71

it already was reported here that they did not score as many
violations as did those subjects in the H-L condition. Below are
some examples.

L-H Subject #l--"No, she was in a hurry and got messed up,
but she was not cheating."

L-H Subject #2--"No, I noticed a couple of times that she
put down the same words twice, but it was
a mistake. It was not intentional."

L-H Subject #3--"No, even when I saw a double word I just
thought it was a mistake."

Although the above analysis is extremely qualitative, the
author suspects that it lends some positive weight to several
suppositions. First, performance expectations initially were formed,
at least for the L-H condition, and probably for the H—L condition,
although the latter is difficult to discern. Second, once these
performance expectations were formed, they were fixed. This is
more demonstrable for the L-H subjects. It is interesting that in
the L-H condition, even when they beat the confederate by a con-
siderable amount, subjects reported that the score was "pretty
close." If this second assumption has no validity, there should not
be such a high discrepancy between the two conditions in terms of
accurate reporting. Third, there is some connection between per-
formance and procedural expectations, although this is less
demonstrable from the data. Otherwise, the subjects in the L-H
condition would not have differed so much from H-L subjects in terms
of lengthier explanations. Of course, this conclusion would have
much more weight if the variance between the groups had been much

stronger in terms of proportions of yes and no answers. However,

72

it is suspected that this lack of variation is due primarily to
the fact that the violations were moderate, not strong, causing
the majority of subjects to report that the confederate did not
cheat. This issue will be treated more fully in the discussion
section.

DISCUSSION AND INTERPRETATION
OF THE RESULTS

 

The discussion below will be organized according to the
following format: responses to violations of expectations, bene-
fit of the doubt, feelings toward the confederate, and performance
expectations. Where applicable, suggestions will be made for re-
finement of the design and the theory.

Responses to Violations of

Procedural Expectatibns:
Critical Trial Data

The results shown in Tables 5.1, 5.2, and 5.3 appear quite
consistent with the theory's predictions. Those tables indicate
that the L-H and H-L subjects differ, and in the directions pre-
dicted. The H-L subjects responded to more violations, and earlier,
than did L-H subjects.

A possible explanation for the slightly higher reporting
of violations on the last two trials in the L-H condition can be
suggested. First, as the theory predicts, the L-H subjects, even
if they noticed a violation, were more reluctant to report it.

Therefore, a latent reaction might be occurring: When L-H subjects

73

finally admitted that blatant violations were being committed,
they may have felt they had to score a violation as a compensation
for all those that they had not scored previously. However, this
explanation is tentative because the theory makes no predictions
for it and because the discrepancies are minimal on the last
trials. Perhaps a more important aspect that this problem points
up is improvements that can and should be made in the experimental
design. One of these now will be discussed.

Although the researcher cannot determine, in advance, the
number of people who will register a violation, some recommendations
for future improvements can be made on the basis of the frequency
and kinds of violation that were detected during this particular
experiment. It would appear that the panel's opinions about the
blatancy of a violation, discussed in Chapter IV, are incongruous
with the actual results of the experiment. That is, the design was
set up according to the supposed blatancy of violations, which was
intended to increase over time during the critical trials. The
results of this research indicate that the general trend is as
predicted, but it does not hold over all trials. It appears that
some violations were not sufficiently blatant. Whether or not this
is true, the n in some cells is so small that a definite interpreta-
tion cannot be offered confidently. For example, the n in trial 17
for the L-H condition is l, and for the H-L condition, it is 0. Of
course, the small number of n's in many of the categories cannot be

prevented entirely, since the theory predicts that the less blatant

74

violations are less likely to be scored, especially in the L-H
condition. In conclusion, although the differences between the
conditions are in the direction predicted, as mentioned in
Chapter III, only "moderate," or what were considered moderate,
violations were used. In the future it would be fruitful to run
experiments with varying magnitudes of violations. For example,
with more or less intense violations, there might be higher
differences in responses between the categories on some trials.

Responses to Violations of

Procedural Expectations:

The Postexperimental
Interview

In the postexperimental interview the question was asked:
“At any time during the experiment, did you think that the opponent
was cheating?“ (Although not a test of the theory, it was hoped
responses might lend some heuristic insights.) The replies to this
question were negative and were difficult to interpret. Perhaps
subjects generally were reluctant to use such a strong word as
"cheat" to a stranger or someone of "higher status." If the sub-
jects' responses are taken literally, then possibly the violation
was so moderate that subjects did not perceive that outright cheat-
ing was occurring and therefore were reluctant to make a public
allegation. Another explanation is that they simply did not feel

the opponent was cheating. Again, if this was the case, then we

75

would expect responses to be much more intense to a stronger
violation. Future research concerning intensity of violations

should help determine the accuracy of these interpretations.

Benefit of the Doubt

Although the data concerning the benefit of the doubt
also do not directly test any derivation of the theory, the theory
does contain the implicit assumption that people of lower status
will give more benefit of the doubt to people of higher status
than people of higher status will give to those of lower status.
This assumption is one of the bases for the predictions that L-H
subjects will be slower to react overtly to a violation of an
expectation and that they will react to fewer violations. In other
words, subjects in an L-H condition will form higher performance
expectations as well as higher procedural expectations for their
opponent than will H-L subjects, and L-H subjects therefore will
be more likely to question themselves than an opponent of higher
status.

Table 5.4 illustrated that L-H subjects gave the benefit
of the doubt to their opponents more often than did H-L subjects.
Again, however, the n in each cell is small, and although the trend
is in the direction anticipated, the data would be far more reliable
if the cell number were higher. Further research might include, as
a part of the experimental design, opportunities for subjects to

give the benefit of the doubt to their opponents. For example,

76

subjects might be instructed that, in the scoring process, if they
have any doubts and are not sure whether or not to score a word,
they can put a question mark beside it; however, the choice about

awarding points would remain solely their decision.

Feelings toward the Confederate

The data shown in Table 5.5 supported the prediction that
L-H subjects will feel much less positive toward their opponents
than will H-L subjects. Although data analysis indicated that L-H
subjects did feel six times more negative toward the opponent than
did H-L subjects, it is entirely possible that this difference will
increase with increased violation severity. Future research could
indicate whether or not this prediction is valid.

An unexpected finding was that L-H subjects had signifi-
cantly higher performance output than did H-L subjects, which might
affect the data. In the L-H condition, many subjects won the game
by a significant amount, but they generally were very reluctant to re-
port this fact. In other words, the initial performance expectations
held throughout the game despite the feedback they received to the
contrary.4 It is interesting that the many L-H subjects, with one

exception, who reported outcomes inaccurately were those who beat

 

4Note that Berger and associates do not include feedback
during the experiment in their work; therefore, the data are not
comparable.

77

the confederate by a large margin.5 (The exception refers to a
subject who won by 2 points, but reported that the confederate
won.) Examples of their comments in response to this question
were: "I scored a little higher, but my opponent won," or "I

don't know. I thought it was a tie." These subjects won by an
average of 26 points per game. However, other L-H subjects who won
by a wide margin did report accurately; these subjects beat their
opponents by an average of 60 points per game, which made it
extremely difficult for them to report otherwise. (No L-H sub-
jects who lost the game reported they had won.)

As was mentioned, the fact that H-L subjects lost and
most L-H subjects won may affect the data. It is possible that
those L-H subjects who won unexpectedly and reported that they
won would have felt even more negative toward the opponent if they
had lost, as they most probably expected to do. The differences,
then, between the L-H and H-L conditions might have been even
stronger if the L—H subjects had lost as expected. By the same
reasoning, the H-L subjects who lost to the confederate, and the
majority of them did, probably were not expecting to do so. (Their
responses to the question concerning their performance expectations

of the confederate appear to support some kind of dissonance

 

5This result indicates even more support for the assumption
that initial performance expectations for self and other remain
stable throughout the task. Despite evidence to the contrary, when
L-H subjects won, they could not report the fact that they did.

78

reduction about losing.) Therefore, the discrepancy between the
two conditions with regard to feelings toward the confederate might
have been even greater if the H-L subjects had won as expected,
because their reactions to their opponent, had they won, might
have been even more positive and less negative. Future research
might readjust the task in such a way to ensure that the H-L
subjects, in general, do better than their opponents, and the
L—H subjects, not as well, in order to obtain a more accurate
measure of positive and negative feelings. It is suggested that
a better measure would demonstrate that feelings toward the
opponent would be even more marked if a subject's initial per—

formance expectations for P' and 0 were fulfilled.

Performance Expectations

The data support the assumption that the L-H subjects
formed initial performance expectations for their opponent and
themselves in the direction predicted. They expected their
opponent to do better. The data for the H-L subjects were inde-
terminate, and some possible explanations will be offered.

It is more than likely that the status difference between
a college student and a professor, the L-H condition, and a college
and a high school student, the H-L condition, is not equal. The
researcher was aware of this possibility before the experiment began,
but felt that a more equal distance, for example, a high school and

a grammar school student, would not be believable to subjects due

79

to the nature of the task - word construction. The other alternative
was to use a high school and a college student for the H-L condition,
and an undergraduate and a graduate student for the L-H condition.
This initially was rejected on the intuitive ground that college
seniors (as some of the subjects were) would not feel a large status
differential between themselves and a graduate student, although

the author is aware that some experimental studies have used these
criteria. In future research it might be desirable to set the
experiment up as follows: L-H condition--college student, graduate
student; H-L condition--college student, high school student; or

L-H condition—-college student, professor; H-L condition--college
student, junior high school or freshman high school student. One
very apparent reason for making this change is that the H-L subjects
did not report the initial performance expectations for their
opponent that might have been anticipated from the results of other
expectation experiments. If the scoring mechanism were not altered
(as was proposed in the last section), future research could indicate
whether this aberrant H-L reporting of initial performance expecta-
tions was due to the dissonance H-L's experienced at losing the

game to a status inferior. In other words, did losing cause H-L's

to reduce dissonance by not reporting definite initial performance
expectations for themselves or other, or was the initial status dis-
crepancy insufficient in the L7H condition to produce differential

initial performance expectations?

80

The connection between performance and procedural
expectations, although not directly tested in this experiment,
can be inferred from some of the data. (1) The data all point
in the direction predicted by the theory, namely, that the higher
P's performance expectation for O, the higher P's procedural
expectation for 0. Thus, when 0 violates a procedural expectation,
the L-H P will feel less positive toward 0 than if P's performance
expectations for 0 were lower than those for P'. (2) Subjects in
the L-H condition, as opposed to those in the H—L condition, re-
sponded to the question concerning their feelings about whether
or not their opponent cheated with much lengthier answers. The
author's speculation is that they needed to make congruent their
initial performance and procedural expectations by going into
lengthy discourse about the impossibility of a higher status person
violating a procedural expectation. An alternative explanation
would be that they had to "balance" their noncongruent feelings;
however, the theory contains no predictions for this speculation.

This chapter first described the inclusion and exclusion
criteria that were used to obtain the data for analysis. It then
presented and discussed the results and confirmation status of the
tested derivations. All derivations were found to be in the pre-
dicted direction. Finally, suggestions were made for future

theoretical and methodological improvements.

CHAPTER VI

SUMMARY, CONCLUSIONS, AND SUGGESTIONS
FOR FURTHER RESEARCH

SUMMARY

The work described here was heavily influenced by the
Expectation Theory of Joseph Berger and associates and by Harold
Garfinkel's and Peter McHugh's ethnomethodological research on
expectation violations.

The influence of the Berger group (1966) revolved around

their conceptions, to be discussed below, of performance expecta-

 

tions, external status characteristic, status characteristic, and

 

diffuse status characteristic. In a task situation where P and O
are committed to "successful" completion of the task, it is

assumed that the actors will feel pressured to determine their
relative abilities at the task. These beliefs concerning relative
task ability are labeled performance expectations. If P and O
differ in an external status characteristic (such as profession or
sex), and if these criteria are not directly relevant to the task,
but if there are no other available criteria for determining their
differential expertise at the task, then beliefs associated with the
status characteristic will become the basis for assignment of per-

formance expectations.

81

82

A status characteristic stands for an evaluation by an
actor that he or she is “better" or "worse" than another actor.
However, there are apparently two kinds of evaluations: specific
and general or diffuse. First, judgments of specific abilities
are concerned with evaluations that are made with respect to
specific traits associated with the status characteristic; for
example, professors are considered good writers. Second, these
specific evaluations become generalized or diffuse; for example, a
professor is considered not only a good writer (a specific evalua-
tion), but also to be capable generally. The specific evaluations
that become generalized or diffuse represent the abstract defini-
tion of a diffuse status characteristic. Because it is assumed that
status characteristics determine the assignment of performance ex-
pectations, and because performance expectations are supposed to
determine the observable power and prestige order of the group, the
initial status differences between P and 0 should come to correspond
to the differences in power and prestige between them:l

In addition to extending the theory of Berger and associates,
this work investigated a new aspect, namely, assumptions about viola-

tions of procedural or normative expectations. The latter were

 

1As a variation of the theory and experiment proposed here,
it would be substantively worthwhile to determine how far a violator
must go in order to break the stability of an established power and
prestige order. The theoretical literature, as reviewed in Chapter
II, is very divided on this issue.

83

defined as the beliefs that P and 0 have for self and each other in
terms of their following the task rules. The theory posited that
performance and normative expectations are related, and are related
in degree: The higher the performance expectations that P holds for
O, the higher the normative expectations that P holds for O. The
theory also predicted that when a high status 0 breaks the normative
order, a lower status P will respond later, registering fewer occur-
rences, and less positively to the violation than will a P of higher
status interacting under the same conditions with a lower status 0.

The other major influence on this work derived primarily
from the research of Garfinkel and McHugh on the topic of violations
of procedural expectations. However, their efforts lack precision
of conceptualization and testable derivations, and an attempt to
avoid these shortcomings was made here.

In order to test the derivations of the theory, a labora-
tory test was designed and executed. The two independent variables
were the high or low performance expectations that the subjects held
for their opponent, and the procedural expectations that were violated.
The two dependent variables were the subjects' responses to violations,
and their feelings toward their opponent. Fifty subjects were run in
the experiment, 27 assigned randomly to one condition and 23 assigned
randomly to the other. Ten subjects were excluded from the analysis
since they did not meet the scope conditions of the theory, leaving
an n of 20 in each condition. At the beginning of the experiment,

subjects were introduced either to a high or low status confederate.

84

Their task was to play a word game with their opponent which in-
cluded seven critical trials. These latter were defined as those
in which a procedural violation occurred. The proportion of times
the subjects in each condition detected and/or scored a violation
was considered a measure of their acceptance or rejection of the
other's performance outputs.

After the experiment, subjects were administered an
adjective test. The proportion of times that subjects responded
favorably or negatively toward the opponent, in each condition, was
considered a measure of their feelings toward their opponent.

The data demonstrated that the derived predictions tend
in the direction expected. Subjects in the L-H condition responded
to violations later during the sequence of trials and fewer times,
and they felt less positively toward the other than did subjects in
the H-L condition.

In addition to the data relating to the major tests of the
theory, also discussedwere serendipitous findings, interpretations
of some unexpected data, evaluation of the experimental setting, and
suggestions for future changes in the experimental design and exten-

sion of the theory.

GENERAL CONCLUSIONS

Reeponses to Violations of
Procedural Expectations

The data appear to support the theoretical assertions made

here. In general, the derived predictions basically were confirmed

85

in terms of the expected differences between conditions. The follow-
ing conclusions seem justified:

(l) The differential performance and procedural expecta-
tions that P holds for P' and 0 seem to produce an
observable behavioral difference; and

(2) the differences between the conditions in terms of
reactions to violations of normative expectations
are substantial and in the predicted direction.

Recommendations for Revisions
of Future Research

A brief discussion of some present design aspects that
could be changed and other design possibilities now will be presented.

First, it is possible that the status discrepancy between
the H-L and L-H conditions was not equal. Future revisions of the
experiment could include changing the status differentials between
the L-H and H-L conditions so that they become more equal in their
discrepancies. For example, one could leave the L-H condition as
it is and make the H-L condition a graduate and high school student.

Second, at present there is no direct measure of the benefit
of the doubt phenomenon. The scoring mechanism could be changed so
that a more direct measure of any benefit of the doubt that is given
could be obtained. For example, subjects could be instructed to put
a question mark by a word if they have any doubts about it, but the

decision to score the word would be solely theirs.

86

Third, performance outputs for the conditions were very
discrepant and might have influenced the data. The task could be
changed so that L—H and H-L subjects do not have such discrepant
performance outputs. At present, a way to accomplish this goal is
not well formulated by the researcher.

Fourth, the violations did not always reflect an increase
in magnitude. From the data we have here, the violations could be
altered to reflect more accurately any desired increase in intensity
of violations.

Fifth, the design could be changed in order to test the
following assumption: The less expected a behavior, the stronger
the reactions to it. One way to accomplish this would be to run
two different experiments with L-H and H-L subjects. In one the
instructions could include ways to deal with violations, for
example: "If you are unsure of your opponent's words you need not
score them," or, "There is no cheating allowed. If one of you per-
ceive cheating behavior you are not to give your opponent any score
for that trial." The other could include no instructions on cheat-
ing, as was the case with the experiment presented here. Comparison
of the results from these two experiments should yield some data on
the assumption stated above.

Finally, various procedural rules and their differential
effect could be incorporated into the design. This is both a design

and a theoretical problem since the kind of procedural rule that is

87

violated needs to be precisely defined to ensure the validity of
generalization concerning reactions to violations of procedural
rules.

Recommendations for Future
Extensions of the Theory_

In reference to the above, the theory could attempt to
determine which procedural rules actors consider weak, moderate,
or strong. Can murder and cheating at a game be considered equal
violations? More abstractly, will different populations differentially
evaluate loadings of importance in different procedural rules? Some
of the deviance and cross-cultural literature indicates that different
populations do just that.

What effect do positive violations of procedural expecta-
tions have upon people? Some distributive justice formulations
attempt to deal with this issue, and a modification of the theory
suggested here could be of use in answering this question.

An extension of the theory also could include the role of
possible positive or negative sanctions. Will different sanctions
modify a subject's response to violations? Another aspect would be
the relevance of the violation to the subject. For example, if a
high status violator steals a pen from a low status subject, but the
subject has a free, indeterminate supply of pens, would the subject
react as strongly as to a more "relevant" violation?

The theory might be extended to explain the discrepant per-

formance outputs between the two conditions. For example, an

88

assumption could be added stating that the L-H subjects will have
higher performance outputs than H-L subjects because they will per-
ceive a need to work harder at the task.

The theory could take into account the significance of
the ”source." For example, how important is a significant other,
the significant other being the violator, and what effect does the
significance of the source's relationship to the subject have in
terms of determining responses to violations of procedural rules?

Another aspect that merits attention deals with the
duration and nature of a relationship. What effect does the length
and kind of relationship have on reactions to violations? Would
the fact that a relationship is a longstanding one mitigate re-
actions to violations? Would a husband and wife and a student and
a professor react differently to the same violation? Would some-
one's reaction to a violation by a stranger differ from his reaction
to a violation by an acquaintance?

Perhaps the most interesting extension of the theory, both
substantively and in an applied sense, would be to determine how
various violations affect a relationship over time, that is, does
the relationship qualitatively change? The results could be related
to such variables as duration and nature of the relationship.

The theory could be extended to include sex differences.
Will reactions to a violation differ if the offender is female or
male? Is there a "combining mechanism," as Berger and Fisek (1970)

suggest, that will mean, for example, that a high status female will

89

not receive the full status of her position because a female usually
carries a low status designation? Berger and Fisek postulate that
in this situation the statuses will combine. Would a reaction to a
violation differ if the violator were a female of high status as
opposed to a male of high status? An interesting sidelight here is
that a few of the subjects in the L-H condition mentioned that they
knew their opponent was a female professor, but they kept wanting

to think of her as a male.

Finally, in order to clarify the divided literature on
reactions to high status violators, a future assumption of the theory
might deal with the relationship between the degree of a violation
emitted by a high status person and the eventual breakup of a
stabilized power and prestige order. In other words, ask the question:
At what point in an established interaction can a high status vio-
lator lose the "benefit of the doubt" that he or she has accrued?

The research presented here is in an embryonic state. So
little study has been done on this topic that incorporation of the
recommendations mentioned above should be a slow, but very rewarding,
endeavor. The researcher believes that relationship problems, such
as those between husband and wife, employee and employer, and student

and professor, frequently stem from violated expectations.

APPENDICES

APPENDIX A

COMPLETE EXPERIMENTAL PROCEDURES

APPENDIX A
COMPLETE EXPERIMENTAL PROCEDURES

Experimental Instructions

Host greets S and leads S to her seat. S is told that she is
number 1 and that she will be playing with an opponent who is
designated as number 2. She is told to watch and listen to the
television in front of her.

(Host thanks S and leaves and shuts the door)
(Experimenter's voice starts)

Hello, I am a member of the research team of social
scientists who are conducting this study. We would like to thank
you for joining us today. We hope that you will find this to be
an interesting experience.

We and other social scientists are currently conducting
a large number of studies to find out whether groups using new
kinds of communication equipment can work as well together as do
face-to-face groups working on similar problems. Today you are
participating in one of these studies. The two of you will work
on a series of problems, and you will communicate to each other
about the solutions to these problems by way of the communication
network we have designed. The closed circuit television we are
using now is a component of this system.

First, let us introduce you to each other. (S and C are
shown to each other). Participant number 1 is in an undergraduate
program at Michigan State University. Participant number 2 (show
appropriate film according to condition) is a professor at the
University of Michigan (for L-H condition) a student at Lansing
High School (for H-L condition).

(Pause 15 seconds)
(Fade video)
(Dialogue begins again)

The task that you are about to start is labeled by social
scientists the "Word Cognitive Ability Test.“ It involves being able

9O

91

to construct as many words as you can within a limited time period
from the letters which are projected on the screen. We find that

the ability is a unigue (stress unique) attribute of an individual

and that, for instance, abilities such as verbal ability or mechani-
cal ability do not determine the person's success at the task. We
also know that other specialized abilities like being good at mathe-
matics and having artistic talent are also unrelated (stress unrelated)
to word cognitive ability.

We are now ready for the experimental instructions. You
are about to participate in a study of problem solving in two-person
groups. The task involves your viewing letters on the screen in
front of you and generating as many words as possible from the
letters contained on the screen. A letter can be used in each new
word as many times as the letter appears on the screen during that
trial. Suppose that the screen contains the following letters: A,
B, C, D, E, F, G, H, I, T. (flash letters) From these letters the
following words could be constructed: (show words) HIT, HAT, BAD,
BET, BAIT, and so forth. (fade video) You are allowed to list any
words which you can think of (show rules) except foreign words,
abbreviations, slang words, words which would ordinarily be capitali-
zed, and plurals. (fade video)

Please look at your desk. On the desk you will find a pad
labeled "Word Sheet." (show word sheet) Throughout the experiment
you will record your words on this pad. When the buzzer goes off in
your room you will have forty seconds to copy down the letters from
the screen onto your word sheet and make up as many words as possible
from these letters and list them one under the other.(point to words
on the word sheet, illustrating with a pencil that they are written
one under the other) In the place provided for this on the word
sheet, (point to words on word sheet) please print all words so that
they are easily readable (point to legibily written words) for your
opponent. (fade video) Please look at the pad now for a few seconds
and acquaint yourself with this procedure. (pause for 10 seconds)
(dialogue resumes) When the buzzer goes off the second time this means
that you are to stop writing the words and put your word sheet on the
designated area of the desk so that you can see each other's sheets
on the television screen. Simply detach each sheet from your pad
and put it in the box. (illustrate detaching the word sheet and
putting it in the box) It is important that you place the word sheet
with the bottom facing the camera to your left and leave it there.
(illustrate with a pencil pointing to the bottom of the sheet and
the bottom of the box) This will allow the camera to transmit the
information on your word sheet to your opponent. (fade video) Look
at the desk and the taped area which reads "Word and Communications
Sheet." (pause 5 seconds) As soon as the buzzer goes off to signify
the trial is over put your sheet in this box. Again, this will allow
you to see each other's words so that you can score them.

92

The scoring process will begin when your opponent's words
appear on the screen. Please start this phase by recopying the
letters for that trial onto your scoring sheets. (illustrate the
copying of letters onto the scoring sheet) (fade video) This is
necessary because we are interested in discovering the average
number of possible words that can be made up from the different
letter combinations. After copying the letters from the screen,
you will copy your opponent's words onto your scoring sheet (illus-
trate the copying of words onto the scoring sheet), compute her
scores, and enter the total score for your opponent on your scoring
sheet. (point to the scoring column and the place for opponent's
score on scoring sheet) Please take a few seconds to look at your
pad called "Scoring Pad." (fade video) (pause 15 seconds)

Notice on your scoring sheet that only three-letter words
and above are allowed and that three-letter words equal 1 point,
four-letter words equal 2 points, five-letter words equal 3 points,
six-letter words equal 4 points, and seven-letter or more words are
worth 5 points per word. (point to scoring box on scoring sheet
and point to the word and points columns starting with three and
ending with seven) (fade video) Remember you are not allowed to
make up words with less than three letters in them.

To avoid confusion please do not turn the pages of any of
the scoring sheets, either backwards or forwards, until the appro-
priate trial comes up. Look only at the sheets relevant to the
trial you are working on.

When you have completed the total score for your opponent
you will write her score on the communications sheet and put it in
the box labeled "Word and Communications Sheets." (point to
communications sheets, rip one off of the pad, and put it in the
"Word and Communications Box") Look at your desk for a few seconds
and locate the communications pad. (pause for 5 seconds) (fade video)
Please use a new communications sheet on each trial. The communica-
tions sheet is simply placed over the word sheet. It is placed
directly over where the words are written. (illustrate placing
communications sheet directly over the words on the word sheet) It
is important for you to place your communications sheet and word
sheet within the appropriate taped area on the desk and facing the
camera to your left so that the camera can pick these communications
up and transmit them to your opponent. Again, please make sure that
your sheets face the camera. (point to sheets facing camera) (fade
video

After your opponent has scored you, your score will appear
on your television screen, and you are to enter it on your scoring
sheet so that you will have a record of how both you and your part-
ner did on each trial. (point to place on scoring sheet where S is
to enter her own score) (fade video) The entire scoring process

93

will take one minute. The scoring process starts as soon as you
see your opponent's word on the screen and ends with a buzzer ring-
ing twice. The next trial will begin when the buzzer rings again.

Please use sheets consecutively. In other words, for
Trial 1 use word sheet number 1 and score sheet number 1, etc.
(point to both sheets) (fade video) For reference, the word and
scoring sheets have a summation of the rules that we have already
talked with you about. Notice that there is also a summation of
the allowable word rules on your desk. (point to summation of
allowable word rules) (fade video)

You are to use the pen that is provided for you on the
desk. All words are to be legibly printed.

To avoid confusion, please do not turn the pages of any
sheets, either backwards or forwards, until the appropriate trial
comes up. Look only at the sheets relevant to the trial you are
working on.

Again, please use the scoring sheets consecutively and
fold them over after each trial is completed. Do not detach these
sheets. (emphasize)

There are two things which make this task very important
to social scientists. We know from previous studies that the capa-
city to make up as many words as possible within the time allotted
is not (emphasize not) necessarily related to specialized skills that
the individual might possess, such as mechanical, artistic, verbal,
or mathematical ability. That is, those people who might have high
skills artistically, mathematically, mechanically, or verbally are
not necessarily better at this task than anybody else. Secondly,
the success at this task seems to be dependent on a unigue (emphasize
unique) ability of the subjects to work under the time allotted and
their ability to intuitively discern patterns.

Before we start some practice trials let us review the
procedure. When the buzzer goes off you should start, and when it
goes off again forty seconds later you should stop and start to
score your opponent, then enter the score on the communications
sheet, put your communications sheet in the box, and enter your
score on your own score sheet.

You are scored on speed, accuracy, and quantity of words
that you make up. The person who wins the most points wins the game.
At the end of the twenty trials the experimenter will collect your
scoring pads and compute your total scores in order to determine who
won the game. This information will be given to you later.

94

In this particular experiment you are in separate and
identical rooms with no communication link between the rooms; hence,
if you talk you will not be able to hear each other. Remember, only
words which contain three letters or more are allowed. You are not
allowed to use foreign words, abbreviations, words which are normally
capitalized, slang, or plurals. Also remember that you can only use
each letter from the screen once in each word that you make up, and
you can only use the letters on the screen from the trial that you
ar: wgrking on to make up your words. (show rules again) (fade
v1 eo

Let us try a few practice trials. These two trials will
get (emphasize not) count on your score. Notice that the two top
word and scoring sheets are labeled "Practice." Use these for your
practice trials. When you hear the buzzer begin.

(Buzzer)

(Show letters)

(Wait forty seconds)

(Buzzer)

(Show practice confederate words)

(Score subject)

(Put subject's score in box)

(When sixty seconds have elapsed buzz twice)

Let us practice once more.

(Repeat above)

Are there any questions? If there are please press the
buzzer in your room on your desk and an experimenter will assist you.

(Experimenter goes into subject's room whether they press the buzzer
or not and asks if there are any problems or questions. Experimenter
checks the experimental material to make sure that they are following
directions. Experimenter says "All right" and leaves and shuts the
door

(The twenty trials are conducted in the same sequence as reported above)

(When the experiment is over, the experimenter enters subjects' room,
collects their materials, and gives them a semantic differential test
to fill)out. They are told to open their doors when they have finished
with it

95

WORD SHEET
A B C

1. FORM WORDS FROM LETTERS ON SCREEN FOR 40 SECONDS

 

1 COPY LETTERS FROM SCREEN

 

 

 

FORM WORDS BELOW

2. TO BEGIN SCORING DETACH AND PLACE THIS SHEET IN BOX.

3. NOW PROCEED TO SCORING SHEET.

96

SCORING SHEET

A B C

SCORING INSTRUCTIONS

 

 

COPY LETTERS FROM SCREEN

 

 

 

COPY OPPONENT'S WORDS BELOW

SCORE OPPONENT'S WORDS USING TABLE ON THIS SHEET.

MULTIPLY NUMBER OF WORDS BY POINT VALUE, E.G., 3 (NO. OF WORDS)
X 1 (POINT VALUE FOR A THREE-LETTER WORD) = 3.

 

 

 

 

 

 

 

POINT NO. OF

WORDS VALUE WORDS POINTS
3 LETTER 1 PT.
4 LETTER 2 PTS.
5 LETTER 3 PTS.
6 LETTER 4 PTS.
7 AND MORE

LETTERS 5 PTS.

TOTAL POINTS

 

 

OPPONENT'S SCORE

 

YOUR SCORE

 

 

 

 

 

97

A B C Experimenter

 

 

 

 

Postexperimental Interview

(Host enters and collects test)
(Host leads S to debriefing room)

Thank you for completing the test.

I would like to talk with you about your reactions to today's
study before we discuss the results of the study. There are factors
that affect the results, and I want to talk with you about some of
them.

Your first name is? (pause)

 

And what is your major field of study?

 

And your age?

 

1. In general, what are your feelings about the study? (probe)

 

 

2. How did you feel about your opponent? Why?

 

 

3. How did you feel about the task? (probe)

 

 

4. Have you ever participated in a study like this before?

 

(if yes, probe for a description of the study)

 

5. Have you ever read or heard about a study like this one be-
fore? (if yes, probe for a description of the study)

 

 

10.
ll.

12.

13.

98

How well did you expect to do on the "Word Cognitive
Ability" task before you actually began taking it?
Why?

 

 

 

At the same time, how well did you expect your partner to

do? Why?

 

 

Can you guess what the specific purpose of the experiment

was?

 

Can you guess what the hypotheses of the experiment might

have been?

 

Who do you think won the game?

 

Did you notice anything unusual about the experiment?

(probe, if appropriate)

 

There was deception in the experiment. Can you guess what
it was? When did you come to

that conclusion?

 

At any time during the experiment, did you think that your

opponent was cheating?

 

99

A B C Experimenters

 

 

 

wad

 

Posttask Questionnaire

Instructions: The following page contains a standard rating
scale. Your task is to use the scale to rate your reactions to a
concept. This may be a difficult task, but please do your best --
even if it means guessing on the scale. All ratings will be kept
in strictest confidence. Instructions for using the scale appear
below. Thank you for your cooperation.

 

An example of the kind of scale which appears on the follow-
ing page is the following:

PLEASANT X : : : : : : : UNPLEASANT

 

If you think that the concept that you are rating is extremel
pleasant, you are to mark an "X" in the appropriate extreme intervaI
of the scale, as illustrated above. Conversely, if you feel that
the concept is an extremely unpleasant concept, then mark the scale
gnlthe extreme interval at the other end of the scale, as illustrated

e ow:

PLEASANT : : : : : : X : UNPLEASANT

 

 

Alternatively, if you think that the concept that you are rat-
ing possesses an intermediate degree of that characteristic, then
check the appropriate intermediate interval of the scale. For
example, if you think that the concept is quite pleasant, although
not extremely so, then check the scale as follows:

 

PLEASANT : X : : : : : : UNPLEASANT

 

 

 

Or, if the concept seems only slightly_pleasant, an interval
closer to the center of the scale would be checked:

 

PLEASANT : : X : : : : : UNPLEASANT

 

The center interval should be checked only when the words at
the two ends of the scale describe your feelings toward the concept
egually well, or when the scale is completely irrelevant to the
characteristics of the concept being judged.

 

100

Please work through the scale as rapidly as possible. Make
only one mark on each scale. We are interested in your first
impressions on the scale, so please don't delay unduly on any one
item. On the other hand, please don't be careless, because your
judgments should be as accurate as possible. Please answer every
item, and always put your "X" directly over an interval (as illus-
trated above). Please turn to the next page and begin work.

101

ON THIS PAGE, PLEASE RATE YOUR OPPONENT DURING THIS STUDY.

 

 

 

 

 

 

 

 

PLEASING : : : : : : : ANNOYING
KIND : : : : : : : CRUEL
STRONG : : : : : : : WEAK

FAIR : : : : : : : UNFAIR

POSITIVE : : : : : : : NEGATIVE
FAST : : : : : : : SLOW
CLEAN : : : : : : : DIRTY

HAPPY : : : : : : : SAD

LARGE : : : : : : : SMALL

 

GOOD : : : : : : : BAD

 

102

Debriefing

Now I would briefly like to explain in a little more detail
what we are trying to study in the tests today. We are studying
the effect of how people react to unexpected events. To set up
this type of situation we need to arran e your interactions with
your opponent. In the "Word Cognitive ability" test, we were
interested in your reactions to your opponent. Her answers were
arranged so that she would be giving unexpected answers some of the

time. We are interested in your reactions to her behavior. Is this
clear so far?

This, briefly, is the reason why we had to arrange the experi-
ment this way. If you had expected her behavior we would not have
found out the information that we were interested in. Since the
study involves some fictions we would greatlyheppreciate it ifpyou
did not tell anybody_about it. They might be tested later, and,
as you can see, such information would greatly change the way they
perform and therefore the experiment could not be involved in our
data collection. Can I have_your word that you won't disclose this
information? (pause) If anyone asks you about the test,’it is all
rTght to tell them that it was a test concerning word formation.
But don't tell them about the rest. O.K.?

Thank you very much.

Obviously, because your opponent's answers were arranged, no
one really won the game.

(Ask question 13 on postexperimental interview)

We will pay you two dollars for your participation today.
(Thank subjects again and give cash and receipt)

APPENDIX B

PROCEDURE FOR RECRUITMENT OF SUBJECTS

APPENDIX B
PROCEDURE FOR RECRUITMENT OF SUBJECTS

Speech for Laboratory Study

Would you like to earn money as a participant in a sociology
lab study?

An instructor in the sociology department is looking for
participants for a laboratory study that is being conducted this
term. The study will take approximately 1-1/4 hours of your time,
and you will be given $2.00 for your participation.

The study should be a learning experience for the partici-
pants involved, and it involves no danger or physical pain.

If you would like to participate and earn $2.00, please sign
up on the sheet that is being distributed. We would really
appreciate your participation.

Thank you.

103

104

AD IN PAPER AND ON BOARDS

NEED FEMALE STUDENTS FOR
SOCIAL PSYCHOLOGY STUDY

EARN $2.00

FOR APPROXIMATELY AN HOUR OF YOUR TIME

CALL 353-5179

 

105

Hi.

This is the sociology laboratory. I'm sorry you have
to be answered by a machine, however, there's no one here right
now. If you're calling in regard to the study please leave your
name and phone number when you hear the beep, and we'll get in
touch with you.

Thank you for your response.

106
Scheduling

Hello, may I speak to (first and last name), please?

Hello (first name). I'm calling from the sociology labora-
tory. You indicated an interest in participating in one of our
studies on the form that you filled out in your sociology class.
Are you still interested?

I will need some information before I can schedule you.
What is your major? (If it is sociology or psychology, thank and
...... ) Have you ever participated in any sociological or psycho-
logical experiments before? (If yes) Could you tell me about it?
(If it was a deception study, thank her and tell her that you might
be scheduling her for a study in the future, not this quarter, and
hang up)

Fine, we have an opening on (Dex) at (Time). Can you
make it then? (If no) We have another opening on (Dex) at (lime).
Can you make it then? (If no again) What times are possible for
you? (As soon as she mentions a time that is acceptable, stop her
and place her in it)

The experiment will take place in 300 Berkey Hall. All
right? Fine, then we'll see you on (Dex) at (Time) in 301 Berkey
Hall.

LIST OF REFERENCES

LIST OF REFERENCES

Adams, J. Stacy, and Patricia Jacobsen. 1964. "Effects of Wage
Inequities on Work Quality." Journal of Abnormal and
Social Psychology 69: l9-25.

Anderson, Bo, and Robert K. Shelly. 1970. "A Replication of the
Adams Experiment and a Theoretical Formulation." Acta
Sociologica 13: l-lO.

Berger, Joseph. 1974. "Expectation States Theory: A Theoretical
Research Program." In Expeetation States Theory: A
Theoretical Research Program, edited by Joseph’Berger,
Thomas L. COnner, and M. Hamit Fisek. Cambridge, Mass.:
Winthrop Publishers.

Berger, Joseph; Bernard P. Cohen; and Morris Zelditch, Jr. 1966.
"Status Characteristics and Expectation States.“ In
Sociological Theories in Progress, edited by Joseph
Berger, Morris Zelditch, Jr., and 80 Anderson. Volume 1.

Berger, Joseph, and M. H. Fisek. 1970. "Consistent and Inconsistent
Status Characteristics and the Determination of Power and
Prestige Orders." Sociometrx 33 (September): 278-304.

Blau, Peter. 1960. "Pattern of Deviation in Work Groups." Socio-
metrx 33 (September): 245-61.

Cantril, H., and research associates. 1944. Gauging Public Opinion.
Princeton: Princeton University Press.

Festinger, Leon. 1957. A Theory of Cognitive Dissonance. Evanston,
111.: Row, Peterson.

Garfinkel, Harold. 1962. "Common Sense Knowledge of Social
Structures," In Theories of the Mind, edited by Jordan
M. Scher. New York: *FreeTPress of Glencoe.

1963. "Conceptions of and Experiments with 'Trust' as a
Condition of Stable Concerted Actions." In Motivation and
Social Interaction, Cognitive Determinants, edited by O. J.
Harvey. New York: The Ronald Press.

 

. 1967. Studies in Ethnomethodology. Englewood Cliffs,
N.J.: Prentice-Hall.

107

108

Heider, Fritz, 1946. "Attitudes and Cognitive Organization." Journal
of Psychology 21 (January): 107-12.

1958. The Psychology of Interpersonal Relations. New York:
John Wiley and Sons.

Hollander, E. 1958. "Conformity, Status and Idiosyncratic Credit."
Psychological Review 65 (March): 117-27.

1960. "Competence and Conformity in the Acceptance of
Influence." Journal of Abnormal and Social Peychologx
61 (November): 365-69.

Homans, G. 1950. The Human Group. New York: Harcourt, Brace and
World.

 

1961. Social Behavior: Its Elementary Forms. New York:
Harcourt, Brace and World}

Hughes, E. 1946. "The Knitting of Racial Groups in Industry."
American Sociologjcal Review 11 (October): 512-19.

Katz, 1.; J. M. Robinson; E. G. Epps; and P. Waly. 1964. "The
Influence of Race of the Experimenter and Instructions
Upon the Expression of Hostility by Negro Boys." Journal
of Social Issues 20 (April): 54-59.

Krech, D.; R. Crutchfield; and E. Ballachey. 1962. Individual in
Societx. New York: McGraw-Hill.

Kruse, Joan. 1972. "Effects of Experimenter's Perceived Evalua-
tions." Ph.D. dissertation prospectus, Department of
Sociology. Stanford University.

McHugh, Peter. 1968. Defining the Situation: The Organization of
Meaning in SociaT’Interaction. Indianapolis: Bobbs-Merrill.

 

Maddi, S. R. 1961. "Exploratory Behavior and Variation-Seeking in
Man." In Functions of Varied Experience, edited by D. W.
Fiske and S. R. Maddi. Homewood, Ill}: Dorsey.

Milgram, Stanley. 1974. Obedience to Authority: An Experimental
View. New York: Harper and Row.

Newcomb, T. 1953. "An Approach to the Study of Communicative Acts,"
Psychological Review 60 (November): 393-404.

1961. The Acquaintance Process. New York: Holt, Rinehart
and Winston.
Osgood, C. 1960. "Cognitive Dynamics in the Conduct of Human Affairs."
Public Opinion Dearterly 24 (Summer): 341-65.

109

Osgood, C., and P. Tannenbaum. 1955. "The Principle of Congruity and
the Prediction of Attitude Change." Psychological Review
62 (January): 42-55.

Rosenberg, M., and R. Abelson. 1960. "An Analysis of Cognitive
Balancing." In Attitude Organization and Chan e, by M.
Rosenberg et al. New Haven: Yale University ress.

Rosenthal, Robert. 1966. Experimenter Effects in Behavioral
Research. New York: Appleton-Century-Crofts.

Sabath, G. 1964. "The Effect of Disruption and Individual Status
on Person Perception and Group Attraction." Journal of
Social Psychology 64 (October): 119-30.

Sherif, M., and C. Sherif. 1956. Social Psychology. New York:
Harper and Row.

1964. Reference Groups. New York: Harper and Row.

1965. "The Adolescent and His Group in Its Setting:
Research Procedures and Findings." In Problems of Youth:
Transition to Adulthood in a Changing World, by M. Sherif
and C. Sherif. Chicago: Aldine Press.

Verba, S. 1961. Small Groups and Political Behavior. Princeton:
Princeton University Press.

Wahrman, Ralph. 1970. "High Status Deviance and Sanctions."
Sociometry 33 (December): 485-504.

. 1972. “Status, Deviance and Sanctions." Pacific
Sociological Review 13 (Fall): 229-40.

Williams, F., and H. Cantril. 1945. "The Use of Interviewer
Rapport as a Method of Detecting Differences between
'Public' and 'Private' Opinion.” Journal of Social

Psxchologx 22: 171-75.

Williams, J. A. 1964. "Interviewer-Respondent Interaction: A Study
of Bias in the Information Interview." Sociometny 27
(September): 338-52.

 

  

Typed and Printed in the U.S.A.
Professional Thesis Preparation
Cliff and Paula Haughey

144 Maplewood Drive

East Lansing. Michigan 48823
Telephone (517) 337-1527

"IIllllllllllllllllll