BEHAVIORAL STRATEGIES
IN WAGE DISTRIBUTION

Thesis for the Degree of Ph. D.

MICHIGAN STATE UNIVERSITY
IRVING MARK LANE '
' 19 70

‘ LIBRARY
Michigan State
' University

  
     

  

THESIS

This is to certify that the
thesis entitled

BEHAVIORAL STRATEGIES
IN WAGE DISTRIBUTION

presented by
IRVING MARK LANE

has been accepted towards fulfillment
of the requirements for

M; degree in Mg Y

 

Date June 25, 19‘

0-169

 

 

.I

 

 

ABSTRACT
BEHAVIORAL STRATEGIES
IN WAGE DISTRIBUTION
By

Irving Mark Lane

The present research investigated the effects of differences
in work inputs on the way subjects with total power distribute rewards
to themselves and co—workers for a task they both have performed.

The two major theoretical positions in the area of wage
distribution, the equity hypothesis and the expectancy hypothesis,
make different predictions concerning how an individual would dis-
tribute such rewards. The equity hypothesis predicts that persons
try to obtain fair wages based on their perception of both their own
and their co-workers' inputs and outcomes. On the other hand, the
expectancy hypothesis predicts that persons try to obtain thegreatest
possible rewards for themselves.

In order to determine the conditions under which the equity
and profit maximization motives tend to dominate, four research

situations were used.

Irving Mark Lane

Conditon 1. The person determining the wages (the
chooser) had high inputs, and his co-worker (the receiver)--who
received whatever wages the chooser allotted to him--had low
inputs.

Condition 2. The chooser had high inputs and the receiver
also had high inputs.

Condition 3. The chooser had low inputs and the receiver
also had low inputs.

Condition 4. The chooser had low inputs and the receiver
had high inputs.

First, the subjects in the four conditions completed an
Industrial Opinion Questionnaire. Half the subjects were given a
form of the questionnaire that required 1% hours to complete; these
subjects had high inputs. The remainder of the si1bjects, those with
low inputs, were given a shorter form of the questionnaire requiring
% hour to complete.

After the subjects completed the Industrial Opinion Ques-
tionnaire, they were given the Wage Distribution Booklet. This
booklet contained six sets of two alternative distributions. In a
given set, the two alternatives (the standard distribution and the
alternative outcome) each designated payments to both the chooser

and the receiver. The standard distribution always had one person

 

Irving Mark Lane

receive $4 and the, other person $1. In three of the sets (Series X),
the chooser received the $4 payment if he selected the standard
distribution; in the remainder (Series Y), the standard distribution
awarded $4 to the receiver. The alternative outcome always dis-
tributed the payments more equally than the standard distribution.
In addition, the alternative outcomes differed systematically in
terms of relativewelfare. (Is the payment of the chooser greater
than, equal to, or less than that to the receiver?)

All subjects--both choosers and receivers—~completed the
Wage Distribution Booklet. Each chooser knew that one of his
choices would be selected at random to determine both his payment
and that of the receiver with whom he was paired. The receivers
knew that their choices did not influence rewards.

The four situational factors systematically varied were:
(a) inputs of the subject, (b) inputs of his co-worker, (c) real versus
hypothetical choices, and (d) order of presentation (whether the
chooser responded first to the three Series X alternatives or first
to the three Series Y alternatives).

The responses to the sets of alternatives in the Wage Dis-
tribution Booklet were combined to produce scores for each subject
on three personal dispositions: (a) self—interest-—the degree to

which the subject maximized his own reward, (b) altruism--the

 

Irving Mark Lane

degree to which the subject maximized his co-worker' s reward, and
(c) role symmetry--the degree to which the subject was concerned
with a general principle of allocation (e. g. , equality) rather than a
specific distribution of rewards.

Subjects made no altruistic responses. Therefore, role
symmetry and self—interest were perfectly negatively correlated.

Taken together, the results for inputs of the subject and
inputs of his co-worker provided insight into those conditions which
generate concerns with either fairness (the equity hypothesis) or
profit maximization (the expectancy hypothesis). There was strong
evidence that a person, in distributing rewards to himself and
another individual, only considered the other' 5 inputs. Subjects
tended to give themselves as much of a reward as possible unless
they perceived that their partners had worked hard enough to deserve
more than minimal pay. It was as if the inputs of the co—worker
had to exceed a threshold value before a subject felt that his partner
deserved more than this minimum. If this threshold value was
exceeded, then equity motives tended to dominate. On the other
hand, if the inputs of the co-worker were less than this value, sub-
jects tended to maximize their profits. Thus, these findings have
significant implications for both the equity and expectancy hypotheses.

The variable of order of presentation had a significant

effect on the distribution of wages. Subjects who were presented

Irving Mark Lane

Series Y first tended to be less role symmetric. On the other hand,
the variable of real versus hypothetical choices had very little
effect. In addition, the results of the Guttman scale relating to
relative welfare indicated that it was easiest for subjects to give up
some of their potential wages to their co—workers when the resultant
distribution still gave them more money.

Thus, the results indicated that co-worker input is an
important determinant of the way in which rewards are distributed.
However, they also indicated that other factors besides inputs must
be considered if a comprehensive theory concerning the distribution

of rewards is to be developed.

Approved: ($6,th ”(/7 .. /
Date: 725% #2742

Dissertation Committee:
L. A. Messé, Chairman
C. Hanley
E. H. Jacobson
J. L. Phillips

J. H. Wakeley

BEHAVIORAL S TRA TE GIE S

IN WAGE DISTRIBUTION

By

Irving Mark Lane

A THE SIS

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

DOCTOR OF PHILOSOPHY

Department of Psychology

1970

 

To my wife, Myrna

 

ACKNOWLEDGMEN TS

I sincerely wish to thank my chairman, Dr. Lawrence A.
Messé, for the encouragement, guidance, and understanding he has
provided while this research was being conducted. In addition to his
very significant contribution to both this research and my professional
training, his altruistic behavior has been a constant inspiration.

I would also like to thank the other members of my disser-
tation committee, Dr. Charles Hanley, Dr. Eugene H. Jacobson,
Dr. James L. Phillips, and Dr. John H. Wakeley. Not only have
they provided help during this research, but each one has made a
significant contribution to some phase of my graduate education.

I would also like to thank Dr. Lawrence 1. 0' Kelly for pro-
viding the necessary funds to conduct this research.

In addition, I would also like to thank Ronald Cohen and
Melvin Goldstein for running the subjects.

Finally, I would like to express appreciation to both my
wife, Myrna, and to my chairman' 5 wife, Susan Messé, for their
continued understanding regarding the long periods of time they

were without their husbands as a result of this research.

 

TABLE OF CONTENTS

LIST OF TABLES
Chapter
I. / INTRODUCTION

Review of the Literature
/ The Present Research

II . ME THOD

/ The Sample

/ Instruments .

/ Variables and Experimental Design
I Procedure . . . .

III. RESULTS

,. Effects of Situational Variables on the
Manifestation of the Personal
Dispositions . .

/ Further Analyses of the Effects of
Inputs of Subject and Inputs of
Partner . -

/ The Effect of Relative Welfare .

IV. DISCUSSION .
1 Inputs of Subject and Inputs of Partner

Relative Welfare .
/ Effect of the Control Variables

iv

Page

Vi

77
84
93
94

94
100

. 101

 

 

Chapter Page

V. CONCLUSION . . . . . . . . . . . . . . . . . . 106

,4 Implications of the Results for Wage
Distribution Theory . . . . . . . 106
, Implications of the Present Methodology . . . . 107
, Suggestions for Further Research . . . . . . . 109
/ A Concluding Statement . . . . . . . . . . . . 111
/REFERENCES...................... 112

Appendix

I. WAGE DISTRIBUTION BOOKLET . . . . . . . . . 115
II. ATTITUDE QUESTIONNAIRE . . . . . . . . . . . 122
III. INDUSTRIAL OPINION QUESTIONNAIRE . . . . . . 125
IV. STUDENT BACKGROUND QUESTIONNAIRE . . . . 139

V. GUTTMAN SCALED ROLE SYMMETRY
RESPONSES................... 140

Table

LIST OF TABLE S

The choice situations .
A sample work situation

Summary of the analysis of variance for
role symmetry scores

Cell means for role symmetry scores

Percent scale-point responses for work
quantity tabulated within each experimental
condition

Percent scale-point responses for work
quality tabulated within each experimental
condition

Percentage of role symmetry and self-
interest responses

Number of subjects manifesting a majority
of each personal disposition classified by
experimental condition

t_va1ues for comparison of percentages of
role symmetry responses occurring
between experimental conditions .

Page
57

58

79

80

84

85

88

89

92

I. IN TRODU C TION

Adams (1965) notes that the just distribution of wealth,
power, goods, and services in society traditionally has been the
concern of philosophers, political scientists, politicians, jurists,
and economists. Through 1966, psychologists and sociologists had
displayed little interest in this area. 1 Since then, however, a
growing number of empirically—oriented social scientists have
turned their attention to situations where rewards of a number of
kinds, but especially wages, are distributed. This growing interest
no doubt stems from some degree of appreciation of the possible
benefits, both to basic theory and to applied problems, which
knowledge in this area could provide. For example, such knowledge
should contribute to the basic understanding of the process of social
exchange which underlies all human interaction. On the other hand,
such knowledge when applied to problems in industry should also
provide guidelines for wage distribution plans that employees would

consider fair. Valid guidelines would be of great benefit in this

 

1
Homans and Adams are two notable exceptions to this
generalization.

era where employee satisfaction has become a management objec-
tive.

The present study is part of this growing interest, since
its focus was the investigation of the senses of fairness in wage dis-
tribution. To help understand the nature of this research, imagine
the following situation: An individual is asked to distribute the
wages to himself and another individual for a task they both have
performed. The question arises of how the individual would distrib-
ute the wages if given the opportunity to do so. The present research
was designed to provide an answer to this question.

Such an answer is important for assessing the influence of
the sense of fairness on the distribution of wages, a problem of
some theoretical interest at the present time. As will be seen, the
two major hypotheses in the area of wage distribution, the equity
hypothesis and the expectancy hypothesis, make different predictions
concerning how an individual would distribute the wages if given an
opportunity to do so. The present research, however, is not seen
as a definitive test of the two hypotheses. Rather, it was performed
to determine to what extent and under what conditions each of the
hypotheses contributes to the understanding of how persons distrib-
ute rewards.

In order to better understand the contribution of the present

research to the psychology of wage distribution, a review of the

theoretical work in this area is presented below, followed by a

summary of the research relevant to each theoretical position.

Review of the Literature

 

Theoretical Work in the Area
of Wage Distribution

Relative deprivation

 

Most of the research concerning fairness in wage distribu-
tion is an outgrowth of the studies on relative deprivation. This
concept was initially proposed as an explanatory tool in the first
volume of The American Soldier (Stouffer e_t_a_11. , 1949). Stouffer

et a1. did not formally define the concept, nor did Merton and Kitt

 

(1950), who analyzed the implications of the concept for reference
group theory. However, its essential meaning can be inferred from
the example given below.

Despite the objective fact that soldiers with a high school
education had better opportunities for advancement in the Army,
they were not as satisfied with their prospects for advancements
as were less educated men. This apparent paradox can be explained
by assuming that the better educated men had higher levels of aspi-
ration. These levels of aspiration presumably were based on what

would have been realistic advancement expectations in civilian life.

Therefore, they were relatively deprived of opportunities for

 

advancement .

Adams (1965), in reviewing the survey and experimental
data concerning relative deprivation, noted that four main conclu-
sions emerge. They are:

1. Dissatisfaction is a response to perceived injustice, and is
not a direct response to perceived relative deprivation.
Relative deprivation is a condition which elicits the feelings
of injustice. In turn, these feelings of injustice trigger the
expression of dissatisfaction. Therefore, injustice medi-
ates the effects of relative deprivation.

2. That which is perceived to be just is based on relatively
strong expectations.

3. A comparative process is inherent in the development of
expectations and the perception of injustice. This is
implied by the term relative deprivation.

4. Felt injustice is a response to a discrepancy between what

is perceived to be and what should be.

Distributive justice

 

The existence of relative deprivation raises the question of

the fair share of rewards, or in other words, distributive justice.

The concept of distributive justice was formulated in economic
terms by Homans (1961). The concept of distributive justice is
most easily understood in terms of a dyadic relationship between
Person and m. Person is any individual. M is an individual
with whom Person is in an exchange relationship, or with whom
Person compares himself when both he and 9113 are in an exchange
relationship with a third party (e. g. , an employer). Oier is usu—
ally a different individual but may be Person in another social role.
According to Homans, distributive justice for Person and
OM is realized when the profits they each receive are propor—
tional to their investments. A state of distributive justice for

Person and Other can be represented by the following equation:

Person' 8 Profits Otherl s Profits

 

 

Person' 8 Investments : Other' 5 Investments

Homans defined profit as that reward which is received in
an exchange relationship minus the cost incurred. A cost is defined
as those alternative rewards which are forfeited by participating in
an exchange relationship. Investments in an exchange are the rele-
vant attributes that each individual brings to the relationship.
Examples of investments in an industrial setting are the employee' s
previous job experience, educational background, and the amount of

physical effort he expends while performing his job.

Distributive justice for Person and @323, when they are
two individuals being rewarded by a third party (e. g. , an employer),
is realized when the profit each receives is proportional to his
investments.

Homans stated that when distributive justice is not realized,
a state of injustice (or inequity) exists; and if this injustice is to
Person' 3 disadvantage, Person will display the emotional behavior
called anger. Adams (1965) claimed that what Homans meant by
anger is really the feeling or expression of dissatisfaction. Further—
more, Homans stated that if distributive justice is not realized and
is to Person' 8 advantage, Person will feel guilty. Homans asserted
that the threshold for displaying dissatisfaction is lower than that
for displaying guilt. Therefore, conditions of inequity must be
greater when it is favorable to Person before he realizes it than
when it is to his disadvantage.

The concept of distributive justice is a significant advance—
ment over the concept of relative deprivation, since it specifies to
a greater degree the variables which mediate injustice. According
to distributive justice, Person feels injustice not merely because
his rewards are less than he expects (or feels is fair), but because
when he compares his profit—investment ratio to Other' 3, he per—

ceives that his ratio is smaller.

The equity hypothesis2

Adams' (1963a, 1965) equity hypothesis, although consid-
erably more. elaborate, is built upon distributive justice, and is
consistent with it. Adams noted that whenever two individuals
exchange anything, there is the possibility that one or both of them
will find that the exchange was inequitable. One frequent case of
perceived inequity occurs when a man exchanges his services for
pay. On the man' 8 side of the exchange are his work experience,
intelligence, training, sex, etc. These are what the man perceives
as his contributions to the exchange; and for them, the man expects
a just return. Adams used the term ”inputs" to refer to what
Homans called "investments"; that is, those qualities which are
perceived to be a man' 8 contribution to the exchange. Adams empha-
sized that these inputs are as perceived by their contributor; they
are not necessarily identical with those perceived by the other
party in the exchange.

Adams introduced two distinct characteristics of inputs:
recognition and relevance. The existence of an attribute may be

recognized only by its possessor, only by the other individual in the

 

2There have been attempts on the part of some to elevate
both this and the expectancy hypothesis (see below) to the status of
formal theories. However, strictly speaking, both are of such
limited scope that it seems reasonable to refer to both of them as
hypotheses.

 

exchange, or by both. Whether or not such an attribute is considered
an input is dependent upon the possessor' s perception of its relevance
to the exchange. If he perceives it to be relevant, and therefore, if
he expects a just return for it, then it is an input. Problems of
inequity arise if the possessor of an attribute considers it relevant
to the exchange, but the other party considers it irrelevant and acts
accordingly.

On the other side of an exchange is an individual' 8 receipts.
Adams called these outcomes. In an employee-employer exchange
they include pay, job status, fringe benefits, etc. In a manner
analogous to inputs, outcomes are as perceived by their receiver.
They are also characterized in terms of recognition and relevance.

Inputs and outcomes are correlated, but imperfectly so.
It is because they are not perfectly correlated that there is a con-
cern with inequity. There exists normative expectations of what is
a "fair correlation” between inputs and outcomes. These expecta-
tions are learned during the process of socialization and are based
upon observation of the correlations for a reference person or
group. Although Adams acknowledged that it is important to specify
those variables which determine the appropriate reference group or
person, he conceded that such a specification was beyond the scope

of his hypothesis.

When the normative expectations of the individual making
social comparisons are violated, i. e. , when he finds his inputs and
outcomes are not in balance in relation to those of his comparison
individual (Qt—ha), feelings of inequity result.

Adams (1965, p. 280) states that "inequity exists for E
SE whenever he perceives that the ratio of his outcomes to inputs
and the ratio of Other' s outcomes to Other' 5 inputs are unequal. "

Schematically, inequity exists whenever:

Perceived Outcomes Person Perceived Outcomes Other

 

<
Perceived Inputs Person Perceived Inputs Other
or whenever:

Perceived Outcomes Person Perceived Outcomes Other

 

>
Perceived Inputs Person Perceived Inputs Other
Likewise, a condition of equity exists whenever:

Perceived Outcomes Person Perceived Outcomes Other

 

Perceived Inputs Person : Perceived Inputs Other

Up to this point the equity hypothesis is largely a clarifica—
tion of distributive justice. However, Adams also broadens the
theoretical basis of distributive justice by specifying the conse—
quences of perceived inequity in terms of dissonance theory (Fes—

tinger, 1 957).

 

10

Adams notes that the cognitive changes, "anger, " ”guilt, "
or ”dissatisfaction, " generated by perceived inequities should be
accompanied by some behavioral manifestation. Moreover, the
intensity of these manifestations is determined by mechanisms sum—
marized in two postulates (derived from dissonance theory).

1. The presence of inequity in Person creates tension in him.
The tension is proportional to the magnitude of the inequity
present.

2. The tension created in Person will motivate him to elimi-
nate or reduce it. The strength of the motivation is pro—
portional to the tension created.

From these postulates, and from Festinger' s theory of
cognitive dissonance, Adams has identified six methods by which
Person can reduce or eliminate the dissatisfaction associated with
inequity.

1. Person altering his inputs. Person may vary his inputs,
either increasing them or decreasing them, depending on
whether the inequity is advantageous or disadvantageous.
When the inequity is to Person' 3 disadvantage, he may,
for example, engage in restriction of output. When the
inequity is to Person' 3 advantage, he may increase his

productivity or the quality of his work.

 

Person altering his outcomes. Person may attempt to vary
his outcomes, either decreasing or increasing them depend-
ing on whether the inequity is advantageous or disadvanta—
geous. An example of an attempt to increase outcomes is
the wage demands in union-management negotiations.
Adams stated that decreasing outcomes to reduce inequity
is an unlikely occurrence.

Person distorting his inputs and outcomes cognitively.
Person may cognitively distort his inputs and outcomes.
The direction of the distortion would be the same as if he
had actually altered his inputs and outcomes. Adams stated
that since most individuals are influenced by reality, sub-
stantial distortion is generally difficult. However, there
is substantial literature relating to cognitive dissonance
that has demonstrated that individuals do distort reality in
order to mediate cognitive conflict.

Person leaving the field. Leaving the field involves cutting
off the social relationship. Quitting a job or absenteeism
are two frequently used forms.

Person acting on M. Person may attempt to alter or
cognitively distort Other' 3 inputs and outcomes, or he may

try to force Other to leave the field.

12

6. Person changing the object of his comparison. Person may
change 9m with whom he compares himself when he
experiences inequity and he and 913% are in an exchange
relationship with a third party. This-method of reducing
inequity is not applicable when Person and M are in a
direct exchange. Changing the object of comparison when
Person and M are in a direct exchange reduces to

Person leaving the field.

The expectancy hypothesis

 

Lawler has proposed the expectancy hypothesis as an alter-
native to the equity hypothesis. The expectancy hypothesis, de—
scribed below, is considerably different from distributive justice
and the equity hypothesis, and the predictions derived from it are
in most cases directly opposite to those derived from the other two.

Lawler (1968b) stated that the expectancy hypothesis and
drive theory are the two most widely accepted theoretical positions
in the area of human motivation. Drive theory stems from the work
of Hull and Spence, while the expectancy hypothesis can be traced
back to the works of Tolman and Lewin. More recent, refined
forms of the hypothesis have been advanced by Porter and Lawler
(1967), and Vroom (1964). Drive theory and the expectancy hypothe—

sis are very similar. Atkinson (1964), after an extensive literature

review, concluded that in almost every situation they lead to
similar predictions.

As Lawler (1968b) notes, the expectancy hypothesis empha-
sizes that the motivation to perform an act is a function of both the
strength of the expectancy that the act will be followed by a given
outcome (reward or goal) and the value (valence) of that outcome to
the individual. To him expectancies are subjective probabilities.
They have values ranging from 0 (no likelihood that the act will be
followed by outcome) to 1 (certainty that the act will be followed by
outcome). Expectancies are subjective in nature and, therefore,
may not be in accord with actual probabilities. Outcomes take
values ranging from +1 (very desirable) to -1 (very undesirable);
like an expectancy, valence is subjective in nature. Therefore, a
particular outcome may very well have different valences for dif—
ferent people, and it may vary in valence for the same person from
time to time. Expectancy and valence combine multiplicatively in
determining motivation. Thus, both expectancy and valence must
be greater than zero in order for motivation to be present. It is
also important, as Lawler notes, to consider all the possible out-
comes that might be perceived to depend upon a given act. The
products of all the Expectancy X Valence combinations need to be
summed in order to determine the total amount of motivation

present in a given situation.

14

Lawler (1968b) presented an example of the kind of predic—
tion the expectancy hypothesis makes in a specific situation. Accord-
ing to the expectancy hypothesis, a person should be motivated to
work hard and produce to the extent that he feels working hard is
likely to result in his receiving outcomes that he values. Therefore,
in a properly designed piece rate situation, where large rewards are
related to productivity, a person should be motivated to work hard
provided that (a) he values the reward (money) and (b) he sees that
there is a connection between his hard work and his receiving the
reward. However, it is necessary to consider all the outcomes that
might be related to working hard. The effects of a piece rate system
might be negative if undesirable outcomes (e. g. , social ostracism)
are seen as being related to high productivity.

This point is critical for the expectancy hypothesis. In a
piecework situation, an individual can be predicted to be a high
producer if money has a positive valence for him or a low producer
if money has a negative valence for him. Therefore, contrasting
results can be explained post hoc in such a way that they all appear
consistent with the expectancy hypothesis. From this it would seem
that no evidence could be found which would invalidate the expectancy
hypothesis, and therefore, its usefulness as a predictive tool is

highly questionable. However, Lawler (1968) attempted to increase

15

the usefulness of the expectancy hypothesis, at least in the area of
wage distribution, by providing more structure to the concept of
valence, which he points out has been notoriously vague. He
assumed that an increase in the amount of a reward (like money)
should always increase its attractiveness, and therefore, an incre—
ment of reward can never have negative valence.

As an overall summary, the expectancy hypothesis empha—
sizes a motivation to maximize positive outcomes. In situations
where rewards are being distributed, this motivation leads persons
to attempt to maximize their wages. In contrast, as noted above,
the equity hypothesis emphasizes persons trying to obtain fair
wages based on their perception of an input-outcome ratio compari-
son.

Research Relevant to the Hypotheses
in the Area of Wage Distribution

Research support for
the equity hypothesis

 

 

There is a considerable body of research related to both
Homans' concept of distributive justice and Adams' equity hypothe—
sis. These studies fall into two broad categories: naturalistic
studies conducted in real life industrial settings; and experimental

studies, for the most part conducted in the laboratory.

16

Naturalistic studies

An early study relevant to distributive justice was conducted
by Homans (1953). The study was concerned with two jobs within
one division in a company. One was cash posting, a single repeti—
tive job requiring little thought but some degree of physical mobility.
The other job was that of a ledger clerk which entailed performing
a number of non—repetitive tasks requiring some thought but little
physical mobility. Ledger clerks were older than cash posters and
had worked longer with the company. The company considered
being a ledger clerk a higher level job, although being promoted
from cash poster to ledger clerk brought no increase in pay.

The study was conducted in three phases.

Phase 1. The researcher received permission from both
union and management to observe activity in the division. He then
moved into the back of the room in which the division worked and
spent a month allowing the workers to become used to his presence.

Phase 2. A systematic observation of interaction in the
room was undertaken for a two—week period.

Phase 3. Each member of the division was interviewed for
approximately two hours.

The results indicated that the ledger clerks felt that they

deserved to have higher pay than the cash posters, since their jobs

17

were more responsible. In other words, the ledger clerks felt
inequity. Their inputs were higher than those of the then cash
posters. They were older, more experienced and had a more re-
sponsible job. However, some of their outcomes, such as salary,
were not higher. The ledger clerks attempted to reduce their per-
ceived inequity by complaining to their union representative. In
other words, they tried to alter their outcomes by asking for an
increase in salary.

Clark (1958) studied checkout counters in a chain of super—
markets. These counters were manned by two persons: ringers,
who operate the cash register; and bundlers, who put the goods in
bags. Within the social structure of the supermarket, ringing was
considered a higher status, better paid job that was usually handled
by full-time employees. Bundling was of lower status, and this task
was usually performed by part—time employees, frequently students.
During rush hours in the store, a bundler often would be directed
in his work by a ringer, whose status in non-job related areas, such
as education, was lower than his own. This created in many bun-
dlers feelings of inequitable treatment. These bundlers felt they
had higher inputs than the ringers but that their outcomes were
lower. Through interviewing, Clark determined that the inequity

that existed was actually perceived by both ringers and bundlers.

18

To reduce the inequity, the higher status individuals purposely
slowed down their rate of bundling. By bundling slowly they reduced
their inputs and by doing so succeeded in reducing their feelings of
inequity.

Sayles (1958) stated that fairness of wage distribution for
industrial work groups was figured on the basis of the following
equation.

Our importance in the plant _ Our earning§_
Any other group' s importance - Their earnings

 

 

When equality was attained, satisfaction was experienced; on the
other hand, an inequality between the ratios produced dissatisfaction
and pressures for redress. Sayles' findings provide support for the
equity hypothesis.

Patchen (1961) conducted a field study in which he inter-
viewed approximately five hundred non-supervisory workers. He
performed the research with the belief that comparisons of pay are
not made in a cognitive vacuum. Rather, he felt that such compari-
sons become meaningful to individuals only when they also compare
themselves to others on dimensions that are believed to be bases of
pay. Examples of these dimensions are skill, seniority, education,
etc. In effect, Patchen believed that the individual makes a cognitive

relation of the following type:

My position on dimensions
related to pay
His (their) position on
dimensions related to pay

My pa
His (their) pay

 

compared to

Within the framework of Festinger' s theory of cognitive
dissonance, two major hypotheses were generated.

Hypothesis 1. Wage comparisons which are objectively
dissonant will be judged as unsatisfactory by the person who makes
such comparisons.

Hypothesis 2. Men who are satisfied with specific wage
comparisons will explain their satisfaction in terms of a consonance
between relative wage standing and relative standing on attributes
relevant to pay; men who are dissatisfied with specific wage com—
parisons will explain their dissatisfaction in terms of a dissonance
between relative wage standing on attributes relevant to pay.

These two hypotheses were confirmed.

Another part of Patchen' 5 study was concerned with deter-
mining the factors that influence an individual' 3 choice of a wage
comparison individual. While this concern is relevant to the equity
hypothesis, his results in this area were not statistically significant,
but were promising enough to indicate future research would be
desirable.

Jacques (1961, 1962), in conducting interviews with indus—

trial workers, noted a characteristic pattern of responses as a

20

function of their level of payment. He stated that a worker whose
compensation is at the equitable level feels he is receiving fair pay-
ment. By contrast, deviations in payment below the equitable level
are usually accompanied by feelings of dissatisfaction which become
stronger as the deviation increases. At ten percent below the
equitable level of payment, there are active grievances and an actual
desire to leave the job or take collective action. On the other hand,
deviations above the equitable level are accompanied by feelings of
being relatively well off as compared with others. At ten to fifteen
percent above the equitable level of payment, there is a sense of
preferential treatment and a feeling of unease about the relatively
advantageous position. These results suggest that it is not neces-
sarily true that each worker is out to get as much as he can for
himself. Instead, workers appear to have a strong desire that each
one should earn the right amount, a fair and reasonable amount
relative to others.

The studies cited above offer preliminary evidence to sug-
gest that the desire to achieve an equitable state appears to also

hold for wage distribution in industrial settings.

Experimental studies

The next three studies to be discussed were performed by

Adams, the formulator of the equity hypothesis. Since the method

21

employed in each was quite similar (and has since been adapted by
many other researchers in the field), only the first will be described
in considerable detail.

Adams and Rosenbaum (1962) studied the consequences of
perceived inequity on quantity of productivity. The experiment was
performed to test two hypotheses derived from the equity hypothesis.

Hypothesis 1. If Person is paid by the hour, his produc-
tivity will be greater when his outcomes are perceived as too great
than when identical outcomes are perceived as equitable.

The rationale for hypothesis 1 is as follows: if other means
of reducing inequity are unavailable, increasing productivity will
increase Person' 5 perceived inputs and thus bring them into line
with both his perceived outcomes and Other' s perceived inputs and
outcomes.

Hypothesis 2. If Person is paid on a piecework basis,
his productivity will be smaller when his perceived outcomes are
too large than when the same conditions are perceived as fair.

The rationale for hypothesis 2 is as follows: because
inequity is associated with each unit produced, inequity will increase
as work proceeds. Therefore, Person will strive not so much to
reduce inequity as to avoid increasing it and he will, therefore,

restrict production.

22

The two hypotheses were tested in the following manner.
Students were hired through the Placement Service for part—time
temporary work as interviewers. The subjects were given the
impression that interviewing would take place for an extensive
period of several months. When the subjects reported to their
prospective employer they were randomly assigned in equal num-
bers to the experimental condition (inequity condition) or to the
control condition. They were asked to fill out a questionnaire
requesting, among other things, their previous employment history.
The experimenter then studied the subject' s background in his
presence and, depending on the condition to which he had been
assigned, induced the perception that the subject was qualified or
that he was not.

The subjects in the inequity condition were told, ”You don' t
have any (nearly enough) experience in interviewing or survey work
of the kind we're engaged in here. I specifically asked the Place—
ment Service to refer only people with experience. . .I can' t under-
stand how such a slip—up occurred. . .I guess I' ll hire you anyway
. . .I' 11 just have to pay you at the advertised rate of $3. 50 per hour.
On the other hand, the subjects in the control condition were told,
"We can use you for this work. You meet all the qualifications for
the job. . .We pay $3.50 per hour. This rate is standard for work

of this kind performed by people of your qualifications. "

23

The experimenter then informed each subject about the
interviewing task. Interviewing was chosen because it was brief
enough so that many interviews could be obtained during the sched-
uled work period of 2. 5 hours. When the subjects returned from
the interviewing they were questioned by the experimenter. The
questioning revealed that the subjects uniformly believed that they
had been hired and had worked on a real job. The results of the
experiment confirmed the predictions of hypotheses 1 and 2, thereby
providing substantial support for the equity hypothesis.

Adams (1963b) noted that the subjects in the piecework
condition in the previous study (Adams and Rosenbaum, 1962) could
have produced less when overpaid than when equitably paid for a dif-
ferent reason than originally believed. It was believed that the
lower productivity was an attempt by the overpaid pieceworkers to
avoid increasing inequity by restricting output. It is entirely pos—
sible, however, that subjects reduced inequity by increasing their
effort on the production of each unit; that is, they increased the
quality of their work. This would have the effect of increasing the
production time per unit and, therefore, overall productivity would
decrease. In terms of the equity hypothesis, this explanation
assumes that pieceworkers reduce their inequity by increasing
their inputs , in a manner similar to the hourly workers in the

Adams and Rosenbaum study.

 

24

However, the task used in the Adams and Rosenbaum study
did not permit measuring quality of work. Adams (1963b) replicated
Adams and Rosenbaum' 5 study except that the interview was designed
to permit measurement of both quantity and quality of work.

The specific hypothesis tested was: Pieceworkers who per-
ceive that they are overpaid will produce less work of better quality
than pieceworkers who are paid the same rate and perceive they are
equitably paid. The results supported the hypothesis.

Adams explained the findings of the two previous studies in
terms of the arousal of inequity and its subsequent reduction. How-
ever, Adams and Jacobsen (1964) stated that the experimental design
previously used did not rule out the possibility that the results
obtained were a function of perceived job insecurity and subsequent
attempts to reduce it. Therefore, they designed an experiment to
assess the importance of feelings of job insecurity.

One independent variable manipulated by Adams and
Jacobsen was dissonance. This variable has three levels.

1. High dissonance (HD). Subjects in this condition were
induced to perceive that they were comparatively unquali-
fied. They were told that they would nevertheless be paid
at the advertised piece rate of $. 30.

2. Reduced dissonance (RD). Subjects in this condition were

told the same thing as subjects in the HD condition except

25

that they were told that they would be paid at a piece rate
of $. 20.

3. Low dissonance (LD). Subjects in this condition were
induced to perceive that they were fully qualified for the
advertised piece rate of $. 30.

The other independent variable manipulated was job pros-
pects. This variable had two levels.
1. High prospects (HP). Subjects in this condition were told
that their work would continue for several months and
therefore they might be employed again.
2. Low prospects (LP). Subjects in this condition were told
that the job was a one time only affair-with the implication
that the subject' 5 service would not be required again.
The experimental method used in this study was very
similar in all respects to that used by Adams in his two studies
described above. The major difference was that the task the sub-
jects performed was changed to correcting page proofs. Work
quality was measured by the number of errors detected.

The hypotheses tested were as follows:

Hypothesis 1. Subjects in the HD conditions will produce
higher quality work than subjects in either the LD or RD conditions.
Subjects in the RD and LD conditions will produce work of similar

quality.

26

This first hypothesis is consistent with the equity hypothesis.
The rationale behind hypothesis 1 was previously discussed in ref—
erence to the Adams and Rosenbaum (1962) study.

Hypothesis 2. Subjects in the HP conditions will produce
higher quality work than subjects in the LP conditions. Also, there
will be a significant interaction between job prospects and disso-
nance. Specifically, the HD —HP condition subjects will produce a
higher quality product than the HD—LP condition subjects.

This second hypothesis is consistent with the belief that
the subjects were motivated to reduce their feelings of job insecu-
rity.

Adams and Jacobsen stated that subjects reported that the
manipulations of the two independent variables were effective. The
results of the study were clear cut. Hypothesis 1 was confirmed.
Subjects in the HD conditions found a significantly greater number
of errors per page than the subjects in the RD or LD conditions.
Hypothesis 2 was not confirmed. No significant main effect of job
prospects was found, and no significant interaction effects were
found. The results of this study support an equity hypothesis
interpretation of the results obtained by Adams in his other studies,
and failed to support the alternative interpretation that the results

obtained were a function of the subjects' insecurity about job tenure.

27

Arrowood (1961) designed an experiment in which subjects
were either overpaid or equitably paid on an hourly basis, and per-
formed their work under either public or private conditions. The
equitable-inequitable pay distinction was achieved using the same
method as Adams. The public-private distinction was achieved by
having subjects either submit their work to the ”employer" (the
experimenter) or mail it in a preaddressed envelope to New York.
In the latter case, the subjects were under the impression that the
experimenter would never see their work. The results obtained
were entirely consistent with those of Adams and Rosenbaum. In
both public and private conditions, overpaid subjects produced sig-
nificantly more than equitably paid subjects. There was no signifi—
cant interaction between the inequity—equity and public-private
dimensions.

Unlike the previously cited studies which produced inequity
by the manipulation of perceived job inputs (qualifications for the
job), Andrews (1967) produced inequity by the manipulation of
assigned piece rate. Through pretesting, equitable, overpay and
underpay piece rates were determined. Two major hypotheses
derived from the equity hypothesis were tested. They were:

Hypothesis 1. As compared to workers assigned an

equitable piece rate, workers assigned to a lower rate will tend to

 

28

produce more work of lower quality, whereas workers assigned to
a higher piece rate will tend to produce less work of higher quality.

Hypothesis 2. Within groups, the more that present wage
potential exceeds previous high wage, the lower the work quantity
and the higher the work quality. The more the present wage poten-
tial is less than previous high wage, the higher the work quantity
and the lower the work quality.

The basis for hypothesis 2 was the assumption that a
worker' s perception of his present pay is influenced by pay rates
he has received on other jobs. This emphasizes Adams' position
that inputs and outcomes must be evaluated in terms of the worker' s
perception of them, rather than in terms of reality.

The results of Andrews' study generally confirmed his two
hypotheses. The results were in the predicted direction, but not all
the differences were significant.

Goodman and Friedman (1968) tested three major hypotheses
derived from the equity hypothesis.

Hypothesis 1. Subjects who perceive they are overpaid and
who perceive quantity of production as the basis for equity resolu-
tion will produce more units in the hourly condition than subjects
who perceive they are equitably paid. This is Adams' original
hypothesis, but the task used, questionnaire scoring, was different

from that used by him.

 

29

Hypothesis 2. Subjects who perceive that they are overpaid
will increase quality (or quantity) of their inputs if quality (or
quantity) is perceived to be the basis for equitable payment.

Hypothesis 3. The introduction of ”equitable" production
rates will reduce production variance in both overpaid and equitably
paid groups relative to comparable groups without production rates.

Goodman and Friedman' s method was very similar to the
one Adams employed. In general, the results supported the hypothe-
ses.

Messé (1968) conducted an experiment which tested the
predictions of the equity hypothesis. This study was the basis for
the present research and so is described below in detail. Messé
identified three types of personal dispositions an individual can
follow in distributing wages to himself and another individual.
These personal dispositions were adapted from Sawyer' s (1966)
interpersonal orientations. These personal dispositions are mani—
fested in the form of behavioral strategies. The personal disposi-
tions and their behavioral counterparts, the stragegies, are pre-
sented below. For the following discussion let X represent the
wages of Person and Y the wages of 9m.

Self—interest. Person may be primarily concerned with

how much reward he can get for himself. If Person followed this

 

30

disposition of self-interest, his strategy for distributing wages
would be to maximize X.

Altruism. Person may be primarily concerned with how
much reward he can get for %. His personal disposition would
be altruism and his strategy for distributing the wages would be to
maximize Y.

Role symmetry. Person may equally weigh the reward of
himself and Other. This personal disposition is role symmetry and
the major role symmetric strategy is equality. Motivated by role
symmetry, Person may feel that both M and himself should
receive the same reward. If Person followed his disposition of
equality his strategy in distributing the wages would be to minimize
Ix - YI .

In Messé' 5 study the investments of Person and M
were identical. As a result, equity in his study was equivalent to
the personal disposition of role symmetry. Therefore, Messé
hypothesized that the most frequently manifested personal di sposi-
tion in his study would be role symmetry.

Messé' s procedure was somewhat complex. First, he had
groups of subjects fill out personality inventories. When they were
through with this task, each subject filled out a "wage distribution

booklet. " The 18 items of this instrument asked each subject to

 

 

31

choose between two ways of distributing wages to himself and
another individual who had also filled out the personality inven-

tories. A sample item from the wage distribution booklet follows:

Alternative A Alternative B
You get $4.00 You get $2.50
and and
He gets $1. 00 He gets $2. 50

Which alternative set of payments would you
rather see used? A or B?

Half the subjects in Messé' s study were told before they
responded that their choices on the wage distribution booklet would
actually determine their own wages and the wages of one other
individual. The responses of these subjects were called real
choices. The remaining subjectswere told that their choices on
the wage distribution booklet would have no effect on their own
wages or the wages of anybody else. The choices of these subjects
were said to be hypothetical.

Responses on the wage distribution booklet were used to
compute self—interest, altruism and role symmetry scores for each
person. The results confirmed Messé' s hypothesis. The most
frequently manifested personal disposition was role symmetry,
followed by self—interest and then altruism. In addition, subjects

making hypothetical choices made significantly more role symmetric

 

32

responses. He concluded that behavior guided by the personal
dispositions can be explained and predicted by the equity hypothesis.
The studies cited above offer evidence which suggests that
the desire to achieve an equitable state is a motivation present in
a broad range of settings, both naturalistic and laboratory.
However, the equity hypothesis has been criticized on a
number of grounds by both Weick and Lawler. Weick (1966) has
pointed out some possible limitations to the equity hypothesis.
These limitations can be classified into three major areas.
1. The degree to which inputs and outcomes are independent,
both in a theoretical and empirical sense, is ambiguous.
Often it is difficult to classify a given behavior as an input
or outcome. The nature of the experiments relevant to the
equity hypothesis have drawn attention away from this
simple but important problem. The settings used in these
studies have been sufficiently controlled that subjects have
focused on their effort as the salient input and on their
salary as the salient outcome. However, in everyday work
situations there are considerably more behaviors which
can be recognized in determining the input—outcome ratio,
and many of these can be assigned to either side of the

ratio. This is certainly an issue to which the empirical

 

33

tests of the equity hypothesis have not addressed them-
selves.

Adams himself has noted that inputs and outcomes often
are intercorrelated. Even when this relationship is only
moderate, it would be difficult to reduce inequity. This
difficulty arises because any attempt to change one side of
the ratio also changes the other side to the extent the two
are intercorrelated. Thus, changing the inputs of a person
in an attempt to make them more congruent with those of

a comparison individual could also change his outcomes.
Therefore, a new source of inequity would be created. This
is another issue which has been ignored by the proponents
of the equity hypothesis. Rather, it seems that they as—
sumed that equity can be achieved quite easily when an
individual makes an attempt to do so.

The equity hypothesis assumes that the standard by which
a given person evaluates his input—outcome ratio is always
external. However, as Weick points out, it is possible for
an individual to evaluate his ratio in terms of some com-
posite internal standard which is not directly tied to any
external group or person. If an individual uses internal

standards in place of social standards, then he might feel

 

34

his outcomes are not commensurate with his inputs, even

though his ratio is congruent with that of the external refer—

rent. Thus, feelings of inequity could result even when the

input-outcome ratio seems congruent with external criteria.

Weick and Nesset (1968), in elaborating on this last point,
introduced two terms: Own equity and Comparison equity. Own
equity occurs when Person' 8 inputs and outcomes are identical.
Comparison equity occurs when Person' 5 input—outcome ratio is
identical with his social referrent' s input-outcome ratio. Weick
used a questionnaire to present his subjects with pairs of hypotheti—
cal work situations. The subjects were asked to select the work
situation in which they would be the most comfortable. Among
Weick' s findings was that Comparison equity was preferred to Own
equity. However, he found that Own equity was also a significant
factor that could not be ignored.

To summarize, although open to some criticism, the equity
hypothesis has a substantial amount of empirical support which
attests to its predictive value across a number of settings.

Research support for
the expectancy hypothesis

 

 

The experimental support for the expectancy hypothesis is

somewhat less than that for the equity hypothesis. However, those

 

35

studies which have examined the expectancy hypothesis, in respect
to the distribution of rewards, are summarized below.

Lawler and O' Gara (1967) report research that was more
supportive of the equity hypothesis than the expectancy hypothesis
but did provide some support for the latter. They noted that most
research in inequity ignored underpayment. According to the equity
hypothesis, people underpaid on a piece rate basis should produce
large amounts of low quality work when compared with people paid
an equitable piece rate. Producing large amounts of low quality
work would raise the underpaid workers' outcomes (wages) without
raising their inputs (effort), and therefore this would help establish
equity. On the other hand, the expectancy hypothesis would predict
no difference in the quality or quantity of work between equitably
paid and underpaid piece rate workers. According to the expectancy
hypothesis, equitably paid and underpaid piece rate workers should
both be expected to produce as much as possible in order to maxi-
mize their wages.

As previously reported, Andrews (1967) found that subjects
underpaid on a piece rate basis when compared to subjects equitably
paid on a piece rate basis produced somewhat more of a lower
quality product. However, only the difference in quantity was sig-

nificant. Lawler and O' Gara felt that the lack of a significant

 

36

difference in quality may have been due to the relatively small
piece rate difference that Andrews paid his two groups (equitably
paid and underpaid). Therefore, the major objective of the Lawler
and O' Gara study was to provide a more definitive test of the
equity hypothesis by increasing the piece rate differences between
the two groups.

Unlike Adams' studies which produced inequity by the
manipulation of perceived job inputs (qualifications for the job),
Lawler and O' Gara, using the same procedure as Andrews, pro-—
duced inequity by the manipulation of assigned piece rate. Through
pretesting, they determined that $. 10 per interview would be per—
ceived as too low a wage and $.25 per interview as an equitable
wage. However, the $. 15 difference in piece rate payment between
the two groups was much larger than the $. 05 difference that
Andrews used.

The results obtained by Lawler and O' Gara were consistent
with the equity hypothesis. The underpaid subjects produced a sig~
nificantly greater amount of a significantly lower quality product.
Further support for the equity hypothesis was obtained in a post
experimental attitude questionnaire. Lawler and O' Gara concluded
that the equity hypothesis was helpful for understanding the effects
of piece rate underpayment. However, the expectancy hypothesis

was of little value.

 

37

The only support for the expectancy hypothesis in the
Lawler and O' Gara study was the finding that the importance of
money to the subject was significantly associated with his trying to
increase his productivity when underpaid.

Lawler (1968a) also investigated the effects of hourly over-
payment. He noted that Adams had found that overpaid subjects do
produce more during a 2%—hour work period than do equitably paid
subjects. These findings are consistent with the equity hypothesis.
However, these findings are inconsistent with the expectancy hypothe—
sis, which predicts that equitably paid and overpaid subjects should
produce the same amount in a hourly pay rate situation. The ra-
tionale for this is as follows: In an hourly pay rate situation the
amount an individual produces is not related to the pay he receives.
Therefore, there is no reason to expect overpaid subjects to pro—
duce more than equitably paid subjects. Thus, according to the
expectancy hypothesis, the high productivity found by Adams for his
overpaid subjects should be a transitory phenomenon, if it exists
at all.

To provide a more definitive test of the two hypotheses,
Lawler replicated Adams' overpayment hourly situation. However,
he collected data beyond the initial 2%-hour period that Adams used.

Instead, his subjects worked three 2—hour periods.

 

38

Furthermore, Lawler noted that even if the effects of
overpayment did disappear after one session, the problem of
accounting for the initial high productivity for the overpaid group
would still remain. One explanation for their high productivity that
is congruent with the expectancy hypothesis and does not require the
equity hypothesis has to do with the method Adams has used to pro—
duce the feelings of overpayment in his subjects.

Adams told his overpaid subjects that they were unqualified
for the job. Presumably, the subjects who are told they are unquali—
fied will feel inequity because they will see their input—outcome
ratio as higher than their comparison individual' 8 ratio. However,
Lawler stated that instead of producing feelings of overpayment,
Adams might have challenged his subjects' self-esteem. If this is
so, then the high initial productivity that Adams found can be
explained by assuming that feelings of esteem have positive valence.
Thus, it might have been that overpaid subjects, who were insulted,
used high productivity as a means of raising their self-esteem. On
the other hand, equitably paid subjects, who were not insulted, found
this unnecessary. Some evidence to support this interpretation was
provided by Friedman and Goodman (1967), who found that their sub-
jects were motivated to maintain their self-image.

One way to determine the cause of productivity differences

between the overpaid and equitably paid subjects is to try to induce

 

 

39

overpayment in a way that does not attack the subjects' qualifications.
Lawler did this. He told his subjects that, because of a grant he
had received, he was required to pay them at a higher rate than that
received by people like themselves who did comparable work. Ac—
cording to the equity hypothesis, if these subjects felt inequitably
overpaid, they should be as high producers as the subjects who are
made to feel overpaid by being told they are unqualified. However,
according to the expectancy hypothesis, it would be expected that
they should behave the same way as the equitably paid subjects since
their feelings of esteem were not attacked. As a result, no addi-
tional rewards were related to good performance.

The method Lawler used was very similar to that of Adams,
with the addition of a group of subjects who were overpaid but not
told they were unqualified. Also, as previously stated, he continued
collecting data beyond the 2%—hour period that Adams had used.
Lawler found little difference in productivity between the equitably
paid group and the overpaid—by—circumstance group. This result
supports the expectancy hypothesis and is contrary to the equity
hypothesis. Lawler also found a tendency for the unqualified over—
paid group to produce more than either of the other two groups.

This was a replication of Adams' findings. The high productivity

of the overpaid unqualified group was not a transitory phenomenon,

 

 

40

but it continued over all three sessions. However, after the first
session, the high productivity was accompanied by a decrease in
work quality, a result contrary to the prediction of the equity
hypothesis. However, this decrease is consistent with the expec-
tancy hypothesis which states that unrewarded effort should decrease
over time. It appears that in an hourly overpayment situation the
decrease in effort occurs in the form of a decrease in work quality.

As a result of this study, Lawler concluded that the expec-
tancy hypothesis explains the effects of hourly overpayment better
than the equity hypothesis.

Gordon and Lowin (1965) performed a study in which subjects
in one group were told that they were unqualified and therefore their
pay was being reduced. Subjects in another group were told that they
were unqualified but their paywas not reduced. Both groups pro-
duced at a similar rate. Taken together, the results of this and
Lawler' s (1968a) study indicate that, for hourly overpayment, just
feeling overpaid does not lead to high productivity. On the other
hand, having one' s qualifications threatened does, and this occurs
even when feelings of overpayment are not aroused. These results
suggest the attack on qualifications is the crucial variable producing
differences, a position consistent with the expectancy hypothesis.

Lawler _et_a_l_. (1968) investigated the effects of piece rate

overpayment. As previously stated, Adams had found that piece

41

rate overpayment leads to a lower level of productivity of a higher
quality product than does equitable piece rate payment. This is
contrary. to the expectancy, hypothesis, which predictsthat people
equitably paid and people overpaid on a piece rate basis should both
be high producers, since, for both, each additional unit produced
leads to higherwages.

Lawlergt_al. replicated Adams' overpayment piece rate
situation, but continued to collect data from the subjects'past the

2-hour time period Adams used. Lawler et al. had their subjects

 

work for three 2-hour'sessions. They found that for session 1 the
overpaid piece rate subjects produced significantly less of a sig-
nificantly higher quality product than the equitably paid subjects.
This replicated Adams' finding and it is consistent with the equity
hypothesis. However, they also found that these differences in work
quantity and work quality dissipated over the next two sessions. By
session 3 the productivity of the two groups was very similar. This
result is consistent with the expectancy hypothesis. Therefore,
Lawler 3321. concluded by stating that neither the expectancy
hypothesis nor the equity hypothesis alone can explain the results
in the piece rate overpayment situation. Rather, a combination of
the two hypotheses is necessary.

Thus, there is some experimental support for the expec-

tancy hypothesis. However, Adams (1968) has challenged Lawler' s

vr

 

42

(’1968a)‘ study. Adams (1968) stated that the thrust of Lawler' s
challenge to equity is based principally on the performance of sub-
jects in his overpaid-by-circumstance group. As previously stated,
the overpaid-by-circumstance groupwas made to believe that they
were qualified for the job. In addition, these subjects were told that
their wage rate was to be considerably higher than that paid to I
similar subjects on past projects. The subjects were told their
high wage rate was due to a grant the researchers had received.
The results of Lawler' 5 study indicated that the overpaid-by-
circumstance group produced no more and no better a product than
did subjects in an equitably paid group.

In replying to this study, Adams conceded that the overpaid-—
by-circumstance group did feel overpaid. However, he stated that
feelings of overpayment are not necessarily the same as feelings of
inequity. Lawler assumed that the comparison individuals for the
overpaid—by-circumstance group were less well-paid subjects on
past projects. Adams clained that it seems more likely that the
comparison individuals for the overpaid subjects were other over-
paid subjects.working on the same project. Adams felt that these
other overpaid individuals provided a more relevant and immediate
comparison. If Adams' contention is correct, then the overpaid-by-

circumstance subjects would experience no feelings of inequity, and

43

their performance would, therefore, not be expected to differ much
from that of the subjects in the equitably paid group.

Adams stated that the reasoning he had used to explain
Lawler' s findings in terms of the equity hypothesis was based on his
own speculative judgments. However, he claimed that Lawler' s
explanation of the results was also based on speculative judgments.
Clearly, further experimentation is needed to assess the contribu-
tion of both the equity hypothesis and the expectancy hypothesis to
the understanding of hourly rate overpayment.

Adams (1968) concluded his rebuttal with an appeal to the
adherents of both major hypotheses. He stated that the contrast
which has been drawn between the equity hypothesis and the expec-
tancy hypothesis has implied that performance in work situations
must be accounted for by m one or the other. Multiple motiva—
tional states may influence behavior, though one state may be
dominant in an individual at a particular time. Under particular
conditions the motivation to achieve equity may dominate; under
others maximizing gain may be salient. Adams doubted if anyone
could seriously question the fact that the desire to maximize out-
comes is a powerful determinant of behavior, and he stated that now
there is also sufficient evidence that the desire to achieve justice

has a substantial influence on behavior. According to him, the

 

 

44

question of greatest importance is not whether the equity hypothesis
is better or worse than the expectancy hypothesis, but rather, under
what conditions do equity motives and gain-maximizing motives

account for certain proportions of observed performance variance.

The Present Research

 

Following Adams' suggestion, the purpose of the present
study‘was not to determine which of the two hypotheses is best, but
rather to determine what each of the two hypotheses contributes to
understanding behavior in a multitude of situations. As previously
stated, the present research investigated how an individual would
distribute an allotted sum of wages to himself and a co-worker when
given the opportunity to do so.

In Messé' 3 study cited above, there was one major experi-
mental group, where all the subjects had identical inputs. For
reasons described in Chapter II, it was felt the method Messé used
was in some ways superior to those of Adams and Lawler. There-
fore, it was decided to use Messé' 3 basic method, but to employ a
wider range of conditions for the study than he had.

To facilitate the following discussion, Person and 91313,
which have already been used by Homans, will be defined in a slightly

different manner. The term Person will represent an individual

 

45

who distributes wages to himself and a co-worker of his. This
co-worker, the one who receiveswhatever wages Person allots to

him, will be referred to as Other. 3

 

The two major variables in the present study were the

inputs of Person and the inputs of Other. Each variable had two

 

levels, low and high, systematically varied such that four experi-

mental conditions were created.

Condition 1. Person had high inputs and Other had low inputs.

 

 

Condition 2. Person had high inputs and Other had high inputs.

 

Condition 3. Person had low inputs and Other had low inputs.

 

Condition 4. Person had low inputs and Other had high inputs.

 

Condition 2 was a replication of the situation Messé used
in his study. The other three conditions were an extension to his

research.

Hypotheses

The hypotheses tested in the present study are presented

in two sections. Both sections are concerned with the predictions

 

3Note that the number of underlines is used throughout the
present report to differentiate the Person of Homans from the Person
in the present research who distributes the wages. Thesame 1s
also true of Other and Other.

 

 

 

46

that the equity hypothesis and the expectancy hypothesis make con-
cerning the way Person‘will distribute an allotted sum of wages to

himself and a co-worker, Other. Specifically, the hypotheses are

 

concerned with predicting which personal dispositions Person will
manifest while distributing the wages. The first section of hypothe-
ses are predictions about the manifestation of the personal dispo-

sitions within each of the four experimental conditions. The second

 

section are predictions about the manifestation of the personal dis-

positions between each of the four experimental conditions.

Within gfliup hypotheses

 

Condition 1

The equity hypothesis. For Persons in Condition 1, equity
and self—interest are equivalent. Therefore, according to the equity
hypothesis, the most frequent personal disposition should be self-
interest, followed by role symmetry and then altruism. 4

The expectancy hypothesis. For Persons in Condition 1,

maximum wages for themselves can be obtained by manifesting the

 

4The reason why role symmetry was predicted to be more
prevalent than altruism under Condition 1 was that role symmetry
responses have a self—interest component to them, while altruism
responses have none. Therefore, role symmetry should always be
more prevalent than altruism when self-interest is predicted to be
the dominant response.

47

personal disposition of self-interest. Therefore, according to the
expectancy hypothesis, the most frequent personal disposition should
be self-interest, followed by role symmetry and then altruism.
Under this condition, both the equity hypothesis and the
expectancy hypothesis make identical predictions about the way

Persons will distribute a given amount of potential'wages.

Conditions 2 and 3

The equity hypothesis. For Persons in Conditions 2 and
3, equity and role symmetry are equivalent. Therefore, according
to the equity hypothesis, the most frequent personal disposition
should be role symmetry, followed by self—interest and then altru-
ism. 5

The expectancy hypothesis. For Persons in Conditions 2
and 3, maximum wages for themselves can be obtained by manifest-
ing the personal disposition of self-interest. Therefore, according

to the expectancy hypothesis, the most frequent personal disposition

should be self-interest, followed by role symmetry and then altruism.

 

5The reason why self-interest was predicted to be more
prevalent than altruism under Conditions 2 and 3 is that the equity
hypothesis postulates a differential threshold between inequity in
one' s favor and inequity against oneself. This implies that a person
will beconcerned with the welfare of both himself and others, but
self-interest will be somewhat stronger than interest in the other.
In addition, Messé (1968) found empirical support for this aspect of
the equity hypothesis.

48

For'Persons in Conditions 2 and 3, the equity hypothesis
and the expectancy hypothesis make different predictions concerning

the distribution of wages.

Condition 4

The equity hypothesis. For Persons in Condition 4,
equity and altruism are equivalent. Therefore, according to the
equity hypothesis, the most frequent personal disposition should

be altruism, followed by role symmetry and then self-interest. 6

The expectancy hypothesis. For Persons in Condition 4,
maximum wages for themselves can be obtained by manifesting
the personal disposition of self-interest. Therefore, according to
the expectancy hypothesis, the most frequent personal disposition
should be self-interest, followed by role symmetry and then altru-

ism.

For Persons in Condition 4, the equity hypothesis and the
expectancy hypothesis make different predictions concerning the

distribution of wages.

 

6The reason why role symmetry was predicted to occur
more frequently under Condition 4 than self-interest was that role
symmetry responses have an altruistic component while self-
interested responses have none.

49

Between conditions hypotheses

 

Self-interest

The equity hypothesis. According to the equity hypothesis,
self-interest is equivalent to equity for Persons in Condition 1.
Therefore, self-interest should be most frequent in Condition 1,
less-in Conditions 2 and 3, and least in Condition 4.

The expectancy hypothesis. According to the expectancy
hypothesis, self-interest should occur with equal frequency in all
four experimental conditions. The rationale for this is that in all
four groups Person should be equally likely to try to maximize his

own wages.

Role symmetry

The_equity hypothesis. According to the equity hypothesis,
role symmetry is equivalent to equity for Persons in Conditions 2
and 3. Therefore, role symmetry should be most frequent in these
conditions, less in Condition 4, and least in Condition 1.

The expectancy hypothesis. According to the expectancy
hypothesis, role symmetry should occur with equal frequency in all
four experimental conditions. The rationale for this is that there
is no reason to believe that Persons in the four experimental con-

ditions will distribute the wages differently.

50

Altruism

The equity hypothesis. According to the equity hypothesis,
altruism is equivalent to equity for Persons in Condition 4. There-
fore, altruism should be most frequent in Condition 4, less in Con-
ditions 2 and 3, and least in Condition 1.

The expectancy hypothesis. As with self-interest and role
symmetry, the expectancy hypothesis predicts that altruism should
occur with equal frequency in all four experimental conditions. The
rationale for this is that there is no reason to believe that Persons
in the four experimental conditions will distribute wages differently.

The following chapter presents a detailed account of the
research situation, the instruments used to test these hypotheses,

and the variables examined within this situation.

II. METHOD

The present chapter is a discussion of the method used to
test the hypotheses presented in Chapter I. The four sections of
this chapter are (a) recruitment and characteristics of the subjects,
(b) instruments used, (c) variables and experimental design, and
(d) procedure. As noted above, the method used was very similar

to that employed by Messé (1968).

The Sample

 

Recruitment of Subjects

It was essential that the subjects were interested in a
monetary reward so that their responses, which determined how
much they and a co-workerwould be paid, were meaningful for
them. Therefore, it was necessary that the subjects not only volun—
teered, but volunteered to make some money. The experimenter
told the students in his introductory industrial psychology class that
an industrial opinion survey was being conducted on campus, and
that male volunteers were being sought. They were told the survey

would take from one to two hours of their time, and that they would

51

52

receive no extra credit for the course by participating in the research;
instead each subject would be paid up to four dollars for paritipating.
Further, anyone interested in making some money should sign his
name on one of the sign-up sheets that were passed around.

The sign-up sheets were then distributed. They listed
several convenient periods of time for which a subject could volun—
teer. From the sign-up sheets, four times to run subjects were
selected in terms of the number of subjects that would be available.

In the class period following the one in which sign-up sheets had
been passed out, those subjects who had volunteered were given a

card indicating the time and room they should be at for the study.

Characteristics of the Subjects

Fifty-eight male students at Michigan State University
served as subjects. The reason that all the subjects were of one
sex (male) was that the sex of an individual' s co-worker has been
shown to affect the way in which the individual distributes the wages
to himself and that co-worker (Messé, 1968). An investigation of
the combined effects of sex of the co-worker and differential inputs
did not seem worth the cost of running additional groups of subjects.

The subjects ranged in age from 19-35, with the large

majority being either 20 or 21 years old. Most of them were either

53

juniors or seniors, and they had a wide range of majors. However,

more than half were majoring in either social science or business.

Instruments

 

There were fourinstruments used in the present research:
a wage distribution booklet, an attitude questionnaire, an industrial

Opinion questionnaire, and a student-background questionnaire.

Wage Distribution Booklet

The subjects' responses to the Wage Distribution Booklet
(presented in Appendix I) were the dependent variable. The Wage
Distribution Booklet used in the present research was a shortened
version of the one used by Messé. Since it is the instrument used
to measure the dependent variable, it is described in considerable
detail below. 1 The first part of the discussion is concerned with the
selection of the six items used in the Wage Distribution Booklet.

Let two persons perform an identical task. Their employer
could then say to one of them, "Suppose for the work you both have
done, we are to pay you $4.00 and him $1. 00. " This base set of
payments to which an alternative set is to be compared will be called

the standard distribution. "Would you rather have me give you each

 

1The following discussion borrows heavily from the one
which Messé used to describe his booklet.

54

$2. 50 instead?" This is the alternative outcome to the standard
distribution. "What would you say if the choice-were between $4. 00
for you and $1.00 for him (standard distribution) versus $2.00 for
you and $3.00 for him (alternative outcome)?"

In both cases, an individual motivated by either altruism
or role symmetry would choose the alternative outcome, while a
person motivated by self-interest would choose the standard distri-
bution. In the first case, the standard distribution provides the
chooser with a greater reward ($4.00 versus $2. 50). On the other
hand, the‘alternative outcome gives a greater reward to his co-
worker ($2. 50 versus $1.00), and yields a more equal distribution
of rewards ($0. 00 difference versus a $3.00 difference) than does
the standard distribution. In the second case, the standard distri-
bution gives the chooser a greater reward ($4. 00 versus $2. 00) than
the alternative outcome, and a greater reward than that of the co-
worker ($4. 00 versus $1.00). The alternative outcome gives the
co-worker a greater reward ($3. 00 versus $1.00), and distributes
the rewards more equally (a $1.00 difference versus a $3.00 dif-
ference) than does the standard distribution. However, the alter-
native outcome gives the co-worker more than the chooser ($3. 00
versus $2.00).

These differences in rewards to both are due to two inde-

pendent dimensions of the payoff structure. A person--if he is to

55

make a rational choice—-must address himself to these dimensions,

since their various combinations determine just what his alternatives

are. These two dimensions are:

1. Relative welfare of the standard distribution

 

Does the standard distribution give the chooser:
(a) more than the co—worker?
(b) the same as the co-worker?

(c) less than the co-worker?

2. Relative‘welfare of the alternative outcome

 

If the chooser selects the alternative outcome,
does he obtain:

(a) more than the co-worker?

(b) the same as the co—worker?

(c) less than the co-worker?

Since these dimensions are independent, nine unique situa-
tions can be generated by varying the possible combinations of dif—
ferent cases which occur in each. The number of situations can be
reduced to six by eliminating case (b) of the standard distribution
dimension. Case (b) of the standard distribution was eliminated
because it was felt that the standard distribution case which gave

both persons the same reward was, by far, the least interesting;

 

56

the more extreme cases of being either ahead or behind the other
were thought to present the subjects with more meaningful and more
difficult choices.

Table 1 presents these six situations modified by the follow-
ing restrictions which further define their structure. (a) The sum
of the standard distribution rewards is always $5. 00, divided into
either $4. 00 for the chooser and $1. 00 for the co-worker, or $1. 00
for the chooser and $4.00 for the co-worker. (b) The sum of the
alternative outcome rewards was also always $5.00, divided into
either $3.00 for the chooser and $2. 00 for the receiver, $2. 50 for
the chooser and $2. 50 for the receiver, or $2. 00 for the chooser
and $3. 00 for the receiver. (c) When the two rewards in an alter—
native outcome are unequal, they always differ by $1. 00. It was
hoped these restrictions would give each subject a standard frame
of reference, and therefore enable him to make more meaningful
choices.

Thus, Table 1 indicates the standard distributions and
alternative outcomes presented to the subjects. For example, in
the first row of Table 1 (series X situation (1)), if the individual
determining the wages selected the standard distribution, he would
receive $4. 00 while the co-worker whose wages he was determining

would get $1.00. However, if the individual selected the alternative

57

outcome instead, he would get $3. 00 and his co-worker would

receive $2. 00.

 

 

 

Table 1. -- The choice situations

Standard Distribution Alternative Outcome
Series X

(1) $4.00 $1.00 $3.00 $2.00

(2) $4. 00 $1. 00 $2. 50 $2. 50

(3) $4.00 $1.00 $2.00 $3.00
Series Y

(1) $1.00 $4.00 $3.00 $2.00

(2) $1.00 $4.00 $2.50 $2.50

(3) $1.00 $4.00 $2.00 $3.00

 

 

The outcomes presented in Table 1 in terms of standard
distribution and alternative outcome were given more neutral labels
--Alternative A and Alternative B, respectively--and put in the form
of a six page Wage Distribution Booklet. This booklet is presented
in Appendix I. It had two forms, A and B. In Form A the Series X
items were presented first, while in Form B the Series Y items
were presented first. For convenience, a sample page of the Wage

Distribution Booklet is presented as Table 2. The order that the

58

Table 2. -- A sample choice situation

 

 

Alternative A

You get $4. 00
and
He gets $1. 00

 

 

 

Alternative B

You get $3. 00
' and
He gets $2. 00

 

 

\I/

1. Which alternative gives you
more? A or B? (Check one)

 

2. Which alternative, if either,
gives the greatertotal to the
two of you, together? Check
A or B. If the two give the
same total, check "Equal. "

\V

 

 

 

3. Which alternative gives the
two of you more nearly the
same? A or B? (Check one)

 

 

SKIP THE QUESTION BELOW: GO ON TO THE NEXT PAGE

 

 

CHOICE X1:

Which alternative set of payments
would you rather see used? A or
B? (Check one)

 

 

Remember: If this choice is the one drawn, you and he will be paid

just the amounts you now choose.

(X1)

 

59

Series X and Series Y items appeared in the Wage Distribution Book—
let was the same order in which they appeared in Table 1.

As can be seen from Table 2 and Appendix I, each page of
the booklet contained one of six choice situations. Underneath each
choice situation was a set of three questions which were identical
on all pages of the booklet. The questions served two purposes:

(a) they insured that the subject was aware of the specific structure
of each choice; (b) they allowed the experimenter to check if each
subject understood the structure of the choices. None of the answers
given by any subject on these three questions were in error. Beneath
each set of questions was another question. This asked each subject
to select which of the two sets of payments (Alternative A or Alterna-
tive B) he preferred to see used.

From a person' 5 choices on the Wage Distribution Booklet
it was possible to infer the strengths of the various personal dispo-
sitions: self—interest, role symmetry, and altruism. This was
done by comparing the individual choices in the three item pairs
which were mirror images of each other. Item 1 in Series X and
Item 3 in Series Y are a mirror image pair. Likewise, Item 2 in
Series X and Item 2 in Series Y are a mirror image pair, as are
Item 3 in Series X and Item 1 in Series Y. These three pairs of

items are mirror images because for each pair the values in the

 

60

standard distributionand alternative outcome are interchanged from
one to the other. The method by which the strength of each of the

personal dispositions was computed is presented below.

The index of self-interest

 

If a person chooses Alternative A in the Series X member
of a mirror image pair, but Alternative B in the Series Y member,
he is displaying self-interest. This is because his choices indicate
that he is concerned only-with obtaining as much money as he can
for himself. Each such display contributes one point to his self-
interest score. Since there are three mirror image pairs, a per-

son's self-interest score can range from zero to three.

The index of role symmetry

 

If a person chooses either Alternative A or Alternative B
for both members of a mirror image pair, he is displaying role
symmetry. This is because his choices indicate that he is concerned
with both the wages of himself and the co-worker for whom he is
determining the wages. As with self-interest, a person' s role sym-

metry score could range from zero through three.

The index of altruism

 

If a person chooses Alternative B in the Series X member

of a mirror image pair, but Alternative A in the Series Y member,

X

61

he is displaying altruism. This is because his choices indicate that
he is concerned only with obtaining as much money. as he could for
the co—worker for whom he was determining the wages. As with
both self-interest and role symmetry, a person' s altruism score

could range from zero through three.

Relative welfare

 

In addition to providing a measure of the strength of each
of the personal dispositions, the six items on the Wage Distribution
Booklet permitted the examination of the effects of the reward
structure variable of relative welfare on the manifestation of the
personal dispositions. Relative welfare refers to whether an
individual receives more than, less than, or the same amount of
money as the co-worker whose wages he is determining. As can be
seen in Table 1, for Series X item (1), both the standard distribu-
tion and the alternative outcome give the chooser more money than
his co—worker. Therefore, both alternatives have positive relative
welfare for the chooser. However, for Series X item (3), the
standard distribution gives the chooser more money than his co-
worker, while the alternative outcome gives the co-worker more
than him. Therefore, the standard distribution has positive relative
welfare for the chooser and the alternative outcome has negative

relative welfare for him.

62

Attitude Questionnaire

The Attitude Questionnaire (see Appendix II) contained five
rating scale items and two open—ended questions requiring brief
answers. The main purpose of this questionnaire was to determine
the degree of success that had been attained in manipulating the
independent variables. Two of the questions (items 2 and 3) are
concerned with the degree to which subjects perceived differences
in work quantity and the extent to which these differences affected
choices on the Wage Distribution Booklet. Two other questions
(items 4 and 5) asked for the same information, but with work quality
as the relevant variable. The remaining questions were concerned
with other variables that might have affected choices on the Wage
Distribution Booklet, as well as the student' 5 interest in and per-

ception of the study.

Indust rial Opinion Que sti onnai re

The Industrial Opinion Questionnaire (see Appendix III) had
two forms, Research Form 1 and Research Form 2. Form 1 con-
tained four short essay questions which asked the subjects about
their opinions concerning areas in the business world. Form 2 con-
tained twelve such questions, including the four from Form 1. The

specific number of questions in each form, four and twelve, was

 

63

selected so that the subjects would work the desired length of time,
30 minutes or 90 minutes. The main criterion for selecting the
particular twelve items used was the apparent easerwith which

written responses could be made to them.

Student Background Questionnaire

The Student Background Questionnaire (see Appendix IV)
contained several biographical items, and a short essay item asking
for the subject' s attitude toward labor unions. As discussed below,
the subject' 5 responses to this instrument were used as the reason

why they were given tasks of different lengths.

Variables and Experimental Design

Four situational conditions were manipulated. Two of these,
(a) hypothetical versus real choices and (b) order of presentation of
the choices series, were manipulated mainly for reasons of control.
The other two variables, (0) inputs of the subject and (d) inputs of

his partner, were the main focus of the research.

Situational Variables

Hypothetical versus real choices

All subjects, after completing the Industrial Opinion Ques-

tionnaire, filled out the Wage Distribution Booklet. Half the subjects

64

were each told that their responses on the booklet would determine
not only their'wages but thewages of a co-worker. These subjects
were told that they would be making r_ea_l choices on the booklet. The
other half of the subjects were told that their responses. on the Wage
Distribution Booklet would have no effect on the wages they received
nor on the-wages anyone else received. These subjects were told

that they would be making hypothetical choices.

 

Order of presentation

 

Messé (1968) found that the order of presentation of the six
items in the Wage Distribution Booklet (whether the Series X or
Series Y items were presented first) had both a significant effect on
the responses overall and interacted with a number of other variables.
Although-this phenomenon was more or less irrelevant to the specific
theoretical concerns of the present research, Messé' s findings
implied that this variable should be manipulated for reasons of con-
trol. Therefore, half the subjects were given booklets which had
the three Series Xitems first, followed by the three Series Y items.

The remaining subjects were presented the series in reverse order.

Inputs of the subject

 

Before each group of subjects began, they were informed,

supposedly on the basis of their responses to the Student Background

 

65

Questionnaire, if they were expected to produce high or low quality
answers to the items on the Industrial Opinion Questionnaire. The
subjects were told that those persons expected to produce high
quality answers were being given a longer questionnaire than those
expected to produce low quality responses. The more extensive
questionnaire required three times as long to complete. Therefore,
the variable inputs of the subject had two levels, low and high.
Those subjects who had low inputs worked one-third as long as those
with high inputs, and they were told that their answers were expected
to be of low quality. Those subjects who had high inputs worked
three times as long as those with low inputs, and they were told that

their answers were expected to be of high quality.

Inputs of partner

 

Subjects were not told the specific identity of the other indi-
vidual (their partner), but they were told whether those in the other
group had high inputs (long Industrial Opinion Questionnaire and ex-
pected to produce high quality answers) or low inputs (short Industrial

Opinion Questionnaire and expected to produce low quality answers).

Experimental design

 

As noted above, four situational conditions were system-

atically manipulated. This produced a factorial design whose

 

 

66

dimensions were 2 (inputs of subject—-high versus low) X 2 (inputs
of partner--high versus low) X 2 (choices--real versus hypothetical)
X 2 (order of presentation——Series X first versus Series Y first).

Thus, the design required sixteen cells.

Procedure

 

There were two major features of the experimental pro-
cedure: (a) one subject determined the reward for himself and
another subject; (b) the subjects were run in groups, and not indi-
vidually, as has been customary in this area of research. The
rationale for these procedures is presented below.

It is rare in ”real life” for one individual to freely deter-
mine the rewards for himself and another individual. In real life
most encounters take the form of interactions. In an interaction,
each person involved partially determines the rewards of all.
Therefore, the factor of reciprocity enters into real life situations.
However, if an individual is given the opportunity of determining
rewards for both, he can more easily express any personal dispo-
sition he wishes, since he does not have to be concerned with the
actions of the other person. In the present research, the individual
distributing the wages can do what he wants without considering the
other' 8 reactions. Thus, as with most experimental studies, the

justification for the sacrifice of realism is that it makes the

 

67

behavioral expression of the personal dispositions more directly
accessible to investigation.

As previously stated, the subjects were run in groups.

The procedure generally used in equity and expectancy research
involved the subjects working on a task in the absence of co—workers.
The subject, in previous research, was first told that he was either
qualified or unqualified for a job, and therefore, entitled or not
entitled to the pay associated with the job. The subject was then
hired for the job and he went out and performed it in the absence of
co-workers.

Running subjects in groups has some distinct assets. First,
it is more efficient in terms of data collection. Second, it provides
a different experimental setting for the research than those pre—
viously used. The use of different experimental settings is desirable
in terms of testing the generality of theories. Consistent results
obtained using different procedures give additional support for a
theory beyond that provided by a simple replication. Inconsistent
results from studies with different experimental settings frequently
generate reasons for these discrepancies which often lead to a bet-
ter understanding of the variable in question. Third, in real life
individuals often work in the presence of co-workers. Therefore,
the procedure used in the present research provides an element of

realism not present in previous studies.

 

68

Perhaps the most important advantage of running subjects
in groups pertains to the area of reference individuals. In Adams'
and Lawler' 5 research, the reference individual for the subject was
hypothesized to be a co-worker described to the subject by the
experimenter. This co—worker was never seen by the subject. In
the present research the reference individual (the individual for
whom the subject is determining the wages) is a visible co-worker.
In the present research the subject is able to see the amount of work
his co—worker actually does, while in previous research the behavior
of the hypothetical co—worker had to be imagined.

As previously stated, there were four experimental condi—
tions. The procedure used in the four conditions was identical
except for two features. First, the part of the oral instructions per—
taining to what quality answers the experimenter said he expected
the subjects to produce was different for the four conditions. Also,
the time the subjects spent filling out the Industrial Opinion Ques—
tionnaire was different for the four conditions. The procedure used
for the group where the individuals working 1% hours determined the
wages for themselves and for their co—workers who worked % hour
is described in detail, below.

The experiment was run in a classroom at Michigan State

University. The room was arranged so that the 48 seats were

 

 

69

placed in two groups of 24. Each group of 24 seats was arranged
into three rows of eight. A large aisle separated the two groups of
24 «seats. In each group, subjects were seated in the two rows
closest to the aisle. This was done so that each subject could see
the remaining subjects in the room.

When a subject arrived for the study, the experimenter
asked the subject for his name. He then told the subject which
group of seats he was to sit in, and he also told the subject to sit
in one of the two rows of seats closest to the aisle. The subjects
had been randomly assigned to one of the two groups. Prior to
arriving for the experiment, one group of subjects had been told
they were to fill out a questionnaire for 1% hours, while the other
group had been told they were to fill out a questionnaire for % hour.
When the subjects arrived for the study they did not know that there
would be other subjects who were filling out a questionnaire for a
different length of time than they. 2

Five minutes after the time the subjects were told to arrive
the experimenter began with the instructions. 3 The oral instruc-

tions for the subjects were as follows:

 

2The individual running the subjects was not the experi-
menter. The reason the experimenter did not run the subjects was
that they were his students and his presence might have had an
effect on the subjects' responses.

3One subject arrived an hour late. He was told to go home
since the experiment was almost over.

 

70

”Each of you has volunteered to participate in this industrial
opinion study. This morning you will fill out a questionnaire asking
your opinions about some important industrial issues. Because you
have varied backgrounds pertaining to your work experiences, grade
point, etc. , you have been separated into two groups.

”We have two different questionnaires. People on the left
side of the room will fill out their questionnaire for 1% hours.
People on the right side will fill out a shorter questionnaire for %
hour.

"Based on the questionnaire you filled out in your Psychology
255 class, we feel that the people in the 1% hour group should be
better able to answer the longer questionnaire, and to produce
higher quality answers on it. However, we are also interested in
the responses of the people in the % hour group.

"I will periodically indicate the amount of time remaining
on the board in front of the room. When the time elapses, I will
collect the questionnaire. You will be required to remain the full
time. The group on the left will remain 1% hours and the group on
the right % hour. Please answer all questions and return at one
o' clock when you will be paid. Please put your name on the first
page of the questionnaire. Do you have any questions ?”

When % hour had elapsed the experimenter said, ”The group

on the right please stop and hand in your papers. I will see you back

 

 

 

71

here at one o' clock when you will be paid. The group on the left
still has an hour remaining. " When the remaining hour had elapsed
the experimenter then said, "Please stop writing and hand in your
papers. I will see you at one ol clock when you will be paid. "

When the subjects arrived at one o' clock they were told to
sit in the same group of seats as they did in the morning. At 1:05
all the subjects had arrived. The experimenter then read the fol—
lowing instructions, 4 "Now we are going to determine what your
wages will be. Your wages will be determined by your answers on
the Wage Distribution Booklet which I am now passing out. Please
put your name on the front of the booklet. Now open your booklet
and you will notice that there are six pages. Each page contains
two alternative ways of distributing the wages. You will notice that
you are asked to distribute wages to yourself and one other individ-
ual. The answers on this booklet of those individuals who worked
1% hours this morning will determine the wages for themselves and
one individual who worked % hour this morning. Even though the
choices of the people on the right side of the room (experimenter

pointed to those who had worked % hour) will not have any effect on

 

4The oral instructions were pretested by having 20 people
listen to them one at a time and then answer the question, "What is
unclear about these instructions ?" Many modifications of the original
instructions were made.

 

72

the way the wages are to be distributed, we are interested in how

you would have distributed the wages if you were given the oppor-

tunity to do so. Therefore, you should also fill out the booklet. "

The experimenter then went through the following example

which indicated how the wages were to be distributed.

1.

The subjects were told to look at each page of the booklet
and to notice that they were asked to select one of two ways
of distributing wages to themselves and another individual.
They were also told that the other individual they were asked
to determine the wages for was someone in the room, but

he was in the other group. They they were informed that
the experimenter had paired each subject in one group with
another. individual in the other group for the purpose of
distributing the wages. The specific identity of the sub-
ject' s "partner" was not told to him, but as previously
stated, he knew this individual was seated across the room
from him.

The subjects were reminded that the responses on the Wage
Distribution Booklet of those who had worked 1% hours in

the morning would determine their wages and the wages of
their "partner" in the other group, while the responses of
those who had worked % hour would have no effect on the way

the wages were to be distributed.

 

73

They were then shown that each page of the Wage Distribution
Booklet had a letter (X or Y) and a number (1, 2 or 3) at the
bottom right hand corner. They were told that all the Wage
Distribution Booklets had the same six pages (X1, X2, X3,
Y1, Y2 and Y3) but that they were in different orders.

The experimenter then held up six cards. Each card con—
tained either X1, X2, X3, Y1, Y2 or Y3. These six cards
each corresponded to a page in the Wage Distribution Book—
let.

The subjects were told that after they had all completed
their Wage Distribution Booklets that someone who had
worked % hour would select at random without looking one
of the six cards (X1, X2, X3, Y1, Y2 or Y3). The experi-
menter held up the six cards again. The subjects were told
that the card selected would determine which choice situa-
tion would actually be used to distribute the wages. This
one card would serve as the indicator of the choice situa—
tion for all "partners" but the actual distribution of wages
for each pair of subjects would be determined by the re—
sponses of each individual who had worked 1% hours in the
morning.

It was then pointed out that only one choice situation would ,

be used to determine the wages for a pair of subjects, but

74

that each situation had the same probability of being picked.

Therefore, all the choices were important.

7. The experimenter then selected one of the six cards at
random and he then showed what the possible wages would
be if this choice were selected.

8. He then asked for and answered questions concerning the
distribution of wages.

The subjects completed the Wage Distribution Booklet,
after which the experimenter then collected the booklets and passed
out the Attitude Questionnaire, which the subjects completed.

A subject was then called up to select one of the six cards
at random, and his choice was announced to all subjects. The
experimenter explained what this choice represented in terms of
the alternative ways the wages could be distributed. The experi—
menter examined the Wage Distribution Booklets of the individuals
whose choices determined how the wages were to be distributed.
The wages were then given to all the subjects in sealed envelopes.

The entire procedure——the experimenter reciting the
instructions, the subjects filling out both the Wage Distribution
Booklet and the Attitude Questionnaire, and the experimenter deter-
mining and distributing the wages-—required only about 25 min-

utes.

 

75

As previously stated, the procedure used in the four experi—
mental groups was very similar. The procedure cited above was
for the group in which the subjects working 1% hours determined the
wages for themselves and their co—workers who worked % hour. The
only difference in procedure in the four experimental groups involved
the length of time the subjects worked and the part of the oral instruc-

tions involving what the experimenter said about the quality of the

 

answers he expected the subjects to produce. These differences
are further described below.

To the group of subjects where everyone worked 1% hours
the experimenter said, ”Based on the questionnaire you filled out
in your Psychology 255 class, we feel that all the people in this
group should be able to answer this long questionnaire and to pro—
duce high quality answers on it. ”

To the group of subjects where everyone worked % hour the
experimenter said, "Based on the questionnaire you filled out in your
Psychology 255 class, we feel that all the people in this group will
produce low quality answers and therefore we are giving you a short
questionnaire. "

The procedure for the group in which the subjects who
worked % hour determined their wages and those of a co-worker who

worked 1% hours was identical to that of the group in which the subjects

76

who worked 1% hours determined their wages and those. of a co-
worker‘who worked% hour, except the individuals determining the
wages were the ones who had worked % hour.

The following chapter presents the results obtained by

using‘the method described in this chapter.

III. RESULTS

The present. chapter is divided into three sections. The
first section presents an examination of the situational variables
discussed in Chapter-II. The second section is concerned with
further analyses of the effects of inputs of subject and inputs of his
partner. The third section is concerned with the effect of the
variable of relative welfare.

Effects of Situational Variables on the
Manifestation of the Personal Dispositions

 

 

Since there were no altruistic responses, role symmetry
and self-interest were perfectly inversely related. For this reason,
. for the most part, only role symmetry is referred to in the follow-
ingsections. The reader should keep in mind, however, that the
opposite of any statement made concerning role symmetry is true
of self-interest.

The effects of the four situational variables, inputs of the
subject, inputs of his partner, real versus hypothetical choices, and

order of presentation, on the manifestation of role symmetry were

77

78

determined by a four-way unweighted means analysis of variance,
summarized in Table 3.

The summary table indicates that the third order inter-
action was significant at the .01 level. Therefore, tests of simple
effects were performed to determine under which conditions each
of the four situational variables had a significant effect on the mani-
festation of role symmetry. Table 4 presents the cell means for
role symmetry scores; for convenience and ease of identification in

later discussion, each cell mean has a letter next to it.

Inputs of Subject

The main effect for inputs of subject was essentially zero
(F = .07). Also none of the first or second order interactions
involving the variable were significant, but as noted earlier, the
third order interaction was significant. Exploration of the contri-
bution of inputs of subject to this interaction through analyses of the
simple effects revealed one significant difference between means
((1) and (1) (see Table 4). This difference indicates that, when a
subject' 3 partner had high inputs and the subject made hypothetical
choices starting with Series Y, the subject tended to be more role
symmetric when he had high inputs than when he had low inputs

(F = 5.59; d.f. = l, 42; p< .05).

79

Table 3. -- Summary of the analysis of variance for role symmetry

scores

 

 

 

 

 

 

 

Source of Variation d. f. M. S. F
Inputs of subject 1 O. 07 O. 07
Inputs of partner 1 17. 06 17. 58**
Real versus hypothetical choices 1 2. 66 2. 74
Order of presentation 1 4. 71 4. 85*
Inputs of subject X inputs of partner 1 0. 00 O. 00
Inputs of subject X real versus
hypothetical choices 1 2' 34 2' 41
Inputs of subject X order of presentation 1 1. 42 1. 46
Inputs of partner X real versus
hypothetical choices 1 0' 46 0' 47
Inputs of partner X order of presentation 1 6. 16 6. 35>:<
Real versus hypothetical ch01ces X order 1 O. 42 0. 43
of presentation
Inputs of subject X inputs of partner X
. . 1 1. 56 1. 61
real versus hypothetical ch01ces
Inputs of subject X inputs of partner X 1 0. 14 0. 14
order of presentatlon
Inputs of subject X real versus
hypothetical choices X order of 1 0. 56 O. 58
presentation
Inputs of partner X real versus
hypothetical choices X order of 1 4. 89 5. 04*
presentation
Inputs of subject X inputs of partner X
real versus hypothetical choices X 1 9. 84 10. 14**
order of presentation
Error 42 . 97
* p < . 05

>I‘*p< .01

80

.m 3 0 89C 53> Eco omHBox: 3:00 3an wit 5 swim monoom hypogahm 39H
smog 9: .m 8. 0 89¢ Eco hawk, 3:00 HooEHHm Some .HoM mouoom hugger? 20h o5 oochw

 

3:9: 33

 

 

 

A3 rm .o on oo H A5 rm .0 Ads hm .o was hoauhmnm
$59: .33
93 H835.
. . . . 9.:qu HHNHHH
E 8 H 3 on N 3 8 H 3 ms N 92H Satan
. . . . 3:9: 33
3 E o AB 8 H CH 8 H H8 8 0 m2 esteem
335 SE
was H836
. . . . mBSHHHmHHH
:3 3 N 3 ON N 3H 8 o 3 N N

was sunfish

 

 

Hmsa poEomonQ
V moEwm

 

«mum popcommsm
N moiom

“mum Uwuqomoug
W moflwm

 

“mum pmucomona
M moEom

 

moo Hone Hmo 3.230%:
moxma Hocflnsm

 

mooHoHHo Boa
mmxma wooflnsm

 

 

 

.m

monoom buofifihm 29H .HocH madman Sou I .v 2an

81

Inputs of Partner

The main effect for inputs of partner was significant (F =
17.58; d. f. = 1, 42; p < .01). One first order interaction, inputs
of partner X order of presentation, was also significant (F = 6. 35;
d.f. = 1, 42; p < .05), as was one second order interaction, inputs
of partner X real versus hypothetical choices X order of presenta-
tion (F = 5.04; d. f. = 1, 42; p < .05). Exploration of the significant
third order interaction through analyses of the simple effects
revealed four significant differences (see Table 4). They were
between means (a) and (e), (d) and (h), (i) and (m), and (k) and (o).
The difference between (a) and (e) indicates that when subjects have
high inputs and they make real choices starting with Series X, they
tend to be more role symmetric when their partners have high inputs
than when their partners have low inputs (F = 9. 24; d. f. = 1, 42;
p < .01). The difference between ((1) and (h) indicates that when sub-
jects have high inputs and they make hypothetical choices starting
with Series Y, they tend to be more role symmetric when their
partners have high inputs (F = 7. 87; d.f. = 1, 42; p < .01). The dif—
ference between (i) and (m) indicates that when subjects have low
inputs and they make real choices starting with Series X, they are
more role symmetric when their partners have high inputs (F = 7. 87

d. f. = 1, 42; p < .01). The difference between (k) and (0) indicates

.

1

 

82

that when subjects have low inputs and they make hypothetical
choices starting with Series X, they tend to be more role symmetric

when their partners have high inputs (F = 4. 09; d. f. = 1, 42; p < .05).

Real Versus Hypothetical Choices

The main effect for real versus hypothetical choices was
non-significant (F = 2. 74; d.f. = 1, 42; p > .05). None of the first
order interactions were significant. One second order interaction,
inputs of partner X real versus hypothetical choices X order of pre-
sentation, was significant (F = 5.04; d. f. = 1, 42; p < .05). Explo-
ration of the significant third order interaction through analyses of
the simple effects revealed one significant difference between means
(b) and (d). This difference indicated that when subjects have high
inputs, their partners have high inputs, and they start with Series Y,
these subjects tend to make more role symmetric responses when

their choices are hypothetical (F = 13. 79; d. f. = 1, 42; p < .01).

Order of Presentation

The main effect for order of presentation was significant
(F = 4. 85; d.f. = 1, 42; p < .05). One first order interaction,
inputs of partner X order of presentation, was significant (F = 6. 35;
d.f. = 1, 42; p < .05), as was one second order interaction, inputs

of partner X real versus hypothetical choices X order of presentation

 

83

(F = 5.04; d. f. = 1, 42; p < .05). Exploration of the significant third
order interaction through analyses of the simple effects revealed
three significant differences (see Table 4). They were between
means (a) and (b), (i) and (j), and (k) and (l). The difference
between means (a) and (b) indicates that when subjects have high
inputs, their partners have high inputs, and they make real choices,
those who start with Series X are more role symmetric than those
who start with Series Y (F = 9.24; d. f. = 1, 42; p < .01). The dif-
ference between means (i) and (j ) indicates that when subjects have
low inputs, their partners have high inputs, and they make real
choices, those who start with Series X are more role symmetric
than those who start with Series Y (F = 5.59; d. f. = 1, 42; p < .05).
The difference between means (k) and (1) reveals that when subjects
have low inputs, their partners have high inputs, and they make
hypothetical choices, those who start with Series X are more role
symmetric than those who start with Series Y (F = 4. 09; d. f. = 1, 42;
p < . 05).

The situational variables of real versus hypothetical choices
and order of presentation were included in the present research
mainly as controls. The focus of the present research, as noted in
Chapter I, was on the variables of inputs of the subject and his

partner. Therefore, both the degree to which their manifestation

 

 

 

84

was successful and their effects on choices were explored through

additional analyses, the results of which are presented below.

Further Analyses of the Effects of Inputs
of Subject and Inputs of Partner

 

Validity of Manipulati ons

As noted in Chapter II, the degree of success attained in

manipulating the variables of inputs of the subject and his partner

 

was measured by the subject' 8 responses to questions 2 and 4 of the
Attitude Questionnaire. These responses are summarized below
in Tables 5 and 6.

Table 5. -- Percent scale —point responses for work quantity tabulated
within each experimental condition

 

 

 

 

 

Scale—point responses
As compared to his partner,
Experimental a subject reported working
cond1t1on 1 2 3 4 5
4 times about 2 about about %
as much to 3 times the % to% as much
or more as much same as much or less
1 (1:30 — :30)a 14 86 o o o
2 (1:30 — 1:30) 0 6 94 0 0
3 ( :30 - :30) 0 0 100 0 0
4 ( :30 — 1:30) 0 7 33 60 0

 

 

 

 

 

 

aThis indicates that the 1% group determined the wages for
themselves and their partners in the % group.

 

85

Table 5 presents the percent of scale-point responses for work

quantity tabulated within each experimental condition. Table 6 pre-
sents the same information pertaining to work quality.
Table 6. -- Percent scale-point responses for work quality tabulated

within each experimental condition

 

 

Scale—point responses

 

As compared to his partner, a subject felt
the quality of his work was

 

 

Experimental
cond1tion 1 2 3 4 5
Consider- Some- About Some— Consider—
ably what the what ably
better better same worse worse
1 (1:30- :30) 14 50 36 0 0
2 (1:30 — 1:30) 0 17 71 12 0
3 ( :30 — :30) 8 0 92 0 0
4 ( :30 — 1:30) 0 7 86 7 0

 

 

 

 

 

 

These results are then discussed below for each experimental con-

dition.

Condition 1

(Subject had high inputs and
his partner had low inputs)

All fourteen subjects in this condition indicated that they

felt they did at least two to three times as much work as their

partners.

In addition, nine of the fourteen subjects felt that the

 

 

86

quality of their answers was either somewhat or considerably better
than that of their partners. These results indicate that all the
subjects in Condition I felt that they had higher inputs than their
partners and, therefore, the independent variables were success-
fully manipulated.

Condition 2

(Subject had high inputs and
his partner had high inputs)

 

Sixteen of the seventeen subjects indicated that they felt
they worked about the same amount as their partners. Also, twelve
of these seventeen subjects felt that their answers were of the same
quality as those of their partners. These results indicate that most
of the subjects in Condition 2 felt that they had approximately the
same inputs as their partners and, therefore, the independent vari—
ables were successfully manipulated.

Condition 3

(Subject had low inputs and
his partner had low inputs)

A11 twelve subjects in this condition indicated that they felt
they had worked about the same amount as their partners and eleven
of these subjects felt the quality of their answers was the same as
their partners. Therefore, it seems that almost all the subjects in

Condition 3 felt that they had almost identical inputs as their partners

87

and, as a result, the independent variables were successfully
manipulated.
Condition 4

(Subject had low inputs and
his partner had high inputs)

Nine of the fifteen subjects indicated that they felt they
worked about % to% as much as their partners, while five of the
remaining six subjects felt they worked as long as their partners.
On the other hand, almost all (13 or 15) the subjects felt that their
answers were of the same quality as those of their partners. There—
fore, about half the subjects in Condition 4 felt their inputs were as
high as their partners, while the other half felt their inputs were
only somewhat lower than their partners. Since the independent
variables were successfully manipulated in the other three experi—
mental conditions, the difference in Condition 4 seems to reflect a
cognitive distortion on the part of the subjects.

Tests of the Predictions Generated from
the Equity and Expectancy Hypotheses

The hypotheses tested in the present research were pre-
sented in two sections of Chapter I. The first section was concerned
with the manifestation of the personal dispositions within each of the

four experimental conditions, while the second section was concerned

 

88

with the manifestation of the personal dispositions between the
experimental conditions. The results obtained to test these hypothe-
ses are presented below.

Tests of the within
group hypotheses

The results obtained relating to the within group hypotheses
are summarized in Tables 7 and 8, presented below. Table 7 sum—
marizes the percentage of role symmetry and self-interest responses
found in the four experimental conditions, while Table 8 classifies
each subject in a given experimental condition by the personal dis-

position which was his majority responses.

 

 

 

Table 7. —- Percentage of role symmetry and self—interest responsesa
Percentage Percentage
Condition Inputs of Inputs of of role of self—
the subject his partner symmetry interest
responses responses
1 High Low 21 79
2 High High 61 39
3 Low Low 25 75
4 Low High 62 38

 

 

 

 

 

aThe category of altruism has been omitted from this and
subsequent tables because no altruism responses occurred in any
condition.

 

89

sition classified by experimental condition

Table 8. -- Number of subjects manifesting a majority of each dispo—

 

 

 

Number of subjects Number of subjects
manifesting manifesting
Condition self—interest role symmetry p
more frequently than more frequently than
role symmetry self—interest
1 12 2 . 01
2 5 12 14
3 8 4 28
4 5 10 30

 

 

 

 

 

The following discussion is concerned with the results pre-

sented in the two tables above.

Condition 1

Seventy-nine percent of the responses made on the Wage
Distribution Booklet by subjects in Condition 1 were self-interest
responses, while the remaining twenty—one percent were role sym-
metry responses. In addition, a significant number (12 of 14) of
subjects acted in a self-interested manner on at least two of the
three mirror—image pairs of items on the Wage Distribution Book-
let (p < .01; sign test). The results obtained in Condition 1 were

consistent with both the equity and expectancy hypotheses.

90

C onditi on 2

Sixty-one percent of the responses made on the Wage
Distribution Booklet by subjects in Condition 2 were role symmetric,
while the remaining thirty-nine percent were self—interested.

Twelve of the seventeen subjects in the present study acted in a role
symmetric manner on at least two of the three mirror-image pairs
of items on the Wage Distribution Booklet (p < . 14,- sign test). These
results are more consistent with the equity hypothesis than with the

expectancy hypothesis.

Condition 3

Seventy—five percent of the responses made on the Wage
Distribution Booklet by subjects in Condition 3 were self-interest
responses, while the remaining twenty-five percent were role sym-
metry responses. In addition, eight of the twelve subjects acted in
a self-interested manner on at least two of the three mirror-image
pairs of items on the Wage Distribution Booklet (p < .28; sign test).
These results are more consistent with the expectancy hypothesis

than with the equity hypothesis.

C onditi on 4

Sixty-two percent of the responses made on the Wage Dis-

tribution Booklet by subjects in Condition 4 were role symmetric,

 

91

while the remaining thirty-eight percent were self-interested. In
addition, ten of the fifteen subjects tended to act in a role symmetric
manner on at least two of the three mirror-image pairs of items on
the Wage Distribution Booklet (p < . 30; sign test). These results

are not consistent with either the equity or the expectancy hypotheses.

Tests of the Between Group Hypotheses

Role Symmetry

 

As noted above, Table 7 indicates the percentage of role
symmetric and self—interested responses occurring in the four ex-
perimental conditions. To determine if the frequency of personal
dispositions differed as a function of experimental conditions, sub-
jects in each condition were classified according to whether the
majority of their responses were role symmetric or self-interested
(see Table 8). A chi square test performed on these data indicated
a significant difference in the majority response as a function of
experimental condition (X2 = 12. 96, p < . 005). To explore this
finding further the mean role symmetry score for a given condition
was compared with the mean score for each other condition by means

of individual t_ tests. 1 The results of the t_ tests are summarized in

 

1Individual t_ tests were performed instead of one of the
multiple comparison methods because of the unequal number of sub-
jects in the four experimental conditions.

92

Table 9, below. A positive t_ value in Table 9 indicates that the
equity hypothesis correctly predicted the experimental condition in
which role symmetric responses would be more frequently mani-
fested, while a negative t_ value indicates that the equity hypothesis
generated an incorrect prediction.

Table 9. --_t_ values for comparison of percentages of role-symmetry
responses occurring between experimental conditions

 

 

 

 

Condition
1 2 3 4
Condition 1 3. 52>:=>:< _ 35 3. 17*);<
Condition 2 ~2. 32>:< . 13
Condition 3 2. 35*
Condition 4

 

 

 

 

 

* p < . 05
** p < . 01
On five of the six comparisons, equity theory correctly pre—
dicted in which condition role symmetry would be more frequently
manifested. Three of the differences were statistically significant,

two at the . 01 level and one at the . 05 level.

93 "
Altruism

As noted earlier, no altruism responses occurred in any of

the four conditions.

The Effect of Relative Welfare

 

As noted above, relative welfare was the one reward struc-
ture variable systematically examined. As previously stated, the
relative welfare of the reward structure is the difference between the
payment for the individual distributing the wages and the payment of
his partner; that is, does the individual distributing the wage receive
more than, the same as, or less than his partner?

An attempt was made to Guttman scale the data in respect
to relative welfare. Results of this attempt (presented in Appendix V)
indicate that the data order on this dimension quite well. All but one
of the fifty—eight subjects'i responses scaled perfectly.

These results are discussed in the following chapter,

especially in respect to the equity and expectancy hypotheses.

IV. DISCUSSION

[The present chapter is divided into three sections. The
first section focuses on the variables of inputs of the subject and
his partner, especially in respect to the equity and expectancy hy-
potheses. The last two sections present an examination of the
variables less central to the focus of the study: (a) relative‘welfare

and (b) order of presentation and real versus hypothetical choices.

Inputs of Subject and Inputs of Partner

As noted above, the present research was mainly concerned
with the effects of the variables of inputs of the subject and his part-
ner. The equity hypothesis postulates that these variables are major
determinants of the way in which wages are distributed. Because of

this, these variables were the focus of the present study.

Inputs of Subject

The equity hypothesis states that persons should attend to
the inputs of all relevant people in a work situation, including them-

selves. This is so because the fair distribution of wages is a function

94

95

of every individual' 3 input-outcome (ratio. On the other hand, the
expectancy hypothesis states that one' 3 inputs should be irrelevant
since each person should attempt to maximize his own wages regard-
less of differences in inputs.

The (results of the present research, for the most part,
support the expectancy hypothesis. In the analysis of variance, only
one group of subjects was influenced by the manipulation of their own
inputs, and these made hypothetical choices. Moreover, this one
difference, which was attributable to the effects of inputs of the sub-
ject, was contrary to the predictions of the equity hypothesis. In
this instance, subjects with high inputs were more role symmetric
than subjects with low inputs, a finding which neither theoretical
position predicts nor can readily explain.

This lack of support for the equity hypothesis is further
emphasized by the additional analyses performed to examine in detail
the effects of inputs of the subject and his partner. Individual com-
parisons of the experimental conditions through t_ tests (Table 9,
page 92) indicates that there was no difference between those con-
ditions where inputs of the partner were the same but the inputs of
the subject were different. Comparison of Conditions 1 and 3 yielded
a t_ Value of only . 35. Likewise, a comparison of Conditions 2 and 4

yielded a t_ value of . 13.

96

These findings, taken together, imply that the complexity of
the equity hypothesis is not justified, at least for the range of input
conditions examined in the present research. Inputs of the subject
just did not matter. Contrary to the equity hypothesis, subjects, in
choosing among distributions of wages, did not take into account
their own inputs. However, they did, as noted below, take into
account the inputs of their partner.

A partial explanation of why inputs of the subject failed to
have the effect predicted by the equity hypothesis lies in the ability
of the subjects to distort reality. Adams (1965), as noted in Chap-
ter 1, indicated that distortion of reality is one mechanism by which
persons can alleviate feelings of dissatisfaction produced by per-
ceptions of inequity. This mechanism seemed to be operating in
Condition 4, where subjects worked % as long as their partners.
These subjects' answers to questions 2 and 4 of the Attitude Question-
naire (seen in Tables 5 and 6) indicated that at least 40 percent of
them felt that they did at least as much work of the same quality as
did their partners. Since the amount of role symmetry displayed in
this condition was about the same as that displayed in Condition 2-—
where both the subject and his partner had high inputs--subjects in
Condition 4 seemed to have overestimated their inputs.

Comparison of the role symmetry scores for subjects in

Condition 4, who distorted their inputs, with those who were realistic

97

provides additional support for the contention that such distortion
influences choice behavior. Those subjects who reported an accurate
perception of their inputs had a, mean role symmetry score of 2. 1
(out of a possible mean of 3.0). On the other hand, subjects who
perceived their inputs as higher than they actually were had a mean
role symmetry score of 1. 4. Moreover, the mean role symmetry
score for subjects in Condition 2 was 1. 8. Although none of these
differences is significant--probably due to the low number of sub-
jects in each comparison--these results tend to support the equity
hypothesis. Such a distortion, however, at most can be only a par-
tial explanation of why the variable of inputs of the subject did not
have the effect predicted by the equity hypothesis. First, the equity
hypothesis also did not correctly predict the behavior of subjects in
Condition 3, where both the subject and his partner had low inputs.
In this condition the perception of inputs was not distorted. Second,
as discussed in detail below, the subjects did attend to the inputs of
their partner; that is, they tended to be more role symmetric when
their partners had high inputs. Thus, it is difficult to explain why,
if subjects distorted at all, they only distorted the perception of their
own inputs.

In summary, the results of the present research-—including

the findings which indicate that some subjects distorted the perception

98

of their inputs—-imply that the variable of inputs of the subject has
less of an effect on choice behavior than predicted by the equity
hypothesis. On thewhole, these results supported the expectancy

hypothesis.

Inputs of Partner

As noted a number of times, the equity hypothesis postulates
that subjects should attend to the inputs of their partners when dis-
tributing wages. On the other hand, the expectancy hypothesis states
that this variable should be irrelevant. The findings of the present
research provide strong support for the equity hypothesis.

The results of the analysis of variance (Table 3, page 79)
indicate that of the seven comparisons which involve inputs of part—
ner, four were statistically significant. Moreover, only one addi—
tional comparison (the main effect for order of presentation) was
also significant. These results imply that inputs of partnerwas the
crucial variable examined in the present research. Further, these
findings, when examined in greater detail through analyses of the
simple effects, indicate that all of the four significant differences
attributable to inputs of partner were in the direction predicted by
the equity hypothesis. That is, subjects tended to be more role

symmetric when their partner' 3 inputs were high than when they

99

were low. In addition, differences in three of the fournon-significant
simple effects were also in the direction predicted by the equity
hypothesis.

Further support for the equity hypothesis is provided by the
additional analyses performed to examine in detail the effects of
inputs of the subject and his partner. Examination of Tables 7 and 8
reveals that when inputs of the subject are the same, more role
symmetry is displayed when the inputs of the partner are high than
when they are low. That is, there is significantly more role sym—
metry in Condition 2 than in Condition 1, and in Condition 4 than in
Condition 3. Moreover, when just inputs of partner are considered,
these results, presented in Table 8, also support the equity hypothe-
sis. In Conditions 1 and 3, where inputs of the partner are low, the
number of subjects whose major response is self-interest exceeds
those whose majority response is role symmetry. (For Conditions 1
and 3 combined, X2 = 7. 54; d. f. = 1; p < .01). In Conditions 2 and
4, where inputs of the partner are high, the number of subjects
whose major response is role symmetry exceeds those whose
majority, response is self—interest. (For Conditions 2 and 4 com-
bined, X2 = 4.50; d.f. = 1; p< .05).

In summary, the present research provides strong support

for the contention that persons both attend to the inputs of others and

100

weigh differences in these inputs when choosing among distribution
of rewards. This is precisely the prediction that the equity hypothe-
sis makes. On the other hand, the findings in respect to inputs of
partner are contrary to the predictions generated by the expectancy
hypothesis.

Taken together, the results for inputs of subjects and inputs
of partner provide insight into those conditions which generate con-
cerns with either fairness (the equity hypothesis) or profit maximi-
zation (the expectancy hypothesis). There. is strong evidence that a
person, in distributing wages to himself and another-individual, only
considers the other' 3 inputs. Subjects tend to give themselves as
much money as possible unless they perceive that their partners have
worked hard enough to deserve more than minimal pay. It is as if the
inputs of the other must exceed a threshold value before a subject
feels that his partner deserves more than this minimum. If this
threshold value is exceeded, then equity motives tend to dominate.

On the other hand, if the inputs of the other are less than this value,
subjects tend to maximize their profits. Thus, these findings have

significant implications for both the equity and expectancy hypotheses.

Relative Welfare

 

The results of the attempt to Guttman scale responses as a

function of relative welfare indicate that it was easiest for subjects

101

to give some of their potential wages to their partners when the
resultant distribution still gave them more money. Moreover, sub—
jects found it easier to give uprsome money-when the resultant dis-
tribution gave both persons the same than when it gave their partner
more than themselves. These findings are consistent with those of
Messé (1968).

These results imply that most subjects have a competitive
orientation which affected their concerns with equity and profit
maximization. That is, in making their choices, they seemed to
take. into account the difference in rewards to themselves and the
other that each distribution of wages provided. This implies that
subjects were motivated to stay ahead or at least not let the other
receive more than they. Thus, there is at least one, and probably
more motivating factors, in addition to concerns with equity and
profit maximization, which affect the distribution of rewards.
Therefore, those theorists who exclude these other potential factors
from consideration may be presenting a somewhat simplistic model

of what happens in those situations where rewards are distributed.

Effect of the Control Variables

 

As noted above, two of the situational variables, order of

presentation and real versus hypothetical choices, were manipulated

102

primarily for reasons of control. However, these variables were
found to affect choice behavior. Although these results are beyond
the focus of the present research, they are discussed below since

they havesome interesting implications.

Order of Presentation

The results of the analysis of variance (Table 3, page 79)
indicate that of the seven comparisons which involve order of pre-
sentation, four-were statistically significant. Moreover, only one
additional comparison (the main effect for inputs of partner) was
also significant. These results imply that although order of presen-
tation was a control variable, it still had a major-impact on choice
behavior.

Results of the analyses of the simple effects indicate that,
for those differences which were significant, subjects were more
role symmetricwhen they began with Series Y choices. These
findings are congruent with those of Messé (1968). A possible
explanation for these results is related to concepts of response set
and equity. For the subjects in the conditions where significant
simple effects occurred, equity is equivalent to role symmetry,
since most of these subjects perceived that they had approximately

the same inputs as their partners. When Series X was presented

 

103

first to these subjects, choosing the alternative outcome was con-
sistent with equity and role symmetry but not with self—interest.
Consistent with this set, most subjects then chose the alternative
outcome in Series Y, the choice consistentwith role symmetry and
equity. However, when Series Y was presented first, selecting the
alternative outcomewas consistent not only'with equity and role
symmetry but also withself-interest. Therefore, in some subjects
the set of self-interest may have been aroused and continued into
Series X choices. As a result, presenting Series Y first more likely

aroused the set of self-interest.

Real Versus Hypothetical Choices

The control variable real versus hypothetical choices had
less of an effect on responses than did order of presentation. How-
ever, one simple effect was significant. For pairs where both
members had high inputs, subjects making hypothetical choices on
Wage Distribution Booklets "which began with Series Y were more
role symmetric than subjectsmaking real choices. These results
replicated the findings of Messé (1968). It is interestingto note that
the situation in the present research where real versus hypothetical
choices made a difference--where both subject and partner had high
inputs--was the one which most closely approximated the conditions

of his study.

104

Since hypothetical choices did not have an effect on the
actual distribution of wages, any difference due .to this variable
should have been in the direction found; that is, subjects making
hypothetical choices should have beenmore role symmetric than
subjects making real choices. However, in the present research
this occurred only once. An explanation for this result is not readily
apparent. However, it is possible that this variable affected choice
behavior in the conditions it did because of the potential difference
with which subjects making real or hypothetical choices attended to
order of presentation. Under conditions of high inputs the equity
hypothesis predictsthat role symmetry should be dominant. It is
only-under these conditions that order of presentation had an effect.
It could be that the set toward self-interest induced by exposure to
Series Y first only occurred when subjects made real choices, since
these subjects might have paid closer attention to such subtle differ-
ences in the Wage Distribution Booklet. This explanation is highly
speculative, especially since no such difference in role symmetry
occurred as a function of real versus hypothetical choices when sub-
jects had low inputs and started with Series Y but their partners had
high inputs.

Taken together, the results for the control variables indi-

cate that othersituational factors besides inputs influenced wage

 

 

105

distribution behavior. , It seems reasonable that variables—~both
motivational, such as the competitive orientation, and situational,
such as the sequential order of relevant experiences-~otherthan
those specifically-related to the equity or expectancy hypotheses also
must be considered if a comprehensive-theory concerning wage dis-

tribution is to be developed.

V. CONCLUSION

The concluding chapter contains four brief sections. They
are (a) implications for wage distribution theory, (b) implications
of the present methodology, (c) suggestions for further research,

and (d) a concluding statement.

Implications of the Results
for Wage Distribution Theory

 

 

The stated purpose of the present research was to delineate
those circumstances in which equity motives and profit maximization
motives tend to dominate. The results indicate that inputs of the
subject' 3 co-worker is a major determinant of behavior in wage dis-
tribution situations. When the inputs of the subject' 3 co-worker
were high, equity motives tended to dominate; when inputs of the
subject' 3 co-worker were low, profit maximization motives tended
to dominate. It is as if a threshold value on the variable of inputs
of the subject' 8 co-worker must be exceeded before a subject feels
that his co-worker deserves more than minimum pay. Once the sub-

ject feels that this threshold value is exceeded, equity motives come

106

 

107

into focus and then exert a significant effect on behavior. If the
threshold value is not exceeded, profit maximization motives tend
to dominate.

However, despite the fact that equity motives tend to
dominate in certain situations while profit maximization motives
tend to dominate in others, in all four situations studied in the
present research, a substantial quantity of responses consistent with

the less dominant motive in that situation occurred. Therefore, it

 

appears that neither the equity hypothesis nor the expectancy hypothe-
sis alone can explain all the results obtained pertaining to wage dis-
tribution. Moreover, results indicate that other factors, both
motivational and situational, also affect choice behavior. Therefore,
the results of the present research, taken as a whole, suggest that
further theoretical work in the area of wage distribution should focus
on the multi-dimensional nature of man' s motivations and the situa-

tional cues which interact with them.

Implications of the Present Methodology

 

One finding in the present research questions the validity
of the procedure used by Adams and Lawler. The results in
Condition 4 (subject determining the wages has low inputs and his

partner has high inputs) of the present research indicated that a

108

substantial number of subjects distorted their inputs in relation to
those of their partner. In the procedure used by Adams and Lawler,
subjects should find it very difficult to distort inputs. The compari-
son individual is a hypothetical person described to the subject by
the experimenter. The subject is told that this hypothetical individ-
ual is more qualified than he is for a job the subject is hired to per-

form. However, the subject never sees him. Therefore, it should

 

be more difficult to distort the inputs of the other-individual in
Adams' and Lawler' s procedure than in the present research or

in ”real life" situations. Thus, there is evidence that the procedure
used in the present study is a productive tool that could be used
effectively in further research.

An additional methodological point concerns the use of the
post-session Attitude Questionnaire. Most research in experimental
social psychology uses some sort of post-session measurement
designed to assess the degree to which manipulations were success-
ful. The studies of Adams and Lawler employed such a procedure.

The present study extended this technique somewhat. For
example, the subjects in Condition 3 accurately reported on the
Attitude Questionnaire that they and their partners had the same
inputs. However, they also reported that these inputs, since they

both were low, had very little effect on their choices. These

 

 

109

responses seemed valid since subjects in Condition 3 tended to be
self-interested. While the responses indicating the effectiveness of
the manipulations were not formally presented in this respect, they
did, however, aid in the understanding of choice behavior. There-
fore, such questions might be productive additionsto post-session
measurement devices, especiallysince they aid in formulating

theoretical explanations for the observed behavior.

Suggestions for Further Research

 

The present study suggested a number of issues which could
be investigated through further research; three possibilities are dis-
cussed below.

The present research indicated that when the subject' 3
partner worked 1% hours, even though equity motives tended to
dominate, a substantial number of profit maximization responses
occurred. Research could be conducted to see if extending the
length of time that both the subject and his partner worked would
lower the number of self-interested responses. Such a test would
require two conditions.

Condition 1. Both the subject and his partner would work
10 hours.

Condition 2. Both the subject and his partner would work

1% hours.

 

 

110

It is hypothesized that subjects in Condition 1 would be
significantly more role symmetric. This would demonstrate the
strength of the equity motive under the high input conditions typically
found in industry.

Subjects in the present research displayed no altruism.
Therefore, a third condition might be introduced into this proposed
study in an attempt to induce subjects to make at least some altruistic
responses. In this condition, the subject would work 1%- hours and
his partner 10 hours. If such responses did occur, as predicted by
the equity hypothesis, they would seem to indicate that the equity
motive is relevant to those situations, such as found in industry,
where inputs are high.

Another possibility for future research would be to have
subjects perform a task requiring different degrees of co-operation
between them. This would experimentally manipulate group cohe-
siveness and simulate the work group atmostphere frequently found
in industry. It is hypothesized that the equity motive would be mani-
fested more frequently as cohesiveness increased.

A third possibility for future research would be to present
the Wage Distribution Booklet, modified appropriately, to workers
in a number of industrial settings. Since the present study found

very little difference between real and hypothetical choices, the

111

proposed research could present the alternative distributions in the
questionnaire form typically used in industrial field research. Using
an instrument, such as the Wage Distribution Booklet, in both field
and laboratory settings would provide a degree of comparability

between findings which has been lacking to date.

A Concluding Statement

 

The present research and those studies previously reported
provide evidence that persons have two motivations which are to
some degree incompatible. Clearly, man is to some extent moti-
vated by self-interest and this is the basic premise of the expectancy
hypothesis. On the other hand, under certain circumstances, as the
equity hypothesis predicts, man is also concerned with the-welfare
of others. It would be difficult for man to function as a social animal
without the presence of norms such as equity. In the absence of such
norms, it is doubtful that men could live and work together in a

mutually productive and satisfying relationship.

 

 

 

REFERENCES

REFERENCES

Adams, J. S. Toward an understanding of inequity. Journal of
Abnormal and Social Psychology, 1963a, fl, 422-436.

 

 

Adams, J. S. Wage inequities, productivity and work quality.
Industrial Relations, 1963b, 3, 9-16.

 

Adams, J. S. Inequity in social exchange. In L. Berkowitz (Ed.)
Advances in experimental social psycholggy. Vol. 2.
New York: Academic Press, 1965.

 

Adams, J. S. Effects of overpayment: two comments on Lawler' 3
paper. Journal of Personality and Social Psychology,
1968, 12, 315-316.

 

Adams, J. S. , and Jacobsen, P. R. Effects of wage inequities on
work quality. Journal of Abnormal and Social Psychology,
1964, 29, 19-25.

 

Adams, J. S., and Rosenbaum, W. B. The relationship of worker
productivity to cognitive dissonance about wage inequities.
Journal of Applied Psychology, 1962, Q, 161-164.

 

Andrews, 1. R. Wage inequity and job performance: an experi-
mental study. Journal of Applied Psychology, 1967, _5__1_,
39-45.

 

Arrowood, A. J. Some effects on productivity of justified and
unjustified levels of reward under public and private con-
ditions. Unpublished doctoral dissertation, University of
Minnesota, 1961.

Clark, J. V. A preliminary investigation of some unconscious
assumptions affecting labor efficiency in eight super-
markets. Unpublished doctoral dissertation, Harvard
University, 1958.

112

 

 

 

113

Festinger, L. Atheory of cogitive dissonance. Evanston: Row,
Peterson, 1957.

 

Friedman, A. , and Goodman, P. Wage inequity, self-qualifications,
and productivity. Organizational Behavior and Human Per-
formance, 1967, 3, 406-417.

 

 

Goodman, P. , and Friedman, A. An examination of the effects of
wage inequity in the hourly condition. Organizational
Behavior and Human Performance, 1968, 3, 340-352.

 

 

Gordon, B. F. , and Lowin, A. Qualifications, co-worker character-
istics and productivity--An extension and critique of Adams'
inequity theory. Technical Report, December 15, 1965,
Bell Telephone Laboratories.

 

Homans, G. C. Status among clerical workers. Human Organiza-
tion, 1953, _13, 5-10.

 

Homans, G. C. Social behavior: its elementary forms. New York:
Harcourt, Brace and World, 1961.

 

Jacques, E. Equitable payment. New York: Wiley, 1961.

 

Jacques, E. Objective measure for pay differentials. Harvard
Business Review, 1962, 4_0_, 133-138.

 

Lawler, E. E. Effects of hourly overpayment on productivity and
work quality. Journal of Personality and Social Psychology,
1968a, 12, 306-314.

 

Lawler, E. E. Equity theory as a predictor of productivity and work
quality. Psychological Bulletin, 1968b, 19, 596-610.

 

Lawler, E. E. , et al. Inequity reduction over time in an induced
overpayment situation. Organizational Behavior and Human
Performance, 1968, _3_, 253-268.

 

 

Lawler, E. E. , and O'Gara, P. W. Effects of inequity produced by
underpayment on work output, work quality, and attitudes
toward the work. Journal of Applied Psycholog, 1967,
_5_1_, 403-410.

 

114

Merton, R. K. , and Kitt, A. S. Contributionsto the theory of
reference group behavior. In Continuities in social research.
R. K. Merton and P. F. Lazarsfeld (Eds.) Glencoe,
Illinois: Free Press, 1950.

 

Messé, L. A. Parameters of fair play: an investigation of differ-
ences in the just distribution of rewards. Unpublished
doctoral dissertation. University of Chicago, 1968.

Patchen, M. The choice of wge comparisons. Englewood Cliffs,
New Jersey: Prentice-Hall, 1961.

 

_Porter, L. W. , and Lawler, E. E. Managerial attitudes andjer-
formance. Homewood, Illinois: Irwin, 1968.

 

 

Sawyer, J. The altruism scale: a measure of cooperative,
individualistic, and competitive interpersonal orientation.
The American Journal of Sociology, 1966, LXXI, 409-420.

 

Sayles, L. R. Behavior'of industrial work groups: prediction and
control. New York: Wiley, 1958.

 

Stouffer, et al. The american soldier: adjustment during army
life. Vol. 1. Princeton, New Jersey: Princeton Univer-
sity Press, 1949.

 

 

Vroom, V. H. Work and motivation. New York: Wiley, 1964.

 

Weick, K. E. The concept of equity‘in the perception of pay.
Administrative Science Quarterly, 1966, _13, 414-439.

 

Weick, K. E. , and Nesset, B. Preferences among forms of equity.
Organizational Behavior and Human Performance, 1968,
3, 400—416.

 

 

APPENDICES

 

APPENDIX I

WAGE DISTRIBUTION BOOKLET

 

115

 

WAGE DISTRIBUTION BOOKLET

 

 

 

 

 

116
Alternative A Alternative B
You get $4.00 You get $3.00

and and
He gets $1.00 He gets $2.00

1

 

 

 

 

 

1. Which alternative gives you
more? A or B? (Check one)

 

 

2. Which alternative, if either,
gives the greater total to
the two of you, together?
Check A or B. If the two give
the same total, check ”Equal."

 

 

Equal
3. Which alternative gives the
two of you more nearly the
same? A or B? (Check one)

 

 

SKIP THE QUESTION BELOW; GO ON TO THE NEXT PAGE

 

 

CHOICE Xl:

Which alternative set of payments
would you rather see used? A or
B? (Check one)

 

Remember: If this choice is the one drawn, you and he will be paid
just the amounts you now choose.

 

 

 

117

'Alternative A‘-
You get $4.00 YOu get $2.50
and and
He gets $1.00 He gets $2.50

I

 

 

Alternative B

 

 

 

 

 

 

 

1. Which alternative gives you
more? A or B? (Check one)

 

 

2. Which alternative, if either,
gives the greater total to
the two of you, together?
Check A or B. If the two give
the same total, check "Equal."

 

 

 

Equal
3. Which alternative gives the
two of you more nearly the
same? A or B? (Check one)

 

 

SKIP THE QUESTION BELOW; GO ON TO THE NEXT PAGE

 

 

CHOICE X2:

Which alternative set of payments
would you rather see used? A or
B? (Check one)

 

 

Remember: If this choice is the one drawn, you and he will be paid
just the amounts you now choose.

(X2)

 

 

 

 

 

 

118
Alternative A Alternative B
You get $4.00 You get $2.00

and and
He gets $1.00 He gets $3.00

 

 

 

 

 

 

Which alternative gives you
more? A or B? (Check one)

 

 

Which alternative, if either,
gives the greater total to

the two of you, together?
Check A or B. If the two give
the same total, check "Equal.”

 

 

 

Equal
Which alternative gives the
two of you more nearly the
same? A or B? (Check one)

 

 

SKIP THE QUESTION BELOW: GO ON TO THE NEXT PAGE

 

 

CHOICE X3:

Which alternative set of payments
would you rather see used? A or

B?

Remember: If this choice is the one drawn, you and he will be paid

(Check one)

 

just the amounts you now choose.

(X3)

 

 

 

 

 

 

119
Alternative A Alternative B
You get $1.00 You get $3.00

and and
He gets $4¥00 He gets $2.00

 

 

 

 

 

 

1. Which alternative gives you

more? A or B? (Check one)

 

 

2. Which alternative, if either,
gives the greater total to
the two of you, together?
Check A or B. If the two give
the same total, check ”Equal."

 

 

Equal
3. Which alternative gives the
two of you more nearly the
same? A or B? (Check one)

 

 

SKIP THE QUESTION BELOW: GO ON TO THE NEXT PAGE

 

 

CHOICE Yl:

Which alternative set of payments
would you rather see used? A or
B? (Check one)

 

Remember: If this choice is the one drawn, you and he will be paid
just the amounts you now choose.

(Y1)

 

 

 

 

 

 

120
Alternative A Alternative B
You get $1.00 You get $2.50

and and
He gets $4.00 He gets $2.50

 

 

 

 

 

 

1. Which alternative gives you
more? A or B? (Check one)

 

 

2. Which alternative, if either,
gives the greater total to
the two of you, together?
Check A or B. If the two give
the same total, check ”Equal.”

 

 

Equal
3. Which alternative gives the
two of you more nearly the
same? A or B? (Check one)

 

 

SKIP THE QUESTION BELOW; GO ON TO THE NEXT PAGE

 

 

CHOICE Y2:

Which alternative set of payments
would you rather see used? A or
B? (Check one)

 

 

Remember: If this choice is the one drawn, you and he will be paid
just the amounts you now choose.

(Y2)

 

 

 

 

 

121
Alternative A Alternative B
You get $1.00 You get $2.00

and and
He gets $4.00 He gets $3.00

. l

 

 

 

 

 

 

1. Which alternative gives you
more? A or B? (Check one)

 

 

2. Which alternative, if either,
gives the greater total to
the two of you, together?
Check A or B. If the two give
the same total, check ”Equal.”

 

 

Equal
3. Which alternative gives the
two of you more nearly the
same? A or B? (Check one)

 

 

SKIP THE QUESTION BELOW; GO ON TO THE NEXT PAGE

 

 

CHOICE Y3:

Which alternative set of payments
would you rather see used? A or
B? (Check one)

 

 

Remember: If this choice is the one drawn, you and he will be paid
just the amounts you now choose.

(Y3)

APPENDIX II

A TTI TUDE QUE STIONNAIRE

 

122

ATTITUDE QUESTIONNAIRE

 

 

 

,,

123

Please check the best answer for the following questions.

1. I found this study to be

very somewhat neither somewhat very
interesting interesting interesting uninteresting uninteresting
nor unin-
teresting

2. In comparison to the people in the other group how much work
do you feel you did?

 

4 times as about 2 to about the about 1/3 to 1/4 as much
much or more 3 times as same 1/2 as much or less
much

3. How important was the difference in the amount of work between
yourself and the people in the other group in influencing your
choices on the wage distribution booklet?

 

 

 

 

 

It was £§g_major It was one of It was a factor It was a factor

factor influenc- several impor- but not as im- of little or no

ing my choices tant factors portant as other importance in—
influencing my ones in influ- fluencing my
choices. encing my choices choices

4. In comparison to the peOple in the other grogp I feel the
quality of my answers were

 

 

 

 

considerably somewhat about somewhat considerably
better better the same worse worse

 

124

5. How important was the difference in quality of the work between
yourself and the people in the other group in influencing your
choices on the wage distribution booklet?

 

 

 

 

It was £23 major It was one of It was a factor It was a fac-
factor influenc— several impor— but not as im- tor of little
ing my choices tant factors portant as other or no impor—
influencing my ones in influ— tance influ-
choices encing my choices encing my
choices

6. What other factors beside quantity and quality of the work af—
fected your choices on the wage distribution booklet? Explain

your answer.

7. Please write below any comments you may have concerning the
study.

APPENDIX III

 

INDUSTRIAL OPINION QUESTIONNAIRE

 

125

 

INDUSTRIAL OPINION QUESTIONNAIRE

Research Form 2

 

126

Name

Age?

Instructions: Please spend about 8 minutes on each question.
Try not to spend too much time on any particular question. For
your convenience, the time remaining will be written periodi-
cally on the board in the front of the room. The questionnaire
will be collected in l 35 hours. Please try to answer all 12
questions.

 

127

What do you feel will be your job when you finish your
education?
What were the major factors influencing your career choice?

What college courses do you feel were the most beneficial
to you?

 

128

How much confidence do you place in psychological tests as
a device for selecting employees for a job?

How would you feel about having to take a psychological
test before you were hired?

 

129

If you were asked to take a psychological test for a job
you wanted how likely is it that you would fake your
answers to try to make a very favorable impression?

What personality traits do you feel are generally desired
in prospective employees?

Why do you feel these qualities are desired in prospective
employees?

To what extent do you feel you possess these qualities?

 

130

While working on your job you are reprimanded by your super—

visor for making an error in your work. The error was really
your associate's.

a. What would you do in such a situation?
b. Why would you do this?
c. What factors would influence your decision?

d. wa do you feel the average workingman would react to
such a situation?

 

131

While working for a large company you make a suggestion to

your supervisor which he reports to a vice-president as his

suggestion. The suggestion proves to be a valuable one. It

saves the company a considerable amount of money.

a. What would you do in such a situation?

b. Why would you do this?

c. What factors would influence your decision?

d. How do you feel the average workingman would react to
such a situation?

 

132

Yen are praised by your supervisor‘s boss for an idea that

was your associate's.

a. What would you do in such a situation?

b. Why would you do this?

c. What factors would influence your decision?

d. How do you feel the average workingman would react to
such a situation?

 

133

What factors do you feel are the most important ones for
employees being satisfied with their jobs?

Which factor of the ones mentioned in part a. do you feel
is the most important one?

Why did you select this one factor? Explain fully.

 

134

YOur company introduces a new plan. Whenever any suggestion

by a worker saves the company money Eli the employees share

in the benefits.

a. Do you like this plan better than one in which only the
person who thought of the suggestion gets the bonus?

b. Explain the reasons for your answer in part a.

c. Which plan do you feel would be more successful?

 

135

In deciding which of two companies to work for what three
factors would be the most important ones influencing your

decision?

Why did you select these three? Explain fully.

 

10.

136

Some companies train their promising young executives by

having them work two years on a job and then transferring

them to another job. These job changes often involve

moving to another part of the country;

a. Would you accept such a job?

b. Why?

c. What were the most important factors influencing your
decision?

 

 

 

11.

137

Some companies have unions that are not affiliated with the
larger national unions.
a. Do you feel these local unions can be as effective in

representing their members as the ones who are affil-
iated with national unions?

b. What did you base your answer to part a. on?

 

 

 

 

———_______ 7”

12.

138

It is widely believed that most employees now place little

pride in their work.

a. Do you feel this is true?

b. If you feel this is true what do you feel is responsible
for this occurring? If you feel this is not true, why
do you believe so many people feel that most employees
place little pride in their work?

 

APPENDIX IV

 

STUDENT BACKGROUND QUESTIONNAIRE

139

Psychology 255

Instructor: Mr. Lane

Student Background Questionnaire

1. Name
2. Sex
3. Age
4. Major

5. Grade Point Average

6. Previous Job Experience
Type of Job How Long Employed

a.

b.

c.

d.

e.

f.

7. Please answer the following question. Spend about 10-15 minutes

on your answer.

Do you feel labor unions have become too powerful at the
present time? Explain your opinion on this issue.

Icheb 12-27-68

 

APPENDIX V

 

GUTTMAN SCALED ROLE SYMMETRY RESPONSES

140

Personal Disposition Manifested When Series X Alternative Outcome

Gave Subject Less Wages, Equal Wages, and More Wages

Than His Partner

 

 

Subj e ct Numbe r

Series X Alternative Outcome Gave Subject

 

Less Wages Equal Wages More Wages

 

 

 

 

 

141

 

 

Series X Alternative Outcome Gave Subject

 

Subject Number
Less Wages Equal Wages More Wages

 

17

18

19

20

 

21

22

23

24

25

26

27

28

29

30

31

32

33

34

NNNMNNNNNNN

m
NNNNNN

35

 

 

 

142

 

 

Subject Number

Series X Alternative Outcome Gave Subject

 

Less Wages

Equal Wages

More Wages

 

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

 

NNNNNNX‘NNNNN

 

NMNNNN

NNNNNNNNNNN

 

NNNNNNNNNNNNNNNNNN

 

143

 

 

Series X Alternative Outcome Gave Subject

 

 

 

Subject Number
Less Wages Equal Wages More Wages
SS4 X X X
855 X X X
S56 X X X
S57 X X X
S58 X X X

 

 

 

 

3A blank box represents a self-interested response while
an X represents a role symmetric response.

S 42 responses were the only ones that did not Guttman
scale.

 

 

 

_!__

.< L

 

”71111131112!

  
       

TATE UNIV

93 030

 

 

 

 

 

 

 

 

 

 

   

fllflillfllgfiflml“