LIBRARY

Michigan State
University

 

 

 

PLACE N RETURN BOX to roman thin chockout «on: your record.
TO AVOID FINES Mum on or baton data duo.

T—_———————"
DATE DUE DATE DUE DATE DUE

Nov???“ ”gm 33"
JUL 2 3 2007 u H

 

 

 

 

 

 

  
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

usu loAnNfirmaflvo Mon/Emu Oppomnuylmmwon
m

 

THESIS

3/

NIVERSI

IIIIIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIII

3 1293 01570 8492

 

 

 

 

This is to certify that the

dissertation entitled

CONCEPTUALIZING THE PSYCHOLOGICAL CONTRACT CONSTRUCT:
THE IDENTIFICATION AND EMPIRICAL INVESTIGATION
OF CRITICAL ISSUES

presented by

Mark Vincent Roehling

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Business Administration

 

 

/J/Z/%

Date f//C/;7

MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771

'

w

r r—-'w 1" 1 v

.'

r'— r- -‘ -'

‘ ‘W—

 

LIBRARY

Michlgan State
University

 

 

 

; PLACE IN HEI'URN BOX to remove thie checkout from your record.
' TO AVOID FINES return on or before date due. i

DATE DUE DATE DUE DATE DUE

INVIIEIWW
I

JUL 2 3 2007

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I nu
— _____IE
fiI—Tm

MSU le An Affirmative ActioNEquel Opportunity Inetitulon

 

 

 

 

 

 

 

 

 

 

 

 

 

THESIS
{‘9

J2

”IIIIIIIIII]IIIIIIIIIIIIIII "

3 1293 01570 8492

This is to certify that the

dissertation entitled

CONCEPTUALIZING THE PSYCHOLOGICAL CONTRACT CONSTRUCT:
THE IDENTIFICATION AND EMPIRICAL INVESTIGATION
OF CRITICAL ISSUES

presented by

Mark Vincent Roehling

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Business Administration

 

 

/J/Z/%

Date f//C/97

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

LIBRARY

Michigan State
University

 

 

 

PLACE Ii RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU le An Affirmative ActiorVEquel Opportunity Inetltuion

 

I: Eligi" \II .I

«I

CONCEPTUALIZING THE PSYCHOLOGICAL CONTRACT CONSTRUCT:
THE IDENTIFICATION AND EMPIRICAL INVESTIGATION
OF CRITICAL ISSUES.
By

Mark Vincent Roehlin g

A DISSERTATION
- Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Management

1997

ABSTRACT
CONCEPTUALIZING THE PSYCHOLOGICAL CONTRACT CONSTRUCT:
THE IDENTIFICATION AND EMPIRICAL INVESTIGATION
OF CRITICAL ISSUES
By

Mark Vincent Roehling

This dissertation has three main components. First, an extensive review of the
psychological contract (PC) literature is conducted and four issues that, it is argued, are
critical and need to be addressed in future conceptualizing are identified and discussed.
Second, an integrative conceptualization of the PC construct is presented. Third, a field
study that investigates one of the four issues that were identified is reported and discussed.
The study uses survey data collected from 1052 employed participants to investigate the
equivalence of three measures of the employee PC construct that are based on
conceptualizations of the construct that incorporate alternative belief elements
(obligations, expectations, or promises). The results of multiple group confirmatory factor
analyses suggest that the measures have similar factor patterns. The pattern of
correlations between the respective measures and external criteria indicates that while the
measures behave Similarly in many regards, depending on the measurement purpose,
results may vary significantly depending on whether the PC measure focuses on

Obligations, expectations or promises.

DEDICATION

This dissertation is dedicated to my beloved family, Pat, Alex, and Julia.

iii

ACKNOWLEDGMENT

I would like to acknowledge my gratitude and appreciation for the constructive
criticism, advice, and support I received from all of the members of my dissertation
committee: John Hollenbeck, Georgia Chao, Dan Ilgen, and Alison Barber. I am I
particularly obliged to John Hollenbeck for his role as my chairperson, his training in
research methods, and his manner of challenging me while providing the necessary
support.

I would also like to acknowledge my gratitude and appreciation for the enormous
support and encouragement provided by my Spouse, Patricia Roehling. Her support

came in many forms, and I feel greatly obliged.

iv

TABLE OF CONTENTS

List of Tables ............................................... ix

List of Figures .............................................. xi

INTRODUCTION ............................................. 1
Introduction ........................................... 1
Overvrew . . ..3

CHAPTER 1: REVIEW OF THE LITERATURE .......................... 6
Theoretical Origins of Psychological Contracts Concept ................ 6
Origins of the Term ”Psychological Contract” ..................... 10
Early Developments .................................... 13
The Generalization of the Psychological Contract Concept .............. 23
Rousseau and Colleagues ................................. 24
Other Recent Developments ................................ 40
Areas of Agreement Regarding the Nature of .................... . 44
Psychological Contracts

CHAPTER II: CRITICAL ISSUES IN CONCEPTUALIZING THE ........... 46

PSYCHOLOGICAL CONTRACT CONSTRUCT

The Level of the PC Construct .............................. 47
The Nature of the Beliefs ................................. 51
The Discrete ("Transactional”)-Relational Continuum .............. 58

V

DEDICATION

This dissertation is dedicated to my beloved family, Pat, Alex, and Julia.

iii

ACKNOWLEDGMENT

I would like to acknowledge my gratitude and appreciation for the constructive
criticism, advice, and support I received from all Of the members Of my dissertation
committee: John Hollenbeck, Georgia Chao, Dan Ilgen, and Alison Barber. I am
particularly obliged to John Hollenbeck for his role as my chairperson, his training in
research methods, and his manner of challenging me while providing the necessary
support.

I would also like to acknowledge my gratitude and appreciation for the enormous
support and encouragement provided by my spouse, Patricia Roehling. Her support

came in many forms, and I feel greatly obliged.

iv

TABLE OF CONTENTS

List of Tables ............................................... ix

List of Figures .............................................. xi

INTRODUCTION ............................................. 1
Introduction ........................................... 1
Overview ......................................... ‘. . .3

CHAPTER 1: REVIEW OF THE LITERATURE .......................... 6
Theoretical Origins of Psychological Contracts Concept ................ 6
Origins of the Term "Psychological Contract” ..................... 10
Early Developments .................................... 13
The Generalization of the Psychological Contract Concept .............. 23
Rousseau and Colleagues ................................. 24
Other Recent Developments ................................ 40
Areas of Agreement Regarding the Nature of .................... . 44
Psychological Contracts

CHAPTER II: CRITICAL ISSUES IN CONCEPTUALIZING THE ........... 46

PSYCHOLOGICAL CONTRACT CONSTRUCT

The Level of the PC Construct .............................. 47
The Nature of the Beliefs ................................. 51
The Discrete ("Transactional")-Re1ational Continuum .............. 58

V

The Role of Individual Differences ........................... 70

CHAPTER III: AN INTEGRATIVE CONCEPTUALIZATION OF THE ....... 77
PSYCHOLOGICAL CONTRACT
Defining the Construct ................................... 77
Level of the PC Construct ................................ 79
Relevant Relational Level Variables ......................... 80
Nature of Beliefs ...................................... 83
Antecedents of EPCS ..................................... 97
The Usefulness of the Discrete-Relational Continuum ................ 101
Relational Employment, Relational Obligations, Relational EPCS ....... . 106
CHAPTER IV: RESEARCH QUESTIONS AND HYPOTHESES ............ 113
The Humpty Dumpty Approach to Defining Constructs ............. 113
Measurement Equivalence ................................ 1 14

Obligations versus Expectations: A Difference That Makes a Difference?. . . .116

Hypothesis 1A ................................... 119

Hypothesis 1B .................................... 119
Obligations versus Promises: Again...A Difference That Makes a ........ 119
Difference?

Hypothesis 2A .................................. . 120

Hypothesis 2B .................................... 120
Investigating the Pattern of Relations with External Criteria ............ 120

Hypothesis 3A .................................. 123

Hypothesis 3B .................................... 124

vi

Hypothesis 4A .................................. 124
Hypothesis 4B ................................... 124

Hierarchieal Nature of Perceptions: Obligations, Expectations, Promises .. . . 124

Hypothesis 5A .................................... 126
Hypothesis SB ................................... 126
Hypothesis 6A ..................................... 126
Hypothesis 6B ..................................... 126
Fulfillment Measures and the EPC Nomological Network .............. 126
ObligationsVersusExpectationS.........................127
Hypothesis 7 ..................................... 131
Hypothesis 8 ..................................... 132
Hypothesis 9 .................................... 133
Obligations Versus Promises ........................... 133
Hypothesis 10 .................................... 133
Hypothesis 11 ..................................... 134
Hypothesis 12 ................................... 134
CHAPTER V: METHODS .................................... 134
Overview of Methods ................................... 134
Sanqfle ........................................... 135
Measures ......................................... 136
Procedures ........................................ 151
Analytic Procedures .................................. 154

vii

H I

CHAPTER VI: RESULTS ..................................... 164

Internal Consistencies of Measures .......................... 164
Effectiveness of Randomization Procedure ...................... 164
Preliminary Factor Analysis .............................. 167
Multiple Group CFA ..................................... 172
Correlations with External Criteria ........................... 180
Hypothesis 5A and 5B ................................... 186
Hypothesis 6A and 6B ................................... 188
Hypothesis 7 through 12 ................................ . . 197
CHAPTER VII: DISCUSSION .................................. 206
Analysis of Latent Variables .............................. 207
Patterns of Correlations with External Criteria .................... 209
Post Hoc Regression Analysis .............................. 220
Hierarchical Nature of Obligations, Expectations, and Promises ........ 223
EPC Fulfillment and Job Satisfaction, Trust and Turnover Intentions ..... 225
Summary of Evidence Regarding Equivalence and Implications ......... 228
Limitations .......................................... 236
Summary of Contributions ............................... 238
Future Research ...................................... 240
APPENDDC A: Employee Attitudes and Beliefs Survey ................... 244
APPENDD( B: Anecdotal Illustration of Promise Framework Limitations ....... 258
LIST OF REFERENCES ..................................... 260

viii

Table 1:

Table 2:

Table 3:

Table 4:

Table 5:

Table 6:

Table 7:

Table 8:

Table 9:

Table 10:

Table 11:

Table 12:

LIST OF TABLES

Occupations ....................................................................... 137
Description of Subsamples by Source ......................................... 139
Alpha Coefiicients, Means, and Standard Deviations of Measures ........ 165

Assessing the Efi‘ects of Randomization: Results of One-Way
ANOVAS ............................................................................................. 166

Principle Axis Factor Analysis Employee Side Items-Obligation
Group Only...... ................................................................................ .168

Principle Axis Factor Analysis, Employer Side-Obligation
Group Only ........................................................................................... 170

Result of LISREL Confirmatory Factor Analysis-Obligation
Group Only.. ........................................................................................ 172

Results Of LISREL Multiple-Group Analysis of Equivalence:
Employerltems... 176

Results of LISREL Multiple-Group Analysis Of Equivalence:
Employee Items ..................................................................... 178

Means, Standard Deviations, and Coefficient Alpha for EPC
Measures: Total Item Scores and Sub-Scales ................................ 181

Correlations Between EPC Measures and Equity Sensitivity,
Work Centrality, Education, Tenure, Trust, Gender, and Age ............ 183

Planned T-Test Comparisons to Assess Differences in Mean
Number of Items in EPC Across Forms ....................................... 187

ix

Table 13:

Table 14:

Table 15:

Table 16:

Table 17:

Table 18:

Table 19:

Table 20:

Table 21:

Planned T-Test Comparisons to Assess Overall Level
Differences Across Form3189

Planned T-Test Comparisons to Assess Mean Factor Level
Differences Across Forms ...................................................... 191

Planned T-Test Comparisons to Assess Mean Individual Item
Differences Across Forms ........................................................ 193

Correlations Between Perceived EPC Fulfillment Measures and
Global Job Satisfaction, Trust, and Intention to Turnover ................. 198

Correlations Between Perceived EPC Fulfillment Measures and
Job Satisfaction Subscales .................................................................... 201

Hierarchical Regression Of Job Satisfaction, Trust, and Turnover
Intentions on Fulfillment Measures ............................................ 204

Correlations Between Employer side EPC and Measures of Trust ......... 211

Antecedents of EPC Employer Side Factors: Hierarchical .

Regression Results ............................................................... 221

Antecedents of EPC Employee Side Factors: Hierarchical

Regression Results ............................................................. 222
x

 

Figure 1:

Figure 2:

Figure 3:
Figure 4:
Figure 5:
Figure 6:
Figure 7:

Figure 8:

LIST OF FIGURES

Characterization of Obligations in Rousseau (1990) and

Robinson, Kraatz, and Rousseau (1994) .................. 32
Conceptualizations of Psychological Contract Construct by

Level of Construct ............................... 49
The Nature of the Belief Element of Psychological Contracts ..... 52
Antecedents of Psychological Contracts ............... '. . 100
Models of Equivalence and Factor Analytic Corollaries ........ 117

Approaches Used to Assess Employee Psychological Contracts . . . 143
Measurement Model Employer’s Side ........................ 174

Measurement Model Employee’s Side ........................ 175

INTRODUCTION

The relationship between individual employees and their employers has been
conceptualized as involving a "psychological contract”. At a general level, the term
"psychological contract” is used to refer to a set of beliefs regarding what employees
are to give and receive with respect to their employer. Early references to psycholog-
ical contracts ascribed them a central role in organizational behavior. Psychologieal
contracts were described as "dominating” the relationship between employee and
employer (Argyris, 1960, p. 96), and as Operating ”powerfully as determinants of
behavior” in organizations (Schein, 1965, p. 11).

Acceptance of the concept of a psychological contract has grown steadily and is
being reflected in a variety of forums, including: scholarly journals (Guzzo, Noonan,
& Elron, 1994; Rousseau, 1990; Van Dyne, Grahm & Dienesch, 1994), management
books (Bowditch & Bruno, 1982), textbooks (French, Kast, & Rosenzweigh, 1985;
Noe, Hollenbeck, Gerhart, & Wright, 1993; Schuler, Bentell, & Youngblood, 1989),
practitioner journals (Baker, 1986; Dunahee & Wangler, 1974; Goddard, 1988;
Tomow, 1988), and the pOpular press (Yates, 1993).

It has been argued that the violation of a psychological contract can have important
individual and organizational consequences, including deep, long lasting feelings of
betrayal and resentment, anger and frustration, decreased employee motivation, job
dissatisfaction, reduced employee commitment, turnover, employee initiated litigation,
and unionization efforts (cf. Cole, 1981; Rousseau, 1989; Schein, 1980; Tomow,

1988). Organizations are being told that the management of psychological contracts is

2

essential to the organization's successful functioning (cf. Covey, 1988; Goddard, 1988;
Herriot, 1988; Kotter, 1973). Human resource professionals are being urged to
explicitly consider the implications of the psychological contract construct for their
organization's human resource practices (Guzzo & Noonan, 1984; Rousseau & Greller,
1994; Sims, 1944), and schemes have been offered which prescribe the type of
employee psychological contract that is likely to be most appropriate for an
organization based upon the organization's business strategy (defenders, prospectors,
analyzers, responsives; cf. Rousseau & Wade-Benzoni, 1994).
Acceptance of the concept of a psychological contract and the importance attributed it
has greatly outpaced developments in the explication and empirical assessment of the
psychological contract mnslmni. Efforts to formally explicate a psychologieal contract
construct are scarce. As the existing literature moves from discussing the general
notion of a psychological contract to specifying the elements necessary to fully
articulate a theory, there is considerable disagreement regarding a number of
fundamental issues such as the level at which psychological contracting occurs (e.g.
individual, dyadic) and the nature of the beliefs that constitute the psychological
contract (e.g. expectations, obligations, likelihoods, hopes). These and other
important issues are seldom explicitly addressed, and it is not unusual for the writings
of a Single person to reflect inconsistent treatment of these issues.

The relatively few attempts that have been made to empirically assess the psycho-
logical contract construct have done little to resolve these disagreements or otherwise

promote movement toward a conceptual consensus. The empirical studies have

3

typically adopted a particular conceptualization, with little or no discussion of compet-
ing conceptualizations. Similarly, while there are several thought provoking non-
empirical articles that discuss psychological contracts in organizations (e.g. Rousseau,
1989), no published article provides an in depth discussion of competing conceptual-
izations of the construct or the critical issues that should be addressed in any
conceptualization Of the construct.

The psychological contract construct is a promising vehicle for integrating theory
and findings from a number of established literatures that have been found to be useful
to explain behavior in organizations (e. g. social exchange theory, equity, procedural
justice). AS a construct, however, the psychological contract construct is still in a
relatively early stage of development. The evolution of the construct from an
intuitively appealing concept that means very different things to different people to a
construct that may assume a ”key role” (Schein, 1970) in understanding behavior in
organizations requires that attention be focused on current competing conceptualizations
and the issues that are critical to future theorizing.

Overview

The focus of my dissertation is the psychological contract construct. I seek to
address what I perceive as the lack of conceptual clarity that characterizes much of the
thinking about psychological contracts (PC) that appears in the literature. The focus of
my dissertation reflects the influence of Ilgen and Hollenbeck (1991), Van Dyne,

Cummings, & McLean Parks (1994), and others who have articulated the importance

4

of attending to the theoretical development of ”the constructs themselves” (Ilgen &
Hollenbeck, 1991, p. 199).

I view my dissertation as having three main components. First, using an inductive
approach, I conducted a comprehensive review of the literature and identified four
issues that I argue are critical and nwd to be addressed in all future conceptualizing
about employee PCs. The results of the review are reported in Chapter 1. Non-
empirical works that did not appear to make an original contribution are not reported.
Nonetheless, the review of the PC literature that is provided is substantially more
comprehensive than any review that currently exists in the literature (no attempt to
provide a comprehensive review exists in the literature). Areas of agreement among
those writing about PCS are also identified in Chapter I. The four critical issues that
should be addressed in future conceptualizing are discussed at length in Chapter II.

The second major component of my dissertation is the development of an integrative
conceptualization (or re-conceptualization) of the employee PC construct. My
conceptualization and supporting arguments are presented in Chapter III. Among other
considerations, the four critical issues that I have identified are specifically addressed.
With further time spent on articulating the PC's nomological network, my effort in this
regard might be characterized as working ”Toward a Theory of Psychological
Contracting in Organizations".

The third main component of my dissertation is a field study that focuses on the
investigation of one of the critical issues that were identified. The study investigates the

equivalence of three measures that are based on alternative conceptualizations of the PC

5

construct. The Specific hypotheses that were investigated are set forth in Chapter IV.
In addition to those hypotheses, in the process of investigating the equivalence of the
three measures, evidence was collected that is relevant to a number of other hypotheses
that have been repeatedly asserted in the literature but, based on my review, never

empirically investigated.

6
CHAPTER I

Review of the Literature
Theoretical Origins of the Concept of Psychological Contracts

Social contracts. It has been suggested that the PC concept is an extension of what
has been written by philosophers about social contracts (Schein, 1980 p.22). With
origins traceable to ancient Greek philosophers, social contract theories and theorists
(e.g. Hobbes, 1651/1946; Locke, 1690/1967; Rousseau, 1762/1988) experienced their
”heyday” in the seventeenth and eighteenth centuries. Gough (1978) distinguishes
between two kinds of contracts referred to under the name "social contract”. One kind
of contract, which deals with the origin of the state, supposes that a number of
individuals who were living in a state of nature voluntary consented to form an
organized society. Commonly associated with this kind of social contract is the notion
of ”natural rights” that belong to individuals as a result of their agreement to form an
organized society.

The second form of social contract may be called the ”contract for government“ or
the ”contract for submission” (Gough, 1978, p. 3; cf. Barker, 1980). This form of
social contract presupposes the existence of a state, addressing the terms on which the
state is to be governed. In this sense, the social contract is an ”agreement" regarding
the reciprocal rights and duties of the state and its citizens. For example, the governed
promise to pay taxes, obey the laws, and share the risk of defense in exchange for

security, protection, and opportunity for development provided by the state.

7

While the concept of a social contract originated in the context of philosophical
inquiries aimed at prescribing principles and rights, more recently, some writers and
researchers have used the term social contract descriptively, to refer to shared,
collective beliefs regarding appropriate behavior in a social unit (cf. Rousseau &
McLean Park, 1993). Further diffusion and translation of the term can be found in the
management literature. For example, Goldberg (1994) defines an ”implied social
contract" between employer and employee that is said to consist of an unwritten set of
rights and obligations that each expects the other side to live up to. This definition
sounds very much like the employee PCs described by Schein (1980), and others,-
suggesting that the PC literature is now influencing thinking about "social contracts".

Employer/employee relations as involving exchange. Bamard's (1938) theory of
equilibrium employs an exchange perspective in describing the conditions under which
an organization can get its members to continue their participation. According to the
theory, each employee receives from the organization inducements (pay, etc.) for
which the employee makes contributions. An employee will continue his participation
so long as the contributions offered him are as great or greater (measured in terms of
the employee's "values") than the contribution the employee is asked to make. The
inducement-contribution model discussed in March & Simon's (1958) book
Organizations, built upon Bamard's equilibrium theory, further detailing the nature of
the exchanges influencing an employee's decision to participate in an organization.
March and Simon allude to the concept of unwritten contractual obligations between the

organization and employee: " In joining the organizations, he (the employee) accepts an

II

8

authority relation, i.e. he agrees that within some limits (defined both explicitly and
implicitly by the terms of the employment contract) he will accept as the premise of his
behaviors orders and instructions supplied by the organizations." (1958, p. 90).

Menninger (1958). A chapter in Karl Menninger's book WM
Technique (195 8), provides a general discussion of "two-party contracts" as a preface
to his discussion of the psychotherapist-patient contract in particular. Menninger
emphasized that contracts and contractual relationships require that the exchange
between the parties result in the reciprocal satisfaction of the parties' needs in order for
the contractual relationship to be continued:

[I]n any engagement between two individuals in which a transaction occurs, there is

an exchange, a giving and a gain of something by both parties with a consequent

meeting of the needs in a reciprocal way, mutual way. When this balance is not

achieved, either because one does not need what the other has to offer or because

one does not give what the other needs or because there is a feeling on the part of

one that the exchange is not a fair one, the contracts tends to break up

prematurely.(Menninger, 1958, p. 21)
Using Simple examples such as a man purchasing apples from a vendor, and the barber-
client relationship, Menninger illustrates how the parties' needs are met in a variety of
tangible and intangible ways (e.g. the pleasure of companionship). This discussion is
credited with having made a substantial contribution to the origination of the PC
construct (cf. Levinson, et al., 1962, discussed at length, below).

The reciprocity norm. The norm of reciprocity has been incorporated into theories
of social exchange, and later, conceptualizations of PCs. Gouldner's (1960) article is

the standard reference provided in discussions of the norm of reciprocity. Gouldner

distinguished between reciprocity as a pattern of exchange and the general "moral”

9

norm of reciprocity. As a pattern of exchange, reciprocity may be defined as the
mutually contingent exchange of benefits between two or more units. Beyond
reciprocity as a pattern of exchange, there is a generalized norm of reciprocity which
defines certain actions and obligations as repayments for benefits received. The
generalized norm of reciprocity leads us to experience obligation to others because of
what they have previously done for us (i.e. because of the history of previous
interaction we have had with them; Gouldner, 1960, p. 171). The norm of reciprocity
plays a central role in exchange theories, and exchange theories, in turn, play a central
role in most conceptualizations of PCS.

Social exchange theories. Social exchange theories (e. g. Homans, 1961; Blau,
1964) extended the exchange perspective of human behavior by explicitly distinguishing
between social exchange and strictly economic exchange. Two primary distinctions are
made. First, social exchange entails unspecified obligations, while economic exchange
is said to require specified obligations. Second, since there is no way to assure the
appropriate performance of an unspecified obligation (e. g. the return of a favor), social
exchange both requires and promotes trust among the exchange partners, strictly
economic exchange does not (Blau, 1964). Social exchange theories point to the
potential role of non-economic, ”social exchange commodities" that may motivate
behavior in organizations. For example, it is argued that a basic reward that people
seek in their associations is social approval (Blau, 1964). The conscious or unconscious
desires to obtain this basic reward may explain helping behaviors in organizations that

do not lead to economic rewards.

10

Origins Of The Term "Psychological Contract"

Both Argyris (1960) and Levinson (1962) have been given credit for introducing the
"psychological contract" terminology. Argyris (1960) used the term "psychological
work contract” to describe an implicit understanding between a group of employees and
their foreman that arose as a result of a particular leadership style. He observed that
the foremen at a plant in which he was conducting field research had a ”passive” or
"understanding" leadership style. Argyris attributed this to the fact that all of the
foreman had come up through the ranks, and in the process, they had been influenced
by the informal employee culture. As a result, the foreman realized that the way to get
the employees to behave in the desired manner was to maintain the informal employee
culture and not to behave in a way that violates the culture's norms. Argyris
hypothesized that one result of the passive or understanding leadership style was an
employee-management relationship that was "dominated" by the "psychological work
contract". He described the contract as follows:

Since the foreman realize that this system will tend to produce optimally under

passive leadership, and since the employees agree, a relationship may be

hypothesized to evolve between the employees and the foreman which might be
called the ”psychological work contract“. The employee will maintain the high
production, low grievances, etc., if the foreman guarantees and respect the norms of
the employee informal culture (i.e. , let the employees alone, make certain they

make adequate wages, and have secure jobs)" (Argyris, 1960, p. 97).

Argyris observed that the psychological work contract between the employees and their
foremen was deemed to have been violated when the foremen were required, by upper

management, to implement a budgeting system that usurped employees sense of control

over their work.

11

The parties to the psychological work contract identified by Argyris were a group of
rank and file employees who shared certain norms and their immediate supervisors,
who were at least aware of the employees' norms. Argyris seems to suggests that the
informal employee culture was the driving factor behind the formation of the
psychological work contract. He states that the "predispositions” of employees which
were the basis for the contract were Shaped by the employee culture. But, Argyris
aclmowledged that the extent to which the relevant predispositions were brought by
employees to the work place versus were shaped by the work place was uncertain.

Although it may have been unspoken, the PC that Argyris described involved»
actual mutual agreement; the parties had the same understanding of what they were
obliged to do in order for the contract to be maintained. The extent to which Argyris'
conceptualization of the psychological work contract was bound to the particular
context he was investigating, or whether it was intended to have greater
generalizability, is unclear.

Levinson's similarly termed "psychological contract" concept was clearly intended
to generalize across people and organizations. In a book titled Men,_Management._and
MentaLHealth (Levinson, Price, Munden, Mandl, & Solley, 1962), Levinson et a1.
introduced their conceptualization of employee psychological contracts. While earlier
works of Argyris were cited, Argyris' (1960) concept of the psychological work
contract was not referenced in the Levinson et al. (1962) book. Rather, Levinson's
thinking about PCs apparently evolved out of a study funded by the Menninger

Foundation in which 874 employees at a large utility were interviewed in order to

12

investigate the effects of the work experience on mental health. In the course of
conducting the interviews, they observed that when people Spoke about their work, they
Spoke of expectations, and that these expectations seemed to have an obligatory quality,
”as if the company were duty-bound to fulfill them" (p. 20). This observation, they
report, reminded them of Karl Menninger's ( 1958) discussion of the intangible aspects
of contractual relationships, out of which they evolved the concept of the psychological
contract.

Levinson et a1. defined PCS as "a series of mutual expectations of which the parties
to the relationship may not themselves be dimly aware but which nonetheless govern
their relationship to each other. " (p.21). The expectations of both the individual
employee and the company were conceived of as "components” of the PC. Levinson et
al. describe the expectations as being mutual in the sense that each Side to the PC
tacitly agrees to the other side's expectations (often described as "demands"). The
mutual expectations that make up the PC were described as having two characteristics:
1). they are largely implicit and unspoken, and 2). they frequently antedate the
relationship of the person and the company.

Levinson et a1. identified a number of different types of employee expectations,
both unconsciously and consciously held, that may make-up the employee's Side of the
PC. Unconscious expectations included those having to do with psychological issues,
such as nurturence. They also identified more explicit expectations having to do with
job performance, the use of Specific skills, social relations in the work place, job

security, and economic rewards.

13

The company's expectations were described as arising out of the company's history
and business environment. They suggested that inferences regarding the company's
expectations may be drawn from the circumstances under which it operates, it polices
and practices, it values, statements at managers’ meetings, and its evaluations of
employee job performances. Levinson et al. deduced that the company in which they
conducted their interviews expected employees to: be good citizens in the community,
be concerned about cost control and efficiency, acquire Skill to advance to supervisory
level, be dedicated to the principal of free enterprise, and demonstrate a certain amount
of employee flexibility

While expectations were described as frequently antedating the employment
relationship, Levinson et a1. explicitly recognized the dynamic relationship of PCs.
They viewed PCS as often evolving or changing over time as a result of the changing
nwds of the individual or the organization, and through the reciprocal interaction of the
two parties.

Finally, Levinson et al. noted that in addition to the PC between the individual and
the organization, there are also in effect PCs "of lesser proportion” between people in
the organization, within work groups, and between groups and the organization. They
argued that these PCS may be viewed as ”collateral agreements that have bearing on the
person-organization relationship" (Levinson, et al. , 1962, p. 38).

Early Developments in Thinking About the Concept of a Psychological Contract

Schein (1965, 1970, 1980). Schein's belief that the psychological contract is a key

concept in understanding behavior in organizations is reflected in the extensive

14

treatment the concept receives throughout all editions of his book ”Organizational
Psychology” (1965, 1970, 1980). Schein originally described the PC as follows:

The notion of a psychological contract implies that the individual has a variety of

expectations of the organization and that the organization has a variety of expecta-

tions of him. These expectations not only cover how much work is to be performed
for how much pay, but also involve the whole pattern of rights, privileges, and
obligations between worker and organizations. For example, the worker may
expect the company not to fire him after he has worked for a certain number of
years and the company may expect that the worker will not run down the company's
public image or give away company secrets to competitors. Expectations such as
these are not written into any formal agreement between employer and organization,

yet they operate powerfully as determinants of behavior (1965, p. 11)

The second edition of Schein's book contains the same description of the PC.
Schein also notes that the concept of a psychological contract relates to the inducement-
contribution model of March and Simon (Schein, 1970, p. 13). The third edition of the
book expanded the description somewhat, making a linkage between psychological
contracts and organizational roles, and elaborating on the nature of the respective
expectations.

According to Schein, individual employees forge their expectations from their inner
nwds, what they have learned from others, traditions and norms which may be
operating, their past experiences, and ”a host of other sources" (Schein, 1980, p. 24).
Schein also States, however, that ultimately the relationship between the individual and
the organization is interactive, unfolding through mutual influence and mutual
bargaining to establish a workable psychological contract. The psychological contract

is viewed as changing over time as the organization's and employee's needs change,

requiring that the contract be constantly renegotiated.

15
Gibson (1966). Gibson's (1966) theory of absence takes into account the

psychological contract, incorporating the view that absence behavior was ”very closely
bound up with contractual relationships between the worker and organization” (p. 113).
According to Gibson, the need-oriented individual and the goal-oriented organization
negotiate a "work contract" that Specifies the various rights and duties of the worker
and the organization. In describing the work contract, a distinction is made between its
formal contract and "quasi-contract" components. The formal contract is explicitly
agreed upon, usually reduced to writing, and it specifies relatively few rights and duties
of the parties. The quasi-contract involves an unwritten understanding of the rights and
duties of parties regarding when there is a ”consensus between the parties" (p. 117).
Gibson describes the term "psychological contract" as being a less formal term that
alludes to the individual's perception of the quasi-contractual aspect of the work
contract.

In Gibson's theory of absence, the work contract and the perception of how it is
being fulfilled influences absences in at least two ways. First, the contract specifies
which kinds of absences are legitimate and those that are not. The Significance of this
categorization is that legitimate absences would not be the basis for punishment, and
the worker may be entitled to compensation. Second, workers' perceptions of
management's adherence to the work contract is thought to influence their assessment
of their own obligations. Gibson argues that if management is perceived as not meeting

its obligations, workers feel relieved of their obligations, allowing them to legitimize a

16

greater range of behavior (including absences) outside or at variance with the formal
contract.

Jurek (1968). Based upon my review of the literature, Jurek's (1968) unpublished
dissertation reports the earliest effort to quantitatively assess the psychological contract
concept. The purpose of the study was to assess the relationship between the strength
of the psychological contract of route salesmen (i.e. the extent to which a psychological
contract existed) and their sales performance. According to Jurek's conceptualization,
a psychological contract between the worker and his company existed to the extent that
there were agreed upon expectations W. The psychological contract
was assumed to develop along the motivator and hygiene nwds of the employee, as
described in Herzberg‘s two-factor theory of motivation (Herzberg, Mausner, &
Synderman, 1959).

Subjects in the study were 107 salesmen for eight laundry and dry cleaning
companies and 29 supervisors. The nature of employee psychological contracts and the
extent of their existence was assessed using questionnaires that addressed the six
motivator and nine hygiene job factors stated by Herzberg et al. (1959). Each party to
the contract (the routeman and his supervisor) first indicated the importance to
themselves of selected motivator and hygiene items, and then each indicated the extent
to which the other was meeting their needs. The extent to which a psychological
contract existed was ”established” by comparing the extent to which each party met the
other’s needs. The percentage of needs met by the other party was calculated for each

party in the dyad, and the lower of the two percentages was taken as the ranking of the

17

extent to which a psychological contract existed. Simple correlational analysis indicated
that sales performance was greatest when the psychological contract was of moderate
strength and reasonably well met by the parties. There did not appear to be a linear
relationship between the psychological contract and sales performance, prompting the
researcher to conclude that the relationship appears to be dependent on other factors.
Obvious limitations in the conceptualization and design of the Jurek study make it
difficult to draw any firm conclusions from its modest findings.

Lawless (1972). lawless (1972) described the psychological contract as having
"the most general and all pervasive influence over a person's behavior while he carries
out his work as a member of his organization” (p. 144). His thinking about
psychological contracts appears to have been heavily influenced by the writings of
Schein (1965). Lawless describes the PC as involving the employee and organization's
mutual expectations, and as being dynamic. Lawless, however, incorporates the notion
of self-fulfilling prophecies into the psychological contracting process, focusing on
management perceptions and how those perceptions influence management action and
workers actions. In Lawless' view, management makes certain assumptions about
workers according to some general philosophy. These assumptions lead management to
treat workers in certain ways. The worker perceives how the organization perceives
him and responds accordingly: "He is like little Johnny who is perceived by the parents
as their "problem child". He, perceiving their perception, rises to the challenge and
meets their expectation. This is the self-fulfilling prophecy in operation again” (p.

146). Consistent with this view, lawless explicitly states that because management is

18

the dominant force in the management-worker interaction, management's side of the
contract has more influence on the outcomes of the relationship (p. 148).

Kotter (1973). Kotter defines the psychological contract as “an implicit contract
between an individual and his organization which specifies what each expect to give
and receive from each other in their relationship. "(p. 92). While Kotter's
conceptualization of the PC incorporates the expectations of both the employee and
employer, unlike conceptualizations that define the PC as involving agreed upon
expectations, Kotter allows incongruent employee-employer expectations within the
psychological contract. He introduces notion of "matching" to describe the situation
where the employee and organization agree regarding a given expectation, and the
extent to which the PC is comprised of ”matched” or ”mismatched” expectations is
treated as an independent variable of interest.

Kotter (1973) discusses the findings of an unpublished master thesis that applied his
matching concept to investigate the role of the psychological contract at early stages of
the employee's involvement with an organization. Ninety randomly selected, relatively
recent graduates of the Sloan School of Management's master program completed a
questionnaire that asked them to indicate the extent to which they expected to give
certain things to their employer (e.g. ability to work with group, conformity) and the
extent to which they expected to receive certain things from their employer (e.g.
interesting work, personal development Opportunities). Responses were provided using
a five point Likert scale, l=not expected, 5 =strongly expected. The individual

employee expectations were purportedly matched with employer expectations regarding

19

giving and receiving the same things. Kotter reports that psychological contracts which
were made up primarily of matches in expectations were related to greater job satisfac-
tion, productivity, and reduced turnover (p. 92). Kotter emphasized it was the
matching of expectations that was important, claiming that mismatches that gave
employees more than they expected caused as many "problems” as those that gave the
employee less than had been expected (p.94).

It was stated that the employee-employer expectations were "purportedly” matched
because, although many details of the study are not reported in the article, it appears
that in addition to indicating their own expectations, the employee-respondents were
also asked to indicate what their employer's expectation were pertaining to them.
Thus, what appears to have been ”matched” was an employee's expectations and the
employee's perceptions of his/ her employer's expectations, both being assessed
retrospectively. If this is the case, the validity of the interesting findings Kotter
discusses might be challenged for a variety of reasons (e. g. levels of analysis issues,
mono-method bias, causal ambiguity).

Kotter (1973) also briefly discussed a case study in which the expectations of new
employees, their supervisors, and senior managers from the same organization were
measured, and matches and mismatches in expectations were identified. A workshop
was conducted in which misunderstandings and misperceptions were “aired and
resolved” (p. 97). In a footnote provided in the third edition of his book, Schein

(1980, p. 22) indicates that case study Kotter (1973) reported ”tested” his idea of a

20

psychological contract. This footnote suggests that Schein viewed Kotter's
conceptualization of the PC as being at least consistent with his own conceptualization.

Portwood and Miller (1976). Portwood and Miller defined the PC as "an implicit
agreement negotiated between the employee and employing firm, usually at the
employee's time of entry, and it is a recognition of mutual obligations to be fulfilled by
both parties in the course of their association'(p. 109). They present a model of the
psychological contracting process that posits that individual's expectations about their
employment relationship are influenced by factors such as their individual needs,
attitudes toward work, relevant job knowledge and experience. An organization's
expectations pertaining to the individual and individual's job are said to become
formalized into policies and management practices, creating a ”job reality". The extent
to which the individual's expectations match this job reality is positively related to the
individual's job satisfaction, commitment, and productivity.

The Portwood and Miller (1976) study focused on the relationship between the
extent to which employees' expectations were met, which they referred to as the degree
of ”job integration", and two dependent variables: a) job satisfaction; and b) a measure
of employee' s satisfactoriness to the organization (task competence, commitment,
compatibility). Job integration was operationalized as the difference between job
expectations about outcomes at the time of hire and the extent to which the outcomes
were perceived to have been fulfilled after taking the position. The study was
conducted in a midwestem retail firm with 43 locations in the Michigan area. They

found a correlation of .37 between job integration and job satisfaction, and a

21

correlation of .25 between job integration and job satisfactoriness. Portwood and
Miller conclude that their findings indicate that organizations’ compliance with
employee PCS is positively related to employee job satisfaction and work behaviors.
Holtz (1978) . Holtz' S unpublished dissertation reports an exploratory, qualitative
study that investigated the PCS of 13 managers working in a single, large organization.
She was particularly interested in assessing how PCS change as a function of personal
development over one's life cycle and career development. Holtz defined the
psychological contract as ”an ongoing, implicit agreement between an individual and a
company, made up of a variety of expectations that specifies the needs and obligations
of each and, thus, dictates the relationship between the two parties. "(p. 23). The
individual's expectations were assessed through interviews that were "open-ended but
not unstructured" (p. 68). The company's expectations were assessed through
interviews with the 13 managers' superiors and members of the personnel department,
and from a review of company written philosophy, policies , and practices. Based upon
her findings, Holtz concluded that: a.) the PCS of young managers (n=4) were
characterized by expectations for technical training, education and broad experience,
frequent job moves with increasing challenge and responsibility, and adequate
supervisory support; b.) the PCS of middle-age managers (n =5) were characterized by
a review, reassessment, and resetting of expectations based on their past experiences
with the company and in anticipation of their futures; and c.) the PCS of older
managers (n =4) were characterized by a sense of finality, a resolute acceptance of the

past, and a determination to make the most of the final stage of their careers. They

22

expected recognition (including advancement within limits), opportunities for continued
contribution, and security.

Weick (1979). In his influential book, "The Social Psychology of Organizing”
(1979), Weick's discussion of individuals' reasons for joining organizations
incorporates the PC concept. Weick appears to accept both Schein' s definition of PC
and Schein's belief that PCS play a central role in organizational behavior. Schein's
(1965) definition of the PC is restated with apparent approval, and in Speaking of PCS,
Weick states: ”Satisfaction, productivity, interpersonal ties, and the likelihood of
leaving are all dependent on the terms of the contract and its fate at any given moment
of time." (Weick, 1979, p. 19).

Nicholson & Johns (1985). Building on the work of Gibson (1966), Nicholson and
Johns employed the concept of PC to provide a new theoretical framework for
understanding absence behavior. They described the psychological contract as being
comprised of mutual assumptions about employment rights and obligations (p. 400),
and as being ”the essence of the individual-organizational linkage, because employment
entails an implicit exchange of beliefs and expectations about what constitutes
legitimate actions by either party. " (p. 398).

Nicholson and Johns argue that one's PC is likely to influence interpretations of
absence events and absence behavior. The level at which they conceptualize the PC
construct, and its relation to organizational culture, is, however, somewhat ambiguous.
In presenting a typology of absence cultures based on the degree of salience of the

culture and the level of trust inherent in the PC, and in stating that the trust inherent in

23

the PC influences the content of the absence culture (p. 402), Nicholson and Johns
appear to be defining the PC as either an antecedent or part of the organization's
culture. But, they also describe psychological contracts as emerging through interaction
and communication (which derive from culture), and as the "psychological mechanism
by which collective influence is translated into individual behavior” (p. 398). The
latter descriptions suggest that PCS are influenced by organizational culture, mediating
the influence of organizational culture on individual behavior. Finally, although they
describe the PC as the essence of the individual-organizational linkage, they seem to
assume that classes of employees (e. g. by occupational class) share the same PC, .
indicating a group level construct. As Rousseau (1989) has suggested, it appears that
Nicholson and Johns (1985) are describing issues of psychological contracting at some
level of conceptualization above the individual level.
The Generalization of the Psychological Contract Concept

While the concept of a PC was originally employed to describe employees'
relationships at work, it has Since been argued that the concept generalizes to a variety
of relationships, including those between renter and landlord (Radford & Larwood,
1982), consultant and client (Boss, 1985), husband and wife (Dunahee & Wangler,
1974), and student and teacher (Kolb, Rubin, & McIntyre, 1984). In fact, Menninger,
whose discussion of the psychological aspects of ”two-party" contracts (Menninger,
1958) influenced Levinson et a1. 'S adoption of the term, revised his discussion of two-
party contracts so that the second edition of his book (Menninger & Holzman, 1973)

expressly adopts the term psychologicalcontract to describe the relationship that exists

24

between the client and the therapist. Given the application of the term to a variety of
relationships, it will be increasingly important for theoretical statements to indicate
whether they purport to apply to PCS in general, or whether the statements are limited
to PC involving certain kinds of relationships (e.g. employer-employee).
Rousseau and Colleagues

In recent years, no one has been more influential in the psychological contract
literature than Denise Rousseau. In terms of generating interest in employee PCs, it
would be difficult to over-state the contribution of the work she has published,
individually and with others. This section focuses on the work of Rousseau and her
colleagues. I will begin by discussing what may be described as Rousseau's seminal
article regarding PCS, a theoretical piece that appeared in a 1989 issue of the Emplnyee
W. The article is frequently referenced in
contemporary writings about PCS, including Rousseau's current work. Next, three
Studies conducted using a Single group of MBA students, and two data collection times,
1987 and 1989, will be discussed. The studies were reported in three separate
publications: Rousseau (1990), Robinson, Kraatz, & Rousseau (1994), and Robinson &
Rousseau (1994). The findings of these studies have been cited to support a variety of
claims. In addition to discussing the Studies' main findings, I will identify
inconsistencies that have apparently gone unnoticed, calling into question the claims
based on the studieS' findings. Further review and critical analysis of the work of
Rousseau and colleagues will be provided at various places in subsequent sections of

my proposal (Chapter III and Chapter IV).

25

Rousseau (1989). Rousseau initially described the PC as follows:

The term psychological contract refers to an individual's beliefs regarding the terms

and conditions of a reciprocal exchange agreement between the focal person and

another party. Key issues here include the belief that a promise has been made and

a consideration offered in exchange for it, binding the parties to some set of

reciprocal obligations. (p. 123) I
The general statement provided in the first sentence identifies the level at which the PC
was conceptualized. Rousseau (1989) explicitly distinguished between
conceptualizations of contractual construct at the level of the individual level versus the
level of the relationship (eg. dyadic, inter-organizational). Her conceptualization of PC
as involving individuals: subjective beliefs is at the individual level. The parties to the
contract, the employee and the employer, need not agree. In contrast,
conceptualizations that view PC as existing to extent that agreement exists between the
parties are at relational level.

Although the first sentence of Rousseau's description of PCS (quoted above) refers
to beliefs generally, the qualifying statements made in the second sentence of the
quoted definition reflect Rousseau's focus on beliefs about obligations. More

particularly, Rousseau's conceptualization focuses on a specific kind of obligation:

those that are W. The central role of promise in Rousseau's

 

‘ The role of consideration in contract law is fraught with intricacies and contro-

versy. It has been described as a "tortuous and extraordinary concept" that is too ad hoc
to be of explanatory value, and, is so complex that it offers little guidance (Levin &
McDowell, 1983, p. 29). W (1968) defines over twenty ”types” of
consideration. I would argue that what value the concept of consideration has in the law
of contracts (as a device for sorting contracts that Should be legally enforceable from
those that Should not) does not transfer to the process of psychological contracting,
particularly where the process is viewed as involving subjective perceptions.

26

conceptualization of PCS is clearly conveyed in her writings. For example, Rousseau
(1989, p. 126) describes PCS as involving ”an individual's belief that a promise of
future return has been made, a consideration or contribution has been offered (and
accepted), and an obligation of future benefit exists”, and Rousseau (1990) states that
”psychological contracts differ from the more general concept of expectations in that
contracts are promissory and reciprocal” (p. 390; emphasis added).

The focus on promised based obligations sets Rousseau's conceptualization apart
from every conceptualization of the PC that preceded her. Others have focused on
obligations (e. g. Gibson, 1966; Nicholson & Johns, 1985); other conceptualizations
have emphasized the importance of individual subjective perceptions (e. g. Stagner,
1964; Dunahee & Wangler, 1974); and, most conceptualizations have recognized that
promises play a role in Shaping the PC. But no one prior to Rousseau (1989) suggested
that perceived promises were the sole basis for the beliefs that constitute the PC. It is
quite likely that the implications of using a promise framework in conceptualizing PCs
are not readily apparent to the reader. They will be discussed at length in Chapter III
(at pp. 87-96).

Subsequent work by Rousseau maintains the focus on promise, but it also contains
general definitions of PCS, such as that provided in the first sentence of the definition
set forth above. This is significant because researchers describing Rousseau's (1989)
conceptualization of employee PCS, including some who purport to adOpt it (e. g.
Guzzo, Noonan, & Elron, 1994), have focused on her general statements and left out

the most distinctive aspect of her conceptualization - the focus on promise.

27

Rousseau (1990). In the first of three studies conducted using the single group of
MBA students, Rousseau (1990) imports the notion of a transactional-relational
continuum to the PC literature (the terms were briefly alluded to in Rousseau, 1989, p.
135). According to Rousseau, Macneil (1985) argues that there are two types of
contracts, anchoring opposite ends of a continuum underlying contractual arrangements
(p. 391)”. Transactional contracts are described as involving Specific monetizable
exchanges between parties over a Specified period of time. Relational contracts, in
contrast, are described as involving open-ended agreements to establish and maintain a
relationship that involves both monetizable and non-monetizable exchanges.

The 1990 study investigated whether recruits' perceptions of employee and
employer obligations were interrelated, and if so, whether the pattern was consistent
with the two types of contracts. The subjects were 228 MBA students who had
accepted job offers. The subjects' PCS were assessed by asking them to indicate what
they believed were their obligations to the particular employer whose job offer they
accepted, and what they believed were their future employer's obligations to them.
Using a five point Likert scale (1 =not at all obligated, 5 =very highly obligated),
subjects indicated to what extent their employer was obligated or owed them:

1. promotion and advancement

2. high pay
3. pay based on your current level of performance (merit pay)

 

Macneil does not argue that the two contract ideals actually exist, and he has repeat-
edly emphasized that modern contractual relations involve mm discrete (what
Rousseau describes as transactional) and relational elements cites. This important
point will be discussed more fully in Chapter H.

28

training

long-term job security

career development (development)
support with personal problems

>195“?

Beliefs about employee obligations were assessed using the same type of scale, with
subjects indicating the extent to which they believe that they were obligated to:

work extra hours (overtime)

loyalty

volunteer to do non-required tasks (extra role behavior)
advance notice if you were taking a job elsewhere (notice)
willingness to accept a transfer

refusal to support your organization's competitor
protection of proprietary protection

Spending a minimum of two years in the organization

9°99??pr

Canonical analysis was conducted to determine the underlying relations between
Specific employee and employer obligations. The study derived its measures of
psychological contract obligations through canonical factor scores weighting employee
obligations in terms of perceived employer obligations, and vice versa. Analysis of
sets of employee and employer obligations yielded two significant canonical functions.
The first function linked employee obligations to work overtime, to volunteer to do
non-required tasks, and to give notice before quitting, along with absence of an
obligation for company loyalty, with a set of employer obligations to provide high pay,
performance based pay, training, and career development. The second function
revealed relationship between employee obligations (loyalty and minimum length of
employment) with employer obligations to provide job security.

Rousseau characterized the first function as reflecting an empirical relationship

between hard work on part of employees and exchange for high extrinsic reward,

29

consistent with the terms of a transactional contract. The second function was
characterized as reflecting the relationship between loyalty and continued membership
and job security, and was said to be consistent with a relational contract (Rousseau,
1990, p. 395). It Should be noted that the items were not characterized on an a priori
basis as being either indicators of either transactional or relational contracts, and there
was little discussion of the theoretical basis for the characterizations of canonical
functions that were made. Also, the basis for considering Specific obligations as
transactional versus relational are not entirely obvious. For example, Rousseau
characterized the employer obligation to provide employees career development .
opportunities as an extrinsic reward that was consistent with the terms of a transactional
contract (p. 395). Such an obligation is relatively general, and it may be viewed as
implying that there is an on-going relationship between the parties that is expected to
continue in the future, indicating a relational contract (Macneil, 1980, 1985). The same
might be said of the obligation to provide training, depending on whether the obligation
pertained to specified training for a current job, versus training to improve Skills and
prepare for future positions (the item did not distinguish between these levels of
training).

Robinson, Kraatz, & Rousseau (1994). The second study involving the Rousseau
( 1990) MBA subjects sought to examine: a.) the factor structure of the items used to
assess EPCS; b.) changes in employee PCS over time; and c.) violations of PCS and
how they relate to transactional and relational contracts. The questionnaire that was

used in 1987 was mailed in 1989 to the same MBAs (now graduated), and a total of

30

216 responded. Of those who responded to the first questionnaire, 128 responded to
the second.

The factor structure of items at time 1 (1987) and time 2 (1989) was evaluated using
exploratory factor analysis. At both times, employer obligations to provide promotion
and advancement (described in the report as ”rapid advancement“), high pay, and pay
based on performance loaded onto one factor. They described this factor as reflecting
”transactional" obligations in that the obligations are specific, monetizable, and of a
quid pro quo nature that requires them to be fulfilled at a definite time'(p. 143).
Employer obligations to provide long-term job security, training, and career
development also loaded on a single factor at both times. This second factor was
described as reflecting ”relational obligations in that the obligations serve to maintain
the relationship and are not of a quid pro quo nature but rather, require fulfillment over
an indefinite period during the course of the relationship.” (pp. 143-144). One
obligation, support with personal problems, loaded on the so-called relational factor at
time 1, but did not load on any factor at time 2.

Employee obligations to work extra hours, to be loyal, and to volunteer to perform
non-required tasks loaded on the same factor at time 1 and time 2. This factor was said

to represent relational obligations, but no explanation for this characterization was

 

The use of adjective "rapid" appears to be a post hoc characterization. The actual
instrument Simply asked respondents to indicate the extent to which they believed that
their employer's obligations to them included "promotion and advancement”, and the
adjective "rapid" was not referred to in Rousseau's ( 1990) description of the same
obfigafion.

31

provided. At both time, employees' obligations to give advance notice, to be willing to
accept transfer, to refuse to support the organizations competitors, and to protect
proprietary information loaded on a Single factor that was described as reflecting
"transactional obligations in that the obligations concern legal issues and possibly
constitute formal rules that are necessary in a transactional relationship in which the
focus iS not long-term trust." One item, spending a minimum of two years with the
organization, loaded on the so-called transactional factor at time 2, but did not load on
any factor at time 1.

Robinson et a1. (1994) conclude that ”in sum, the factor analysis results reveal that
both employee and employer obligations can be categorized as either relational or
transactional. " (p. 145). But, the careful comparison of the factor analysis findings and
their interpretation with the arguments made and findings reported in Rousseau ( 1990)
would seem to require a different conclusion. It appears that the factor analysis findings
fail to support Robinson et a1. conclusions.

I have prepared Figure l to support the Statements made in the preceding paragraph.
The table compares how the identical obligations where characterized in the first study,
Rousseau (1990), versus how they were characterized in the second study, Robinson,
Kraatz and Rousseau (1994). Recall that each obligation was measured by a single
item, identical items were used in the two studies, and the subjects in the two studies
were from a Single sample of MBA students. The left hand column of the table reports
how obligations were characterized in Rousseau (1990). Under each of the two main

headings, "Employer Obligations” and "Employee

32

Figure 1

Characterization of Obligations in Rousseau (1990) and Robinson, Kraatz, & Rousseau (1994)

 

 

 

  
 

 

 

 

  

 

Rousseau (1990) Robinson, Kraatz & Rousseau (1994)
We; __p_xc1__Is__Em lo Obl' Idiom
A. Transactional (lst canonical function) A. Transactional
hish pay hish I»)!
pay based on current performance pay based on current performance
. . 8
career development
promotion and advancement ‘
B. Relational (2nd canonical function) B. Relational
long-term job security long-term job security
training
career development
support with personal problems
C. “Other“ Items (not characterized in the study) C. “Other Items (not characterized in the study)
promotion and advancement
support with personal problems
fimflusdfiflnflma Emmuaflfimnflma
A. Transaeuonal A. Transactional
work extra hours
volunteer to honorequircd tasks
advance notice (of taking job elsewhere) advance notice
willingness to accept a transfer fl
refusal to support your organization competitors
protection of proprietary interest

spendingaminimumoftwoyears

 

B. Relational B. Relational
lord!!! low“!
spending a minimum of two years
work extra hours
volunteer to do non-required tasks
C. '0ther'(notcharacterizedinthesmdy) C '0ther‘(notcharacterizedintheetudy)

 

willingress to accept a transfer
refusal to support your organization's competitors
protection of proprietary interest\

 

 

 

 

33

Obligations", the obligations are listed in three categories: "Transactional",
”Relational" , or "Other" . The Other category includes obligations that were not
characterized in the Rousseau (1990) study because they were not included in either of
the canonical functions. The right column categorizes the same items based on how
they were characterized in Robinson et a1. (1994). Because all of the items used to
assess PCs were characterized as either relational or transactional in Robinson et a1. ,
the right column of Figure 1 does not list any items under the "Other" category.

Arrows have been drawn in Figure l to indicate those items that switched categories
from the first study to the second study. For example, in Rousseau (1990), the ,
employer's obligation to provide training and the employer's obligation to provide
career development were both characterized as being transactional obligations. But as
the arrows indicate, in Robinson et a1, those same items were characterized as being
relational items. The employer obligation to provide promotion and advancement was
not characterized in Rousseau (1990), but was characterized as being a transactional
obligation in Robinson etal.(1994)

There was even more switching among the employee obligations. Employee
obligations to work extra hours and volunteer to do non-required tasks were described
as transactional obligations in Rousseau (1990). It was argued that these obligations and
their canonical counterpart reflected relations of hard work on the part of employees for
high extrinsic rewards. In Robinson et al. , however, these two obligations loaded on

the same factor as the obligation to be loyal, and it was simply stated that "this factor

34

represents relational obligations" (Robinson et a1. 1994, p. 144); no further explanation
was provided.

The employee obligation to spend a minimum of two years that was characterized as
relational by Rousseau (1990) was characterized as transactional by Robinson et a1.
(1994). Finally, as Figure 2 indicates, three employee obligations that were not
included in either relational or transactional canonical function in Rousseau (1990)
were subsequently characterized as being transactional. The theoretical basis for their
categorization is not readily apparent. I would argue that the nature of these items are
particularly likely to be contextually bound; it depends on what, if anything, was.
expressly communicated between the parties.

The amount of recharacterization between the two reports is considerable. Of the
ten obligations that were characterized as being either relational or transactional in
Rousseau (1990), fifty percent (5 of 10) were characterized as being exactlxjhe
opposite in Robinson et al. (1994). Again, the studies used identical items to assess the
perceived obligations, and the subjects came from the sample. There may be sound
reasons for the recharacterizations that occurred. The difficulty presented for those
reviewing the PC literature is that no reasons were provided.

In addition to investigating the factor structure, Robinson et a1. (1994) also sought
to investigate changes in PCs over time, and, the impact of contract violations on
relational and contractual obligations. To assess changes in PCs over time,

MAN OVAs were computed for both employer and employee obligations, treating the

time 1 obligations (1987) as the dependent variables and the time 2 obligations as the

35

covariates. Significant MANOVA findings were followed up by conducting paired-
sample t-tests to assess differences between the individual obligations at time 1 and time
2. The findings indicated that most perceived employee obligations decreased over
time, and most employer obligations increased over time.

Robinson and Rousseau (1994). The data for the third study using the Rousseau
( 1990) MBA students was apparently collected at the same time as that analyzed in the
second study. Subjects were the 128 graduated MBAs who responded to both the first
and second questionnaires. The impact of violations on a number of dependent variable
was assessed both quantitatively and qualitatively. Contract violations were
quantitatively assessed two ways. First, a single item asked subjects to indicate the
extent to which employee fulfilled promised obligations, l=very poorly fulfilled,

5 =very well fulfilled. Second, a single dichotomous item asked subjects to indicate
whether their employer had ever failed to meet the obligations that were promised.

The qualitative assessment of violations involved an open response format question that
asked subjects to explain the nature of violations that occurred.

Fifty-four percent of the subjects responded that their employer had, at sometime,
violated an obligation owed them. The dichotomous measure of violation and the
continuous measure of the extent of contract fulfillment were cross-tabulated, and the
results indicated that some people who reported that an obligation of violated contracts
reported a significant degree (moderate to high) of fulfillment. There are a number of
possible explanations for this finding. Fundamentally, the finding raises a question

regarding the reasonableness of using the obligation fulfillment scale as a measure of

36
violations, and the reasonableness of treating an employer failure to keep any promise
as a measure of contract violation. It would seem plausible that some subjects
recognized that a promise had not been kept, but did not treat the unmet promise as a
violation because they perceived that there was an explanation for failure.

Other analysis found that the occurrence of violations correlated positively with
turnover, and negatively with trust, satisfaction, and intentions to remain with the
employer. Robinson and Rousseau acknowledged their measurement of violations at a
single time did not allow them to establish causality of violations impact on trust; lower
trust may lead to greater likelihood to perceive violations.

Other, non-empirical work by Rousseau and colleagues. Rousseau and
colleagues have published a number of other, conceptual articles that provide further
evidence of their view of PCs. In a chapter in the Researclrinmganizationalfiehayjor
series titled ”The contracts of individuals and organizations", Rousseau and McLean
Parks (1993) argue that obligations between employee and employer can come from
one of two forms of contract: social or promissory. Rousseau and McLean Parks
defined social contracts as follows, providing the citations indicated:

Social contracts are normative, addressing shared collective beliefs regarding

appropriate behavior in a social unit (often discussed at the level of the organization

or society by anthropologists and philosophers, Gough, 1963; Homans, 1961).

Norms of exchange, reciprocity, and prosocial behavior can create obligations

between work or within organizations. Such social contracts are in effect cultural.
As indicated earlier, at page 6, this descriptive use of the term social contract is

significantly different than philosophers' use of the term to prescribe rights and

responsibilities. They describe promissory contracts as being ”essentially paid for

37

promises where commitment of future behavior is offered in exchange for payment"
(Rousseau & McLean Parks, 1993, p. 4). Promissory contracts are said to have three
components: promise (a commitment to do something); payment, a commitment
offered in exchange for the promise - which they identify as being the equivalent of
”consideration" in the law of contracts; and acceptance, consent to the terms of the
contract reflecting the contract's voluntariness. They describe PCs as "an individual's
beliefs regarding terms and conditions of a reciprocal exchange between that person and
another party (Rousseau, 1989)”, but they go on to state that their use of the term
focuses on an individual's belief in and interpretation of a promissonroontraot
(Rousseau & McLean Parks, 1993, p. 19).

A discussion of transactional-relational continuum by Rousseau & McLean Parks
(1993), and its apparent re-characterization of earlier findings, further obscures and
undermines the claims that there is empirical support for the classification of PCs as
transactional or relational. The reader will recall that Rousseau's (1990) canonical
analysis of employer and employee obligations yielded two significant canonical
functions. The canonical function said to be associated with transactional contracts was
described as reflecting a relationship between hard work on the part of employee in
exchange for high extrinsic reward (Rousseau, 1990, p. 395). However, Rousseau and
McLean Parks write: " Rousseau (1990) found that two factors characterized patterns of
employee and employer obligations reported by new recruits, a transactional one
emphasizes fasmraok advancement and pay for performance and a relational one

including loyalty for long term commitments. " The re-characterization is misleading

38

because based on the results of the eanonical analysis, the item regarding promotion
and advancement was not even among those included in the canonical function that was
described as being transactional (the item was not included in either canonical function;
see Rousseau, 1990, p. 395).

Rousseau recently served as a guest editor of a special issue of Humanfiesouree
Management that was devoted to employment contracts. She co-authored two articles
that appeared in the issue. In the first article titled "Human resource practices:
Administrative contract makers” Rousseau and Greller (1994) argued that a major
function of an HRM department is to foster the appropriate PC. In this article, PCs
were defined as "an individual's system of beliefs, shaped by the organization,
regarding the terms of the exchange agreement between himself/herself and the
organization (Rousseau, 1989; Rousseau & Parks, 1993)" (Rousseau & Greller, 1994,
p. 384). The article makes a strong statement regarding the antecedents of PCs.
Rousseau and Greller state: "employee interactions with the organization are the source
of information about the contract" (emphasis added, p. 386). In emphasizing the
importance of employee-organization interactions as the source of PCs, the role of
individual differences or societal influences is relegated to the distant background.

Rousseau's second article in the special issue was titled: ”Linking strategy and
human resource practices: How employee and customer contracts are created.“
(Rousseau & Wade-Benzoni, 1994). In this article, a definition of PCs that emphasizes
Rousseau's focus on promise was provided: ”Psychological contracts refer to beliefs

that individuals hold regarding promises made, accepted, and relied upon between

39

themselves and another. "(p. 466). The article describes what its claims are four
”common" types of PCs in United States firms. To arrive at the four types of PCs,
two contact terms that were said to be emphasized in contemporary employment
relations, time frame (the duration of employment) and performance requirements (well
specified or weakly specified), were arranged in a 2 x 2 framework. The resulting four
categories of PCs were labeled and described as follows:

Imnsactinnalmntraets, of limited duration with well specified performance terms

Relatinnalmntzacts, open-ended membership with incomplete or ambiguous

performance requirements attached to continued membership.

Balancedmmraet, open-ended and relationship oriented employment with well-

specified performance terms such to change over time.

Itansitinnalfinnnam, essentially a break-down in contracts, reflecting the absence

of commitments regarding future employment as well as little or no explicit

performance demand or contingent or contingent incentives.

Rousseau and Wade-Benzoni state that there may be other kinds of contracts, but
they state ”we find that most modern organizations contain one or more of the variants
of these contracts" (p. 468); no references are offered in support of their claim. The
article goes on to provide a scheme for matching up types of PCs with specific
organizational business strategy. For example, they argue that ”prospectors" (Miles &
Snow, 1984) have a need for flexibility, and as a result, will develop transactional
employee PCs. Finally, Rousseau and Wade-Benzoni seek to link the type of employee

PC found in an organization with the type of relations the organization has with its

40

customers. For example, they argue that transactional employment relationships tend
to breed transactional customer relations.
Other Recent Developments

Ofori-Dankwa (1991). Virtually all writing and research regarding PCs has
focused on the PC between individual employees and their respective organizations.
Ofori-Dankwa' s unpublished dissertation, however, investigated the PC between
employees and their supervisors. He attempted to assess the relationship of both
individual level and dyadic level ”measures of employee PCs" and three dependent
variables: a.) job satisfaction, b.) employee satisfactoriness (task performance,
commitment, compatibility), and c). employee turnover intentions.

Ofori-Dankwa defines the PC as "the degree to which the employee's expectations
of what his immediate supervisor will provide him and what job-related behaviors he
owes his immediate supervisor match what the supervisor’s expectations are as to what
he is to give and receive from his employer” (pp. 9-10). In further describing the PC,
he states that it " focuses specifically on the rights and obligations of the parties to a
dyadic relationship” (p. 11; emphasis added). Thus, at least initially, Ofori-Dankwa
describes the PC at the dyadic or relational level. Accordingly, his first measure of the
PC assesses the agreement between the employee and supervisor regarding their mutual
rights and obligations. This measure was referred to as the ”across congruence
measure" (p. 31). He also claims to measure the PC at the individual level, describing
the purported individual level measure as assessing the match between the employee's

perception of his or her rights (and the supervisor's obligations) and the employee's

41

perception of what is actually being provided by the supervisor. This measure was
referred to as the "within congruence measure" (p. 31).

Clearly, the purported measures of the PC are assessing different constructs. Given
that Ofori-Dankwa described the PC as existing at the dyadic level (pp. 9-10), it would
be more consistent, and less confusing, to describe the within congruence measure as a
measure of the extent to which the employee perceives that the supervisor has fulfilled
his or her obligations to the employee.

At a general level, Ofori-Dankwa hypothesized that both the across congruence
measure and the within congruence measure would be positively related to the three
dependent variables (identified above). From this general hypothesis, more specific
hypotheses were derived that focused on perceived rights and obligations relating to
supervisors' leadership behaviors. Items from the Leadership Behavior Description
Questionnaire (LBDQ) were used (Fleishman, 1953A; Fleishman, 1953B) to assess
expectations regarding two dimensions of leadership behaviors: initiating structure (18
items), and consideration (28 items). The study was conducted using employees from
19 locations of a retail chain. Questionnaires were administered to 76 employees and
their immediate supervisors. It should be noted that although his hypothesis speak of
rights and obligations, it appears that the questionnaire assessed expectations regarding
leadership behaviors (p. 79).

Separate analyses were performed for two dimensions of the LBDQ, initiating

structure and consideration. The correlation between the across congruence measure

(matching employee-supervisor expectations) on the initiatingmm dimension and

42

employee satisfactoriness was significant (.27). There was a significant correlation
(.24) between the so-called within congruence measure on the initiatingjtruoture
dimension and job satisfaction. There were also significant correlations between the
within congruence measure on the consideration dimension and both employee
satisfactoriness (.24) and tenure intentions (.23). In summary, the results were mixed,
providing support for approximately half the hypotheses that were investigated.
Guzzo, Noonan, & Elron (1994). Guzzo et al. state that they endorse Rousseau's
( 1989) conceptualization of the PC (Guzzo et al., 1994, p.618) in their investigation of
the role of EPCS in mediating the impact of company practices on retention relevant
outcomes, organizational commitment and intentions to leave employer. As previously
indicated, however, they do not address Rousseau's focus on promise. The subjects in
the study were 148 expatriate managers working in 43 different companies. Subjects
provided data on the”status" of their PC in two ways. First, subjects responded to a
questionnaire that asked them to rate both: a) the extent to which each of 43 company
practices was provided by the respective employers (5 =to no extent! not provided,
1=to a great extent); and, b) the extent to which what was being provided was
different from what they though should be provided
(-3=much less, 0=about the same, 3=more than should be). The 43 items (one of
each practice) were grouped by content into three scales: financial inducements, general
support, family oriented support. The ratings regarding the sufficiency of employer
practices were viewed as one set of indicators of the extent to which subjects thought

their PC was being fulfilled (p. 620).

43

Perceptions of organizational support were used as a second set of indicators of the
status of employee PCs. The use of perceived organizational support as a measure of
the extent to which the organization is perceived as fulfilling the employee PC only
makes sense if one assumes that as part of their PCs, the subjects expected or believed
that their employer was obligated (depending on the belief element in the particular PC
conceptualization adopted) to provide the kind of organizational support assessed by the
instrument. Guzzo et a1. appear to make this assumption, stating: "we believe that
expatriates experience broad relational PCs because virtually all parts of their daily
lives are open to employer influence" (Guzzo et al., p. 618). No attempt was made to
measure whether or not subjects actually believed that their employers were obligated
to provide the kind of support assessed by the perceived organizational support
instrument.

No significant correlations were obtained between the measure of how much
organizations provided their expatriates and the retention relevant outcomes. In other
words, organizations that provided employees more did not necessarily realize stronger
loyalty among their overseas managers. Indicators of the fulfillment of the PC was
positively related to organizational commitment and negatively related to intention to
leave. Perceived organizational support was also positively related to organizational
commitment and negatively related to intention to leave. Guzzo et a1. conclude that
their study provides evidence that the status of the employee PCs (the extent to which it
is perceived as being fulfilled) mediates the relationship between employer practices

and retention relevant outcomes.

44

Areas of Agreement Regarding the Nature of Psychological Contracts
At a general level, it is possible to identify a number of areas of agreement among

the conceptualizations of the PC construct that have appeared in the literature. It

should be emphasized that to the extent there is any consensus, it exists at a very
general level. As soon as one probes for details, a divergence of thinking manifests
itself. With this disclaimer, it can be said that all current conceptualizations of
employee PCs:

1. include at least some role for employee subjective beliefs about what the employee
is to give to and receive from his/her employer (in some conceptualizations, ‘
employees subjective beliefs constitute the contract; in other conceptualizations, the
beliefs are matched with employer beliefs to constitute a contract)

2. purport to distinguish PCs from "legal" contracts on the basis that the former are
less formal, involve unwritten or implicit elements, and are not enforceable in court
(as we shall see, however, conceptualizations vary in the extent to which they seek
to incorporate, or are influenced by, traditional contract law)

3. view PCs as dynamic in nature

4. adopt an exchange perspective

5 . view employee PCs as a central, compelling force affecting behavior in organiza-
dons

6. propose that ”violations" of employee PCs can potentially produce strong, negative

affective and behavioral consequences

45

Around and among these areas of agreement, there are many differences, both basic
and esoteric, in how researchers conceptualize the PC construct. In the next chapter, I

will identify and discuss some of the important areas of difference.

46
CHAPTER II

Critical Issues in Conceptualizing the Psychological Contract Construct
There is no shortage of issues that have been raised but remain unanswered by the
work currently appearing in the psychological contract literature. Four issues, in
particular, warrant careful consideration and should be explicitly addressed in future
conceptualizing about the psychological contract construct. These issues may be
generally stated as follows:

1. At what level is (or should) the psychological contract construct be
conceptualized?

2. What is the nature of the beliefs that either constitute, or are an element of,
the psychological contract?

3. What is the nature of the discrete ("transactional")—relational continuum
(Macneil, 1980), and is it a useful framework for understanding psychological
contracts?

4. What is the role of individual differences in psychological contracting?

Some issues, such as the level of conceptualization, are critical in the sense that it
should be addressed in the articulation of any construct. Other issues, such as the
nature and usefulness of the discrete-relational continuum of exchange behavior, have
assumed special importance as a result of the treatment that they have received in the
literature. In other words, in the abstract, one might imagine a reasonable conceptual-
ization of the psychological contract construct that did not address the issue. But, in

light of the attention that the issue has received in the existing literature, the issue

47

cannot be reasonably ignored. The foreseeable consequence of ignoring the issue is an
increasingly fragmented literature. The specific nature of these four issues, and the
basis for my claim that they should be carefully considered and explicitly addressed in
future conceptualizing, will be more specifically developed in the discussion that
follows. Other "sub-issues" related to the four critical issues will also be identified and
discussed.

The Level of the PC Construct

Miller recommends that in all organizational research: ”Every discussion should
begin with an identification of the level of reference, and the discourse should not be
changed to another level without a specific statement that this is occurring" (Miller,
1978, p.25). The explicit identification of level fosters clear conceptual thinking and
helps identify measurement issues that need to be taken into account (cf. Klein,
Dansereau, & Hall, 1994).

Rousseau (1989) is one of the few people writing about PCs that has explicitly
addressed the levels issue. She distinguished between conceptualizations of the PC
construct at the level of the individual versus the level of the relationship (cg. dyadic,
inter-organizational). Conceptualizations of PCs as involving individuals' subjective
beliefs were described as being at the individual level. The parties to the contract, the
employee and the employer, need not agree. In contrasts, conceptualizations that view
PC as existing to the extent that agreement exists between the parties are at the

relational level.

48

In terms of levels issues, existing conceptualizations may be more specifically
categorized by first determining whether or not the conceptualization involves issues at
the relational level (e. g. does it focus on the agreement of the parties?). If the answer
is ”yes”, one needs to then determine the focal level for each party to the PC (e.g.
individual, group, organization). If the construct only involves the subjective beliefs of
one party, it does not involve relational level issues; it may be categorized as ”non-
relational", indicating that relational issues between or among the parties are not
involved. For non-relational conceptualizations, one still needs to ask the question
"what is the focal level of the subjective beliefs, the individual, group, or
organization?”

Unfortunately, since the level of conceptualization typically has not been explicitly
addressed, it is sometimes difficult to ascertain the intended level. Based upon my
review of their respective works, I have characterized the apparent levels at which
selected researchers and writers have conceptualized employee PCs and related
constructs, and I have summarized my characterizations in Figure 2. The Figure is
organized in a 2 x 3 format (non-relational/relational level x individual/ group
lorganization level). Using this framework, Rousseau's conceptualization of PCs as
involving individual subjective beliefs is characterized as being non-relational, and at
the individual level (upper, left-hand cell). In contrast, Ring & Van de Ven's (1994)
description of PCs between organizations involved in cooperative relationships as

involving congruent beliefs regarding each other's prerogatives and obligations is

49

Figure 2: Conceptualizations of Psychological Contract Characterized by Level of
Construct

 

Employee
Beliefs

Indiyitlual Rousseau (1989, 1990) 1. Individual - Individual (e.g. supervisor)
(subjective)
Robinson & Rousseau (1994) Schein (1980) between employee and
Guzzo, Noonan, & Elron (1994) “managers &. others in the organization”
McFarlane Shore & Tetriclt (1994) Ofori-Dankwa (1990) between employee
& supervisor

2. Individual - Group
3. Individual - Organization

Levinson et al. (1962)
Schein (1965, 1970, 1980)*
Gibson (1966)

Lawless (1972)

Kotter (1973)

Dunahee & Wangler (1974)
Portwood &. Miller (1976)

 

Gm Nicholson & Johns (1985) 1. Group - Individual
(shared among

employees) Argyis (1960; hourly employees
and their supervisor)

2. Group - Organization

 

itemization Ring & Van de Ven (1994) between
organizations in cooperative inter -
organizational relationships

 

50

characterized as being relational, and involving organizational level beliefs (lower,
right-hand cell).

Relational conceptualizations require one to determine the focal level on both sides
of the PC. Therefore, further level differentiations are made within the three general
relational cells“. For example, although most relational conceptualizations focus on the
agreement between individuals and their employers (Levinson, et a1. 1962; Gibson,
1966), Ofori-Dankwa's research focused on the agreement between individuals and
their supervisors.

Kotter (1973) presents the most troublesome characterization. On one hand, he
does not appear to require any kind of agreement or mutual understanding for a PC to
exist, describing PCs as being made up of the "matches and mismatches" among
employee and employer expectations (Kotter, 1973, p. 92). Instead, he looks at the
extentmhioh a PC is made up of match expectations, and argues that those PCs that
are made up primarily of matches are related to greater job satisfaction, productivity,
and reduced turnover. On the other hand, unlike the other beliefs, I have characterized
Kotter's (1973) conceptualization as relational because his main interest appears to be

in assessing the extent of shared expectations - but it is not clear that it is a clean fit.

 

To avoid a cumbersome and confusing number of cells, I adopted the following
convention in categorizing conceptualizations involving relational issues. The determi-
nation of whether a conceptualization was placed in the relational x individual,
relational x group, or relational x organization level was made based upon the focal
level of the employee beliefs (for this purpose, managers acting as agents of the
organization capacity were not treated as employees). Once allocated to one of the
existing six cells, where necessary, sub-headings where added to indicate the focal
level of the "other" party to the PC.

51

Figure 2 and the preceding discussion focuses on the level of the focal construct
(i.e. the PC construct). Figure 2 does not purport to discuss the full range of levels
issues likely to be involved in theories about PCs (cf. Klein et al., 1994, for a general
discussion of levels issues in organizational behavior research).

The Nature Of The Beliefs

All conceptualizations of PCs involve a belief element, some kind of belief about
what an employee or group of employees is to give to and receive from his or her
employer. As the review of the literature revealed, the belief element of the
psychological contract has been conceptualized, with varying degrees of clarity, in a
number of different ways. Figure 3 summarizes the beliefs that have been the focus of
selected researchers' conceptualizations. Included among the belief elements identified
in Figure 3 are expectations, which may be viewed as beliefs about future states.

To the extent that early writings had a focus, they tended to focus on employee
expectations. With very few exceptions (e.g. Portwood & Miller, 1976), no attempt
was made to define what was meant by the use of the term ”expectations”. Also,
discussions of expectations, obligations, and other types of beliefs were often
intertwined. For example, Levinson et a1. (1962) spoke of expectations that took on

the quality of "obligations", and although Schein generally speaks of the parties'

52

Figure 3 The Nature of the Belief Element of Psychological Contracts

 

Researcherflriter
Argyris (1960)
Levinson et al. (1962)

Schein (1965, 1970, 1980)

Jurek (1969)

Gibson (1966)

Lawless (1972)

Kotter (1973)

Dunahee and Wangler (1974)

Portwood & Miller ( 1976)

Holtz (1978)

Nicholson & Johns (1985)

Rousseau (1989, 1990)

Ofori-Dankwa (1991)

Robinson, Kraatz & Rousseau
(1994)

Robinson & Rousseau ( 1994)

Robinson (1994)

McFarlane Shore &
Tetr'ick (1994)

Guzzo, Noonan, & Elron
(1994)

Ring & Van dc Ven (1994)

Wu
Employee and supervisor shared expectations (norms)
Employee and company expectations that have obligatory quality

Employee and management expectations concerning a pattern of rights,
obligations, & privileges

Employee and employer expectations

Rights, duties, and obligations

Unverbalized employee and employer expectations
Employee and company expectations

Employee and company expectations

Employee and organization eXpectations; also describes PC as a recognition of
mutual obligations

Mutual expectations that represent a need and a obligation on each side of the
contract

Mutual assumptions about rights and obligations
Employee beliefs regarding mutual employer-employee promise based obligations

Defines PC as involving expectations, but also states he focuses on the parties'
rights and obligations

Employee beliefs regarding mutual employer-employee promise based
obligations

Employee beliefs regarding mutual employer-employee promise based obligations
A person's beliefs regarding the terms and conditions of an exchange agreement
Employee beliefs regarding reciprocal obligations

Unclear - they endorse Rousseau's (1989) conceptualization, but do not
specifically focus on obligations or promises; discusses expectations but suggest

PC involves more than expectations

Sets of congruent expectations and assumptions held by transacting parties about
each other prerogatives and obligations

 

53

expectations, he also states that the psychological contract addresses ”the whole pattern
of rights, privileges and obligations between the individual and the organiza-tion"
(Schein, 1965, p. 11). While a relationship between expectations and obligations was
frequently suggested, it was never explicitly addressed. More recently, Rousseau and
colleagues explicitly distinguished between expectations and promised based obligations
(cf. Rousseau, 1989; Robinson et al., 1994). The distinction they seek to make is
obscured by their practice of sometimes describing PCs generally as beliefs about the
terms of an exchange agreement (cf. Rousseau & Greller, 1994; Robinson, 1994), a
definition that does not distinguish between expectations, obligations, or promise based
obligations.

Expectations. Recently, it was suggested that researchers should be more
concerned with the definition of expectations (Colella, DeNisi, & Wanous, 1994). This
observation is particularly applicable to the PC literature. Some researchers have
distinguished between two types of expectations, normative and probabilistic (e. g.
Higgins, 1992). Normative expectations are beliefs regarding what ought to occur,
based upon a prevailing normative standard. Probabilistic expectations are beliefs
regarding the probability of something (e. g. an event, outcome) occurring. In this
sense, the use of term expectation sometimes takes into account the full range of
probabilities (e. g. calculations in expectancy theory). Sometimes the term expectations
is used in a more limited probabilistic sense, such that expectations are defined as those

things that are thought to be likelyJoooour (e. g. Portwood & Miller, 1976).

54

Normative and probabilistic expectations are related but distinct. Normative
expectations are likely to influence probabilistic expectations, but this relationship is
likely to be moderated by other factors (such as perceived ability and willingness of the
other party to do what they "ought" to do). Probabilistic expectations may lead to
normative expectations over time, for example, with performance or reliance by the
expectation holder. The two types of expectations may relate to other variables of
interest in different ways. Finally, reactions to the perception that another has not met
one's expectations are likely to be different for normative and probabilistic expectations
(cf. Higgins, 1992). Generally, unmet normative expectations that we hold for others
may be more likely to result in a sense of being wronged, leading to emotional feelings
of anger, and resentment. Probabilistic expectations may be easier to cognitively adjust
to, for example, by re-evaluating the reasonableness of the expectation we held for
another.

I have discussed three distinct uses of the term expectations, its normative use and
two probabilistic uses. Colella et a1. (1994) illustrate a fourth way in which the term
expectations has been used, one that may be described as focusing on the level of some
anticipated event or condition. They define job expectations as "the amount of some
attribute that a person believes will be present in his or her job“ (Colella et a1. , 1994,
p. 4). To measure job expectations, they used the Job Characteristics Inventory. The
measure asked subjects to indicated Much of certain characteristics they believed

their job would contain. No effort was made to assess probabilistic or normative

55

expectations. No doubt, the term expectation has been used in ways not described
above.

Obligations. As Figure 3 indicates, a number of researchers have either focused
exclusively on rights and obligations, or they have included obligations among the
belief elements of PCs. Except for Rousseau, those who conceptualized PCs as
involving beliefs about obligations have not attempted to define or explain their use of
the term ”obligation”. This is unfortunate because the term obligation is also used in a
variety of ways. For example, Rousseau's use of the term obligation focuses on a
particular kind of obligation - those based upon W. In contrast, the
term obligation has been more generally defined as ”a duty, contract, promise, or other
social, moral, or legal requirement that compels one to follow or avoid a certain course
of action” (The American Heritage Dictionary, 1993).

Relationship of obligations to expectations. The relationship between obligations
and expectations may be viewed in a number of ways. Obligations may be viewed as a
kind of expectation. This appears to be the view of Rousseau (1989). Although it has
not been articulated, this approach implies a kind of hierarchy of expectations that are
ordered based upon the strength of the consequences associated with expectations that
are not met. At the top are obligations. If they are not met, the expectation holder is
likely to experience strong feelings: anger, betrayal, mistrust (Rousseau, 1989). At the
bottom of the hierarchy are mere "hopes", which may be thought of expectations that
are void of a normative element, and relatively weak in a probabilistic sense. If a hope

goes unmet, the feelings are likely to be weaker (e. g. disappointment instead of

56

betrayal). One might view probabilistic expectations of a moderate magnitude as
falling somewhere near the middle of the hierarchy.

An alternative view of the relationship between expectations is that obligations are a
basis or source for expectations, both normative and probabilistic. In this view, beliefs
about obligations are not higher order expectations; they are a specific category of
beliefs that lead to expectations. Other, somewhat more complex, views of the
relationship are possible, as reflected in the view that will be presented in my
conceptualization of the PC construct in Chapter III.

Summary. The nature of the belief element of PCs is another area of considerable
ambiguity. Future conceptualizations should specifically identify what the belief
element is: obligations, expectations, some combination of the two, or some other
belief. The identified belief element should be explicitly defined, and if the
conceptualization involves expectations or obligations, the hypothesized relationship
between expectations and obligations should be articulated. The possibility that the
theoretical distinction that has been made between obligations and "mere expectations”
by Rousseau and colleagues, and myself, is really "much ado about nothing" should
also be entertained. The assessment of meaningfulness of the distinction that has been
made should be informed by the collection of relevant empirical evidence.

The Discrete ("Transactional") - Relational Continuum

Macneil's notion of a continuum of exchange or contracting behavior has been

widely adopted by researchers writing about PCs since the continuum was imported to

the PC literature by Rousseau (1989, 1990) (cf. Nelson & Quick, 1991; Robinson et

57

al., 1994; Robinson & Rousseau, 1994; McFarlane Shore & Tetrick, 1994; Guzzo et
al. , 1994). Prescriptive schemes have been developed based upon the continuum which
suggest that, depending on the employer's business strategy, the employer should seek
to develop PCs among its employees that fall at a particular location on the continuum
(Rousseau & Wade-Benzoni, 1994).

To date, no work published in the PC literature has questioned either the general
notion of continuum, how it has been translated and applied in the PC literature, or the
extent of its usefulness in helping us understand PCs. On the surface of the PC
literature, the continuum is certainly not being treated as a critical issue - it is not an
issue at all. But, a careful review of the treatment that the continuum has received in
the PC literature reveals a "looseness" that is disconcerting. The continuum is
described in general terms, usually involving a brief (2 or 3 sentences) reference to
Macneil (1985). There is little explanation of why particular obligations should be
considered appropriately classified as representing one or the other ends of the
continuum, and once identified, some obligations have moved from one end of the
continuum to the other (see discussion of Rousseau, 1990, and Robinson, et al., at pp.
31-34 of this manuscript). Also, researchers have provided examples of transactional
and relational contracts which, on their face, do not fit the researcher's own
descriptions of the contract types. For example, Rousseau and colleagues have
repeatedly described transactional contracts as involving higleay or high extrinsic
reward (Rousseau, 1990; Robinson et al. , 1994; Robinson & Wade-Benzoni, 1994).

However, when they recently sought to provide an example of a transactional contract,

58
they selected the job of public accountant which they state involves 'mkforlmpay”

(p. 478)’. Later in the same article, the job of a MacDonald's counter worker, which
clearly involves relatively lolnay, was also described as involving a transactional PC
(pp. 478-479).

If one turns to the parent literature to investigate the origins of the continuum and
how it evolved before its importation to the PC literature, additional grounds for
questioning the PC literature's treatment of the continuum are found. Macneil has been
the author of many detailed works addressing the continuum and relational contracting
theory (cf. Macneil, 1974, 1978, 1980, 1985). The most complete statement of his
theory is provided in a book devoted to the topic (Macneil, 1980). Yet, with few
exceptions, references in the PC literature are to a single article by Macneil (1985).

When the detail, complexity, and nomenclature of Macneil's original works are
sorted through, it becomes evident that the brief references that are made in the PC
literature to Macneil's theory of relational contracting are sometimes mistaken
(discussed below). Also, the way in which Macneil's continuum of exchange behavior
has been incorporated into conceptualizations of the PC is, it will be shown,
fundamentally inconsistent with Macneil's theory of relational contracting. In
summary, it appears that the risk that exists any time a translation is undertaken was in
fact realized when Macneil's continuum was translated in the PC literature: something

got lost in the translation.

 

Macneil does not identify high or low pay as being characteristic of any type of
contract.

59

Given the frequency with which Macneil has been cited in the PC literature, the
widespread adoption of his continuum, and evidence that his work is being
misunderstood (or at least mischaracterized), it would seem that a more developed
discussion of Macneil's work is warranted - and is probably overdue. The following is
a summary that I believe improves significantly upon the existing treatment Macneil's
work has received in the PC literature in terms of the depth and accuracy of the
discussion. The reader should be forewarned, however, that Macneil's work regarding
relational contracting has been criticized as being difficult to read and as being too
complex (cf. Whitford, 1985). Unfortunately, such criticisms are not totally without
merit.

Macneil's relational contract theory: An overview. Ian Macneil is a lawyer who
has been primarily writing in the legal literature. He introduced the concept of a
relational contract in an article titled: "Contracts: Adjustments of Long—term economic
relations under classical, neoclassical, and relational contract law" (Macneil, 1972; cf.
Goetz & Scott, 1981; Macneil, 1985). Macneil distinguishes between eontraeting
helrayior and contraotlaw. His theory of relational contracting focuses on behavior,
and it has implications for contract law scholarship. Macneil's work may be viewed as
an attempt to influence thinking about contract law in a way that would, ultimately,
make the law more reflective of actual contracting behavior.

Traditional definitions of contract focus on promise (Macneil, 1980), and are likely
to sound familiar to the reader: "A contract is a promise or a set of promises for the

breach of which the law gives a remedy, or the performance of which the law in some

60

way recognizes as a duty.” (Restatementoffiontraets, Sect. 1, 1932). Macneil's

theory adopts a view of contract that contrasts sharply with traditional definitions: ”By
contract I mean no more or no less than the relations among parties to the process
projecting exchange into the future.” (Macneil, 1980, p.4). Noticeably absent is any
reference to "promise", the law, or terms with special legal meaning, such as ”offer”,
"acceptance”, or "consideration". According to Macneil, contracts -in-law (i.e. those
enforceable by the legal system) are just a small fraction of the contracts that exist in
the modern world (Macneil, 1980, p. 5).

Macneil has described the essence of his theory as involving a description of a.)
behavior essential to exchange relations and of ”norms" such behavior yields, and b.)
the variants of these behaviors as they occur in a spectrum of exchange behavior from
"extreme discreteness" to "extremely relational" patterns (Macneil, 1985, pp.523-524).
The spectrum of exchange behavior is the continuum that was imported into the PC
literature by Rousseau (1989, 1990), and that has been referred to above.

As reflected in the polar extremes of his continuum, Macneil distinguishes between
two basic types of exchange behavior: diserete exchange and relational exchange.
Discrete exchange involves the separation of the transaction from all else between the
participants. Its ideal, never achieved in life, occurs when there is nothing else between
the parties, never has been and never will be. (Macneil, 1980, p. 60, note 1). Discrete
exchange is epitomized by the one-time, market exchange between a buyer and seller in

which the parties had no previous relations and no expectations of possible future

61

relations. The participants in discrete exchanges have been likened to microeconomics
"maximizer" (Macneil, 1985, p. 524, n. 187.)

According to Macneil, purely discrete exchange is not possible (Macneil, 1974,
1980). This point warrants emphasis, and therefore, I will be returning to it. For
now, it should be understood that when Macneil refers to discrete contract, he means
relatixely discrete (which is a way of saying relatively less relational). An example of
a relatively discrete exchange would be an out-of-town traveler's one-time purchase of
gasoline from a gas station located along an interstate highway.

In contrast to discrete exchange, relational exchange is embedded in patterns of
interaction among the parties to the exchange. The relations of the parties, and the
anticipated continuation of those relations in the future, influences and constrains the
parties exchange behavior. The concept of relational exchange challenges the view of
exchange as being discrete that has dominated thinking in economics and traditional
contract law (Macneil, 1980; Whitford, 1985).

While what constitutes discrete exchange is relatively easy to describe, what
constitutes relational exchange requires a more involved description. My goal in the
following discussion is to provide a more parsimonious description of relational
contracting, focusing on what I believe are its essential, defining characteristics. This
description is based on a review of Macneil's works (Macneil, 1974, 1978, 1980,
1985), the interpretations, criticisms, and applications of his work (e. g. Barnett, 1992;

Dwyer, Schurr, & Oh, 1987; Kaughman & Stern, 1988; Provan & Gassenheimer,

62

1994; Whitford, 1985), and proposed alternative conceptualization of relational
contracting (Goetz and Scott, 1981).

The essence of relational exchange. Since discrete and relational exchange are
said to represent two ends of a single continuum, it may be generally observed that the
less exchange fits the description of discrete exchange, the more it takes on relational
characteristics. More specifically, relational exchange can be usefully described as
having six primary characteristics“.

First, compared to discrete exchange, relational exchange involves expectations that
the exchange between the parties will be of a relatively longer duration. While the
discrete exchange ideal involves a single transaction requiring single ”performances" by
the parties (one party providing a monetizable good or service, the other providing
monetary compensation), relational exchange involves repeated and continuous
performances (cf. Macneil, 1974, p. 383, where employment is used as an example of
exchange requiring repeated and continuous performances). Related to duration of the
exchange is the amount of contact that exists between the parties. In discrete exchange,
contact is minimal, and is usually instantaneous with performance (Macneil, 1978).

Second, the subjects of relational exchange are difficult or impossible to measure or
specify. In discrete transactions, the subjects of exchange are money on one side, and

a commodity with monetary value on the other, and these will be carefully measured

 

One could reasonably divide the relational pie into more or fewer pieces. The
particular division that was made reflects my judgment regarding the appropriate
balance between preserving the distinctiveness of important characteristics and the
goal of a more parsimonious description.

63

and specified. The difficulty of measuring or specifying what is being exchanged may
be due to the nature of what is being exchanged (its intricacies, complexity, or
intangible nature), uncertain future conditions, and/or the interaction of the two.

Third, in contrast to impersonal discrete exchange, relational exchange involves
relations between the parties that are personal or "primary” in nature. Relying on
Selznick (1969), Macneil describes three characteristics of primary relations: a.) the
parties respond to the ”whole person” (i.e. they respond to many aspects of another's
character and background, permitting feelings to enter into the relationship); b.)
communication is deep and extensive, not restricted to formal and public modes of
interaction; and c.) individual participants derive personal noneconomic satisfaction and
engage in social exchange as well as economic exchange (Macneil, 1974, p. 722; 1980,
p.13).

Fourth, future cooperation between or among the parties to relational exchange is
required. In discrete transaction, almost no future cooperation will be required because
each party produces either the money or the bargained for commodity at the time and
place agreed upon (Macneil, 1980).

Fifth, in relational exchange there is a sharing of benefits and burdens. In discrete
transactions, benefits and burdens are sharply divided. Each party is allocated certain
benefits and certain burdens (with associated risks), and the parties are rigidly held to
the allocation that was made. The sharing of benefits and burdens that is characteristic

of relational exchange can be viewed as arising from the interaction of the foregoing

64

characteristics and the norm of reciprocity, which requires that parties to a relationship
maintain some kind of equivalence in their exchange.

The sixth characteristic that distinguishes relational exchange from discrete
exchange is the nature of the parties' obligations. In extreme discrete contracting,
obligations are specific, foundedonpromise, and monetizable (Macneil, 1978, p. 904;
Macneil, 1980). In contrast, in extreme relational contracting, obligations are
nonspecific, leaning toward the diffuse and ill-defined, and the relation itself develops
obligations, ”with only modest input from individual promise” (Macneil, 1980, p.17).
A discussion of Macneil's view of the role of promise in contracting behavior will give
greater meaning to the preceding statement regarding obligations in relational
exchange.

A fundamental tenet of relational contracting theory, one that is emphasized in all of
Macneil's writings, is that in contracting hehayior, obligations between the parties are
not limited to those based on promise; promise can never cover more than a fragment
of the total situation (Macneil, 1980). Accordingly, Macneil has argued that promise-
centered theories of contracting are inadequate to deal with complex contractual
relations without distortion and omission (Macneil, 1985). Obligations arising out of
the parties' relationship(s) include obligations based on customs, status, habits,
expectations created by the status quo, "and other internalizations" (Macneil, 1980).
Also, for certain relationships in society, some relational obligations may come from

external sources (e. g. the government regulation of employment relationship).

65

It is Macneil's attention to the non-promissory sources of obligations that arise
between or among parties in a relationship that distinguishes Macneil's writing from
traditional contract scholars. Throughout his writings, Macneil contrasts his notion of
relational contracts with the contracts recognized under traditional, promise centered
theories of contract7. Further, he vehemently, and repeatedly, argues that a focus on
promises impedes our ability to understand contracting behavior. For example, after
describing how the focus on contract law instead of contract oehazior inhibits our
ability to understand "contract", Macneil goes on to state: "But an even more serious
hindrance to understanding is the limitation of contract to ‘promise or a set of
promises” (Macneil, 1980, p.5). Also, at the conclusion of his 1985 article that is so
frequently cited in the PC literature, Macneil states that our knowledge of relational
contracting ”is hindered in part by immense intellectual barriers we put in the way of
its acquisition, oartroularlyjhosematodmouradmcttonjopromme.” (Macneil, 1985,
p. 525, emphasis added).

The anchors are ”ideals" that do not necessarily exist. Macneil has also
emphasized that the discrete and relational prototypes are ideals. Macneil has
expressed doubts that ideal discrete or relational types of exchange exists (Macneil,
1980). According to Macneil, if the discrete exchange ideal exists, it is likely to be

found only in highly specialized societies. On the other hand, the ideal relational

 

American contract scholarship has increasingly recognized relational elements of
contracting behavior. Relational concepts such as unconscionability, duress, and good
faith recognize non-promissory obligations among contracting parties (Macneil,

1985).

66

contract is likely to be found only among primitive societies (1980). Thus, according
to Macneil, if the two extreme ends of the exchange continuum exist at all, they won't
be found in the same society.

In contrast to the ideals that may or may not exist in the real world, hutwhiehhaye
menthefoeusofattenfionjnthemmmmre, Macneil describes ”modern contracts“
(i.e. those existing in societies that are not primitive), as involving both relational and
discrete aspects. Like the relational ideal, modern contractual relations generally
involve primary personal relations, tend towards long life, and tacit assumptions
abound. But there are important differences between the relational ideal and modern
contractual relations. First, modern relational contracts "are ridden with measurement
and specificity" (Macneil, 1980, p. 22). But the complexity of modern society calls for
processes and structures that tend to organize even the most specific and measured
exchange into relational patterns, and all modern relational contracts also involve a
great deal of exchange that cannot be or is not measured (e. g. social exchange;
Macneil, 1980, pp. 22-23).

Second, modern relations tend to include both sharp divisions of benefits and
burdens and a sharing of them. For example, employees may receive a precise wage
for performing specific job tasks (dividing benefits and burden), but they may also
share prosperity with management through bonus schemes or more comfortable
working conditions. They may also share hard times through lay-offs, reduced

bonuses, pay cuts etc.

67

Third, modern contractual relations tend to combine both discrete and relational
obligations. The source of obligations are both promises and the relation itself (i.e.
based on customs, habits expectations, status, and "other internalizations").
Obligations are both specific and general. Similarly, sanctions are likely to be both
very specific and of an ongoing, accommodating nature.

Clearing up misleading or mistaken references. There are a number of
references to Macneil's work in the PC literature that should not be perpetuated. I will
identify and briefly discuss four. When Rousseau imported Macneil's continuum into
the PC literature (Rousseau, 1989), she described the anchors as transactional and
relational, referencing her statements to Macneil (1985). Macneil's 1985 article never
uses the term ”transactional contract. " Rather, Macneil focuses on the spectrum of
exchange behavior described above, referring to discrete and relational exchange
(Macneil, 1985, p. 524), and corresponding discrete and relational contracts. If one
were to convert this continuum of exchange behavior to a continuum of contracts or
contracting behavior, which was done by Roderick Macneil, Ian Macneil's son, in
Macneil (1986), the appropriate ends of the contractual continuum would be the
discrete contract, on one end, and the relational contract on the other end.

Where, then, did the transactional contract term that has been used in the PC
literature originate? In an early work, Macneil (1974) contrasts contract transactions
with contract relations, and in the discussion, the term ”transactional contract” is used
on a couple of occasions. In subsequent works, however, as the spectrum of exchange

behavior was developed and refined, the term was abandoned in favor of a different

68

focus and a different terminology. This explains where the term transactional contract
was likely to have originated; how the term was attached to Macneil's (1985) article is
less clear. The result is that the PC literature's use of term transactional contract
appears to be an anomaly. Based upon my review, references to Macneil's theory
outside the PC literature describe the continuum either in terms discrete and relational
exchange or discrete and relational contracts (e. g. Barnett, 1992; Dwyer, Schurr, &
Oh, 1987; Kaufman & Stern, 1988; Provan & Gassenheimer, 1994; Whitford, 1985).
Given the acknowledged, direct linkage of the presently labelled transaction-relational
continuum to Macneil's theory, and to promote consistency and communication across
disciplines, the term discrete should replace transatnional in future references to the
continuum in the PC literature. Accordingly, I will use the terms discrete and
relational to refer to polar extremes of the continuum.

Nelson and Quick (1991, p. 56) make the following claim: ”MacNeil (1985) posits
two types of psychological contracts in organizations: transactional and relational.”
This claim is obviously not based on a review of Macneil's work. Macneil (1985) does
not posit two types of PCs; he makes no reference whatsoever to psychological
contracts. Based on my review, Macneil never explicitly discusses PCs. Nelson and
Quick' s mistaken claim appears to be based on an inappropriate interpolation of
Rousseau's discussion of Macneil's work.

Guzzo et al. (1994) described Macneil as viewing contracts as either transactional or
relational. They went on to offer their view that PCs would be better viewed as

containing elements of both types of contracting. From the discussion provided earlier,

69

it should be clear that Macneil does not view contracts as falling on one end of the
continuum or the other. On the contrary, Macneil has argued, W, that
modern contracts contain both relational and discrete elements.

A final example of a reference that should not be perpetuated comes from Rousseau
and Wade-Benzoni (1994, p. 466), where Macneil (1985) was cited to support the
following claim: ”Two major types of contracts have been common in the work place
and anchor two ends of what has been described as a contractual continuum-
transactional and relational contracts (Macneil, 1985; Rousseau, 1989)”. Macneil
never indicates a belief that there are two types of employment contracts. To the extent
that Macneil discusses employment, employment is described as involving relatively
relational contracting behavior. For example, in the article cited by Rousseau, Macneil
describes employment as ”an extremely relational contract, no matter how strenuously a
party attempts to make its discrete" (Macneil, 1985, p. 492). Supporting the claim that
two ”major” types of employment contracts are ”common” in the work place with a
reference to Macneil (1985), or to Macneil's work more generally, is misleading.

Summary. The assumption that there are two types of PCs must be abandoned in
favor of more rigorous conceptual inquiry regarding the nature of the discrete-relational
continuum and how it might help us know and understand PCs. The outcome of this
inquiry may depend, in part, on how one addresses other important issues regarding
level of PC construct, and the nature of the belief elements. If the continuum is to be
applied to PCs, the conceptualization of the PC construct should at least be consistent

with relational contracting theory. Also, consideration should be given to the fact that

70

the continuum was meant to describe the full range of exchange behavior that occurs in
society. To what extent do employment relations, and employee PCs, reflect that full
range? Can there be such a thing as purely discrete PC, or are employee PCs
inherently relational? Ultimately, one's "conceptual findings" need to be empirically
assessed. Claims that Rousseau (1990) and Robinson et al. ( 1994) provide empirical
evidence that PCs, or particular groups of obligations, can be categorized as discrete or
relational are unjustified, and should not be made.

What is the Role of Individual Differences in Psychological Contracting?

The issue of the impact of employee "predispositions” on PCs has been around since
the earliest writing about PCs (Argyris, 1960). The role of individual differences in
psychological contracting, however, has seldom been formally discussed, and even less
frequently investigated. Nonetheless, in all writings regarding PC, the treatment that
individual differences does (or does not) receive permits inferences regarding the
author's underlying assumptions. I will discuss the treatment that several, influential
researchers have given individual differences. But first, I will make several general
comments and observations that will help frame the discussion of individual differences
in psychological contracting.

General comments. Ackerrnan & Humphreys (1990) identify three broad varieties
of individual differences: a) interindividual differences (differences between
individuals), and two types of intra-individual differences: b.) differences within an
individual over time, c.) intra-individual differences between two or more attributes of

a single individual (e.g. the difference between a person's verbal and spatial abilities).

71

To the extent that individual differences has received attention in PC literature it has
involved two of these varieties of individual differences: interindividual differences and
differences within individuals over time.

Individual differences may play a role at a number of steps or stages of the
psychological contracting process. They may affect the formation of beliefs, the
strength of the relationship between perceived obligations and behavior, the likelihood
of perceiving violations of one's PC, and the nature (e. g. intensity) of one's response
to perceived violations.

One might expect that the level at which the PC construct is conceptualized would
greatly influence the theorized role of individual differences. For examme, Argyris's
(1960) view of PCs as involving a group's shared beliefs with their supervisor, almost
by definition, limits the role of individual differences. The requirement that beliefs
must be shared, first among the employee group and also between the group and the
supervisor, restricts the potential impact of individual differences.

The role of individual differences may also be influenced by the nature of the belief
elements(s) in the conceptualization of the PC construct. For example, a focus on
beliefs about promises made and exchanged is, I would argue, more likely to
emphasize the role of representations and other interactions external to the individual
(e.g. policies, practice) in shaping beliefs than would be the case when there is a more
general focus on expectations or obligations.

The role of individual differences has obvious implications for theory building; it

should be considered an issue that must be addressed in any specification of a

72

nomological network for the PC construct. It also has implications for research design
(e. g. variables to be included as primary independent variables or to control variance,
samples to use) and specific analysis performed (e.g. to assess additive or moderator
effects; cf. Klein et al., 1994). The extent to which employee PCs are influenced by
inter-individual and intra-individual differences has implications for practical advice
regarding the management of employee PCs, something that employers are being urged
to do (e. g. Rousseau & Greller, 1994). For example, if employee PCs, or certain
obligations commonly found within PCs, are strongly influenced by individual
differences that are relatively stable over time, then efforts to manage employee PCs
through the manipulation of organizational communications with the employees may be
misguided. It may be more appropriate to attempt to manage employee PCs by
selection practices, or by adapting organizational policies and practices in ways that
take into account the PCs of the work force.

Selected researchers' treatment of individual differences. Levinson et al. (1962)
viewed employee expectations, which are said to frequently antedate the relationship of
the person and company, as a product of one's needs and motives, tempered by ones
past experience and knowledge of current situation. The work experience, accordingly
to Levinson et al. , can be seen as a process of fulfillment of a contractual relationship
in which both parties seek to continuously meet their respective needs. Levinson et al.
give examples of employees who come to the company with different needs and
motives, that lead to different expectations, which lead to different PCs. Levinson et al.

also incorporate intra-individual difference in the conceptualization of the PC.

73

Employee PCs are viewed as changing over time as a result of the individual's
experience in the work place and their evolving understanding of their individual nwds,
and a number of other factors, such as changes in the employer's requirements.

Schein (1980) argues that both employee and the manager forge expectations from
their inner needs, what they have learned from others (knowledge), traditions, norms
that may be operating, their own past experience, ”and a host of other sources."
(Schein, 1980). Interindividual differences in needs, knowledge, and experience are
important determinant of PCs. Schein's description of changes in individual PCs that
occur over the course of a career, as one's nwds change, also addresses the role of
intraindividual differences. According to Schein (1980), in early career, peoples'
needs and expectations revolve around self-tests. Therefore, they expect the
organization to provide them with challenges. In mid-career, when employees are
typically most productive, they expect most in terms of recognition and rewards. At
later career stages, need for security increases and expectations shift accordingly.

Both Levinson et al. and Schein recognized the role of other factors, such as group
norms and organization needs, in influencing PC. But, both have argued that
employees' interindividual (e. g. nwds, motives, job experiences) and intraindividual
differences have a strong influence on employee beliefs about what they should give to
and receive from their employer, and, to lesser extent, the relational PC that is enacted.

Focusing only on the level of the construct, one might predict that Rousseau's
(1989) conceptualization of PC as involving the individual's subjective beliefs,

"existing in the eye of the beholder", would lead to a greater role for individual

74
differences than Levinson et al. ( 1962) and Schein's (1980) conceptualizations at the
relational level, where at least some degree of objectivity (in the form of agreement) is
required. But, when one compares the treatment these three researchers give individual
differences, precisely the opposite appears to be the case (an explanation for this
occurrence is offered below).

In her seminal article, Rousseau (1989) identifies a number of factors that can
promote or influence employee beliefs about ”paid for promises" (i.e. promise based
obligations): the norm of reciprocity, organization's requests and promises,
organizational policies, normative beliefs associated with the organization, interaction
with other employees, and characteristics of the individual's position within the
organization (e. g. seniority and tenure). Of all of the antecedents of employee beliefs
that are identified by Rousseau (1989), only the characteristics of individual's position
are individual level factors. In sharp contrast to the discussions regarding the
antecedents of employee beliefs provided by Levinson et al. (1962) and Schein (1980),
Rousseau ( 1989) does not mention the potential impact of individual needs, motives,
dispositions, experience or other individual differences on employee beliefs.

Rousseau (1990) does discuss the possibility of a role for individual differences in
shaping employee beliefs regarding obligations, stating:

Employee perceptions regarding contributions and entitlement mightnsomement

reflect individual predispositions, consistent with the interactionist model of

newcomer socialization (Reicher, 1987). . .The role of individual differences in
beliefs regarding obligations merits further research. However, in a recent study of

third party beliefs, no variation in beliefs found to be due to individual differences
(Rousseau and Anton, 1988). (Rousseau, 1990, p. 399; emphasis added)

75

A similar, limited discussion occurs in Rousseau and McLean Parks (1993), where
psychological contracting is described as recognizing individual differences in
conditions, values, and information processing“. However, when discussing the
antecedents of employee beliefs about promised obligations, Rousseau and colleagues
clearly focus on personnel policies, practices, communications and other factors that
may be manifestations of a promise. To the extent that individual differences are taken
into account, they take a distant back seat to factors that involve manifestations of
promise. This is particularly evident in the recent article, Rousseau and Greller
(1994), that makes the strong statement that: ”employee interactions with the
organization are the source of information about the contract“ (emphasis added, p.
386). Here, again, the importance of employee-organization interactions via personnel
polices and practice in particular as antecedents of employee PCs is emphasized, and
the potential role of individual differences is alluded to only indirectly.

The fact that although Rousseau conceptualizes the PC at the individual level -
existing in the eye of the beholder, she places relatively little emphasis on individual
differences as antecedents of employee PCs is paradoxical, but explainable. I would
argue that it is the belief element of her conceptualization that has, either consciously
or unconsciously, greatly influenced her treatment of individual differences. Rousseau's

focus on promise based obligations inevitably focuses attention on policies, practices

 

The possibility is suggested that some predispositions, such as equity sensitivity,
Huseman et al. (1987), could influence the nature of the contracts that individuals
enter into.

76
and other representations, and away from the potential role of individual differences
among employees.

Summary. Consideration of the role individual differences in the psychological
contracting has been in very general terms, and, it has focused on the influence of
individual difference at one stage of the psychological contracting process: the
formation of the beliefs that constitute the PC. Even in this limited scope, very
different views of the magnitude of the role of individual differences are reflected in
the literature. Early writing about employee PCs suggested that individual needs,
desires, and other individual differences have a strong influence on the content of
employee PCs (cf. Levinson et al., 1962; Schein, 1965). More recently, attention has
been focused on organizations' representations and interactions with employees as ”the”
source of the beliefs that constitute the PC. Which of these general views one
subscribes to has important theoretical and practical implications. Future

conceptualizations of the employee PC would benefit from a more explicit and detailed

consideration of the role of individual differences than has typically been the case.

77
CHAPTER III

An Integrative Conceptualization of the Psychological Contract
(or "Towards a Theory of Psychological Contracting in Organizations")

In this chapter I present my conceptualization of the psychological contract
construct. It is an integrative conceptualization in that it incorporates and builds upon:
a.) early thinking about the antecedents of PCs (e.g. Levinson et al., 1962; Schein,
1965, 1979, 1980); b.) Rousseau's view of the PC as involving individual level,
subjective beliefs; c.) a more accurate understanding of relational contracting theory
(Macneil, 1980, 1985); and, what I believe are (1.) original insights regarding the
nature of psychological contracting in organizations.

Defining the Construct

Because the PC construct has been generalized beyond the organizational literature
in which it originated, I will provide both a general definition of the construct and a
more specific definition that focuses on the construct's application in organizational
contexts.

General definition of the ”psychological contract" construct: A psychological

contract is an indiyiduats beliefs regarding the mutual obligations between that

focal person and another party. The obligations are mutual only in the sense that the
focal person perceives obligations flowing in both directions: from the focal person
to the other party, and from the other party to the focal person. heroism!

restrictionthauhrreheagteemenrbetmrtheparties regarding what constitutes

their respective obligations.

78

One may think of individuals within an organization as having a variety of
psychological contracts with different parties, including, but not limited to, the
organization, an immediate supervisor, a co-worker, or a work-group. I have chosen
to focus on individual's beliefs regarding the mutual obligations between the individual
and his or her employer, and I reserve the use of the term "employee psychological
contract" to refer to that set of beliefs

Definition of the employee psychological contract construct (EPC): An EPC is
an indiyidnalemployefis beliefs regarding the mutual obligations between that
mdixidualemploxee and his or her employer. Again, the perceived obligations are
mutual only in the sense that the individual employee perceives obligations flowing in
both directions: from the employee to the employer, and from the employer to the
employee.

The decision to focus on the individual employee's PC with his or her organization
was influenced, in part, by the fact that most researchers in the PC literature have
focused on the employee-organization relationship. The result is that the person-
organization PC captured my attention and interest, framing my thinking about the
construct. The decision also reflects my agreement with Levinson et al. 's (1962)
position that the many other PCs that may exist within an organization may be usefully
viewed as involving collateral relationships that have a bearing on the employee-
organization relationship. While I have chosen to focus on the EPC, I recognize that
there may be circumstances or particular issues where PCs involving other parties, such

as an employee and his or her supervisor or co-worker, would be the more appropriate

79

focus of attention (just as I would expect that those who study organizational
commitment would concede that there may be circumstances that employee
commitment to their supervisor or co-workers, and not the organization, would be a
more appropriate focus of attention).
My conceptualization of the EPC does not involve an assumption that employees'
interactions and relationships with their supervisor or co-workers are unrelated to an
individual employee's PC with his or her organization, nor does it involve an
assumption that employee beliefs about their obligations to their employer, supervisor
or co-workers never overlap. I am, however, making the assumption that when you
ask people about their psychological relationship with their "employer", they are
capable of distinguishing their employer from their supervisor, co-workers, and others
who they may come into contact with in the work place. So, for example, if I asked
my committee members ”what obligations do you believe you owe to mummyloxer",
that stimulus would lead you to identify obligations that you believe you owe to MSU,
and not obligations to the department chair, fellow faculty members, etc. (even though
obligations may overlap). This assumption is relevant to the reliability of the EPC
construct; though I have a definite opinion, no assumptions regarding the discriminant
validity of EPC construct is intended.
Level of the PC Construct

As my definition indicates, I view the EPC as existing at the individual level. This

conceptualization reflects a growing consensus among researchers investigating PCs

(e.g. Rousseau, 1990; Robinson & Rousseau, 1994; Guzzo et al. 1994; , McFarlane

80
Shore & Tetrick, 1994). Individual perceptions have been demonstrated to play an

important role as a relatively immediate precursor of individual behavior, mediating
impact of "objective" circumstances on individual behavior. Even researchers who
have described the PC at the relational level have pointed out the importance of
employee perceptions (e. g. Dunahee & Wangler, 1974). An individual level
conceptualization also has the advantage of enhancing the discriminant validity of the
construct. Formal employment agreements, and legal contracts in general, are typically
thought of as existing at the relational level. Also, work regarding organizational roles
has provided a rich literature that focuses on relational level (dyadic) constructs. It
should be noted that the conceptualization of the PC at the individual level represents
an evolution of the original use of the term. The evolution began as least as far back as
Stagner (1963), and has been advanced most by the work of Rousseau (1989).
Relevant Relational Level Variables

Contract congruence. The extent to which an individual's beliefs correspond with
his/her employer's perspective (measurement issues aside, for the moment) is a
potentially important variable. Kotter (1973) and Ofori-Dankwa (1990) provide
evidence that the degree to which employees' beliefs correspond with employers'
understanding may be related to important organizational outcomes. The extent of
matching can be captured by a construct at the relational level. I suggest that the term
"psychological contract congruence” , or simply ”contract congruence” , be employed to
refer to the extent to which the beliefs of the focal party corresponds with the other

party's belief regarding their mutual expectations or obligations. The general notion of

81

contract congruence is referred to by both Ring & Van de Ven (1994) and Ofori-
Dankwa (1990). In both instances, PCs were conceived of as involving congruent
beliefs; contract congruence was not viewed as a separate construct.

Where both parties are individuals, assessing contract congruence will not pose any
special measurement problems. A measure of agreement can be used. However,
where the other party is the organization (or some collectivity within the organization,
such as a work group), measurement issues arise regarding what are appropriate
indicators of the other party's perception of the agreement (e. g. information from
informants that reflect agreement among the collectivity, evidence of employer
practices). Researchers need to articulate a composition theory supporting their use of
particular information as an appropriate indicator of the organization' s understanding of
the mutual obligations, and empirical evidence needs to be assessed to determine the
extent of support for the composition theory (cf. Rousseau, 1985).

Correspondence versus compatibility. Measures of matching or agreement assess
correspondence, but they do not necessarily assess the extent to which the parties to a
relationship have compatible beliefs regarding their mutual obligations. There may be
circumstances where the lack of correspondence does not indicate or result in a lack of
compatibility. For example, one party may hold itself to higher obligations than the
other party, there would be a lack of congruence, but the congruence may not be likely
to lead to a lack of compatibility. Also, there may be circumstances, such as
conditions as munificence, in which a lack of congruence does manifest itself in a lack

of compatibility. Thus, an employer that has experienced continuous growth may have

82

never believed itself to be obligated to provide a high level of job security, but this
belief was not manifestly incompatible with employers belief in obligation to provide
security so long as the employer was growing and not being required to lay employees
off. These possibilities suggest that there may be situations where, rather than simply
looking at agreement in a global sense, in might be important to assess ”agreement" in
a more circumscribed way.

For the present, I define contract compatibility as referring to the extent to which
the focal person's belief regarding the mutual obligations and the other's beliefs
regarding their mutual obligations can exist in harmony. Contract compatibility is
contextually bound. The absence of perceived violations by the focal person and the
other party (in the case of an organizations, its agents) is an indication that the parties'
beliefs are compatible. Assessment of compatibility will require more sophisticated
measurement than that involved in assessing contract congruence. At the present, I am
not prepared to defend any particular method of assessing compatibility, but raise the
issue to point out the potential limitation of measures of agreement.

Shared psychological contracts. There may also be reasons to be interested in
assessing the extent to which groups or classes of employees possess similar EPCS. For
example, assessing the degree to which employees have the same beliefs may help
identify the antecedents of psychological contracts. Evidence that a heterogeneous
group of employees have very similar EPCS may suggest that organizational factors
have a dominant influence on individual employee EPCs. Conversely, evidence that a

heterogenous group of employees performing the same job in the same organization

83

have very dissimilar contracts would suggest that the PC is being influenced by factors
outside of the organizational content (e.g. individual differences that are independent of
the organizational setting). It might also be hypothesized that the violations of
obligations that are shared amonhg employees will result in more severe negative
consequences because the employee will be less likely to cognitively adjust to the
violation (cf. Olson et al., 1994). Finally, sharredness may be important in predicting
or evaluating the impact of change efforts in organizations. Terms of EPCs that are
widelly shared among employees may create a status quo against which changes will be
compared, leading to greater resistance to change (Rousseau & Greller, 1994).

Nature of Beliefs i

The belief element in my conceptualization of the PC are beliefs about obligations.
I adopt the following general definition of ”obligation": an obligation is a social,
moral, or legal requirement that compels one to follow or avoid a certain course of
action (cf. The American Heritage Dictionary, 1993). Thus, an individual's belief that
an entity is obligated in a certain way involves the subjective perception that the entity
is required by society, morality, or the law to follow or avoid a certain course of
action.

The focus on obligations is, at a general level, consistent with the attention that
Rousseau and colleagues give obligations (Rousseau, 1989; Robinson et al., 1994).
However, I have a very different understanding of the nature of the obligations that are
involved, and in particular, the relationship between promise and obligations. Before

discussing my differences with Rousseau and colleagues' treatment of the relationship

84

between promise and obligation, I will first discuss how I view the relationship between
expectations and obligations and why I have specified obligations as the belief elements
rather than expectations.

The relationship between expectations and obligations. As indicated in my
earlier discussion, the relationship between obligations and expectations can be viewed
as being very simple (i.e obligations are merely higher order expectations, cf.
Rousseau, 1989), or it can be seen as being more complex. In adopting obligations as
the belief elements in the PC, I focus on the role of obligations as a source of
expectations, and I argue that obligation based expectations take on a special
significance. At the same time, I recognize that expectations can play a role in shaping
obligations. In particular, acts carried out in reliance upon a probabilistic expectation
tend to convert the probabilistic expectation into an obligation. The role of
expectations in shaping obligations is of secondary interest to me, one that comes into
play when considering the antecedents of EPCS.

It has been argued that all beliefs imply expectations, in that it is possible to derive
expectations from any belief (cf. Olson, Roese, & Zanna, 1994). In my view, attached
to all beliefs about obligations are expectations, both normative and probabilistic.
Normative expectations and obligations are closely related, and arguably difficult to
distinguish. I would argue that normative expectations include some expectations that,
although they reflect beliefs regarding what “ought to" or "should” be done, do not rise
to level of a duty or responsibility (i.e. an obligation). Therefore, as a category,

normative expectations are more inclusive than obligations. While I can distinguish the

85
two conceptually, I am not convinced that there will be much of a practical difference
between the two. The distinction is not central to my conceptualization of the EPC. I
presume that where obligations are identified, normative expectations of similar
magnitude are attached. Probabilistic expectations, on the other hand, are clearly
distinguishable from beliefs about obligations.

A focus on beliefs about obligations. Any comprehensive theory of psychological
contracting in organizations will need to address both employee expectations and beliefs
about obligations. For a number of reasons, I believe that a focus on beliefs about
obligations in_defining_theEEC_eonstmet will increase the construct's potential for
uniquely contributing to our understanding of employment relationships. .

The attention that the PC literature has historically given the obligatory quality of
employee beliefs distinguishes it from other literatures that deal, more generally, with
expectations (e. g. socialization). This historical focus is reflected in the writings of the
first researchers to provide an in depth discussion of the notion of PC. Levinson et al.
(1962), distinguished their treatment of expectations using the PC framework from the
general literature on expectations by pointing to the obligatory quality of the
expectations they investigated:

There is an extensive literature on expectations. What we emphasize here,

however, is the interplay of expectations and their compelling quality. It is as if

both employees and the company are saying, ‘ You must, for I require it'. When
the employee goes to work, he brings expectations with him. The company tacitly
accepts the expectations when it accepts him. In turn, he tacitly accepts the

company's expectation. This mutuality, with its inherent obligatory quality and
system of rewards, constitutes a psychological contract." (p. 36).

86

A focus on obligations distinguishes the PC construct from the general consideration
of employee expectations by drawing attention to certain expectations that are likely to
have special importance, expectations that are attached to obligations. Expectations
attached to obligations are likely to have special importance for two main reasons.
First, the perceived failure for the other party to meet one's expectations is more likely
to result in strong negative reactions directed at the non-performing party when the
expectation reflects (is ”attached to") a perceived obligation. Though stated somewhat
differently, this is the main argument offered by Rousseau (1989) in support of her
focus on obligations instead of "mere expectations".

Second, expectations regarding one's own performance that are attached to
perceived obligations are more likely to influence employees' day to day behavior (i.e.
when violations are not an issue). Very little has been written about how PCs influence
behavior on a day to day basis. Drawing an analogy to the goal setting literature, one
may look at expectations as being goal-like. Both goals and expectations involve
projections about future states. Like goals, expectations may come from without
(presented by others) or from within the person (self set goals or self generated
expectations). We know that goals are most effective in influencing behavior when
there is commitment to the goals (Locke & Latham, 1992). Similarly, I would argue,
expectations that have attached obligations are more likely to influence behavior. The
expectation exists as a kind of standard, and the greater the individual's sense of duty
or responsibility to meet the standard (i.e. the person's commitment to the standard),

the more likely the person will strive and persist in effort to meet the standard.

87

Finally, I believe defining PC as beliefs about obligations rather than expectations is
likely to sharpen our thinking about the nature of expectations, reducing the risk of
confusing normative and probabilistic expectations. By identifying obligations and
keeping them distinct from probabilistic expectations, we will be in a better position to
evaluate factors such as trust, that may relate differently to normative and probabilistic
expectations.

In summary, a focus on obligations has historical roots in the PC literature, it
appears to have greater potential for predicting behavior than a focus on expectations,
and it enhances the EPCs discriminant validity. Together, these considerations lead me
to conclude that a focus on obligations increases the potential for EPCs to uniquely
contribute to our understanding of employment relationships. A focus on obligations in
drflningtheEPQoonstmot does not eliminate the need to consider the role of
expectations in theorizing about psychological contracting in organizations, but it does
help structure thinking about the role of expectations in psychological contracting.

Obligations or promised based obligations? A major difference between
Rousseau and colleagues' conceptualization of PCs and my conceptualization of the PC
is that Rousseau's conceptualization appears to focus on a specific kind of obligations:
those that are hasedonperceixedpromises. No such focus or restriction is intended in
my conceptualization of PC. The belief element in my conceptualization are
obligations which may have their origins in either social, moral, or legal requirements.
Promise may play a role in the origin of some obligations, but, it is believed, promise

does not play a role in all obligations. Moreover, not all promises result in obligations.

88

If this all is not readily apparent, it should become some when what constitutes a
promise is more specifically addressed.

”Promise" has been defined as a "declaration assuring that one will or will not do
something" (American Heritage Dictionary, 1993), and, similarly, as ”a commitment to
do (or not do) something" (Rousseau and McLean Parks, 1993 p. 6). Promises are
based on some statement or pledge attributable to the promisor. There are, however, a
host of social and legal obligations imposed on us without our declaration or
commitment to conduct ourselves in the obligated manner. Whether a specific
declaration or commitment to do something (i.e. a promise) gives rise to an obligation
depends on the circumstances of the declaration or commitment and prevailing social,
legal, and moral standards. For example, a promise made by a mentally incompetent
or a person under unconscionable duress may not be deemed to create an obligation -
social, legal or moral - to fulfill the promise.

No compelling reason has been offered for restricting obligations to those that are
based on perceived promises. On the contrary, in restricting obligations to those that
are based on perceived promises, Rousseau appears to have been influenced by a
particular, and some would argue - outdated, approach to contract law. Specifically,
Rousseau's ( 1989) conceptualization of the PC incorporates elements that have played a
prominent role in determining contractual relations undenelassicaLandmettelassieal
oontraeLlaw: ”offer", "acceptance”, "consideration", and a focus on promise
(Rousseau, 1989, pp. 125-126). The only references proximally related to her assertion

that PCs are composed of promise based obligations are to contract scholars that can be

89

described as reflecting a neoclassical view (Famsworth, 1982). Rousseau, having
adopted a traditional promise centered view of contracts, appears to be saying,
"contracts are based on promises, therefore, in order to have a psychological contrast,
there must be a belief in promise".

I have been careful to say that it "appears” that this is Rousseau view, because it its
not precisely clear what was intended. At times, it seems as though Rousseau and
colleagues are using the term promise in an extremely broad sense, so that when both
expressed and implied promises (whatever that means) are considered, promise
essentially covers all obligations. While Rousseau and colleagues distinguish between
obligations and mere expectations, they never distinguish between obligations and
promises. On the contrary, in Rousseau and McLean Parks (1993), almost identical
definitions are provided for the terms promise and obligation. They define obligation as
"a commitment to some future action” (p. 3), and, as indicated earlier, they define
promise as ”a commitment to do (or not do) something” (p. 6).

Clearly, Rousseau and McLean Parks (1993) have a much narrower view of the
what constitutes an obligation than the view which I have adopted. Under Rousseau
and McLean Park's view, an employee will believe a party is obligated to act in a
certain way when the employee believes that the party committed to act in that certain
way. Under my view, an employee will believe a party is obligated to act in a certain
way when the employee believes that there is either a social, moral, or legal

requirement that compels the party to act in the certain way.

90

There are a number of problems associated with the use of promise in such a broad
way that promise based obligations are treated as being synonymous with obligations.
The fundamental problem is that promises are not the sole basis for obligations. In
law, including modern contract law, and in social relations, obligations are created and
exist without promises (Macneil, 1980; Blau, 1964). Equating obligations with
promise based obligations is contrary to what appears to be a common usage of the
term obligation to refer to duties or responsibilities that may arise from a variety of
sources: moral, social, legal, or based on promises (cf. American Heritage Dictionary,
1993). Such a use of the term is also inconsistent with my experience as a lawyer,
interviewing dozens of discharged employees seeking legal advice regarding their
possible rights. These individuals typically showed the symptoms said to accompany
violations of EPC (e.g. anger, outrage, betrayal). All felt that their employer had
treated them wrongfully, doing something or failing to do something the employer was
obligated to do (fired them without good reasons, did not give them adequate warning).
But when asked whether their employer had ever directly or indirectly promised to treat
them in the relevant specified manner (e.g. ”did your employer ever promise only to
fire you for good reasons?"), more often than not, they would respond ”no, but I
thought that....".

One might ask, if we all agree that when we speak of promises, we really mean all
obligations, what harm would result (i.e. what harm is there in forcing all obligations
into a promise based framework)? There are a number of potential negative

consequences associated with employing such a fiction. In discussing how the use of

91
different terms to describe relations among organizations are likely to affect one's view
of the relations, Simon (1991) observed:

The choice of name matters a great deal. The name can affect the way in which we

describe its institutions, and the order of description can affect the theory. In

particular, it may strongly affect our choice of variables that are important enough

to be included in a first-order theory of the phenomena (p. 28).

A promise framework shifts the focus of inquiry from the full range of antecedents
traditionally discussed (needs, motives, personnel practices, etc.) to a narrower focus
on the specific representations of parties and their interpretation. It also implies a very
rational psychological contracting process, particularly when a focus on promise is
combined with legalistic terms such as offer, acceptance, and exchange of
consideration, to define the PC construct.

Using the term promise to include all perceived employer and employee obligations
could obscure important distinctions between obligations based on perceived
representations originating with the employer versus those obligations that arise from
other considerations, such as one's moral standards or understanding of employment
law. There is also the danger of potentially misleading employers, and researchers,
regarding the malleability of EPCS. The framing of EPCS as involving the exchange of
promises suggests that employers can readily manipulate employee EPCs through the
promises the employer makes. The magnitude of this risk is not presently known; we
do not know enough yet about the extent to which employee EPCs are driven by

relatively stable internal factors versus external factors, such as employer

representations.

92

Finally, there are some perceived obligations that simply cannot be reasonably
”crammed into a promise framework” (Macneil, 1980, p. 54). For example, Roehling
& Winters' (1994) findings indicated that even in the face of an explicit statement from
an employer indicating that it reserved the right to discharge employees with or without
cause, many people still believe that the employer is obligated to only discharge the
employee for good reasons. It is simply unreasonable to talk of such perceived
obligations as being based on employer promises. For other examples of obligations
that cannot be fit into a promise framework, see Macneil (1980).

In summary, there is no basis for asserting that all beliefs that individuals have
regarding the mutual obligations between the individual and his or her employer are
exclusively based on perceived promises. Any conceptualization that is based on this
assertion would be employing a fiction that has a number of potential negative
consequences, described above. It is also possible that Rousseau and colleagues truly
intend to limit PCs to obligations that are, either expressly or implied (again, whatever
that means), based on perceived promises. That is, of the total set of employee beliefs
regarding mutual obligations between the employee and his or her employer, only those
that are actually based on perceived promises constitute the PC. As stated earlier, such
a restricted view of PC represents a radical departure from the origins of the PC
construct, and it has a number of implications for researchers adopting the view.
Researchers adopting this restricted view need to do a number of things that have not

been adequately done to date.

93

First, they need to define what they mean by ”promise" and ”implied promise”.
Promise has many meanings (Macneil, 1980). The meaning of implied promise is
perhaps even less certain. In the field of law, the term implied promise has been
equated with the terms ”fictitious promise" and "promises implied in law" (Black's
Law Dictionary). In this use of the term, no promise was actually or "in fact" made,
but under the circumstances, it is deemed just or fair that one party should have a right
and another party an obligation. In other words, instead of finding a promise that leads
to the imposition of an obligation, the court finds an obligation that leads the invoking
of a fictional promise to reach a fair outcome. This use of the term implied promise
results in the equating of promise with obligation, and therefore, presents all the
potential problems that were described immediately above.

The term implied promise might also be used to describe a promise that was
intended and understood, but never explicitly declared or ”openly uttered”. In this use
of the term implied promise, the promise that is implied is not a fiction. It can be
thought of as a promise recognized based on circumstantial evidence (instead of an
explicit declaration, which would be direct evidence of a promise). Both of these very
' different uses of the same term can be found within the PC literature.

Second, researchers need to specify how promise based obligations are different
from other obligations that are excluded from their conceptualization of the PC. What
criterion was used? References to classical or neoclassical descriptions of what
constitutes a legally enforceable contract should not dictate the content of a construct

that purports to explain psychological processes and human behavior. In my view, a

94

claim that promise based obligations involve more serious consequences when violated
cannot be the sole reason. There are likely to be perceived rights, not based on
promises, that lead to the same strong reactions if violated (e. g. the "right” to be
treated with some dignity; the "right" not to be arbitrarily discharged). Also, there are
promises that can be violated without negative consequences. We expect that some
promises will be broken, and under certain circumstances may view the breaking of the
promise as a necessary or good thing (Macneil, 1980). Also, there are likely to be
some promises that are simply not all that important to the promises, so that their
violation would not result in hypothesized feelings of anger and betrayal.

Third, researchers must be more precise in the definitions of PC they provide.
Instead of describing PCs as involving beliefs regarding mutual obligations between the
parties, or beliefs about the terms of the exchange agreement, it should be clearly stated
that PCs involve mutual promisoohasodobliganons if the intent is to limit the PC to
promised based obligations.

Fourth, efforts to empirically assess PC need to carefully avoid confounding beliefs
in obligations generally, with those based on promises. For example, the
questionnaires that Rousseau and colleagues have used asked subject to indicate the
extent of obligations, not the extent of promises or promise based obligations. If they
conceptualize PC in the narrower sense, limiting PCs to only promised based

obligations, the findings of their studies may have inappropriately confounded

95

perceived obligations based on other factors with those based on perceived promises”.

Fifth, researchers who conceptualize PCs as being composed of only promise based
obligations need to make sure the other aspects of their theorizing about PCs are
consistent with the notion that the PC is made up of only promise based obligations
(i.e. that their theorizing is internally consistent). In particular, attention must be given
to the appropriateness of applying Macneil's discrete-relational continuum to a
conceptualization of PC that is limited to promised based obligations. A promise based
constraint is fundamentally inconsistent with theory of relational contracting that has
been presented by Macneil (1980, 1985). In the event that the support for this
conclusion was not obvious from the earlier discussion of Macneil's work regarding
relational contracting, I will reiterate and supplement relevant portions of Macneil's
work.

Macneil introduced the relational contract concept to address limitations resulting
from traditional contract scholarship's over reliance on promise as the basis of
contracts. He sought to draw attention to actual contracting behavior, which, he
argued, did not limit obligations to those based on promises. Macneil has repeatedly
indicated his strongly held view that a promise framework, or an analysis based on

promise, inhibits our ability to understand contracting behavior involving relations. He

 

The instrument Rousseau and colleagues used has a "priming” paragraph that states
that employers make a variety of promises to employees. But the question used to
elicit responses asks subjects to indicate the extent to which their employer is obli-
gated to do a number of things, and the labels on the Likert scale are stated in terms
of obligations. No reference to promise is made in the eliciting question or the
response scale.

96

has lamented contract scholars "addiction to promise" (Macneil, 1985, p. 525), and has
stated that promise centered theories of contracting "virtually guarantee that we will not
understand highly relational contract behavior" (Macneil, 1985, p. 508). Bluntly
stated, relational contracting theory involves an outright rejection of a promise
framework for understanding contracting behavior. In a very real sense, Rousseau's
application of Macneil's relational contracting concept to her promise centered
conceptualization of PCs involves the use of a concept (relational contracting) to
describe its antithesis (a promise centered conceptualization of contracting).

While Macneil rejects the promise framework, he still recognizes the obvious role
that promises play in contracting behavior. But Macneil argues that promise is the
essence of diserete contracting behavior, and that the more a relationship relies on
promise, the more discrete it is (Macneil, 1983, p.360). Thus, a conceptualization of
PCs that is framed in promise focuses on the discrete aspects of contracting behavior.

Summary of the nature of the belief elements. The theory of psychological
contracting in organizations that I am developing addresses both beliefs about
obligations and expectations. In defining the EPC construct, however, I focus on
individual employee's beliefs about the mutual obligations between the employee and
his or her employer. I share Macneil's view that a promise framework inhibits our
ability to understand contracting behavior in relations, and I would argue that a
conceptualization of the EPC that does not focus on "promise" will have greater
promise. Accordingly, the beliefs about obligations that are at the heart of my

conceptualization of the EPC are not restricted to those based on perceived promises,

97

but include beliefs about obligations that may come from a variety of sources, including
social norms, legal, moral, or ”other intemalizations".
Antecedents of EPC

In this section I will identify and bn'efly discuss factors that are thought to influence
the beliefs that constitute the EPC. I have organized my treatment of antecedents of
PCs into three general categories: societal factors, organizational factors, and personal
factors. The constructs within given categories are clearly related to constructs in other
categories. But, they are believed to be sufficiently distinct to make their separation
useful. The list of constructs is not exhaustive, and there are other reasonable ways of
categorizing the factors. Nonetheless, I believe that, as a beginning, the scheme
provides a useful framework for organizing a variety of factors that have been
discussed or alluded to in the PC literature, but which have received little systematic .
attention.

Societal factors. The kind of psychological contracts that are conceixable in an
organization are ultimately the reflection of societal factors (cf. Schein, 1980; Gibson,
1960). It is useful to think of societal factors as setting the parameters for the nature
and conduct of psychological contracting in organizations. Societal factors include
such things as society level social norms, legislation (the two are distinguishable
because legislation, particularly in the employment area, sometimes shapes social
norms as much as it reflects them), and societal level environmental influences (e.g

competitive business environment, changing demographics).

98

The influence of societal level factors on EPCs is reflected in writings about the
”new psychological contract” that is said to be evolving among American workers. The
perceived change has been attributed to a number of factors that are thought to be
changing at the societal level, including: an increased emphasis on quality of life and
personal growth among American workers (Sims, 1994, p. 372); sky-rocketing
employee benefit costs (Lucero & Allen, 1994, p. 439); increased global competition,
sales, acquisitions, mergers, reductions (Wilhelm, 1994, p. 23); and increasing
mobility, collapse of church and family, and greater search for meaning at work
(Spindler, 1994).

Organizational factors. Organizational factors include factors that influence EPCS
and are directly connected to the organizational setting. Organizational factors operate
at a number of different levels within the organization. Organizational lend factors
would include organizational culture, organizational climate, and company wide
personnel policies and practices. Each of these factors may also be operating at the
group level. Role negotiation between an employee and his or her supervision is an
example of a organizational factor that operates at the level of a dyad. The formal job
requirements are viewed as an organizational factor that is connected an employee's
position within the organization.

Personal factors. Personal factors are indirddnaLloxrl factors that are more
directly related to the individual employee than to the particular employment setting.
At any given time, an employee has a set of personal factors that may influence or

shape the impact of societal and organizational factors on his or her beliefs. One's set

99

of personal factors, however, are likely to have been influenced by societal and
organizational factors (some individual factors more than others). Moreover, over
time, societal and organizational factors will continue to exert an influence on personal
factors (again, some more than others).

The personal factors that have been most frequently identified as antecedents of
employee beliefs about obligations are employee needs, values, and motives (Levinson
et al., 1962; Schein, 1980), and prior work experience (Levinson et al., 1962; Schein,
1980). Other personal factors that have been identified in the PC literature include
equity sensitivity (discussed below), work centrality (Schein 1965 , 1980), and ,
careerism (Rousseau, 1989).

A heuristic representation. Figure 4 provides a heuristic representation of the
antecedents of EPCS, and attempts to illustrate the dynamic nature of EPCs. The figure
depicts both the individual and the organization existing within the parameters of a
given society. Both the individual and the organization are subject to societal
influences, and as society changes, these changes effect both individual and
organizational factors, and come to be reflected in the EPC. Before the individual
becomes a member of the organization, at Time 1, the individual's beliefs about the
mutual obligations between himself or herself and a specific employer are primary
shaped by personal factors (e.g. prior work experiences, needs, wants,
traits, etc.) and communications from the organization. Other organizational factors,
such as the organizational culture, are likely to have an indirect effect, if any, on the

new employee’s beliefs prior to the employee's entry into the organization.

100

. >2§owemzem 2.. menu

wEOW HO ..mzqmaZQ:

 

 

Em ow o ...fifi >1 oz mack otwmmw EB 95mg rim 0%?
.: d as .5
\ J A _ J
mOO—HH>F EVOHO—wm _ —
meson: 2035 ..
roam—mac: — N J
O”O>2—N>H~OZ>F fitPOHO—fim
owOEN>doz — 3:39.. 85 $3303 —
OO§O>507< GEE—.0
www-mOOEEdoz How WanERBoam —

 

 

 

 

 

 

 

wmwmozb. 553—6

101

After the individual enters the organization, Time 2, organizational factors exert a
more direct, stronger influence on employee beliefs, both in an absolute sense and
relative to the influence of personal factors. The circle representing the individual's
EPC is drawn so that it is not fully within the organizational setting to reflect the belief
that the EPC can never be fully explained by organizational factors. Personal and
societal factors continue to influence EPCS in ways not mediated by the organizational
setting. In Figure 4, the EPC circle is drawn so that one-half of it falls within the
organizational context. This is a neutral, symmetrical, and somewhat arbitrary starting
point. To the extent that there is a strong organizational setting (strong culture and
climate, clearly specified and communicated personnel policies, etc.), it would be
useful to think of a greater percentage of the EPC as falling within the organization.

Finally, Figure 4 indicates that as an individual's weer within an organization
unfolds over time, Time 3 and Time 4, the individual's EPC may evolve as result of
changes in individual or organizational factors, both of which may be influenced by
societal changes.

The Usefulness of the Discrete-Relational Continuum

In Chapter II, I noted the widespread importation of Macneil's notion of a discrete-
relational continuum of contracting behavior to the PC literature, and argued that the
question of the usefulness of the continuum for understanding PCs should be addressed.

In this section, I seek to address that question.

102

Employment relationships are inherently relational. It is important to recognize
that the discrete—relational continuum was introduced by Macneil in a much broader
context than the context in which it is being applied in the PC literature. Macneil was
describing the full range of exchanges that occur within society, from one-time, arms
length transactions conducted under market conditions to marital relations. When the
full range of exchange is considered, virtually all, if not all, employment relations fall
closer to the relational end of Macneil's continuum than to the discrete end. Stated
another way, WWW. Support for
this claim can be found in the many specific references Macneil makes to employment.
Macneil frequently uses the employment relationship as an example of relational
contracting (cf. Macneil, 1980). He has identified employment relations as being
among those structures that are "obviously relational in nature” (Macneil, 1978, p.

85 8). Macneil has observed that employment is an ”extremely relational contract, no
matter how strenuously a party attempts to make it discrete” (Macneil, 1985, p. 492),
and in describing the nature of the most discretely organized economy possible, he
states that ”of course, no employment is possible." (Macneil, 1985, p. 489).

By focusing on important elements of relational contracting, identified originally at
pages 57-59, the relational nature of employment relations can be more particularly
demonstrated. Employment relations necessarily involve some uncertainty that
precludes a specification of all of the terms and conditions of exchange. It has been
observed that no job can be completely specified in advance, eliminating all employee

discretion (Fox, 1974; cf. Levinson et al., 1962). Unlike the single performances

103

involved in the discrete exchange ideal, employment, even part-time employment,
involves multiple performances. Commons (1924) has argued that the employment
contract "is not a contract, it is a continuing implied renewal of contracts at every
minute and hour, based on the continuance of what is deemed, on the employer's side,
to be satisfactory service, and, on the laborers' side, what is deemed to be satisfactory
conditions and compensation (Commons, 1924, p. 285). What Commons (1924)
describes has also been referred to as a contingent sequential contract. Goetz & Scott
(1981) discuss contingent sequential contracts as examples of relational contracts.

The relational nature of employment in the United States, and in many other
societies, is further promoted by the set of obligations imposed on the employment
relationship by the government. Employment legislation creates obligations that arise as
a result of the employment relationship (Macneil, 1980). Even in the absence of
promises or agreements between the parties, based upon the existence of a specific
relationship between the parties (i.e. that of employer and employee), employers are
obligated to employees in a variety of ways relating to non-discriminatory treatment,
safety and health, minimum wage, etc.

It has been suggested that employment agencies such as Manpower and Kelly create
"purely transactional agreements with workers' (Rousseau and Wade-Benzoni, 1994, p.
466). Even though this kind of temporary employment contains relatively more discrete
aspects than more traditional employment relations, it still does not come close to
reaching the pure discrete ideal. In contrast to the discrete ideal, even contract

employees‘ work involves multiple performances that cannot be fully specified, in

104

every detail, in advance. Moreover, expectations relating to future exchange are likely
to creep in (e. g. the possibility of the contract being renewed or the employee being
brought on as a ”regular” employee), and many of the relational obligations imposed by
society apply to contract employees

(e. g. health and safety laws).

In a sense, because all employment involves relational elements, the usefulness of
Macneil's discrete-relational continuum suffers from a restriction of range limitation
when applied to the employment relationship. The result is that while the theory of
relational contracting and the continuum it describes can and does contribute to our
understanding of EPCs, its usefulness for specific purposes may be more circumscribed
than the PC literature currently suggests.

The usefulness of the continuum at a general level. Macneil's theory of
relational contracting and its discrete-relational continuum points out the relational
nature of most modern contracting behavior, and employment relationships in
particular, providing converging support for the concept of employee PCs. Writing in a
literature that is distinct from the literature in which the PC concept evolved, and
drawing on a different disciplinary focus (contract scholarship), Macneil fleshed out
details of how parties in exchange relationships develop expectations and related
obligations that cannot be captured by simply examining the formal arrangements or
promises exchanged between the parties.

Macneil's theory of relational contracting is also useful in that its description of the

conditions under which contractual relations are likely to succeed has stimulated

105

thinking about the role of different employee held beliefs, and it provides a set of
reasoned propositions to be tested. Some of these conditions sound remarkably similar
to the kind of organizational conditions that were identified by Levinson et al. (1962)
as being necessary to successful psychological contracting in organizations (e. g. the
parties having a belief in their shared fate and interdependence).

The usefulness of placing EPCS on a continuum. Relational exchange is inherent
in employment relations, and WW
relations. The question remains, is it useful to describe EPCS as being more or less
relational? The answer to this question ultimately depends on the extent to which
theory based relational characteristics can be reliability employed to identify patterns or
profiles of EPCS, and, relationships can be established between those patterns and
variables of interests such as employee trust, commitment, risk taking and other
performance related behaviors. So far, the literature has failed to do these things.
Thus, the classification of specific EPCs as being either relational or discrete that has
occurred in the literature is generally not very useful.

Before the continuum can be meaningfully used to classify EPCS, additional
foundational work needs to be done. Perhaps the most urgent need is a more careful
explication of the concept of relational employee and employer obligations. Current
definitions tend to be perfunctory in nature, referring to obligations regarding a variety
of ”socio-emotional concerns". Although discussions of relational concepts in the EPC
literature typically provide a reference to Macneil, the concepts are only loosely linked

to the theory of relational exchange (Macneil, 1980).

106

Relational Employment, Relational Obligations, and Relational EPCS

In the PC literature, the discrete-relational continuum has been used to characterize
specific employee-employer obligations, EPCs (which are made up of many
obligations), and, more generally, employment. Some authors have shifted back and
forth in their use of these terms, appearing to use them interchangeably. I found that
recognizing distinctions among these concepts helped clarify my thinking about
relational obligations. Therefore, in this section, I will discuss these clearly related but
distinct concepts. First, I will identify and discuss the main characteristics of relational
employment. Once the nature of relational employment is described, then, a general
definition of relational employee-employer obligations will be provided, and, six
general categories of relational employee-employer will be identified. Finally, I will
discuss my view of what constitutes a relational EC.

The nature of relational employment. Relational employment can be viewed as a
particular example of relational exchange. Like relational exchange, relational
employment is a dyadic level concept - it describes what goes on between two parties,
not one party's subjective perception of the relationship. In contrast, the EPC, as I
have conceptualized it, is an individual level construct that assesses one party's
subjective perceptions. Also, the EPC does not purport to assess perceptions of what is
actually going on between the parties; it assesses one party's perception of what ought
to go on if the parties performed their respective obligations to each other.

Although in the full spectrum of exchange behavior, all employment involves some

aspects of relational exchange, it may be possible to describe specific employment

107

relations as being relatively more or less relational in nature. What follows is a
description of what I propose are the main characteristics of relational employment. If
you recall my discussion of the main characteristics of relational exchange, at pages 58-
60, you should recognize the strong - if not direct, linkage between the theory of
relational exchange (Macneil, 1980) and the characteristics that I have identified. I
propose that a given employment relationship is more relational in nature to the extent
that it possesses the following characteristics.

First, the exchange between the employee and employer includes non-specific,
intangible, subjects of exchange. Adopting the employees perspective, by subjects of
exchange I mean those things that an employee receives from employment that are
somehow traceable to the employer and are of some value to the employee, and those
things that the employee contributes to the employers which are of some value to the
employer.

Subjects of exchange that employees receive are more specific and more tangible to
the extent that their value to the employee is readily reducible to a monetary value that
can be presently calculated. Examples of non-specific, intangibles subjects of exchange
include the following if they are of value to the employee: meaningful, challenging, or
interesting work; personal recognition; opportunities for social interaction; and, status
or prestige associated with a position. Examples of subjects of exchange that are
relatively specific and tangible include: an agreed upon hourly wage or salary; a
specific benefit package, particularly where the employer provides a statement of the

monetary cost or value of the benefits; paid holidays.

108

Looking at what the employee contributes to the employer, subjects of exchange are
intangible and characteristic of relational employment to the extent that the details of
the contribution cannot be fully specified in advance. Examples of relatively non-
specific or intangible subjects of exchange contribute by employees in relational
employment relationships include: loyalty; the exercise of employee judgment,
discretion, or initiative in the performance of work; and engaging in general, non-job
specific behaviors for the benefit of the employer. The subjects of exchange that the
employer receives is lacking in relational character if all that the employee is
contributing is strict performance of well specified tasks. A second characteristic of
relational employment is that the employee and employer have a sense of shared fate.
Each believes that their own success, however defined, is in some significant way
linked to the other experiencing success in the relationship.

Third, relational employment is future orientated. There is a shared belief that the
employment relationship will continue in to the future, but, W
duration. An employer's explicit promise or guarantee of life time employment would
result in a relatively specific, promised based obligation, making an explicit guarantee
of life time employer more discrete, and less relational in nature, than an assumption of
future employment that did not have a specific time frame.

Fourth, employer and employee exhibit flexibility and cooperation in their dealings
with each other. The need for flexibility and cooperation arises from the fact that there
are things being exchanged that do not allow for complete, advance planning of every

detail regarding their respective responsibilities. Under these circumstances, if the

109
relation is to continue so that their shared fated may be successfully realized, each party

nwds to exhibit flexibility and cooperation.

Fifth, there is a sharing of benefits and burdens between employer and employee.
This means that the employer and employee do not sharply allocate certain risks or
potential rewards to one or the other party, and then rigidly adhere to the allocation -
letting the chips fall where they may. Rather, in relational employment the parties
attempt to maintain some kind of equivalence in their exchange, so that one party does
not suffer while the other experiences a windfall.

A sixth characteristic of relational employment is that the employer and the
employee have a holistic view of each other. The employee is not narrowly viewed as a
service or skill providing unit. The employee is seen as a whole person, with a work
life, a personal or non-work life, feelings, emotions etc. Similarly, the employee has a
broader, more holistic view of the employer, its component parts, obligations to other
employees, and the demands it faces in its environments.

Relational obligations. Relational employee-employer obligations are those
relatively diffuse obligation that are associated with relational employment. I use the
term "associated” advisedly. Depending on the circumstances, beliefs regarding
relational obligations may be brought to the employment relationship, and serve to
promote relational exchange. In other circumstances, relational obligations way arise
from a pattern of relatively relational exchange. When fulfilled by the parties,
relational obligations serve to maintain a relationship between the employee and

employer in circumstances where, because of what the parties are seeking in the

110

exchange, the terms and conditions of the parties' exchange cannot be fully specified in
advance. Again, building on my understanding of relational exchange theory, I
propose that relational employee-employer obligations include the following
obligations:

a.) a general obligation to take the other party's interest into account when taking
action relevant to the exchange, and not cause the other party to suffer a loss
without good justification;

b.) a general obligation to share the benefits and burdens of the parties' joint
endeavors

c.) a general obligation to cooperate and be flexible in order to maintain the
parties' relationship

(1.) a obligation to recognize individuals in the relationships as whole persons,
with feelings, and non-work lives, and treat them accordingly

e.) general obligations to provide or exchange inducements that are non-specific

or intangible

f.) general obligations that reflect a clear assumption of long term employment
relations, without specifying the period of employment

Relational obligations do not involve specific monetary obligations thaLmasze

presentlysaloulated (e. g. payment of specific hourly wage or salary). This does not
mean that any obligation that relates to money is non—relational. For example, an
employer obligation to share profits that result from the parties' joint endeavor, where

there is uncertainty about the amount, involves a sharing of benefits of relationship and

111

is relational in nature. Similarly, an employee obligation to accept a reduction in pay if
necessary to keep the employer financially afloat involves the sharing of a burden, and
would also be relational in nature.

Relational obligations may or may not be in writing. If reducing employee or
employer obligations to writing increases their specificity, it will tend to decrease their
relational nature. However, reducing a general obligation to take the other party's
interest into account when taldng action relevant to the exchange relationship does not
convert this central relational obligation to a discrete obligation.

Relational EPCS. When discussing relational EPCS, the literature has implicitly
treated beliefs in employer relational obligations and beliefs in employee relational
obligations as if they will always naturally occur together. It is possible, however, that
some employees will have stronger beliefs in employer relational obligations than in
employee relational obligations, or vice versa. For example, if an employee perceived
that his or her obligations have been fulfilled, but that the employer has failed to fulfill
its obligations to the employee, beliefs regarding employee relational obligations may
be negatively affected while beliefs regarding employer relational obligation remains
high. Belief in employee relational obligations may below because the employee
performed, discharging his/her obligations, and there has been no employer
performance that would renew employee relational obligations, projecting them into the
future. Belief in employer relational obligations may remain high because the
employee is, in a sense, a creditor and may not be willing to let the employer ”off the

hook”. Also, research regarding equity sensitivity (cf. Huseman, Hatfield, and Miles,

112

1987) would predict that it is likely that some individuals will have stronger beliefs in
employer relational obligations than in employee relational obligations, or vice versa,
as a result of their general equity orientation.

At the core of relational exchange is the maintenance of some kind of equivalence
between the parties. When this equivalence is not maintained, over time, the norm of
reciprocity - thought to be fundamental across human societies (Gouldner, 1960) - will
be violated and the Lelafianalfimhange relationship will fall apart (if exchange between
the parties doesn't cease altogether, it will become relatively discrete in nature). In
order for an EPC is to be described as "relational", it should reflect this core aspect of
relational exchange. Therefore, I propose that an EPC is a ”relational EPC" to the
extent that it reflects relatively strong beliefs in both employee relational obligations

and employer relational obligations.

l 13
CHAPTER IV
Research Questions and Hypotheses to be Investigated

This chapter identifies and discusses the research questions and specific hypotheses
that were empirically investigated. The discussion provided to support the hypotheses
builds upon that which has been discussed in previous chapters. In particular, Chapter
111 provides the foundation for the hypotheses that I will be investigating. Observations
and arguments from Chapter III are incorporated without repeating the whet, full
discussion.
The Humpty Dumpty Approach to Defining Constructs and the Need for
Investigation

The general approach to defining the PC construct that is reflected in the literature
is definition by fiat. Bach investigator defines the PC construct in some way that he or
she feels is suitable. Gulliksen (1968) described this approach to defining constructs as
the ”Humpty Dumpty approach. " His use of the term was suggested by the following
quotation from Through the Looking Glass (Carroll, 1923, pp. 110-111):

"When I use a word,” Humpty Dumpty said in a rather scomful tone, ”it means just

what I choose it to mean - nothing more nor less.”

"The question is," said Alice, "whether you can make words mean so many

different things. "

"The questions is," said Humpty Dumpty, "which is to be master, that is all."

Gulliksen ( 1968) argued that when you have multiple definitions of a construct, it is
reasonable to insist that the construct be made the subject of some investigation to

determine the extent to which scientists’ ideas about the construct are consistent or

equivalent. Examples of the problems associated with lack of attention to construct

114

definition and other construct validity issues early in a literature’s emergence ean be
found in the literatures pertaining to over-reward inequity (Schwab, 1980), the least
preferred co-worker (Rice, 1978), and organizational citizenship/extra-role behaviors
(Van Dyne, Cummings, McLean Parks, 1995). Attention to construct validity is
essential to avoid eventual confusion and to facilitate the accumulation of coherent
cumulative findings in a literature.

Given the number of definitions of the EPC construct and their potential non-
equivalence, in my opinion, it would be unreasonable not to insist that the different
views of the construct (including my own) be made the subject of investigation. The
present study acted upon this opinion by addressing the following general research
question: To what extent are views of the EPC construct that vary in the belief element
that they focus on (i.e. , obligations, expectations, or promises) equivalent? Before
presenting hypotheses related to this research question, the nature of “equivalence” will
be discussed. The discussion is intended to be an introduction to the topic of
equivalence that will facilitate the presentation of the hypotheses that were investigated.
The topic of equivalence will be discussed further in subsequent chapters.
Measurement Equivalence

Generally stated, in order for two measures to be equivalent, they must behave the
same way (cf. Smith, Kendall, and Hulin, 1969). There are, however, degrees of
equivalence and different approaches to assessing equivalence. The most stringent
definition of equivalence requires that the measures have the properties of parallel tests.

Somewhat different, but consistent criteria for strictly parallel test have been offered by

115

a number of researchers. According to Gulliksen (1968), to qualify as parallel tests,
two measures must have: a.) equal means; b.) equal variances; c.) equal covariances;
and d.) for each specified criterion, the two tests must have the same correlations.
Stated in other words, for two measures to be parallel, the measures must load on the
same construct to the same degree, have equal error, and, they must correlate equally
with an external criterion. However, because error variances are independent (though
of equal magnitude), parallel measures will not correlate perfectly with each other
(Hattrup, Schmitt, & Landis, 1992). Measures which meet the stringent requirements
of parallel equivalency are said to be interchangeable; it does not matter which of the
measures is used in any given situation (Gulliksen, 1968).

Smith et al. (1969) have argued that strict parallel equivalence may not always be
necessary so long as the measures behave in the same way for individual and situational
changes about which generalizations are going to be attempted. Their approach to
assessing equivalence of measures focuses on the similarity of patterns of relationships
within and across different situations and samples, and the resulting tests are
correlational in nature.

Increasingly, researchers have investigated the equivalence of measures by using
confirmatory factor analysis (CFA) to focus on the relationship among latent variables
(e. g., Petersen & Smith, 1995; Rahim & Magner, 1995; Van de Vijver & Harsveld,
1994; Hattrup, Schmitt, & Landis, 1992). Researchers using the CFA approach have

been explicit in differentiating and assessing different levels of equivalence. The levels

116

of equivalence that have been addressed and their factor analytic corollaries are
summarized in Figure 5.

Parallel equivalence has already been discussed. Tau-equivalence requires that
measures have equal true score variances but not necessarily equal error variance or
reliability (Lord & Novick, 1968). Tau-equivalent measures assess the same construct
to the same degree but are not equally reliable.

Congeneric equivalence is the least restrictive model of equivalence. It requires only
that the true scores of the equivalent measures be perfectly correlated, though true
scores may differ by an additive or multiplicative constant. The congeneric model
implies only that equivalent measures load on an a shared construct, although the
loadings may differ in magnitude and one of the measures may have greater reliability.
In testing the hypotheses that follow, different levels of equivalence were assessed and
different approaches to assessing equivalence were employed.

Obligations Versus Expectations: A Difference That Makes a Difference?

Like others (e. g. Rousseau, 1989; Robinson et al, 1994), I have sought to
distinguish between conceptualizations of EPCs that focus on beliefs about obligations
and those that focus on expectations. The reasoning behind my conceptualization of
EPCs as involving beliefs about obligations was provided earlier, at pps. 86-99 . It is
possible that although a theoretical distinction between beliefs about obligations versus
expectations can be articulated in the abstract, as a practical matter, the distinction may
not be meaningful. Employees may not view operationalized measures of the two

conceptualizations as being distinct. Measures assessing employee expectations and

117

Figure 5: Models of Equivalence and Factor Analytic Corollaries

 

 

Models of Equivalence

(Measurement Theory) Factor Analytic Corollary

0 . E . ]

Assumes only that true scores are Requires only that the pattern
perfectly correlated; although they of factor loadings be equal;
measure the same construct, they loadings may differ in magnitude

have unequal metrics and reliabilities

Iau_Equiyalence
Requires that measures have the Requires equal factor patterns .
same true score variance, but not and equal factor loadings

necessarily equal error variance
or reliability

Requires measures to measure the Requires equal factor patterns,
same construct, to the same degree, equal factor variances, and
and have equal error equal error variances

Validities in predicting an external
criterion should be equivalent

Strict parallelism requires that the
above restrictions hold in each sub-
sample from a given population.

 

118

measures assessing employee beliefs about obligations may function as interchangeable
measures, assessing a single, more general latent construct to a similar degree and with
similar reliability.

The question of the extent to which the distinction between obligations and
expectations is meaningful is both a theoretical and an empirical one. Existing
empirical evidence provides little insight. No study has specifically attempted to
address the issue. There are only a handful of reported empirical studies investigating
PCs, and therefore, it is not possible to use a meta-analytic approach to assessing which
conceptualization has been more successful in explaining relevant variance. Moreover,
there are a number of potential confounding factors (discussed below) that severely
limit our ability to compare the findings of existing studies.

Consistent with the distinction that I made between expectations and obligations in
Chapter II, I expect that the measure of beliefs about employer} obligations will elicit
significantly different responses than the measure of what is expected from employers.
I also expect that there will be important differences in responses between items
assessing employee beliefs about their obligations to their employer and matching items
assessing what the employee thinks is expected of him or her. Employees may feel
obligated to do certain things even though it is not expected of them. More likely,
there will be things that employees perceive are expected of them at work, but which
they believe they have little or no obligation to provide. Further, I expect that
differences in responses across forms will not be solely attributable to differences in the

reliability of the different measures. At a minimum, there will be differences in the

119

degree to which the different forms measure underlying constructs. These predictions
are captured in the following hypotheses:
HYPOTHESIS 1A: Items assessing employee beliefs about their employer’s
obligations and the matching items assessing what employees expect from their
employer are not tau equivalent measures of the EPC construct.
HYPOTHESIS 1B: Items assessing employee beliefs about their obligations to their
employer and the matching items assessing employee's beliefs about what they
expect to provide their employer are not tau equivalent measures of the EPC
construct.
Obligations versus Promises: Again...A Difference That Makes a Difference?
Building on the work of Macneil (1980), I have argued that a conceptualization of
the EPC construct that is framed in terms of promise is too limiting. In support of this
argument, I made the assertion that not all perceived obligations are based on perceived
promises. This assertion has been made by Macneil (1980), and is implicit in
definitions which indicate that the term ”obligations" refers to duties or responsibilities
that may arise from a variety of sources: moral, social, legal, or based on promises
(e. g. , American Heritage Dictionary, 1993). I do not know of any published empirical
evidence bearing on this assertion. I propose to empirically investigate this assertion in
the EPC context by asking employees to indicate their beliefs regarding the extent to
which their employer is obligated to do or provide certain things, and, whether they
believe that their employer promised - directly or indirectly - to do or provide those

things. The findings will provide an indication of the extent to which the practice of

120

treating all obligations as being based on perceived promises when describing employee
beliefs involves a fiction, as I have argued it does. The following hypotheses regarding
the relationship between obligation based and promise based measures of the EPC
construct parallel those that were presented above regarding the relationship between
obligation based and expectation measures of the EPC construct.

HYPOTHESIS 2A: Items assessing employee beliefs about employer obligations

and the matching items assessing employee beliefs about employer promises are not

tau equivalent measures of the EPC construct.

HYPOTHESIS 2B: Items assessing employee beliefs about their obligations and the

matching items assessing employee's beliefs about their promises to their employer

are not tau equivalent measures of the EPC construct.
Investigating the Pattern of Relations With External Criteria

In addition to the confirmatory factor analytic investigation implied by Hypotheses
1A, 1B, 2A, and 2B, the equivalence of the different EPC measures, and if present, the
nature of differences between the measures can be further examined by studying the
correlations of the different measures with external criteria. Some studies have used
only the latter approach, and not CFA, to assess the equivalence of measures. For
example, Ironsen et al. (1989) focused on the similarity of patterns of correlations in
their assessment of the discriminant validity of the 116 and JDI scales. Only a few
studies have investigated equivalence by looking at pattern of correlations with external
criteria in conjunction with the CFA of the measures’ latent variables

(e.g., Van de Vijver & Harsveld, 1994).

121

In both of the studies cited in the preceding paragraph, the investigation of the
patterns of correlations was conducted without specific hypotheses regarding the
relationships between the focal measures, respectively, and the external various criteria.
Rather, the researchers sought to compare and assess the general, overall patterns of
correlations. Differences in patterns were viewed as evidence of non-equivalence which
could be evaluated, on a post hoc basis, to gain understanding regarding the substantive
nature of the measures’ differences. This approach was adopted in the present study.

To facilitate the assessment of the EPC measures’ equivalence by examining the
pattern of correlations with external criteria, the present study measured several
variables that have been identified in the literature as belonging in the EPC construct’s
nomological network. The primary variables included for this purpose are work
centrality, equity sensitivity, and trust in employer, and organizational tenure. Each of
these variables and their conceptual link to the EPC construct will be briefly described.

Work centrality has been defined as the beliefs that individuals have regarding the
degree of importance that work plays in their lives (Palluay, Alliger, & Stone—Romero,
1994). Work centrality is presumed to be a relatively stable set of beliefs that do not
fluctuate greatly in response to conditions in the work environment. Schein (1965,
1970, 1980) hypothesized that those who view work as a central concern in their life
will tend to expect and demand more from work. Based on my review, Schein’s often

re-iterated hypothesis has never been tested.

122

Equity sensitivity is a construct that refers to a person’s preferences for outcomes
versus input in a general work situation. It is based on the belief that individuals react
in consistent but individually different ways to both perceived equity and inequity
because they have different preferences for equity (Huseman, Hat field, & Miles, 1985,
1987). Huseman et al. presented a continuum of three preferences, identifying three
classes of individuals represented along the continuum. On one end of the continuum
are "benevolents" who are described as givers because they give while expecting little
in return. In the middle of the continuum are ”equity sensitives", those who conform to
the traditional norm of equity. Anchoring the other end of the continuum are ”entitleds"
who are described as having high thresholds for feeling indebted, accepting whatever
they get as their due, and as a result, "they feel little or no obligation to reciprocate.”
(Huseman et a1, 1987, p. 225). Huseman et al. argue that equity sensitivity is a trait,
but they acknowledge that the question of whether it is a state versus a trait is one that
must be empirically substantiated.

It has been suggested that equity sensitivity is an individual difference that might
directly influence the nature of an individual's EPC (Rousseau, 1990). This implicit
hypothesis would seem to follow logically from the description of the different equity
preferences provided by Huseman et al. (1985, 1987). Based upon my review, no
study has investigated the impact of equity sensitivity on employee beliefs about

employee—employer obligations, expectations, or promises.

123

A number of researchers have argued that trust plays an important role in
psychological contracting. Nicholson and Johns (1985) treated trust as a defining
characteristic of PCs, introducing the notion of ”high trust" and ”low trust"
psychological contracts. According to Rousseau's (1989), "underlying a psychological
contract is mist” (p. 128; emphasis included in original text). In her conceptualization,
trust appears to be explicitly treated as a necessary antecedent of all psychological
contracts (Rousseau, 1989, Figure 1, p. 125).

The belief that organizational tenure is likely to influence employee PCs has been
around since the earliest detailed writing about PCs (cf. Levinson, 1962). Rousseau
(1989) has argued that the longer a relationship endured between employee and
employer, with repeated cycles of contribution and reciprocity, the broader the array of
contributions and inducements that may be involved. More recently, Rousseau and
McLean Parks (1993, p. 32) explicitly proposed that ”As the duration of the
relationship increases, individuals come to believe the contract terms include more
relational, socioemotional terms (e. g. support, interest in personal welfare). "

Hypotheses 1A, 1B, 2A, and 2B imply that the pattern of correlations with external
criteria will be different across obligation and expectation forms, and across obligation
and promise forms. Thus, although specific hypotheses were not postulated regarding
each of the external variables included in the investigation, the following general
hypotheses were formulated:

HYPOTHESIS 3A: Items assessing employee beliefs about employer obligations

and the matching items assessing employee expectations regarding their employer

124

will have different patterns of relationship to external criteria such as trust in
employer, equity sensitivity, work centrality, and demographic variables.
HYPOTHESIS 3B: Items assessing employee beliefs about their obligations to their
employer and the matching items assessing employee's beliefs about what they
expect to provide their employer will have different patterns of relationship to
external criteria such as trust in employer, equity sensitivity, work centrality, and
demographic variables.
HYPOTHESIS 4A: Items assessing employee beliefs about employer obligations
and the matching items assessing employee beliefs about employer promises will
have different patterns of relationship to external criteria such as trust in employer,
equity sensitivity, work centrality, and demographic variables.
HYPOTHESIS 4B: Items assessing employee beliefs about their obligations and the
matching items assessing employee's beliefs about their promises to their employer
will have different patterns of relationship to external criteria such as trust in
employer, equity sensitivity, work centrality, and demographic variables.
Hierarchical Nature of Perceptions: Obligations, Expectations, and Prom'mes
Hypotheses 1 through 4 predict that there will be differences in participants'
responses to obligation, expectation, and promise based conceptualizations of EPCs.
Differences may occur on a variety of dimensions. I expect there to be differences in
two related but distinct aspects of participants' response patterns. First, I expect that
the average number of items included in participants' EPCs will vary across the three

sets of items; perceived promises will have the lowest average number of items and the

125

expectation set will have the highest average number of items. Operationally, an item
will be deemed part of a participant's EPC if the subject's response to the item
indicates that there is at least some obligation, expectation, or level of promise
(depending on the relevant set). Second, I expect that the strength or magnitude of
participants' responses to EPC items will vary across the three sets of items.
Generally, across corresponding or ”matched” items, the average response to the
promise based item will be the weakest (lowest mean), and the average response to the
expectation based item will be the strongest (highest mean).

The basis for these predictions is the belief that promises, obligations, and
expectations are, generally speaking, related in a hierarchical way. Of the three types
of beliefs, promise is the most narrowly defined concept. Promises are a source of
beliefs about obligations. I would argue that for any person that responds that an
employer promised to do or provide a certain thing, there is a high probability that the
person will respond that the employer is obligated to do or provide the thing. The
converse is less likely because there are other, non-promissory sources of beliefs about
obligations (e. g. social norms). Similarly, beliefs about obligations are a source of
expectations. For any person who responds that an employer is obligated to do or
provide a certain thing, there is a high probability that the person will respond that the
employer is expected to do or provide the thing. Again, the converse is less likely
because not all expectations are based on beliefs about obligations. These predictions

are captured in the following hypotheses:

 

126

HYPOTHESIS 5A: The average number of items included in promise based EPCS
will be less than the average number of items included in obligation based EPCS.
HYPOTHESIS 5B: The average number of items included in obligation based
EPCS will be less than the average number of items included in expectation based
EPCS.
HYPOTHESIS 6A: The mean response to expectation based EPC items will be
greater than the mean response to the corresponding obligation based EPC items.
HYPOTHESIS 6B: The mean response to obligation based EPC items will be
greater than the mean response to the corresponding promised based EPC items.
Fulfillment Measures and the EPC Nomological Network. 1
To further investigate whether or not the distinction between obligations and
expectations and the distinction between obligations and promises are meaningful, the
relative extent to which measures of EPC fulfillment that were framed in terms of
either obligations, expectations, or promises resulted in findings that were consistent
with well accepted links in the EPC nomological network was tested. While there is
disagreement among researchers regarding the nature of the beliefs that constitute the
EPC, in their broader theorizing about employee psychological contracting in
organizations, all researchers have argued that violations of ECCS lead to strong
negative employee responses. This is, perhaps, the most universally agrwd upon
proposition in the PC literature. Evidence that violations of an obligation based EPC
were more predictive of negative employee attitudes or intentions than violations of an

expectation based or promised based EPC (or vice versa) would indicate which

127

conceptualization of the MW of EPCs is more consistent with a uniformly
accepted proposition regarding psychological contracting. In other words, such
evidence would suggest which belief element is more consistent with an accepted
”web” in the extended EPC nomological network.

The relationship between obligation, expectation, and promise based measures of
EPC fulfillment, respectively, and three variables which have been identified as being
affected by EPC fulfillment was investigated. The three variables are job satisfaction
(global and specific facets), intention to turnover, and trust. The specific hypotheses
tested are presented and discussed immediately below.

Obligations versus expectations. Previous research provides evidence supporting
the hypothesized relationship between EPC ”violations" and job satisfaction. Portwood
and Miller (1976) found a correlation of .37 between a measure of unmet employee
expectations and job satisfaction. Robinson and Rousseau (1994) found that a single
item measure of the extent to which an employer had violated its obligations to the
employee correlated -.76 with a two item measure of job satisfaction.

Previous research also provides support, albeit mixed, regarding the hypothesized
relationship between unmet obligations and intentions to turnover. Robinson &
Rousseau (1994) found that a single item measure of the extent to which an employer
had violated its obligations correlated -.32 in intentions to remain with the employer.
Guzzo et al. (1994) looked at three groups of practices (financial inducement, general
support, and family oriented support) and the extent to which expatriate managers’

perceptions of what they received from their employers differed from what they

128

thought they "should” have received. Of the three groups, general support was the
only one correlated significantly with a measure of turnover intentions (—26).

The failure of an employer to meet employee's expectations or perceived obligations
is likely to result in a decline of trust (cf. Robinson & Rousseau, 1994). Empirical
support has been found for the hypothesis that a negative relationship exists between
contract violations and trust. Robinson and Rousseau (1994) found a correlation of -
.79 between a single item measure of contract fulfillment/violation and trust in the
employer.

At first glance, the findings of these studies may suggest that the relationship
between contract fulfillment/violation and job satisfaction and intentions to remain with
the employer are stronger when ECCS are conceptualized as involving obligations.
Confounding factors, however, make such a conclusion unreasonable. The subjects in
the study using the obligation conceptualizations were executive MBA students. The
limitations of generalizing from the executive MBA sample has been noted, and it was
suggested that employees with fewer employment options may not react as strongly to
EPC violations (of. Robinson, 1994). The subjects in the Portwood and Miller study
were clerks working in a chain of grocery stores. The difference in the magnitude of
the correlations in the two studies may be due to the differences between the samples.

A second potential confounding factor is the fact that the above studies focused on
different obligations. Robinson and Rousseau (1994) looked at relatively few (7),

generally stated obligations that covered a broad range of the employment relationship.

129

In contrast, Portwood and Miller measured 20, more specific expectations, and Guzzo
et al. (1994) used 48 items assessing specific practices relating to expatriates.

Finally, the single item measure of contract fulfillment used in Robinson and
Rousseau (1994) may be a confounding factor. Generally, we think of single item
measures as limiting the magnitude of its correlations with other measures. It is quite
possible that the single item, global measure of contract fulfillment/violation was
deficient, and that the high correlation is due, in part, to a strong mono method bias.
In both Portwood and Miller ( 1976) and Guzzo et al. (1994), more specific, multiple
item measures were used to assess the extent of fulfillment/violation.

A relatively recent, unpublished study by Robinson (1994) provides an additional,
more direct comparison of the relative relationship of unmet expectations and unmet
obligations to intention to turnover and trust. The study assessed the EPCS of executive
MBAs at three points in time. At Time 1, after subjects had accepted job offers but
before they had started work, subjects were asked to rate how much of each of seven
items they expected to find in their work situation. The seven items were based on the
instrument used in Rousseau (1990). At Time 2, 18 months later (and after subjects
had been on the job) violations of seven perceived obligations were measured. The
content of the obligations was the same as the content of the expectations items, the
main difference being that the items were framed in terms of obligations. At Time 3,
12 months later, subjects were asked to rate the extent to which the seven items were
actually part of their work situation. The difference between the ratings of the

subject's work environment at Time 3 and their ratings of what they had originally

130

expected at Time 1 was used as a measure of unmet expectations. Job satisfaction and
intentions to remain with the employer were also measured at Time 3. The following

correlations were obtained:

Contract violation Unmet

: l l' . l 1: .
Intention to remain -.37** -.23**
Trust -.39** -.30**

** p < .01
Again, at first glance, these correlations seem to support the conclusion that the

relationship between contract fulfillment/violation and job satisfaction and intentions to
remain with the employer are stronger when an obligation based measure of the EPC is
employed. Unfortunately, the purpose of the study did not include the comparison of
the two conceptualizations, and as a result, the design does not permit such a
comparison. Although the use of similar items somewhat reduces concern regarding
one of the potential confounding factors that existed in the previous discussed studies,
the different times at which Robinson (1994) collected data regarding expectations and
obligations is potentially confounding”. The measure of unmet expectations is based,
in part, on expectations that subjects had before they started work with their employer.

In contrast, fulfillment/violation of perceived obligations was measured after the

 

10

The obligation items assessed ”the extent to which" employers had fulfilled
obligations regarding the seven things. The expectations assessed "how much" of the
of the same seven things were expected in the work situation. Thus, the items are
similar but could have been more closely matched if the expectations items had
assessed the extent to which the seven things were expected.

131

subjects had been on the job for awhile and, more importantly, closer to the point in
time that the criteria (intention to remain and trust) were measured. Given the nature
of the constructs, the fact that they were all measured using a paper and pencil self
report, and the times at which the measures were administered, the higher correlations
obtained with the contract fulfillment-obligation measure may be due to the operation
of method bias.

As I have already indicated, I expect that violations of beliefs about employer
obligations will result in stronger negative responses than violations of expectations
regarding one' s employer. Based upon the preceding discussion, the following
predictions are made:

HYPOTHESIS 7: The correlations between obligation based fulfillment measures

and job satisfaction will be significant, positive, and of a larger magnitude than the

correlations between the corresponding expectation based fulfillment measures and
job satisfaction.

HYPOTHESIS 8: The correlations between obligation based fulfillment measures

and intentions to remain with the organization will be significant, positive, and of a

larger magnitude than the correlations between the corresponding expectation based

fulfillment measures and intentions to remain with the organization.

HYPOTHESIS 9: The correlations between obligation based fulfillment measures

and trust in the employer will be significant, positive, and of a larger magnitude

than the correlation between the corresponding expectation based fulfillment

measures and trust in the employer.

132

In addition to investigating differences between expectation and obligation based
conceptualizations of EPCs, the testing of Hypotheses 7, 8, and 9 provides evidence
regarding the generalizability of earlier findings regarding the hypothesized
relationships (e. g. , between contract fulfillment/violation and trust). The present study
involved a substantially different subject pool, and a more comprehensive measure of
contract fulfillment.

Obligations versus promises. I do not know of any empirical evidence that bears
directly on the question of whether violations of an obligation based conceptualization
of EPCs would be more or less predictive of employee attitudes and intentions to
turnover than violations of a promised based conceptualization of EPCs. It could be
argued that because perceived promises involve some perceived representation or
declaration on the part of the employer, on average, a violation of a perceived
employer promise would result in a greater sense of deceit, personal betrayal, and
anger than a violation of a perceived obligation (which may or may not be based on a
representation or declaration by the employer).

It was expected that violations of an obligation based EPC would be more predictive
of negative employee attitudes and intentions than violations of a promised based EPC.
Employee beliefs about employer obligations are often based on very basic, albeit
unspecified, social norms and moral principles (obligations to treat individuals with a
minimal level of respect, dignity, and fairness). Violations of these basic norms or
moral principles, I would argue, may produce negative reactions equal to, and in some

instances, greater than, the violation of perceived employer promises. Also, while

v -*‘ ‘l-"l I

 

133

there may some non-promissory obligations that can be violated without significant
negative reactions by the employee-promisee, there are also circumstances where the
failure to keep a promise is viewed as justified (cf. Macneil, 1980), and as a result, not
likely to lead to negative employee responses.

Even if the violation of promise based obligations, on average, resulted in greater
negative employee responses (i.e. greater intensity of response), I would expect that
violations of promise based conceptualization of EPC would be less predictive of
negative employee attitude and intentions than violations of an obligation based
conceptualization because of the narrower scope of the former and its resulting relevant
deficiency in terms of capturing negative employee response inducing events.
Violations of both perceived promissory and non-promissory based obligations are
expected to produce strong negative reactions. An obligation based measure will
capture both of these kinds of violations. The narrowed scope of the promise based
measure would fail to capture the violation of perceived, non-promissory based
obligations. Based on these considerations, the following was hypothesized:

HYPOTHESIS 10: The correlations between obligation based fulfillment measures

and job satisfaction will be significant, positive, and of a larger magnitude than the

correlations between the corresponding promise based fulfillment measures and job

satisfaction.

134

HYPOTHESIS 11: The correlations between obligation based fulfillment measures

and intentions to remain with the employer will be significant, positive, and of a

larger magnitude than the correlations between the corresponding promise based

fulfillment measures and intentions to remain with the employer.

HYPOTHESIS 12: The correlations between obligation based fulfillment measures

and trust in the employer will be significant, positive, and of a larger magnitude

than the correlations between the corresponding promise based fulfillment measures
and trust in the employer.

Support for hypotheses 10, 11, and 12 would constitute support for an obligation
based conceptualization of EPC. Lack of support for hypotheses 10, 11, and 12 would
not obviate the need to heed the previously stated admonitions regarding the use of a
promise based conceptualization of EPC. For example, in my view, it would still be
inappropriate to use the relational contracting concept to describe its antithesis, a
promise based conceptualization of contract, without, at a minimum, addressing the
manifest departure from Macneil’s view of the distinction between promise and

obligation.

135
CHAPTER V

Methods

Overview of Methods

This field study involved the collection and analysis of survey data provided by
participants who were solicited from a broad range of organizational settings. The surveys
were administered in person (on-site) and via the mail, depending on the source.
Participants were randomly assigned to complete obligation, expectation, or promise
based forms of a questionnaire assessing EPCs. In addition, all participants completed
measures of other variables (e.g., work centrality, job satisfaction) and provided
demographic information that was used in the testing hypotheses. Multiple approaches to
assessing the equivalence of the alternative measures were employed, including: testing
the invariance of the measures’ factor structure across forms using LISREL multiple
group CFA; evaluating the pattern of observed correlations between different EPC
measures and external criteria; assessing the relative “scope” of EPCs measured by the
alternative forms examining their mean differences; and evaluating the relative extent to
which obligation, expectation, and promise based EPC fulfillment measures correlate
with employee job satisfaction, trust, and intentions to turnover.
Sample

Total sample. Data for this study were collected from 1054 participants. All
participants were employed; 74.1% were employed full-time (35 or more hours per
week). Participants had an average of 13.3 years of work experience (SD =9.6).

Tenure with current employer averaged 4.9 years (SD=6.10). Participants were

136

employed in a wide variety of occupations: managerial positions (18.6%), sales and
marketing (5.6%), professionals (9.1%), secretarial and other office help (22.8%),
traditional blue collar workers (35.7%), and other occupations not easily classified as
white or blue collar (e.g., aerobic instructors, day care providers; 8.2%). A more
specific break down of participants' occupations is provided in Table 1.

Sixty-two percent of the participants were female. Their ages ranged from 17 to
66, and the average age was 32.1 (SD= 10.8). The racial composition of the sample
was as follows: 77.6 % White, 8.6 % African American, 5.5% Hispanic, 5.2 %
Asian/Pacific Islander, and 1.3% "Other”. Ninety four percent had a high school
degree, 72.3% had some college, 32% had a four-year college degree, and 10.2% of
the participants had continued their education beyond a four-year college degree.

Sub-samples by source. Participants were recruited from 21 distinct sources.
Descriptions of the sub-samples by source are provided in Table 2. The sources
contributed sub-samples that varied in size from 3 to 298, constituting between .3%
and 28.3 percent of the total sample.
Measures

Employee Psychological Contract Questionnaire (EPCQ). Developing a workable
measure of EPCs is a considerable challenge because to some extent, EPCs are
conceptualized as idiosyncratic to particular employment relations. In particular, it would
be extremely difficult, if not impossible, to develop a widely generalizable measure that
uses standard scale items to assess a broad range of discrete obligations. Discrete

obligations reflect the most idiosyncratic aspects of EPCs. By definition, they are based

137

 

 

 

 

 

 

Table 1 Qeoupatiops
Valid
Occupation Number Percent
Managers
Upper level 39 3 .9
First line supervisors 29 2.9
HRM l l 1.1
Other 107 10.7
Subtotal 186 18.6
Sales and marketing 56 5.6
Professionals
Lawyers 16 1.6
Accountants 1 1 l. l
Engineers 38 3.8
Scientists 15 1.5
Other professionals 11 1.1
Subtotal 91 9.1
Office workers
Legal secretaries 27 2.7
Other secretaries 32 3.2
Receptionists 10 1.0
Administrative assistants 38 3.8
Assistants to professionals 16 1.6
(e.g., para-legal)
Technical support 18 1.8
General oflice help 87 8.7
Subtotal 228 22.8
Traditional blue collar
Production workers 316 31.6
Building & ground maintenance 21 2.1
Other blue collar (e. g., skilled trades) 20 2.0
Subtotal 375 35.7

 

1 3 8
Table (Cont’d).

 

Other occupations
Health care providers 14 1.4
Day care worker 11 1.1
Aerobics instructors/trainers l 1 1.1
Food preparation 25 2.5
Food server (e. g. waitress) 21 2.1
Subtotal 82 82
Missing 54 Missing

 

Total 1054 100.0

139

.320 N E

 

 

.x. cm

8:: ”83:8 wag .\. 9:- o\.. .\o >mo
mocnoo b 8.an 88 .x. .663 AvNooev can 3% SE8 K 89
chmmoanon E q .q mo Eon—Homo: 395 So .2 .a So 3 A8

385952 3 New New 3 Eacaoacoaé Sb NS Se 8.“ A88

Hgmgawmoa 3 Wu 3 325mg»— Cooaev 3o 3 eoh fig 3.8
Hoses.

moon 838 me We 3 Egona GEE 33 we; 8o 3 63
Beige.

Econ..— wam mm mm mm. 038 Auk—.5 ewe Ge care 39 33
02.0% 33838.8 G was;

3.788 can: No Nm So Zgoafi A395 we.» Na: web web Go. 5

938 31$

mammaoonam Am Ad 3o wnomammoaa Amefov So one 3.0 SQ 3.3
m3.

mama—553m— : 80 So H83. «:25: 30.5 :5 sum 8o awe 63
835m m3. 233ng Quake

20:63? 3 EN 3 wagon“: RNXV ab 3.» gab ugh. C #3

race—Hm

140

PB: m3.— E

52 mm: 3

rmfimndnw

maze @185 can

mam—9 0:5

@383. moo—$8

03:8:
nonmaomwnoa

Hamming—5.
8:98.”

mxoocng
25? ..3

989.98
7.5? .5

AN

um

5

so

um

um

A0

Po

uh

5

Pm

we

we

Pm

.3

mo

mo

.3

an

ac

no

mm

m.»

mm

Hue—o 803.8.

038 31$
”125.3 Geeky

meagre 28$

massage Gee

H.939. 393v

moon wagon G 4.5
Enamele— 69.5

>803.“ 5a. 395

68. 88 Sax;

was mama coosv

038 0&5

Ema—mamm— Gmfov

zgwmaaa G woev
mnemommmoaw— 3&5

OBS 3&5

gag—momm— Quake
gomommmoaa 6:5

Kmawmana Gecko
32.8303— 3&5

3.0

00..»

30

owb

me

many

mu.—

0A..»

0.?“

cub

3Q

Nm

.3.—

3..»

Nun

mwe

qwb

so

36

8o

.3

mum»

cub

Bo

0A..»

web

we

umh

we

”SM

”3

3

mm

Am

36

um.m

web

93

3.3

as

3.:

3.5

9
9...:

$3

3. .3

RB

141

€039.me
canonmnam. w _

€029me
Sigma—am. & N

Héoéofl. 05503
005% «8038

3A

mm

H:

30 co
Mm 5o
PM 80

H020 603.3.

0300 Knox;
was 3093

0B8 A3§

3.00030: 3093

:9 3.0 web NN.A

3.0 Nmb 3b NND

Harm mob «NA NM.»

05

0.3

3.5

 

142

on specific perceived promises (Macneil, 1980), and therefore, are likely to be
influenced by the specific promises or representation a given employer makes. In order
to assess the most discrete aspects of EPCs, it will be necessary to tailor the EPCQ to
the specific unit of interest (organizations, jobs), or to use open—ended questions.

Figure 6 summarizes the approaches that have been used to assess EPCs. A quick
review of the figure will confirm that there is no consensus regarding the approach to
measure EPCs. In light of this fact, and because previously reported measures of the EPC
were viewed as possessing significant limitations for my purposes, a new measure that
builds upon previous measures was developed and employed. A primary purpose in
developing the EPCQ was to provide a measure of EPC that balances the desire to have a
measure with wide generalizability and the desire to capture the important idiosyncratic
aspects of EPCs. It may be possible to develop a measure with greater generalizability,
and, it is possible to measure EPCs in a way that may identify more idiosyncratic aspects
of individual EPCs. The EPCQ, however, seeks a balance.

The original or base form of the EPCQ assesses beliefs about employee-
employer obligations to each other by asking participants to what extent their employer
is obligated to provide specific items (e.g. , work that is challenging), and what extent
their obligations to their employer included specific items (e.g., working extra hours
without extra pay). Participants responded using a five-point Likert scale. In the
obligation based form, anchors range from ”not at all obligated” to ”very highly
obligated” . The specific anchors were adapted from previously published measures

(Rousseau, 1990; Robinson, et al., 1994; Robinson & Morrison, 1995).

143

Figure 6 Approaches Used To Assess Employee Psychological Contracts

 

and):
Argyris (1960)
Levinson et al. (1962)

Jurek (1969)

Kotter (1973)

Portwood & Miller (1976)

Holtz (197s)

Rousseau (1990)

Robinson, Kraatz &
Rousseau (1994)

Robinson & Rousseau (1994)

oral-1.13m. (1991)

Guzzo, Noonan, & Biron
(1994)

Barksdale (1995)

 

Field interviews

Field interviews

Questionnaire asking employee and his/her supervisor to indicate the
extenttowhichtheirneedswerebeingmetfitembasedonsix
motivator and nine hygiene job factors posited by Herzberg et al.
(1959).

Questionnaire asking employee and 'employer' to indicate exterl to
whichthey expected certainthings; items (7) appeartobederived
from Berlew and Hall (1966), a study that identified employee-
employer expectations through in-company interviews.

Questionnaire developed by taking the Minnesota Satisfaction
Questionnaire, Weiss et al. (1967), and changing the sedence stem
toaskemployeesabouthowmuchofthe 'thing'wasexpectedand
how muchwas received.

Field interviews and a review of company written philosophy,
policies, and practices.

Questionnaire assessing employee beliefs regarding his/her and
employer obligations; items said to include those obligations
commonly promised by employers of MBA graduates (based on

interviews of HRM managers). The 15 specific items used by
Rousseau (1990) and colleagues are reproduced at pages 26-27 ofthis
text.

Questionnaire assessing employee and supervisor expectation; items
taken from Leadership Behavior Questionnaire (Fleishman, 1953A;
Fleishman, 1953B) for two dimensions of leadership behavior,
initiating structure (18 items) and consideration (28 items).

Two measures were used. First, a questionnaire (43 items) asked
employees to rate the extent to which 43 company practices relating
to expatriate managers was provided by employer, and, the extem
towhichwhatwasprovidedwasdifi'erentfrumwhat 'should'be
provided; items were based a survey of companies regarding their
practices. Second, items from Eisenberger et al. (1986) perceived
organizational supportmeasure wereusedto assess the 'status" of
employee PCs.

Questionnaire assessing obligations, adapted from Rousseau and
colleagues

 

144

The content of the EPCQ, that is, the aspects of employment assessed by the
items composing the EPCQ, was determined based upon a review of the literature and
the findings of a pilot study that was conducted. The following is a narrative
description of the process that was followed.

In conducting a review of the literature and available measures of PCs, those
aspects of employment that were being described as within the subject matter of
employee PCs were identified. Particular attention was paid to identifying aspects of
employment that were consistently included in descriptions of EPCs, across authors,
and aspects identified through empirical investigations (e. g. , Robinson & Morrison,
1995). An initial questionnaire was constructed that contained 37 items assessing
employer obligations and 23 items assessing employee obligations. The questionnaire
included adaptations of items from existing measures, and items that I generated to
assess obligations that were frequently identified and discussed in literature, but not
included in previous measures. For example, a number of authors have discussed the so
called "new employee psychological contract,” and its emphasis on work - life balance
issues. Existing measures, at least those that I have access to, did not contain items
regarding this aspect of employment. Also, items assessing obligations thought to be
associated with participatory employment relationships were included because of the
importance being attributed to participative management by many (e.g. , Reich, 1987;
Coye & Belohlav, 1995) and because the ”new employee psychological contract” is

being described as participatory in nature (e. g., Sims, 1994).

145

The initial questionnaire was pilot tested using 126 undergraduate students. The
students held a variety of full-time and part-time jobs. In addition to the scale items, the
questionnaire included one open format question asking participants to identify any other
obligations that they believed their employer owed them, and one open format question
asking them to identify any other obligations that they believe they owed their employer.
Approximately one in eight participants identified at least one additional employer
obligation. Although there were significantly fewer items in the questionnaire assessing
employee obligations, only one in 15 participants identified at least one additional
employee obligation. Based on the responses to the open format questions, additional
items were generated. For example, the obligation that was most frequently identified in
the open format responses was an employer obligation to treat the respondent employee as
they would other equally qualified employees, regardless of gender or other personal
characteristics.

The items generated up to this point were evaluated to determine if they appeared
to fairly represent the range of relational obligations identified in my conceptualization of
relational employee-employer obligations (pages 109-114). After reviewing the existing
items, a determination was made to generate additional items specifically aimed at
assessing both employee and employer relational obligations. Specifically, the additional
items were intended to assess employer and employee obligations to be flexible and
cooperative in their dealings with each other, a central characteristic of relational
contracting (Macneil, 1980).

In summary, the items included in the EPCQ can be categorized by their source as

follows:

146

a) items contained in previous measures and adaptations of previous items;

b) items assessing aspects of employment identified prominently in the literature, but
which were omitted from previous, available measures;

c) items based on responses provided in a pilot study to open ended questions
regarding employee-employer obligations (support for the items could also be
found in the literature); and

d) additional items generated specifically to assess relational obligations.

The final base version of the EPCQ contained 44 items assessing employer
obligations and 33 items assessing employee obligations. A copy of the EPCQ is
attached as Appendix A. Comparing the content areas of the EPCQ with the facets of
job satisfaction assessed by the popular Job Descriptive Index, it can be shown that there
are multiple items assessing obligations relating to each of the JDIs's five facets. The
EPCQ is, however, broader in scope, including items related to work/non-work life
balance, participative management, mutual flexibility, and shared sacrifice. As indicated
earlier, the EPCQ does not purport to assess all relevant obligations. It does purport to
assess obligations thought to have importance across range of organizations, jobs, and job
holders. Findings regarding the factor structure of the EPCQ items will be presented and
discussed below.

The base form of the EPCQ, which assessed beliefs about employee-employer
obligations, was adapted to create two additional forms which assessed employee
expectations and employee beliefs about employee and employer promises. The

expectation form asked participants to indicate the extent to which they expected their

147

employer is to provide specific items and the extent to which they expected to provide
their employer specific items. The item stems were identical to those assessed in the
obligation form. Participants responded using a 5-point Likert scale, with anchors
ranging from "not at all expected” to "very highly expected.”

The promise form asked participants to indicate their beliefs regarding the extent
to which their employer promised to provide them specific items and their beliefs
regarding the extent to which they promised to provide their employer specific items.
Again, the item stems were identical to the items contained in the other forms.
Participants responded using a S-point Likert scale, with anchors ranging from "not at
all promised" to ”very highly promised.” ‘

To promote uniformity among participants in their understanding of the
operative terms “obligation,” “expectation," or “promise,” each form provided a
definition of the relevant term. Depending on the form, one of the following definitions
was provided in the initial instructions and at the top of each page of the questionnaire
containing items assessing the EPC:

Obligation form: “An obligation is a duty or responsibility that is imposed by law,
society, promises that are made, or morality (“good
con science") . "

Expectation form: “Something is expected to the extent that you believe that it is
likely to occur.

Promise form: “A promise is a declaration (a clear statement) that one is

committed to do or refrain from doing something. A promise may

148

be made orally (spoken), or in writing, or it may be inferred fiom
conduct which clearly indicates that your employer intended to
make a commitment to you.”

Job Satisfaction. Two measures of job satisfaction were employed. A 14 item
measure taken from Hackman and Oldham (1985) Job Diagnostic Survey assessed five
specific satisfactions: pay (two items), job security (two items), social (three items),
supervisory (three items), and growth satisfaction (four items). Responses were made on a
7-point scale. The items measuring each specific facet were averaged to yield five specific
satisfaction scores, and responses to all 14 items were averaged to yield a global
satisfaction score. Global job satisfaction was also measured with two items used in
previous research by Meyer and Allen (1988) and Irving and Meyer (1994): “All things
considered, how satisfied are you with your job?” and “considering all jobs that you might
realistically have at this point in your career, how does your job compare?” Responses to
these items were made on 7-point scales and were averaged to yield global satisfaction
scores.

Turnover intentions. Turnover intentions were measured by three items adapted
from Cammann, F ichman, Jenkins and Klesh (1979). Adaptations of the Cammann et al.
(1979) measure appear prominently in the literature (e. g., Major, Kozlowski, Chao, &
Gardner, 1995; Irving & Meyer, 1994; Ostroff & Kozlowski, 1992). Participants respond
to one item, "I often think about quitting,” using a 7-point scale ranging from 1 (strongly
disagree) to 7 (strongly agree). The other two items assessed the likelihood that

participants would look for a new job and/or leave the organization within the next year.

149

Participants responded to these items using a 7-point scale ranging from 1 (not at all
likely) to 7 (extremely likely).

Trust. Employee trust in their employer was measured using a scale developed by
Robinson and Rousseau (1994). The scale contains 7 items reflecting the bases of trust
identified by Gabarros and Athos (1976). Sample items include "In general, I believe my
employer's motives and intentions are good” and "My employer is not always honest and
truthful.” Participants respond using a 5-point Likert scale ranging from 1 (strongly
disagree) to 5 (strongly agree).

Work centrality. The primary measure of work centrality was the “Work
Involvement Questionnaire” which was developed and validated extensively by Kanungo
(1982). The scale contains 6 items assessing individuals’ beliefs regarding the degree of
importance that work plays in their life. Sample items include “The most important
things that happen in life involve work" and "Work should be considered central to
life.” Participants respond using a 7-point Likert scale ranging from 1 (strongly
disagree) to 7 (strongly agree). Work centrality was also assessed with a single item
adapted from the measure used by Lundberg and Peterson (1994; referencing The
W, MOW, 1987). Participants respond to the item “Considering
work in general (without reference to any specific job), overall, how important is work in
your life?” using a 7-point scale ranging from 1 (not at all important) to 7 (extremely
important). Post hoc analysis revealed that combining the Kanungo’s (1982) 7 item

measure the 1 item MOW measure resulted in a combined measure of work centrality

150

that had psychometric properties superior to either measure alone. Therefore, the
combined measure was used in the analyses.

Equity sensitivity. The Equity Sensitivity Instrument (ESI) is a five-item scale
developed by Huseman et al. (1985) that identifies a person’s preferences for outcomes
versus input in a general work situation. The subject has a choice of two responses for
each item, one representing a benevolent response, and the other an entitled response.
Subjects show their agreement or disagreement with each response by distributing 10
points between the two statements. The instrument is based on the premise that
benevolents will allocate more of their points to the benevolent statements than to the
entitled statements, entitleds will allocate points conversely, and equity sensitive will
allocate their 10 points equally between the statements. ESI scores are the sum of the
points allocated to the benevolent statement in each of the five survey items. Thus, scores
may range from O to 50.

Acceptable coefficient alphas have been consistently reported for the E81: .83
(Husemen et al., 1985); .79 (Miles et al., 1989); .79 and .77 (Patrick & Jackson, 1991;
student and employee samples, respectively); .81 (King, Miles, & Day, 1993); and .81,
.82, .88, .81, .79 (King & Miles, 1994; 1 student and 4 employee samples). Test re-test
reliability, using a three week interval, and was found to be .80 (Nfiles et al., 1989.).
Finally, evidence of convergent and discriminant validity has been provided by a study that
used five samples to assess the correlations between scores on the ES. and other paper
and pencil measures of constructs hypothesized to be related to equity sensitivity and
constructs thought to be distinct (King & Miles, 1994). The findings generally supported

the predicted relationships.

151

Perceived extent to which EPC has been fulfilled by employer. Three different
versions of the measure of contract fulfillment that correspond with the three different
belief elements that are being compared in hypotheses 1 through 12 were created. For
each item, participants responded using a using a 5-point Likert scale with the following
scale anchors:

Obligation form: "much less than obligated" to "much more than obligated"
Expectation form: "much less than expected" to "much more than expected"
Promise form: "much less than promised" to "much more than promise"

Organization tenure. Organization tenure was assessed with a single item: “How
long have you worked for your current employer? __ Years _ Months.”

Education level. Education level was assessed with a single item that asked
participants to circle the highest level of formal education that they had obtained. The
response categories included seven levels of education, from “some high, but did not
graduate,” to “work beyond a masters degree.”

Demographics and other sociological facts. Participants were asked to provide
general demographic information and other sociological facts, including: age, race, gender,
and education level.

Procedures

Participation in the study involved completing a survey that generally required

between 15 and 25 minutes to complete. Participants were randomly assigned to one of

three survey conditions (obligation, expectation, or promise form), and all received the

152

same instructions. Practical constraints, however, required some adaptations in how the
survey was administered to participants across data sources.

Administration of surveys in groups, on site. Most participants (723) were
administered the survey in groups, either at their workplace or in their classroom. Once
the group had assembled in a single room, informed consent forms were distributed and
an announcement was made indicating that the purpose of the survey was to examine
"certain employee attitudes and beliefs" about their relationship with their employers.
The voluntariness and confidentiality of responses were stressed. In addition to the
consent form, participants who completed the survey at their place of work received a
cover letter confirming the announcement, the voluntariness of participation, and the
confidentiality of responses. They were also given a large envelope in which the
completed survey could be sealed before it was returned. Participants were asked to
read the informed consent form, and if in agreement, sign it. Participants were
randomly assigned to one of the three survey conditions, identified above. Executed
consent forms and completed surveys were collected separately. After completing the
survey, participants were given an opportunity to ask questions.

Alternative administration of surveys to groups. When practical constraints
precluded on site administration of the survey in groups, an alternative procedure was
employed. Once the number of potential participants in the targeted group was
identified, individual survey packets were prepared. Each packet included an informed
consent form, one of the three versions of the survey, and a cover letter, which were

all placed in a large, stamped return envelope addressed to Mark V. Roehling,

153

Management Department Michigan State University. The cover letter provided the
same instructions as those provide participants administered the survey on site, except
that instead of returning the questionnaire directly to the researcher, they were
instructed to place the completed questionnaire in the return envelope, seal the
envelope, and return it via the mail. The survey packets were distributed to participants
by their employer.

Administration of surveys to church group. As part of an effort to obtain a
diverse sample, participants were solicited through a church group. Two hundred
survey packets were prepared for distribution to church members based upon estimates
provided by church personnel. The cover letter that was included in the packets was
adapted to reflect the fact that a small financial contribution would be made to the
church for each completed survey that was returned by a church member. On two
consecutive Sundays, an announcement regarding the study was made in church, and
survey packets were made available after the service. Church members who elected to
participate picked up a survey packet from the back of the church after the service had
ended, completed the survey at home, and returned the completed survey to a collection
box placed in the church.

Administration of surveys to individuals. Thirty-eight participants were
solicited individually, through personal contacts. They were given a survey packet
which included the materials described above and a 3" x 5" post card. They received
the same instructions as those participants who were administered the questionnaire in

groups, except that instead of returning the questionnaire directly to the researcher,

154

they were instructed to place the completed questionnaire in the return envelope, seal
the envelope, and return it via the mail. Because feedback through their respective
employers was not available, individually solicited participants were also instructed that
if they desired additional information regarding the study, they could provide the
researcher their address or phone number on the enclosed 3" x 5" card and they would
be contacted (Note: All participants received a consent form that provided them the
researcher's address and phone number and instructed them that they could contact the
researcher if they wanted more information about the study).
Analytic Procedures

Hypotheses 1A -ZB: Overview of LISREL multiple-group CFA analyses. To
test Hypotheses 1A, 13, 2A, and ZB, the factor structure, maximum likelihood
estimates of factor loadings, factor intercorrelations and unique residual variances were
examined with LISREL VIII (Joreskog & Stirbom, 1993). LISREL provides for the
simultaneous assessment of relationships in a hypothesized model. In confirmatory
factor analysis, LISREL allows the examination of models based on various
assumptions about factor patterns, factor loadings, factor intercorrelations, and
uniquenesses that are fixed to equal or specific values or to other parameter estimates,
or allowed to be free to take on any value.

Models for the employer and employee sides of the EPC were evaluated
separately because predictive or causal relationships between employee and employer
sides of EPC are not a focus of the study and, more importantly, such potential

relationships would create specification problems if both sides were simultaneously

155
included in the analysis. General support for this decision based on the identified

concern is provided by Riordan and Vandenberg (1994) and Vandenberg and Self
(1993). I would also note that reported factor analysis of EPC measure have evaluated
employer and employee sides separately (e.g. , Robinson et al. , 1994).

For the multiple-group analyses, a covariance matrix for the relevant EPC items
(employer side or employee side) was calculated for each of the three forms. The
following models were then estimated with LISREL and compared sequentially on the
basis of fit: (a) Model 1, the pattern of factor loadings was held invariant across forms;
(b) Model 2, the pattern of factor loadings and the factor loadings were held invariant
across forms; (c) Model 3, the pattern of factor loadings, the factor loadings, and the
errors were held invariant across forms; and (d) Model 4, the pattern of factor
loadings, the factor loadings, the errors, and the variances/covariances were held
invariant across forms. The four models are nested hierarchically, with Model 1 being
the least restrictive and Model 4 being the most restrictive. Evidence that a less
restrictive model fit the data provided a basis for testing models that incorporated
greater restrictions.

Assessing model fit. LISREL VIII provides a chi-square value indexing the
extent to which the observed covariance or correlation matrix differs from the
reproduced matrix implied by the model parameters. In the case of multiple-groups
confirmatory factor analysis, the chi-square represents the fit in all groups taken
simultaneously. However, because the chi—square value is dependent on sample size,

large samples may result in significant chi-squares even for appropriate models

156

(Joreskog, 1979). This consideration has lead a number of researchers to propose the
use of the ratio of chi-square to degrees of frwdom as a measure of model fit. There
is, however, no consensus regarding how large the ratio can get before requiring the
conclusion that a model is unsatisfactory. Researchers have recommended using ratios
as high as 5 to 1 (Wheaton, Muthen, Alwin, & Summers, 1977) and as low as 2 to l
(Bryne, 1989). The average of these recommendations, a maximum chi-squareldf ratio
of 3.5 to 1, was used as an upper boundary in assessing the fit of the models
investigated in the present study.

When comparing the fit of nested models, researchers using LISREL multiple-
group analysis have uniformly employed the chi-square difference procedure to obtain
incremental chi-square values. The incremental chi-square difference values are not as
prone to problems associated with the over-all chi-square test of model fit (Bollen,
1989; Riordan & Vandenberg, 1994). A significant worsening of fit (i.e., a
statistically significant chi-square difference) would indicate that the equality constraint
placed in a model was inappropriate, and as a result, the given null hypothesis should
be rejected.

A variety of other fit indices have been employed less uniformly in multiple-
group analysis. In additional to the incremental chi—square values, Hattrup, Schmitt, &
Landis (1992) focused on the non-formed fit index (NNFI) described (by Tucker and
Lewis, 1973) to assess the appropriateness of each model they tested. The NNFI is
derived by comparing the chi-square/degree of freedom ratio for each model tested

with that of a null model specifying zero covariances among measures and sample

157

invariant variances to see how much better the model fits compared to the null model.
Hattrup et al. (1992) provided two reasons for their focus on the NNFI. First, the
NNFI has been shown to be relatively insensitive to sample size (Marsh, Balla, &
McDonald, 1988). Second, although increasing the number of parameters to be
estimated generally increases the value of other fit indices (e.g. Goodness of Fit (GFI)),
the NNFI controls for the number of estimated parameters by including the degrees of
freedom of models in its computation (Marsh et al., 1988). As a result, more
parsimonious models will yield higher NNFI values when all else is equal. For these
reasons, the NNFI was one of the indices relied upon in the present study. Other fit
indices used in previous studies employing LISREL multiple-group CFA to assets
equivalence were also taken into account in the present study, but given less weight.
They include the GFI, the root mean square error of approximation (RSMEA), and the
comparative fit index (CFI).

Specification of the EPCQ factor structure. The specification of the factor
structure to be tested in the multiple-group analyses was based on an analysis of item
content and the results of exploratory and confirmatory factor analyses that were
conducted using data from only one of the three groups, the group completing the

obligation form of the EPCQ.” Several published studies have used the results of

 

“ As indicated, above, pilot data from an earlier version of the EPCQ was
previously factor analyzed and the results lead to modifications that are included in
the present form. The earlier version of the EPCQ was further modified based on
considerations other than the results of the pilot testing and the guidance it provided
regarding the items’ factor structure. For example, several items were added to
facilitate the assessment of the replicability of the findings from previous factor
analytic studies (e.g., Robinson et al., 1994).

158

exploratory factor analyses on data provided by a given sample to guide the
specification of a factor structure that was subsequently tested using LISREL multiple-
groups confirmatory factor analysis of data from the same sample (e.g., Peterson &
Smith, 1995; Drasgow & Kanfer, 1985).. The decision not to split the obligation sub-
sample to provide a hold out sample that could be used for cross validation purposes
results in less evidence of the generalizability of the structure beyond the combined
sample used in the present study, but, it increases confidence in the guidance that was
provided by the exploratory factor analysis.

Hypotheses 3A, 3B, 4A, 4B. An additional analysis of the equivalence of the
different EPC forms was obtained by correlating responses to the different forms with
external criteria. This analysis involved three general steps. First, correlational analyses
were conducted within each group to assess the relationship between EPCQ measures
completed by the group and the external criteria. Second, correlations obtained in the
expectation and promise groups were compared with corresponding correlations from
the obligation group, and the significance of observed differences were tested using the
z-test for testing the significance of differences between independent correlations (Glass
& Hopkins, 1984). Third, the resulting evidence of the pattern of relations between
EPC measures and external criteria were examined to assess similarities and differences
across forms.

The following ten EPCQ based variables were calculated and included in the
correlational analyses: overall level of response to employer items (the sum of

responses to items 1 through 44 assessing beliefs about what the employer is to give or

159

provide the employee), overall level of response to employee items (the sum of items 1
through 33 assessing beliefs about what the employee is to give or provide the
employer), scores on four employer item sub—scales (employer core, enriched work,
balance, high pay), and scores on four employee item sub-scales (employee core,
sacrifice, extra role, stay). All sub-scales were based on the results of the factor
analytic procedures described above (the specific results of the procedures are discussed
below). Sub-scale scores were calculated by summing the responses to the EPCQ items
that were identified as loading on the factor that the sub-scale represents. The EPCQ
measures were correlated with the following criteria: equity sensitivity, work centrality,
tenure, education, trust in employer, gender, and age. A

In addition, although not strictly required to test Hypotheses 3A, 3B, 4A, and
4B, hierarchical regression analysis was conducted to investigate the extent to which
equity sensitivity, work centrality, tenure, education, trust in employer, gender, and
age combined to explain variance in the EPC factor-based sub-scales (described above).
The variables were entered in four steps, in the following order: 1) gender, age; 2)
work centrality, equity sensitivity; 3) education, tenure; and 4) trust in employer.
Comparisons of the obtained adjusted cumulative R2 and the patterns of incremental
changes in R2 were made across the three forms of the EPCQ.

Hypotheses 5A & 5B. Hypothesis 5A and SB make predictions regarding the
average number of items included in obligation, expectation, and promise based EPCs.
These hypotheses were tested by calculating the average number of items that

participants identified as being part of their EPC and comparing the relevant means.

160
Operationally, an item was deemed part of a participant's EPC if the subject's response
to the item indicates that there is at least some obligation, expectation, or level of
promise - depending on the relevant item set (i.e. the response was > 1, indicating an
obligation, expectation or promise).

Support for hypothesis 5A will be provided if the average number of items in
the obligation based measure of EPC is greater than the average number of items in the
promise based measure. Support for hypothesis 5B will be provided if the average
number of items in the expectation based measure of EPC is greater than the average
number of items in the obligation based measure. The significance of mean differences
were tested using separate planned t-test comparisons for each hypothesis.

Hypotheses 6A & 6B. To test the hypotheses regarding mean responses to
corresponding obligation, promise, and expectation based measures, responses across
forms were compared at three levels: “overall” response level, the factor or sub-seale
level, and the individual item level. First, group differences in overall response level
were assessed separately for employer side and employee side items. This involved
testing mean group differences on two variables: the sum of responses to all employer
side items (1 through 44), and the sum of responses to all employee side items (1
through 33). The significance of mean differences were tested using two separate
planned t-test comparisons of the relevant group means.

Second, group differences in sub-scale (factor based) scores across forms were
assessed using two multivariate analysis of variance (MANOVAs). The first

MANOVA focused on the employer side; scores on the following sub-scales were

161

included as the dependent variables: employer core, enriched worked, balance, and
high pay. The second MANOVA focused on the employee side; scores on the
following sub-scales were included as the dependent variables: employee core,
sacrifice, extra role, and stay. Univariate analyses were conducted using planned t-test
comparisons to further explain MANOVA findings indicating significant group 3
differences.

The third level of analysis was the individual item level. All items were
used, even those not included in the formation of the sub-scales. Group differences in
responses to individual items were assessed using two MANOVAs. The first focused on
the employer side items, including items 1 through 44 assessing beliefs about what the
employer is to give or provide the employee as dependent variables. The second
MANOVA included items 1 through 33 assessing beliefs about what the employee is to
give or provide his or her employer as dependent variables. Follow up univariate
analyses were conducted using planned t-test comparisons to further explain significant
MANOVA results.

Hypotheses 7 - 12. The testing of Hypotheses 7 ~12 involved two steps. First,
correlational analyses were conducted within each group to assess the relationship
between the EPC fulfillment measures completed by that group and external criteria.
Second, the correlations that were obtained from the groups that completed the
expectation based and promise based fulfillment measures, respectively, were compared
with corresponding correlations from the group that completed the obligation based

fulfillment measure, and the significance of observed differences was tested using the z-

162

test for testing the significance of differences between independent correlations (Glass
& Hopkins, 1984).

Five fulfillment measures were calculated and included in the correlational
analysis. An overall level of fulfillment score was calculated by summing the responses
to all 42 fulfillment items. In addition, four fulfillment sub-scales scores were
calculated by summing the responses to the fulfillment items that corresponded with
the previously described, factor based employer EPC sub-scales: employer core,
enriched work, balance, and high pay. For example, the employer core fulfillment
sub-scale assessed the extent to which the employer was perceived as fulfilling core
obligations (or expectations or promise, depending on the form). ‘

The fulfillment measures were correlated with nine criteria: two measures of
global job satisfaction, trust, intention to turnover, and measures of five specific facets
of job satisfaction (growth satisfaction, pay satisfaction, job security satisfaction, social
satisfaction, and satisfaction with supervision).

While Hypotheses 7 to 12 make predictions regarding individual corresponding
fulfillment measures and their respective relationship to job satisfaction, trust, and
intentions to turnover, the hypotheses imply that the five obligation based fulfillment
measures will combine to explain more variance in job satisfaction, trust, and turnover
intentions than either the combined expectation based or combined promise based
fulfillment measures. Hierarchical regression analysis was conducted to test these

implied predictions.

163

It was reasoned that the relationship between fulfillment and the criterion may
be stronger when the employer is being perceived as not fulfilling its obligations (or
expectations or promises) than when the employer is perceived as meeting or exceeding
them. In other words, the negative effect of an employer being perceived as not
meeting its obligations may be greater than the positive effect (if any) of being
perceived as exceeding its obligations. To test for potential nonlinear aspects of the
investigated relationships, the hierarchical regressions included a polynomial term for
each of the fulfillment variables.

The predictor variables were entered in four steps, in the following order: 1)
core fulfillment variable; 2) work/non-work balance, high pay, and meaningful work
fulfillment variables; 3) overall fulfillment variable (all 42 fulfillment items); and 4)
five polynomial terms ( a square term for each of the five variables entered in steps 1,
2, and 3). Comparisons of the obtained cumulative R2 and the patterns of incremental

changes in R2 were made across the three forms of the EPCQ.

164
CHAPTER VI

Results
Internal Consistencies of Measures (other than EPC based measures)

The internal consistencies of all measures were determined by calculating the
coefficient alphas corresponding to each scale. These values for measures other than
those based on the EPCQ (which are reported later) are provided in Table 3 (p. 218).
All alphas were acceptable (all greater than .70; Nunnally, 1978), with the lowest value
equal to .73. The means and standard deviations of these variables are also presented
in Table 3.

Effectiveness of randomization procedure.

To assist in establishing that any observed difference between the groups were
attributable to the differences in the form of the EPCQ that the groups completed
(obligation, expectation, or promise), group differences on all variables used to test
hypotheses except those intentionally varied by group (i.e. , variables based on the
different forms of the EPCQ) were assessed using one of two types of analyses.
Differences among groups for gender and race were assessed using chi squared tests.
There were no significant differences in the gender composition (x’= 1.28, df =2, ns)
or racial composition (x2: 11.2, df =8, ns) of the three groups. A one-way analysis of
variance testing for group differences was performed for each of the remaining
variables. The results are presented in Table 4. No significant differences were found
at the p < .05 level. The randomization procedure was reasonably effective in avoiding

group differences that might confound results.

Table 3 WWW

 

 

Number

Measure Alpha of items M SD
Equity Sensitivity .74 5 27.6 6.8
Work Centrality .77 7 27.0 7.1
Trust .91 8 39.2 10.2
Turnover intention .83 3 10.2 5.8
Global Job Satisfaction

Job Satisfaction 1 .91 14 66.7 14.1

Job Satisfaction 2 .78 2 10.0 2.6
Job Satisfaction Subscales

Growth Satisfaction .86 4 19.6 5.0

Pay Satisfaction .87 2 9.0 3.1

Social Satisfaction .73 3 16.2 8.9

Security Satisfaction .81 2 10.1 2.8 .

Supervision Satisfaction .89 3 14.8 4.3

 

166

 

 

 

 

 

 

Qua—GA m ..t . 1.0.5 . .r......:.;...... n. o. 1» pro ».
men:
OEEmeo: 9”qung $838

<m1w2o K a MD K a MD K a MD m Bee m Ede.
>ma mu; 3»: 5a an; Aumnv :b 3Q Gm: Se .3 hoe
eon—:8 Ab Gm: Mm Ab Aummv Mm A6 A83 9m .8 .mne
min—:2 mesa. web Gwd 9A 36 A33 9m 3.x ANA—3 any fine .7:
Ham" woo 38v 8m. ”3; Anemv w... mmb A33 3..» .3 ..Ze
mun—common we Que Hm Pu fimd rm Pg 3%; _.m .2 .03
$9..“ 02:. no; 693 ab Bu So: u... 3.x A88 .5 Mum .owo
H5539 S.» Aemuv Sb Sb 3qu 94 8e 3on ab .3 .m3
ace menu; owe Q33 Eb gob ©ro 5.x emu C63 T: .3 boo
Hoe Sewn 8b 3%; we ob A83 Ne 9m A55 Na .3 63
we”. mag: Ge 3qu m; 56 3on Fm Eu A53 Ab .3 .23.
wernmv. mb 3qu mb ob A53 Nb mu finev u: Nmm .omq
mermcomm— :3 3.3V No 3b Annoy No 3.9 ANNE u: #3 .Nmm
murmooenma. :5 A33 Pm z: A53 Nu 0.x Aunev Nm .3 a:
me". $602. In» 3%: Aw 1: A33 a: Se 6&3 Pm .3 .03

 

167

Peliminary Factor Analyses Using Obligation Group Data Only

Principal axis factor analyses. Separate principal axis factor analyses
employing varimax rotation were conducted for items assessing each side of the EPC.
The items that were retained met the following criteria: 1) a largest loading of over
.5, 2) no other loadings over .4, and 3) item content that was consistent with the
content of other items included in the factor. In summary, items that loaded weakly,
cross-loaded, or did not make substantive sense were eliminated.

Using the minimum Eigenvalue of 1.0 criterion, the principal axis factor
analyses of the obligation sub-sample resulted in the specification of four employer
obligation factors that accounted for 53 percent of item variance and four employee
factors that accounted for 45 percent of the item variance. The items which were
identified as indicators of the respective latent factors and their factor loadings are
presented in Table 5 (employee side) and Table 6 (employer side).

CFA. Measurement models for employer and employee sides of the EPC
based on the factor structures indicated in Tables 5 and 6 were assessed using LISREL
VIII CFA. As stated earlier, only data from obligation forms were used in this
analysis. Fit indices for the measurement models are reported in Table 7 . The chi-
square/df ratios and pattern of other fit indices indicate a reasonably good fit between
the models and the data. T-values for all parameters were significant at the p < .05

level. The modification indices suggested further modifications that would modestly

168

Hmd_am....s...m.. "a. r.» .. :L.‘ r, ... ... .....I... ......

 

 

 

 

:9: n :9: v.88: 885m...
_ N w A

02.0 Bingo—.8— oBEo<8 ocmmmnoamw EmE< came—.28 88% 83v
NM wand—.3 m: gammaoa 8c 8&8 8 :5 com: o». <9: 99:3 ..3 .5 .3 .8
no 98$ an me: £01: on mac ..3 -.3 .8 .2
we mo=o£ w: 8:6»3 :58 «be vomaom ..3 .3 Jon -.3
3 wroi cc we: £01: o<o:< am< :5” <0: m8 merge—on. 8 Sen—r mo 83m 8 <0: Ed ..E .8 .8 be

3<mmom=< min 8 no 8
mm E <9: 368%: 58$. m. <0: 33 83233» pace. :5 £m< <0: 26 38m :88.— .a— .8 .3 .3
mun—.58 Aocmmwmoam 8 3&8 «scanner v.38 8 :5 9:33. 889:8 £5. Ea

368558: 8830333

o €01” QB.» :95 «<3 m. m: 88 :9 5875 933 cw< -.oc ..3 .No .. 3
5 Even 8333 mow—m wrong om <9: 9:83— moam on 88o consume—5 .oo .am .8 .NA
5 095:3 <2: coagu— Ew 8 am no: 8 5:: :5 98:0 edema 0». <9: «3232 .3 .am .3 .5
m 7.58 383588 5 <9: 3:83— :mo 35:? 9.3333. moni— uomimomv 8 88» .5 .2 .5 .3

98min 888% 0: <9: ..oc
mm 7.38 38:»— mmoamoom £5: So 8333 mm «88.395 ER 858 -.3 .3 .3 .Na

 

169

and—o m 83..“

 

sauna--33: 825395 8 aamwmo m: warm—$03 ”3.5885. $928 8 c353
8_o 8:33.588 8a :53. ..ocmv

 

N 00:8 cu SE. 503 cm mammammoam 8n 568195 v55. 0828.on 32303 .3 .Nm be .0;
u mag. v55. 8-£o:82 i=0 «8 335 $385. Racism Saw. .8: .5 -.3 ha .NN
_ axe—canon 8 no 95% m" 29.x 9% 98 no» 8335. 89:88 3. «0:13 .oo .3 .3 .3
no book 8.. .82 £35 8 831.850 8 =5 cammaumnoad 288% .3 .3 .8 .9.
:95... nocmmmmoam 8 was. 23. :5 0838303 83535 8838583

0 29 moor w 32 ..oc 25. 2592 «3286». 8 .oam mm v55 2.803 .3 .3 .5 ..3

36.32 mm 385m 5.. 085303 8 we:

S 336 moon 398nm mo“ 523m «55. «3286.. .3 .G .3 ..3
an mun—a w 333:3 o». 28 woman 5 :5 062338: -.oa .3 .8 .am

 

392 _ n mat—38 88

3.92 N n mmonmoo

392 u u :93» 8?:

3.08m A u mag. £3. :8 Ema—swung:

170

ngom ......-...... .. . m... . :..:m ........n... _.,.o.

 

 

 

 

83 k 283 v.88: 2083mm
2 N u .2
008 33:38:8— 83:_0<o_. 028::an EmE< 8:098: 888 .825
an H88 <0: $5: 888" .a— .2 .3 .8
nm 2.83.8 <0: £5. 8 28:8 8 8: <03. man 0m :8 m8Q €28: <0: 8288 .8 .oo .3 .2.
33 <0: .88 c8: :88: ::85< c< <03. 233888 :08 on 882288
mo 2:8 moo: 8808 0a ..moo: 8:8: 2.0:. 83338 <03. 8330383 ..3 .8 .0: .2
3. 5.038 <0: :8 8:8 02:03:38» 38: 0888 £23 <05 09:82:88 3: .3 .oo .8 .3
283.0388 Em8n<
”3 2.838 30819.: :32 332383 80883 8 28.83 <03. 2.0.0 .2 .8 .3 .08
2 5.03.8 «xv—8:08 33 3:8 8:88 <0: £8: 383m 808208 80:» .8 .3 .2 .8
<05 ..0:
5 20838 <0: 0 88 208 2:8 .3 ..3 .02 -.om
a 20838 <0: 8mg: 8288* 80:» :92 <0: :8 283.0338 <03. ..0: .am .20 .00 .3
H388.— ¢<0..r 828883.. 8 @838 £08 :5. 8 3833mm:— 3: arm—882:3
2 20838 <0: £08 33 mm 02828828 .5 .aa .00 .2
A 2.838 <0: £08 33 3.. 38838 .0: .ma .2: .No
S 20838 <0: £08 :3 £58 <0: : 8:8 02. 0:608 .8 ..3 .Na .3
S 20838 <0: £08 83 m2<om <0: : m83 .8:— 0». 88082225 .8 .3 .mo .3

171

.320 a 003.3

 

$0172032<01~ E8 war-:8

 

N 282 <0: 330 on. 3.83 20% 8 203.8388 3 300.30% 00:383.. 82 50 .2 .3 ..3 .8
332.083 8 <0:

3 ma 8088 3 38 50088 3 58.. 8 9:03 <0: 8 c.< 8 38:8 <05 35w .3 .3 .3 .3
33 00808. =<0m

.3 3:8 38 00003: <05 303.301" :8 3035. 88303.33. 300.30? 8.3383 5 .3 .3 .20
92 50 3320033: 8 <0:

3 3 3200033088 <05 35* 88833 8:80» 38: 008320 .3 .3 .3 ..3

Eu: 2.3.

no 20m< <0: w 6033.. 883 00 30 05383303.» 003.0338 .2 .NA .3 ..3

NA 3.038 <0: w 882 0m v~< 88 mm :33: £83 0038.83 8 3:: 300" 032 .3 .3 .3 ..3
033083 50 333m 8030 83.0333 <05 .38 020d

3 038 <0: 8m:_5 0m< 30880 .30 .3 .2 .3

 

3.085 2 u m330<5 008

3.085 N u 9:308; £01”

3085 u u €0~E=08€5w 8838
2.5085 A n 23m: v~<

Table7 1° _ .

'-

172

2.1.

O O
\
- . O

)0 0.1

 

 

Model )8 df NNFI RMSEA GFI CFI
Employer 430.5' 146 .92 .061 .92 .94
obligations
Employee 348.1' 113 .89 .063 .93 .91
obligations

 

’p<.05

173

increase the models' fit, but further modification was rejected based on substantive
considerations.

Figures 7 and 8 present the hypothesized measurement models for the
employer and employee sides of the EPCQ that were derived from the preliminary
factor anlayses procedures outlined above, and which were subsequently assessed
using multiple-group CFA to test hypotheses 1A, 13, 2A, and 23.

Multiple-group CFA

Employer side models. Table 8 shows the results of the multiple-groups
analysis across the three forms of the EPCQ. The overall chi-square value was
significant for each of the four models, which is not surprising given the relatively
large sample size. Only model 4 had a chi-square/df ratio that was greater than 3.5 to
1.

The chi-square/df ratio of 2.56 to l and NNIF value of .90 both indicate that
Model 1 fits reasonably well across the three forms, providing support for the
equivalance of factor patterns across the three forms. It appears that the different
forms of the EPCQ elicit the same conceptual framework.

As a test of the hypothesis that the factor loadings are equal across forms of
the EPCQ, the chi-square of Model 2 was compared with the chi-square of Model 1.
The difference was found to be significant, meaning that the hypotheses that

the factor loadings are equal across forms must be rejected on a statistical basis.
However, the difference in NNIF is relatively small (.01), raising the issue of Model

2's relative fit from a practical standpoint.

174
Figure7M ur ntModlEml ’ i

 

Error Variances Items Latent Factors
Theta Delta
0 _
O -
O _
. _ Employer
. _ Core
0 -
O —
O _
O _
O -
O -
O -
O -
O -
Balance
0 _
O -
O -
High
. Pay

 

175

Figure 8 Mesurement Model Employee’s Side

 

Error Variances Items Latent Factors
Theta Delta
0 _
-

 

O
O i I: ' Employee
Core

0
O——>l:l<
. .
.
W
C

o
O

 

 

 

 

 

176

arc—o m .r .4. .

'-

o
. . .
. .

5

.... I

«2. ,

 

 

Zone x” a». DE. 223 55mm.» Q3 03
was»: wagon IBM. 3m Z> Z> be .3 .3 be
2:83
mag: weapon Scam. 3o 3. 5 .mo .2 .mm .8
2:83 93 535%
we.»— mmaon 32o:— SmPu. Amm 76.x. um .mo .3 .3 .8
59.35 2a and;
mag— mmoSn 25.83 Egg. mum Ebro. wm .mA .8 ..3 .mA
52:5? 233. 25
<3§8Q8<~5~508

 

.nAbm

177

The difference in chi-square between Model 3 and Model 2 was significant,
meaning that the hypothesis of equal errors across forms must be rejected on a
statistical basis. However, there was no difference in NNIF values between the
models, again raising issue of the model’s relative fit from a practical standpoint.

The chi-square of Model 4 was significantly different from Model 3 and there
was a .05 decrease in the NNFI value to .84. Therefore, the hypothesis of equal
covariance among factors across forms must also be rejected. The correlation among
the latent factors were not invariant across the three forms. On the whole, the results
suggest that the model that fits the data best is Model 1, the model which posits
congeneric equivalence.

Employee side models. Table 9 shows the results of the multiple-groups
analysis of the employee side measurement models across the three forms. Again,
the overall chi-square was significant for each of the four models. The chi-square/df
ratio for Model 1 is 2.75 to 1, within the range of ratios that have been viewed as
acceptable in the literature. The .04 RSMEA for Model 1 indicates a close fit
(Brown & Cudek, 1993). However, the NNIF value of .86 is below the .90 value
often used as an indicator of a “good fit” (e. g., Vandenberg & Self, 1993). A review
of studies using LISREL CFA to assess the equivalence of measures found that it is
not unusual for models with indices values somewhat lower than .90 to be judged to
have an adequate, reasonable, or moderate fit (e.g., Peterson & Smith,1995; Rahim

& Magner, 1995; VanDeVijver & Harsveld, 1994). The overall pattern of the fit

 

178

Had—a0 .r. . .

o
.. . o
o . b

 

 

..r t. n. ..u .. .... .1». ...

380— x~ Q. bx” Dam 223 Emma? Q3 03
mag— mmaon 8mg. 39 Z> Z> .mm .3 .3 .8
3.39.:
mag— mmoSn 03.x. ufl Nag. 5 .mm .o... .mm .mm
3:09 25 _om&=mm
mnem— meQ. 3203 33¢. umm 590. “3 .ma .3 .mm .3
535m 3a 333
mag— mwofle 3303 3 Ba. 20 Numb. ”X .3 bu .3 .3
Ema—mam. 2.33. 25
<3§8Q8<m1§8w

 

.bAbm

 

179

indices indicates that Model 1 fits the data fairly well; the fit of model provides
evidence adequate for postulating the same number of factors across forms.

As a test of the hypothesis that the factor loadings are equal across forms of
the EPCQ, the chi-square of Model 2 was compared with the chi-square of Model 1.
The difference was found to be significant, meaning that the hypotheses that the factor
loadings are equal across forms must be rejected on a statistical basis. There was no
change in the NNFI, raising the issue of Model 2's relative fit from a practical
standpoint.

The hypothesis of equal errors across forms must also be rejected from a
statistical standpoint, as the chi-square of Model 3 was significantly different from
Model 2. The difference in chi-square between Model 4 and Model 3 was significant
and there was a .03 decrease in the NNFI value. Therefore, the hypothesis of equal
covariance among factors across forms must also be rejected.

As a whole, the LISREL multiple-groups CFA of the models provides
evidence that obligation, expectation, and promise based forms of the EPCQ loaded
on shared constructs. For both employer side and employee side measurement
models, the model which represents the factor analytic corollary of congeneric
equivalence, Model 1, fit the data best. Model 1 fit the employer side data well.
However, the fit of Model 1 to the employee side data was judged to be only
adequate.

While the LISREL multiple-groups CFA provides evidence that the obligation,

expectation, and promise based forms elicit the same conceptual framework, the

180

findings also suggest that the forms may have differential loadings, specific variance,
or differential reliabilities. Thus, observed scores should be expected to differ across
forms, and firm conclusions about the equivalence of forms in relating to external
criteria cannot be made without additional evidence.

Correlations with external criteria.

An additional analysis of the equivalence of the different EPC forms was
obtained by correlating responses to the different forms with external criteria. As
discussed earlier, ten EPCQ based variables were calculated and included in the
correlational analyses: overall level of response to employer items, overall level of
response to employee items, scores on four employer item sub-scales, and scores on
four employee item sub-scales. All sub-scales were based on the results of the factor
analytic procedures described above.

The evidence supporting a finding of congeneric equivalence among the EPC
forms supports the use of the factor based sub-scales to make comparisons across
forms. The internal consistencies of the EPC based measures were determined by
calculating the coefficient alphas corresponding to each scale by group/ form. These
values, and the scale means and standard deviations, are provided in Table 10.

The results of the analyses investigating the correlations between the EPC
measures and external criteria are presented in Table 11. The most salient differences
and the most salient similarities both involved the same external criterion, trust.
Focusing on employer side EPC measures, of the five correlations between trust and

the five EPC measures that were obtained in the obligation group, four were

181

TablelOtr-r 210:.- .'l:!01‘ 2H .riI-i‘r ; .130 113‘

W
Adjusted
n M SD or a

 

Emnlnxflltzms
Total Item Score (44 items)
Obligation form 522 151.9 29.6 .96 .96
Expectation form 254 156.8 28.1 .95 .95
Promise form 252 126.4 31.5 .95 .95
Sub-Scales
Core (8 items)
Obligation form 534 33.8 5.7 .88 .88
Expectation form 256 39.0 6.2 .88 .88
Promise form 255 25.8 6.7 .84 .84
Enriched Work (4 items)
Obligation form 537 11.1 4.1 .88 .88
Expectation form 260 13.5 3.6 .82 .82
Promise form 256 11.3 4.1 .84 .84
Balance (4 items)
Obligation form 535 12.1 3.6 .74 .74
Expectation form 258 13.5 3.2 .67 .67
Promise 257 12.0 4.4 .82 .82
High Pay (3 items)
Obligation form 536 8.6 3.2 .75 .75
Expectation form 260 9.2 3.0 .61 .61

Promise form 256 7.5 3.0 .49 .50

182
Table 10 con’t

W
Adjusted

n M SD 0: an

Employee Total (33 items)

Obligation form 520 118.7 17.2 .90 .92

Expectation form 249 127.3 19.9 .90 .91

Promise 249 108.7 22.7 .93 .94
Sub-scales

Core (5 items)

Obligation form 535 22.2 2.8 .77 .79

Expectation form 258 22.1 2.6 .72 .75

Promise form 257 20.5 3.5 .75 .77
Sacrifice (5 items)

Obligation form 534 14.1 4.2 .75 .75

Expectation form 260 15.9 3.9 .70 .71

Promise form 257 13.0 4.6 .74 .74
“Extra-role” (4 items)

Obligation form 533 13.7 3.3 .78 .78

Expectation form 260 15.4 2.7 .68 .69

Promise form 257 13.0 3.9 .80 .80
Stay (3 items)

Obligation form 536 7.8 3.1 .67 .67

Expectation form 258 10.4 3 0 .66 .66

Promise form 255 7.3 3.5 .68 .68

 

 

183

 

and 8.
is S.-
83 8.-

23:2.-
23:2.-
35:8.-

8& 8.
$3 :.-
33 8.-

5.5 .2.-
35:2.-
§2 8.

648 8.-
349.: .-
282 3.-

?3 8.
ans 8.
2 m2 ...2.

Ammmv sap
Ahmmvoo-
Aommv no.

#3 8.-
33 8.-
£3 5.-

53 8.
G3 8.
8.6.2 ..3.

82v 8.
53 8.
Ea :2.

RE :2.
A83 2.
52 S.-

83 :2.
GE :5.
A82 8.

away 2.
£48 :2.
52 8.

GE :3.
sea :3.-
scmv :2.

GE ..3:
£5 :3.
an: .8.

€va 8.-
G3 :.-
85:3.-

fa :.-
56 8.-
65:2.-

GE .2.-
fig 8.
:3 8.-

33 .2.-
33:2.-
52 S.-

avu S.-
avmv .3.-
Eé 8.-

Anna .2.-
GE .2.-
$28»..-
cas 8.
and 8.
£2 3.-
3va 8.
:23 :3.
A32 .2.-
Afis S.-
Amms S.-
52 8.-
848 8.-
:3 3.-
22 :2.-

aa: .2.
ca: :.
93 .....N.

A88 8.
cm: 8.
£3 8.

SE :2.
:3 I.
:52 :3.

Ana—V .e_.
a”: 8.
£3 8.-

63 .2.
AS: :.
63 ..2.

35 .2.
ans 2.
A32 8.-

a§ 8.
and 8.
$3 8.-

34% 8.
ans 8.
£2 8.

$3 :.
58 .2.
A32 8.

$3 .2.
GE .3.
52 8.

02:85
couafionxm
83:8
ham
02:85
cowufioqu
83:8
coca—am
02:55
5550on
822830
eczema—m
02:85
cowSooaxm
82330
0.80
83893
02:85
2.388qu
823250

A208 So: 30: .32 =So>o

02m ngo—mEm

 

05:3.

833:8

£8.80

.83

52:23.
beam

 

.... ...-4

...MJ.“

... . ... 22.3

.Amo. v 8 8:20.23 22820830: mam—0:88.200 05 :82 3280588 8.2% 28220.28 :02808: .8 ..20208098: Bomeom ”082

8. v8 : “8.8 .

 

184

 

 

A82 8.- 882 8. 382:8. 882 8- 882 2.. 82 :8. 882 2. 82.88
83 :. 882 8. G82 :8. G82 8.- 882 8.- 822 8. 682 :2. 8858.5
82 .2. A82 8.- 882 :8. 82 8. 82 2.- A82 :8. 682 :2. 8888 .2
5m
:82 8. G82 8. 682 :8. $82 8. $82 8. A82 :8. A82 2. 882.2
9.82 8.- 882 8.- 882 .2. G82 8.- 8.82 8. ca: 8. 83 8. 88888-2
52 8. $82 8.- 82 :2. E2 8. 82 8.- A82 :8. A82 :2. 8888
o—Ounabxm
:82 .8. .82 8.- 882 .2. $82 8. $82 8. A82 .2. A82 :2. 388.2
882:8. 882 .2.- 382 :2. G82 8. 9.82 .2. :22 .2. 83 :8. 8808qu
922.2. 82:2.- 282 :2. :22 ..2. A82 ..2. :82 :8.. A82 :8. 88830
QUE-Sam
:82 8. G82 .2. 882 :8. 882 S. :82 .2.- 82 2. 882 .2. 388.2
882 8. 882 :. 882 :2. 882 8. 882 8.- A8: 8. 83 8. 888qu
82 8. 82:2. :82 :8. A22 8.- 82 :2.- 882 8. 82 ..2. 8888
0.50
mogaam
882 8. 5.2 8. 882 :8. G82 8. A82 8.- 82 :8. 3.2 ..2. 3.882
882:2. A282 8. £82 :8. $2 8. G82 8. 82v 8. :82 :8. 888qu
82:2. 52 8. 82 :8. 82 8. 52 8.- A82 :8. 8:2 :8. 8:38
A088 2:0: 882 85. =80>O
0va 8.0022952
0w< 88:00 825. 08:02. 820033— 8:82.00 8320:0m
83 >85

 

r80 : 032.

185

significantly lower than the corresponding correlation obtained in either the
expectation group or the promise group (i.e., a total of 10 comparisons resulting in 8
significant differences, all in the same direction; p < .05). In sharp contrast, the
correlations between the five employee side EPC measures and trust were remarkably
similar across all three groups. The 10 comparisons revealed no significant differences
across groups.

Work centrality appears to be more strongly related to the obligation based
EPC measures than the expectation or promise based EPC measures. In 18 of 20
comparisons, work centrality’s correlation with the obligation based measure‘is
greater than its correlation with the corresponding expectation or promise based EPC
measure. In 6 of the 18 cases, the difference is statistically significant.

The correlations between education and four of the five employer side EPC
measures (overall level, employer core, enriched work, and high pay) were
significantly different across the obligation and expectation forms. One of the
significant differences involved observed correlations that were in opposite directions.
Education correlated -. 11 (p < .05) with the measure of employer’s obligation to
provide enriched work, and + .20 (p < .01) with the corresponding expectation based
measure, indicating that more educated participants were less likely to view their
employer as obligated to provide enriched work, but more likely to expect their
employer to provide such work. Education consistently correlated in a negative

direction with employer side, obligation based EPC measures.

186

The pattern of correlations for the remaining external criteria, equity
sensitivity, tenure, gender, and age, did not differ markedly across forms. The 20
comparisons made for each of the four variables (10 comparing obligation and
expectation, 10 comparing obligation and promise) revealed a total of five significant
differences between the correlations obtained in the obligation group and the
correlations obtained in the expectation or promise groups (i.e., eighty comparisons
revealing five significant differences). The number of significant differences among
the four criteria is only slightly greater than the number that could be attributed to
random occurrences at the p < .05 level of significance.

Overall, the pattern of correlations with the external criteria were much more
similar among correlations involving employee side EPC measures than among those
correlations involving employer side EPC measures. Except for the comparisons
involving work centrality and one comparison involving age, all of the comparisons
resulting in significant differences involved correlations with employer side EPC
measures.

Hypotheses 5A & 5B

Hypothesis 5A and 5B make predictions regarding the average number of
items included in obligation, expectation, and promise based EPCs. The results of
planned t—test comparisons, reported in Table 12, support hypotheses 5A and 5B. The
average number of items included in obligation based EPC measure was significantly

less than the average number of items in the expectation based measure, and the

187

 

 

 

 

 

 

 

 

Hugo ~N ......w. u ...; 132...... . m .. ..m. .. . .. {..L r ......a . «a... .. ... .
mesa 9333a
mom:
02$. 5... mg. Comm. <m. $33
Ocmmwmoa mxvoofiwaoz 5238 K K
U59. » can... a
<33? K AB MD K as MD K as MD 38 Stan b 88 <95.“ b

 

a mBEowoq
mwomaoam we; Gwd 9a

m mac—860
mwnzoam no.» Gud Pu

So 68v .3

3.. Gas N...

RS Ed 5

nah Awmd m6

3m -w.m .ooo

HQ ha» .95

M; So .95

No Sm .ooo

 

188

average number of items included in obligation based EPC measure was significantly
greater than the average number of items in the promised based measure.
Hypotheses 6A & 6B

To test the hypotheses regarding mean responses to corresponding obligation,
promise, and expectation based measures, responses across forms were compared at
three levels: “overall” response level, the factor or sub-scale level, and the individual
item level. Results of planned t-test comparisons of the relevant group means are
reported in Table 13. They reveal that the expectation group’s mean overall level of
response to both employer side items and employee side items is significantly greater
than the obligation group’s means, providing support for hypothesis 6A. Further, the
obligation group’s mean overall level of response to both employer side items and
employee side items is significantly greater than the promise groups’ means,
providing support for hypotheses 63.

Second, group differences in sub-scale (factor based) scores across forms were
assessed using two multivariate analysis of variance (MANOVAs). The first
MANOVA focused on the employer side; scores on the following sub-scales were
included as the dependent variables: employer core, enriched worked, balance, and
high pay. The second MANOVA focused on the employee side; scores on the
following sub-scales were included as the dependent variables: employee core,
sacrifice, extra role, and stay. MANOVA results indicated significant differences
across forms on employer side sub-scales, E (43,940) = 51.05, p < .000, and

employee side sub-scales, E (32,968) = 8.04, p< .000).

189

 

 

 

 

 

 

 

 

dug—0 Hw . .9...1». I .J ......o. o. o r .. O » OM. q: o 0:.
manna COBB—Ba
men:
Ocmm. <m. @609. Ocmm. <m. $33.
Comma—mo: mxuoognoa $8.38 K K
55.3. H can... a
<53? K as MD K 9v MD K A5 MD 0:00 <wEo n «:8 $55 b

 

mac—360 3mg 3on :.N
:35 88.
maaowon are GNNV nod

:36 8::

5.3. 93 ab

:3 3%: N3

83 A23 NS

53 33 w;

-md -aa .25

Lb -N.N .03

3b ab .25

NM.» 3..» .25

190

Follow up univariate analyses were conducted using planned t-test comparisons
to further explain differences. Table 14 provides the results of the analyses. The
comparison of sub-scale scores on obligation versus expectation forms revealed
significant differences in the predicted direction (expectation > obligation) for six of
the eight sub-scales. The two comparisons that did not conform with predictions were
the comparisons involving the employer core and employee core sub-scales. There
was a significant difference in the employer core mean scores, however, the direction
of the difference was contrary to predictions. No significant difference between
employee core scores was found.

The comparison of sub-scale scores on obligation versus promise forms
revealed significant differences in the predicted direction (obligation > promise) for
six of the eight sub-scales. All four employee sub-scale comparisons conformed with
predicted differences. There were small, non-significant group differences in the sub-
scales assessing the employer’s obligation/promise to provide enriched work and in
the sub-scale assessing the employer’s obligation/promise to provide employees
balance.

The third level of analysis was the individual item level. All items were
included, including those items that were not included in the formation of the sub-
scales. Group differences in responses to individual items were assessed using two
MANOVAs. The first focused on the employer side items, including items 1 through
44 assessing beliefs about what the employer is to give or provide the employee as

dependent variables. The second MANOVA included items 1 through 33 assessing

191.

 

Em - a.

28 «bum - am
cows—8m - an

23 098—qu - t

%& gmfi I a?
353m .. .m

3.83 3:095 .. .N
23 she—gm - .—

 

 

 

 

 

 

 

 

So. can :3 8o. 4.». S- 3 58 S 3 $3 0.2 9n 6an a.» .4
as. Rd 85 08. Sn 5. an 58 92 Em and 4.2 3 55 5.2 .M
so. m2 4: So. an- M2- 3. SE 3: an 808 a: 3. $3 2; ..N
c8. «3 «5 ac. to 2. 3 2.3 now 3 and 3m 3 93 2a ._
ficfiamoomo—afim
25. 8.4 :._ 8o. 2- no- 9m 63 3 on 88v 2 on 93 3 .4
mm». :.o o; 25. mm- 2- 3. fins 92 an 33 «.2 3 G3 _.N_ .m
as. 8.? 2.0- 8o. 3. 3- 3 63 h: an 83 n: 3 53 _.: .N
8o. 3: «3 8o. 5. S 3 $3 MEN 3 63 Sn 3 53 ”.2 L
838m .83—qu
a 31> 35 a 028, 35 an A5 2 a as E an A5 2 9.38..“
a sage H same
2 2 legal 3 Ilflmmfiq
.895 .m> .330 30me .m> EEO
8.8m

 

339.50 £50m

 

a o .o ......o

...... . z 23.

192

beliefs about what the employee is to give or provide his or her employer as
dependent variables. The MANOVA results revealed significant differences across
forms on both employer items, E (3, 1,035) =51, p < .000, and employee side
items, E (3, 1,033) = 21.7, p < .000. Follow up univariate analyses were
conducted using planned t-test comparisons to further explain differences. The results
are presented in Table 15.

Focusing on employer items first, the comparison of individual item scores
on obligation versus expectation forms revealed significant differences in the predicted
direction (expectation > obligation) at the p < .05 level for 15 of the 44 items
assessing what the employer is to give or provide. Of the comparisons that did not
conform with predictions, 8 revealed significant differences in the direction contrary
to predictions. The findings indicated that there were no significant group differences
in mean responses to the remaining 23 employer items.

Comparison of individual item scores on obligation versus promise forms
revealed significant differences in the predicted direction (obligation > promise) at
the p < .05 level for 35 of the 44 employer items. Of the comparisons that did not
conform with predictions, 3 were significantly different in the direction contrary to
predictions, and 6 were non-significant.

Turning to the employee items, the comparison of individual item scores on
obligation versus expectation forms revealed significant differences in the predicted
direction (expectation > obligation) at the p < .05 level for 21 of the 33 items

assessing what the employee is to give or provide his or her employer. All of the 12

 

193

 

 

 

 

Sc. 5.. «a. a3. 2....- S- 2 58 2d 2 .88 a." 2 5.3 8.. 2
.8. :.m on. so. an..- on.-. 2 6va ....m I 88. R...- S 2.3:... 2
8c. 2..- 2.- 8o. 3&- m..- S 6va 3.. o.— .82 on 2 EH. 2." 2
an. «S 2. 2o. 2.". R.- E .33 and a. .82 man 3 5.3 8.. 2
o8. 8.. an. 08. S..- 2.- .1 93 ....N 2 88V 2.. m2 ..3 8a :
m8. .3 R. So. a..- 2.. 2 53 N3 2 .82 3.. 2 EH. .2 2
8o. 22 8.. m... 25 8. 2 22:3 I .88 we. 2 $32.. a
o8. a... 2.. g. 2..- 2.- S 5.3 2d 2 and 8.... S 63 Rd 2.
.8. SN 8. o8. S...- R. 3 Ed 2... 3 .82 on... 2 63 a... .-
8o. ..3 3. 2o. .3- em- 2 ..3 2..” S .82 25 N2 :.an 3.” e
8e. 2.. 2.. E. .3 8. 2 ..3 8a 2 .82 Ea S 63 3.. m
..3. 8.7 2.- 89 an..- 3.. 2 ch. 3." 2 .83 an. S .23 an." 2.
8o. 2... S. «S. 2.7 2.- ... 63 a.” 2 .82 8.. S 5.3 :.m n
2:. :1- 2; 8o. 8..- an. .... :.va ..3 N. 88. 2.. S 53 8“ N
..3. 8.0 S. 80. 3..- $.- 2 5.3 San ... .82 t... 2 $2 .2 2
2:8. ego—58m
a 02.2. 88 a 02? 35 am as 2 an A5 2 dw A8 3 .02

. .525 . .525
2 2 4355' Indian 8:550 88.

.89.. .m> $.50 .80me .m> .wzno

8.5m

 

 

.58qu0 5:0...

 

 

‘I
4

.‘ .. .....-

.
n

o ..4 .-....
I

.45.... .

m H 2an

194

mob.

_ .m.
moo.
ooo.
ooo.
ooo.
moo.
ooo.
ooo.
ooo.
ooo.

mum.
mm...
mum.
moo
ind
:.m-
2.:
mm.”
mm.-
opo—
22m
2.:
mv...
oom—
...-mo
mmd.
Gd
vim
co...
on.
26
3...:
E...
:.m
omo.
mm.m_
no...
and

2..
mm;
8..
Vmo
wv.
3..
8.—

3o.-
mm.
mm.
Va.
o..
m._
..3.

a...
3.
5...

co...
3.
mm.

2..
mo.
hm.
om.
...
3.
mo.

moo.
wvo.

m3.
.8.

ham.

3..
.8.
man.

go.
3m.
vmm.

3....
o3.
moo.
.oo.
moo.
no...
mam.
com.
o3...

mg.
wmfi

bod
3..
mm...
coo
mm...”-
36..
mod
w.m
an..-
:6
mad.
and
8..
.mg
mm...-
:.v
mno-
mm.
mod-
om.m.-
wo.m-
on.
no..-
:.7
osmo-
mo...
wmo-
mm;

3.
mm.
mo.
mm.-
hm.-

om.
:.-
vm.
mo.-
mm.
3.
2.
m..-
on.
no...
so.
am..-
mm.-
on.-
no.
mo...
m..-
ho.-
om.
mo.-
m_.

... 6.3 ....
... 6.3 ....

6.3 3..
6.3 ....
6.3 2..
6.3 6...
6.3 ....
6.3 v...
6.3 2...
6.3 o...
6.3 v...
6.3 ....
6.3 2...
6.3 8..
6.3 E.
6.3 3..
6.3 ....
6.3 v...
6.3 2....
6.3 a...
6.3 v...
6.3 ....
6.3 o...
6.3 ....
6.3 R.
...-3 ....
6.3 v...
6.3 ....

m. 05...-

6.3
6.3
6.3
6.3
663
663
663
663
663
663
663
663
6.3
663
6.3
663
663
663
663
663
663
663
66.3
663
663
6.3
663
663

mad
mmd
«....
vb...”
mud
9..“
mad
3...
on.n
m..v
on.”
.m.v
86
o...
cm...-
on...”
mm.»-
Sim
mom

hm
bad
mo...”
win
ham
Run
3....
Sin
3..”

 

S 6.3 8...
... 6.3 2...

6.3 o...
6.3 2...
6.3 8..
6.3 ....
6.3 8..
6.3 3....
6.3 ....
6.3 ....-
6.3 ....
6.3 .3.
6.3 ....
6.3 N.
6.3 3...
6.3 a...
6.3 ....
6.3 2...
6.3 ....
6.3 ....
6.3 8..
6.3 66..
6.3 ....
6.3 ....
6.3 ....
6.3 6..
6.3 ....
6.3 ....

mv
mw
3.
9.
on
an
hm
on
mm
Vm
mm
mm
.m
on
mm
mm
hm
om
mm
vm
mm
mm
_m
om
a.
m.
C
o.

19S

 

 

 

so. 3; «2 8o. 2+. 2- 2 $3 a: 2 A83 «3 2 33 2a 3
So. m5. m3 ..3. m2. 8.. S 53 one he 83 an... 3 :3 a? mu
o8. o5. ”so a. as 8. a... :3 e3. 3. 83v :3 no 53 m:- mm
«8. 85 «No :5. ohm. :.- 2 5d Sm no 83 8+ .3 5.3 a: 5
8c. 2d :3 E. a: 8. 3 2.3 a: 3 33 one he 63 a? on
8o. «2. Rd 8m. 85 8. 3 :3 an 3 88v «3 mo 68 VS- 2
«8. 2a m3 So. a:- 8- 3 58 8e ho A85 N3 mo 88 a? M:
Sc. 8+ ”No E. «no. 8.- 3 93 NS. 3 A83 Se 3 63 8+ 2
8e. o: «to :5. man- as S Ga 85 mo A83 3v 3 63 8+ 2
So. an 03 So. 8d. 2.- 2 $8 m: 3 88V Re 3 A23 86 2
08. m3 :3 o8. m2.- 8.- Z 33 a: mo and «I. 3 63 a2. 2
so. an «no ooo. 6+. 9..- 1 fins 2:“ 2 A86 m: S 63 8.». 2
8o. 3.” 8.0 Be. own- «a.- S 93 N2 3 83 8.». 2 $3 :3 2
8o. ”2 VS 8o. an. S- S 93 mod 3 83 n3. 3 GE :2“ :
So. :8 «no So. 9%. 8.- Z 33 S.” 3 83 m: 3 63 :.m 2
8m. 8." :.o 8o. 25. K.- 3 Es and S 33 man 2 63 own a
SN. 02 :.o So. 2.? 2.. S €va can 2 88v «2 2 63 RN w
8o. 92 «3 8o. 3a. 8.- I 9va v: 3 83 :.v 3 63 3m 5
vs. 3.. 8- So. 3m- 9.. 3 53 com .1 83v m: 2 GE 92 o
25. and %d to. «.2- 8- mo 5.3 a: 2 ans 2% 2 G3 Ea m
8o. Ea «to to. 37 2.- 2 6va v3 .1 63 25 I $3 an v
8o. «2 and So. 8.? on- 2 :3 n3 3 88V ”3 2 GE 8a M
8o. 3.“ «No ooo. on? 3.- N. as o2“ mo 88V 23 2 G3 8.». N
36. SN- M:- ooo. 3.x- mo.- 2 as :.m 3 808 Sn 2 $3 93 _

was: oozoEEm
me. $._ 03 8o. ”.3- 3- 2 €va ”2 3 33 SM 3 33 :3 S

2 2an

 

196

 

08.
:5.
:6.
So.

ooo.

ooo.

ohm
and
3:0.
om.—
amd
Nmfi
new
Nm.m
and

and
mod
and
N: .o
omd
mod
#00
and
36

who.
ooo.
ooo.
vwh.
ooo.
02..
mm _ .
N8.
000.

mum—-

no.2-

N~ .0-
hm...
0N6.
3.0
Q;-
wvd
ood-

:.-
X”;-
om...
No.-
On...
No.
2 .-
no.
N0.

N4
m4
m;
06
N.—
n.—
v;
A:
N;

83 8+
2.3 8a
53 RN
€va «3
93 mom
53 RM
53 :.m
6va 2+
58 2.”

p28 2 25

no A83
3 $3
2 806
M3 80%
3 83v
3 88v
3 88v
3 88v
2 898

m4»
91r-
mvd
and
mod
5%
mm...»
mm...»
3N

wd
m4
~.~
«.0
0A
o.—
c.—
5.0
_._

€an 3;
53 N3
53 93
£3 Re
53 3m
93 Se
:3 m3
€an m?
GE and

mm
NM
—m
cm
am
mm
5N
0N
mm

 

197

comparisons that did not conform with predictions involved non-significant group
differences; there were no significant difference in the direction contrary to
predictions.

Finally, the comparison of individual item scores on obligation versus promise
forms revealed significant differences in the predicted direction (obligation >
promise) at the p < .05 level for 27 of the 33 items assessing what the employee is to
give or provide his or her employer. There were non-significant group differences
for five items, and a significant differences in the direction contrary to predictions for
one item.

Hypotheses 7 - 12. Alpha coefficient reliability estimates were calculated for
the fulfillment measures that were used in testing Hypotheses 7 -12 and are reported
in Table 16. The results of the correlational analyses involving the global measures
of job satisfaction, trust, and intentions to turnover are also reported in Table 16. A
comparison of the correlations obtained in the obligation group with the
corresponding correlations obtained in the expectation group revealed only one
instance in which there was a significant difference between the matched correlations.
Trust correlated significantly higher with the obligation based enriched work
fulfillment measure (r=38) than with the expectation based enriched work fulfillment
measure (r=22).

A comparison of the correlations obtained in the obligation group with the
corresponding correlations obtained in the promise group also revealed only one

instance in which there was a significant difference between the matched correlations.

Hugo—G ....... "mp-c... .. ... «...... ..... (.5,-v“... .... .. .. . ... P,» L. ....... .

 

 

 

 

 

 

 

 

>6...» 3 go: 3 How H5532.
Damon-”35¢ 00am” momma—ones. mama-women... Ham" 583
mam—553 38550 E MD 3 «Eng n a n a n a n a
:93»? Sewage 3N mnaamv
Ocmmmmos $94 th .3 ho do: So .mm: in .mo: ANN ..3: t;
mango: SN.» NPN .om bu .3: Na .3: MS .3: Now AUX: Bo
5.238 SN; nah be .3 .2: wow .3: So .3: Noa -.3: New
% 3907338 Boa—manna
1 003 Q #953
Ocammmoa Nu...» Mo .mm .mm .mm: 3A ..3: 3o .mm: #3 ..3: 30
wanton-Enos nnd 9m .mu .mu .3: NNm ..3: nma .3: Ba ...B: Be
5038 .mo .3 .3: N: .3: So .3: N3 -.Na: Ba
micron so? AA :35
Commune: 5.; n: .3 .ma .3: 3w .3: Sn .8: sue -.wN:A3
@8885: 53 Nb .3 .3 .mo: mum ..3: nmc .3: 3— Ana: N:
5838 a: we .3 .mm .3: NS ..3: Eu 9:: NB -2.-5..
wagon 3 :35
Comma-no: Eb Wu .3 .mm ..3: .30 .3: 3m .3: 3.. -.om A:
messes :.w 3 .S .3 .3: gm .3: Be .3: N8 «:3:

mono—Ema Hub Wu .3 .3 .AN: NB .3: Sn .3: NM. ...3 NNA

199

Si 6 won—av

Cammmmo: o. ~
mxvoonwmo: mh
3038 m6

Na
Nu
Nu

mg..-
.30.
.8:

Hugo 5 83d

.mo
.mq
.am

.wo: 33
30: N3
.mm: N;

.wq: #3
.3: NNm
.3: NNA

.uA: Amu
.3: N;
.3: NNO

-.uo: Aqw
-.Nm: Nwo
Jug: N8

 

3:362 om :35 u 3
vac—scan o». :35 u N

200

A three item measure of global job satisfaction correlated significantly higher with
the obligation based enriched work fulfillment measure (r=58) than with the promise
based enriched work fulfillment measure (r=37).

The results of the correlational analyses involving the EPC fulfillment
measures and the five job satisfaction sub-scales are reported in Table 17. A
comparison of the correlations obtained in the obligation group with the
corresponding correlations obtained in the expectation group revealed two instances in
which there was a significant difference between the matched correlations. The
correlation between the obligation based core fulfillment measure and social '
satisfaction (r =40) was greater than the corresponding expectation based correlation
(r=23), and the correlation between the obligation based enriched work fulfillment
measure and social satisfaction (r=42) was greater than the corresponding
expectation based correlation (.24).

There were also two instances where correlations obtained in the obligation
group were significantly greater than the corresponding correlations obtained in the
promise group. Those instances involved the enriched work fulfillment measure’s
correlation with social satisfaction (obligation group, r=42; promise group, r= .24)
and with pay satisfaction (obligation group, r=37; promise group, r=21).

In summary, although to the extent that there were significant differences, they
were in the predicted direction, overall, there was marked similarity across forms in
the correlations between the EPC fulfillment measures and the criteria. The results of

the correlational analysis provide very little support for hypotheses 7 through 12.

 

201

.920 3

 

 

 

 

 

 

 

 

... 0. cl’pfl. s r... .. t .4. .. ..Jo. ». .1?
RES 03$? web woos—.5. moomm— ween—Amman
OoaBnmoa momma—one: mwmmmwnaos mwemmwoeoa mmammuawoa mmammwomos
mam—.82: Somme—d a new..." n u n a n a n a n a
..Qer: 308:8 EN :aamv
Oemmwmon be be .3: 3e .3e: 3e .3: 3e .3: 3e .ee: Ame
mxeoameos be be .3: Be .3: Be .ue: 3e .8: Be .em: N—e
5938 be be .em: use .3: Be .3: woe .we: Nee .eb: woe
3307638 3085.2
Geno Q #35
Commune: .me .me .3: 33 ..3: 33 be: 3e.» .8: 33 .ee: 33
mxueogmo: .mu .3 .wm: Mum .ue: NNu .Nq: N318 Rum .mq: mum
3038 .me .we .3: N: bu: 33 .Nm: N: .8: N: .eo: N:
memoran— So? 3 53E
95.3% .mo .mo 2.. SN ..3: .3 .3: SN .3: SN .3: SN
mxboammoa .mu .3 .eu: 3m ..3: Be .3: mum .3: use .3: Se
3238 .me .me .3: Sb .3: Sb .3: Sb .3: 3b be: Eb
wagon A3 ”833
Odmmwmo: .3 .mm .3: aeb .wo: Aeb .uo: B«eb .Ao: Aeb .uw: aeb
968830: .3 .3 .3: Se .3: 5m .3: mum .3: mum .3: Be
323% .3 .3 .3: 5b .5: Sb .3: Eb .3: 3b .3: 3b

202

 

awe—a 3 83d
5% G 3833 .
Oemmmao: .eq .eb be: new .3: aem .Ne: B“em .3. Beem .3: Aem
mxboammoa .ee .eq .3: NM.» .3: MN.“ .8: mug .: N3 .uo: MN.»
3.3838 .ee .ee .3: 3w .eu: 3m .Nb: 3m .3 3m .3: 3m
:bA.o3
.bAbe

208” woamm8e ..axbooSeoa: 9. .6338: 8:65:03 $2.2 «Eamon—Ev: man. :5 gqomuosemam €053.88 .ee—Emacs: 83.0530:

G A .oev.

 

203

Hierarchical regression analysis was conducted to investigate the extent to
which the fulfillment measures combined to explain variance in job satisfaction, trust,
and turnover intentions, respectively, and to test for potential nonlinear aspects of
those relationships. The results of the analyses are summarized in Table 18. A
comparison of the cumulative adjusted R2 3 reveals that the obligation fulfillment
measures explain 7% more of the variance in job satisfaction, 12% more of the
variance in trust in the employer, and 7% more of the variance in turnover intentions
than the expectation based fulfillment measures. Also, the obligation fulfillment
measures explain 5 % more of the variance in job satisfaction than the promise based
fulfillment measures.

In each of the three regressions involving the obligation fulfillment measures,
the entry of the polynomial variables at Step 4 resulted in a significant change in R2 .
For example, in the regression involving turnover intentions, the entry of the square
terms in Step 4 increased R2 .07, from .14 to .21. The significant change in R2 at
Step 4 is interpreted as detecting and capturing the variance attributable to a
curvilinear aspect of the involved relationships (Cohen & Cohen, 1983).

The three regressions involving the promise fulfillment measures also all
resulted in significant changes in R2 at Step 4. In the case of the regression involving
turnover intentions, the increase in R2 at Step 4 was less than ‘A of the magnitude of
the increase that resulted at Step 4 when turnover intentions was regressed on the

obligation fulfillment measures (.03 versus .07). Of the three regressions involving

 

204

 

 

Table18 [.1 I... {‘9‘ no N .201 a. 3.!‘,.ll' ° i'tu‘n
EulfillmenLMeasures
Cumulative
Criterion/ Step 1 Step 2 Step 3 Step 4 Adjusted
Form R2 A R2 A R2 A R2 R2 R2

 

Job Satisfaction

Obligation .29‘ .13‘ .02‘ .05‘ .49‘ .48‘

Expectation .26' .14' .02' .03' .44' .41'

Promise .30' .10’ .02. .04' .45' .43'
Trust

Obligation .29' .05' .05' .02' .40' .39'

Expectation .23‘ .06‘ .02‘ .oo .30‘ .27‘

Promise .34. .03. .01 .02. .40. .3 7‘
Turnover intentions

Obligation .04' .09‘ .01 .07' .21' .19'

Expectation .05' .06' .O3' .01 .16‘ .12'

Promise .06. .09. .02. .03. .21. .17.
Regression Model:

.3199 Xanableumgred

Core fulfillment

Balance fiilfillment, pay fulfillment, enriched work fillfillment

Over-all fulfillment measure

Polynomial variables (a square term for each of the five variables entered in
steps 1, 2, and 3)

AWNv—a

 

' p<.05

205

the expectation fulfillment measures, only the regression of job satisfaction on the

expectation fulfillment measures resulted in a significant increase in R2 Step 4.

206

CHAPTER VII
Discussion

Researchers use the term “psychological contract” to describe a variety of
potentially very different phenomena. Descriptions of “the" psychological contract
vary in the level at which the construct is conceptualized and the antecedents that are
emphasized. Also, while all conceptualizations of the EPC involve some kind of
belief about what an employee or group of employees is to give and receive from his
or her employer, what I have labeled the “belief element”, researchers have
incorporated a number of different kinds of beliefs into their definitions of the EPC
construct (e.g., obligations, expectations, and/or promises; see Figure 3, p. 52).

In Chapters III and IV, I identified and discussed theoretical distinctions that
can be made among conceptualizations of the EPC that vary in the belief element that
is focused on. It was later argued that given the theoretical distinctions that can be
made, it was reasonable to insist that the equivalence of the alternative
conceptualizations be made the subject of empirical investigation, and, it is
unreasonable to continue to simply ignore or assume away potential differences.

The present study acted upon the foregoing argument. The equivalence of
three measures of the EPC that varied in the belief elements that they focused on,
obligations, expectations, or promises, was empirically investigated. Multiple
approaches to assessing equivalence were employed. First, the invariance of the
measures’ factor structure across forms was tested using LISREL multiple group CFA

to examine relations among latent variables. Second, the pattern of observed

 

 

 

207

correlations between different EPC measures and external criteria (e.g. , work
centrality, equity sensitivity) were evaluated. Third, hypotheses regarding differences
in the “scope” of EPCs assessed by the alternative forms were tested by looking at
mean differences across forms at three levels: overall response level, the factor or
sub-scale level, and the individual item level. Finally, the relative extent to which
different EPC fulfillment measures, based on the three alternative conceptualizations
of the EPC, were related to employee job satisfaction, trust, and intentions to
turnover was tested.

In this chapter the findings regarding the equivalence of the obligation,
expectation, and promise based EPC measures are reviewed and discussed. In the
process of investigating the equivalence of the three measure, evidence was collected
which bears on a number of hypotheses that have been repeatedly asserted in the
literature but never empirically investigated. This evidence will be discussed both as
it relates to the primary question of the measures’ equivalence and as it relates to the
specific hypotheses that have been made by others. In addition, limitations of the
study are identified and discussed, and suggestions for future research are provided.
Analysis of Latent Variables

LISREL CFA showed a reasonable fit for a model that postulated equal factor
patterns across forms, or congeneric equivalence, for both employer and employee
sides of the EPC. These findings have theoretical meaning and practical implications

for how the present study was conducted. The findings suggest that obligation,

 

 

208

expectation, and promise based forms elicit the same general conceptual or mental
framework from participants.

The inquiry into the equivalence of the measures, however, cannot stop there.
While the results of CF A provide evidence of one level of equivalence, the results
also provide reason to expect differences in the measures’ observed scores. The
significant change in chi-square from model 1 to model 2 indicates that items load
differently on the factors. It is likely that the measures do not measure the shared
factor pattern to the same degree.

It is possible that differences in observed scores may also result from the less
than perfect fit of the congeneric equivalence model to the data. Applying standards
that have been used in the literature, the model was judged to have a reasonable fit.
While this judgment is sufficient for postulating a model of equal factor patterns, it
does not necessarily rule out all differences in the factor patterns of the latent
variables that might contribute to differences in observed scores. For all equivalence
models, there appears to be little information available regarding the relationship
between the degree of model fit (e.g., “excellent,” “good,” or “adequate”) and relative
differences in observed scores.

In summary, the CFA findings set boundaries on the scope of the equivalence
inquiry in the present study. They clearly indicate that the obligation, expectation, and
promise based EPC measures are not parallel measures in either the strict
measurement sense or in the more liberal factor analytic assessment of parallelism.

The CFA findings also indicated that the EPC measures appear to meet at least the

 

209

minimum threshold requirement for meaningful measurement equivalence, congeneric
equivalence. A practical implication of this finding, one based on theoretical
considerations, is that support was provided for the use of EPC factor-based sub-seale
measures across groups, which would not be justified without evidence of equal factor
patterns across groups. Within those boundaries, the CFA results do not provide a
basis for ruling out or ruling in the possibility that the EPC measures will, in settings
in which we seek to use the measures, “behave” in ways that are sufficiently different
so as to effect research findings, and consequently, our substantive understanding of
the EPC construct. To get a more complete assessment, additional approaches to
investigating equivalence were also employed.
Patterns of Correlations With External Criteria

To further investigate the equivalence of the measures, the patterns of
correlations between EPC measures and external criteria were assessed for similarities
and differences across forms. This analysis includes specific quantitative comparisons
(e.g., assessing the statistical significance of differences between correlations), and
general judgments that are somewhat more qualitative in nature regarding the
significance of the overall patterns of correlations. As indicated wlier, in addition to
providing general evidence regarding the measures’ equivalence, this investigation
provided evidence relevant to a number of specific hypotheses regarding the
antecedents of EPC that have been stated in the literature but, based on my review,
never empirically investigated. I will begin by discussing specific relationships

regarding which hypotheses have been made in the literature. Building on that

 

210

discussion, I will then comment on overall patterns and the evidence that they provide
regarding the equivalence of the EPC measures. Finally, I will briefly discuss the
results of post hoc regression analysis that looked at differences across forms in the
extent to which the seven variables (gender, age, etc.) combined to explained variance
in the EPC variables.

Trust. According to Rousseau (1989), ”underlying a psychological contract
is trust" (p. 128; emphasis included in original text). Trust has been explicitly treated
as an antecedent of psychological contracts (Rousseau, 1989, Figure 1, p. 125;
Nicholson & Johns, 1985). Conclusions regarding the extent to which the present
findings support that claim vary depending on which side (employer or employee) and
which form of the EPC is the focus of attention.

Looking at the measures that assess beliefs about the employer-side EPC and
their correlation with trust in the employer, the findings reveal that the pattern of
correlations obtained with the obligation form is very different from those obtained
with the expectation or promise forms. Of the five correlations between trust and the
five employer-side EPC measures that were obtained in the obligation group, four
were significantly lower than the corresponding correlation obtained in either the
expectation group or the promise group (i.e., a total of 10 comparisons resulting in 8
significant differences, all in the same direction; p < .05). The relevant correlations
from Table 11 are presented below, on the next page. The correlations between trust
and the obligation employer-side EPC measures are relatively small; three of the five

correlations are not statistically significant. Focusing only on the employer side, and

211

looking only at the obligation form, one would conclude that there is little evidence to
support the claim that trust is an antecedent of all psychological contracts. It appears
that whether or not you trust your employer has relatively little to do with what you

believe are your employer’s obligations.

 

Table 19 WW

 

 

 

 

Employer Side EPC Measures
Overall Enriched High
Form level Core work Balance pay
Obligation .09' .13” .08 .06 -.07
Expectation .36" .38" . 19" .27” .10
Promise .36" .40” .10 .29” .17"
' p < .05
” p < .01

Note: Boldfaced “expectation” or “promise” correlations differ significantly from
corresponding boldfaced “obligation" correlation (p,.05).

 

In contrast, there is a significantly stronger association between trust and both
expectation and promise based EPC measures. There would seem to be a rather
straightforward explanation for the correlations between trust in employer and
employees’ probabilistic expectations about what the employer will do or provide.
Generally, the less you trust someone, the less likely you would be to expect them to
provide or do things for you. Even though an employee believes that an employer is

obligated to do something, the level of trust would be expected to affect the

 

212

employee’s assessment of the likelihood that the employer will actually do what it is
obligated to do.

The interpretation of the correlations between trust and the employer-side
promised based EPC measures is more complex. There is an issue of causal
ambiguity. Does greater trust lead to more perceived promises, or do more employer
promises lead to greater trust? Also, it is possible that the correlations between trust
and the promise based EPC measures are spurious in the sense that it may not be an
employer’s making of promises that leads to greater trust, but something associated
with the work environment in which more employer promises are made that ‘
influences trust (e.g., overall better treatment of employees). For purposes of the
present study, the main point to be made is that when looking at the employer-side,
conclusions regarding the relationship between trust and EPC would depend on
whether an obligation versus an expectation or promise based measure was used; the
different EPC forms behave very differently when assessing the relationship between
trust and beliefs about the employer’s side of the EPC.

Turning to the employee side, across all five EPC measures and all three
forms, without exception, greater trust in one’s employer is associated with higher
perceived obligations, expectations, and promises (all 15 correlations are positive and
significant). There is a remarkable similarity across all three groups. The
comparisons between the correlations obtained with the obligation form with those
obtained using the expectation and promise forms revealed no significant differences

across groups (0 out of 10 comparisons). It appears that employees who trust their

213

employer tend to feel more obligation, expect to do more, and perceive themselves as
having promised to do more. These findings provide support for the claim that trust
is an antecedent of EPCs.

Work centrality. The results of the correlational analysis also provide
evidence regarding the long asserted hypothesis that differences in how central work
is in one’s life will influence one’s EPC (Schein, 1965). Support for this hypothesis
varies as a function of the specific sub-scales and EPC forms that are focused on.
Overall, the correlations between work centrality and the EPC measure were higher
for the obligation group than the other groups. Looking at Table 11, the obligation
group's correlation is larger than the corresponding expectation or promise group
correlations in 16 out of 20 comparisons. In six of the comparisons, the obligation
correlation is significantly greater (p < .05) than the corresponding expectation or
promise correlations.

If one were to focus solely on the correlation pattern of the obligation group,
one would conclude that Schein’s hypothesis received substantial support. The
correlations between five of the eight EPC sub-scales were significant, with
correlations ranging from .22 to .31 (p< .01). A focus on the promise group would
provide similar, but significantly weaker support. A focus on the relationships
between work centrality and the EPC measures in the expectation group would,
however, lead to a rejection of Schein’s hypothesis; only one out of eight correlations

with EPC sub-scale measures was significant at p < .05.

 

214

Equity sensitivity. It has been hypothesized that individual differences in
equity sensitivity will directly influence one’s EPC (Rousseau, 1990). The results of
the correlational analysis looking at the relation between equity sensitivity and the
EPC measures provide evidence regarding this hypothesis which varies depending on
which side of the EPC (employer or employee) is the focus of attention, and to some
extent which form of the EPC is the focus of attention.

Again, looking at Table 11, equity sensitivity did not correlate with any of the
five different employer side obligation based EPC measures (correlations ranging
from -.03 to .06); some evidence of a relationship between equity sensitivity and the
expectation and promised based EPC measures is found (2 significant correlations
involving each form). The employer side correlations present a very different picture;
all five of the obligation based EPC measures correlate significantly with equity
sensitivity.

The findings indicate that individual differences in equity sensitivity involve a
focus on the input side of the equity ratio, effecting beliefs about what the employee
is obligated, expected, or promised to provide. This raises questions regarding the
way in which individuals who fall on one end of the equity sensitivity continuum,
“entitleds”, have been described. Entitleds have been described as having high
thresholds for feeling indebted. They are said to accept whatever they get as their due,
and as a result, "they feel little or no obligation to reciprocate." (Huseman et al.,
1987, p. 225). That description is supported by the current findings. The correlations

indicate that entitleds tend to feel less obligated to provide or do things for their

 

 

215

employer. However, more recently entitleds have been described as having a greater
focus on outcomes in the employment relationship than on their own inputs (King et
a1. , 1993). The present findings provide very little support for this description of
entitleds.

Age. The only significant differences across forms in the correlations
between age and the EPC measures involved the employer core sub-scale. Age was
not significantly correlated with beliefs about employer core obligations (r=.03). It
was , however, negatively correlated with employer core expectations (r== -l7, p
< .01) and beliefs about employer core promises ( =-. 13, p< .05) . The findings
suggest that when it comes to the employer core items, increasing age is associated,
albeit modestly, with lower expectations and lower beliefs that promises were made
by their employer; no such association was found regarding beliefs about obligations.

There are two factors regarding which there was strong convergence across
forms indicating relationships between age and EPC beliefs. Age correlates
negatively (significant for all forms, p< .05) with the employer “balance" factor,
which assesses beliefs regarding whether the employer is to be flexible in
accommodating work/non-work life balance issues. Age correlates positively
(significant for all forms, p < .05) with the employee “sacrifice” factor, which assesses
beliefs regarding the employees’ obligations, etc. to be flexible and/or sacrifice in the
interest of the employer.

The findings suggest that younger workers tend to believe that their employer

is more highly obligated, more likely, and promised to a greater extent, to be flexible

 

216

and accommodating of the employee’s non-work life, and at the same time, younger
workers believe that they are less obligated, etc. , to be flexible and accommodating,
at least to the extent of making sacrifices in the interest of the employer. The findings
are consistent with claims that Generation X employees are seeking greater balance in
their work and non-work lives. However, it is not clear that the correlations reflect
the mind set of a new generation. An alternative explanation is that the findings
reflect the relative immaturity or naivety that has always existed among younger
employees. It is possible that when people first enter the work force, they have more
of a sense of entitlement that abates as they get older and acquire more experiences.
These alternative explanations could be investigated further in the present study, but
the cross sectional nature of the data will preclude firm conclusions.

Tenure. It has been hypothesized that organizational tenure has the effect of
expanding the scope of EPCs (Levinson, et al, 1962; Rousseau, 1989). If this
hypothesis is true, one would expect to find evidence of a positive relationship
between tenure and at least some of the non-core employer-side and employee-side
factors. Across all three forms, the findings do not support the hypothesis. To the
extent that there are significant correlations between tenure and the EPC measures (6
of 30 correlations, p < .05), they are, with only one exception, in a negative direction.

The findings suggest that the general hypothesis that tenure leads to an
expansion of the scope of EPCs is, at a minimum, too broad. It may be that in
certain work situations (e.g. , “dead-end” jobs), rather than leading to increased

obligations, increased tenure is associated with a kind of psychological withdrawal,

 

217
which leads to a reduction in the scope of the EPCs. The hypothesis that increased

tenure will lead to an expanded scope of the EPC appears to be based on an implicit
assumption that things generally go well for the employee over time, or if they do
not, the employee will exit the organization - voluntarily or involuntarily. The
present findings and anecdotal evidence obtained in conversations with blue collar
employees and managers during the course of the data collection process suggest that
the assumption often does not hold. There may be constraints operating on the
employee (e. g. , lack of better opportunities) and/ or the employer (e.g. , tight labor
market, fear of litigation) that keep employment relationships going for years even
though the employer or the employee, or both, do not feel that the relationship is
going well. There is little or no reason to expect that under such circumstances
increased tenure would necessarily have the effect of expanding the scope of EPCs.
Education. A comparison of correlations between education level and the five
employer-side EPC measures that were obtained using the obligation form with those
obtained using the expectation form reveals significantly different correlations for
three of the five relationships (p. < 05). The correlations between education level
and belief in employer’s obligation to provide meaningful work provide a dramatic
example of how our conclusions may vary depending on the particular belief element
that is used in assessing EPCs. The findings indicate that more educated participants
were significantly less likely to believe that their employer was obligated to provide

them interesting, meaningful work (r=-.11, p < .05), but were significantly more

218

likely to expect their employer to provide such work (r= .20, p < .01). Each
correlation is significant, but in opposite directions.

The positive correlation between education and the expectation of more
meaningful work would seem to be explained by the fact that more educated people,
on average, tend to occupy the kind of jobs that are commonly viewed as being more
challenging and meaningful. Thus, they are more likely to expect their employer to
provide them meaningful work.

The interpretation of the negative correlation between education level and the
belief in employers’ obligation to provide meaningful work is not as straightforward.
With the exception of the employee sacrifice sub-scale score, education has a
consistent negative effect on beliefs about obligations (9 out 10 correlations, with 4
correlations significant at p < .05). More educated participants were generally less
likely to view their employer as being obligated to do or provide them things. It is
possible that as a result of their education, and/or associated abilities, more educated
participants have a clearer, stronger understanding of the term “obligation” than less
educated participants, and as a result of this understanding, were less likely to respond
that they or their employers were “obligated. "

It is also possible that as a result of their education or associated factors, such
as higher socioeconomic status, more educated participants have a different view of
the extent to which their employment relations involve mutual obligations. In other
words, even assuming that they had the same understanding of what was meant by the

term “obligation” as less educated participants, their personal view or philosophy

219

regarding the employment relationship may be different. For example, more educated
participants may place a greater emphasis on independence or self-reliance in the
employment relations, whereas less educated participants may view the employment
relationship in a more paternalistic fashion.

Comparison of overall patterns. Looking at the overall pattern of
correlations, across all seven criteria and across the three forms, there are several
summary observations that can be made. First, generally, the obligation group’s
pattern of correlations is more similar to the promise group’s pattern than to the
expectation group’s pattern. Second, there is greater similarity across groups among
the correlations with employee side EPC measures than among the correlations
involving employer side EPC measure. Third, there are three “areas” of the
correlation matrix (Table 11) where significant differences in correlations across
groups appear to be more the rule rather than the exception: a) correlations with
work centrality, b) correlations with trust - employer side EPC measures only , and
c) correlations involving education - employer side EPC measures only. Fourth,
significant differences across forms in the correlations involving equity sensitivity,
tenure, gender, and age appear to be the exception rather than the rule. They occur at
a rate that could be substantially explained by chance occurrences at the p< .05 level
of significance.

In summary, there is considerable similarity across groups in the pattern of
correlations between the EPC measure and seven external criteria. However,

correlations involving three of the external criteria, trust, work centrality, and

 

220

education, showed relatively consistent, significant differences from the
corresponding expectation group or promise group correlation. The evidence from
the analysis of correlational patterns indicates that the obligation, expectation, and
promise based measures cannot be considered fully interchangeable. Depending on the
specific relationship that is being investigated, the use of different EPC measures may
lead to different substantive conclusions regarding the EPC construct and the nature of
its nomological network.

Post hoc regression analysis. Hierarchical regression analyses were
conducted to investigate differences across forms in the extent to which the seven
variables (gender, age, etc.) combine to explain variance in the EPC variables and
differences in the pattern of incremental explained variance. A summary of the
results is presented Table 20 and Table 21. Separate regressions were run for each
form (obligation, expectation, promise) of each of the eight EPC factors (four
employer-side factors and four employee-side factors). The order of entry of the
variables is identified in the tables.

Differences across forms in the cumulative variance that is explained and in
incremental variance that is explained at different steps of the regression models
provide some additional support for the general proposition that when investigating
“the” EPC, findings may vary depending on whether obligations, expectations, or
promises are the focus of attention. However, for two reasons, these findings provide
little guidance regarding which specific belief should be made the focus of attention in

defining “the” EPC construct. First, while there is a strong consensus that the

221

 

 

 

 

Table20 51' '0‘; 0 mo ’ r not
Step 1 Step 2 Step 3 Step 4
EPC Factor/ Cum. Cum. Cum. Adj.
Form R2 A R2 1?.2 A R2 1?.2 A R2 R2 R2
Core factor
Obligation .02' .Ol .03' .00 .03' .02' .05' .04
Expectation .04' .01 .05. .01 .06' .10' .16. .13
Promise .02 .04' .06' .01 .07' .13’ .19' .17
Enriched work
Obligation .01 .06' .06' .01 .07’ .01 .08' .06
Expectation .03' .Ol .03 .04' .07' .03' .10' .07
Promise .00 .03' .03 .04' .08' .01 .09' .06
Work/non-wk. balance '
Obligation .05' .00 .05' .00 .05' .00 .06' .04
Expectation .03' .00 .04 .01 .05 .06' .1 1° .08
Promise .04' .02 .06. .00 .07' .07' .14. .11
High pay .
Obligation .04 .05‘ .09‘ .11‘ .20’ .oo .21’ .19
Expectation .02 .01 .03 .01 .04 .00 .05 .01
Promise .00 .06' .06' .02 .08' .01 .09' .06
Regression Model:
.5111: MW
1 gender, age
2 work centrality, equity sensitivity
3 education, tenure
4 trust in employer

 

‘ p<.05

222

‘H {‘1

 

 

 

 

TabICZI il' "'l 0 ,, "'0 " 1'10 \‘u
Step 1 Step 2 Step 3 Step 4
EPC Factor/ Cum. Cum. Cum. Adj.
Form R2 A R2 R2 A R2 R2 A R2 R2 R2
Core factor
Obligation .03' .02' .05' .01 .OS' .03' .08' .07
Expectation .Ol .00 .01 .00 .02 .04° .05 .02
Promise .04’ .03' .06' .01 .07' .05' .11' .09
Extra role behaviors
Obligation .Ol .08' .08' .00 .09' .00 .09' .08
Expectation .01 .00 .01 .00 .01 .01 .02 .00
Promise .00 .04' .04' .01 .05 .05' .09' .07
Sacrifice
Obligation .07' .12' .18' .03' .21' .01 .23' .21
Expectation .05' .07' .12° .02 .14' .01 .16' .13
Promise .09‘ .04’ .13’ .02 .15‘ .03‘ .18‘ .15
Commitment
Obligation .01 .10’ .ll' .01 .12' .02' .14' .13
Expectation .01 .03' .04 .00 .04 .03' .06' .03
Promise .01 .03' .04 .00 .05 .04' .08' .05
Regression Model:
.Stsp Xanahlmntered
1 gender, age
2 work centrality, equity sensitivity
3 education, tenure
4 trust in employer

 

’ p<.05

223

relationships between EPC fulfillment and job satisfaction, trust, and turnover
intentions are important links in the EPC construct’s nomological

network, no such strong consensus exists regarding the relationships that were
investigated in the regression analyses described immediately above. Second, the
results reveal that a great deal of the variance in the EPC variables is left unexplained.
Only two of the 24 cumulative R’s were greater than .20 (both involved the
obligation form). This is not surprising given the number of factors (individual,
organizational, and societal) that might be expected to influence employee beliefs
which are not included in the analyses.
Hierarchical Nature of Obligations, Expectations and Promises.

The hypotheses regarding the hierarchical nature of obligations, expectations
and promises received substantial support. The predicted differences in the average
number of items included in participants’ EPCs across forms were found. The
promise form average was significantly lower than the obligation form average, and
the expectation form average was significantly higher than the obligation form
average.

Predictions regarding hierarchical differences in mean responses across forms
were investigated at three levels: overall response level, sub-scale level, and
individual item level. Generally, mean responses to the promise form were
significantly less than the corresponding obligation form responses, and mean
responses to the expectation form were significantly greater than the corresponding

obligation form responses. The hypothesized general hierarchical nature was not

 

 

224

found when comparing the obligation and expectation forms of employer core factor
and the employee core factor. For these two factors, the mean of the expectation form
was lower than the obligation form meaning.

The fact that the general hierarchical nature of responses did not hold up in
this area of comparison might be explained by the existence of a kind of ceiling effect
and the role of trust in moderating the relationship between beliefs in employer
obligations and the corresponding expectations regarding what the employer is likely
to give or provide. There are instances where although an employee may believe that
an employer is obligated to do something, the lack of trust in an employer’s
willingness or ability to do the thing would result in a lower response to the
expectation item than to the corresponding obligation item. While I expect that this
effect would occur among some participants across the range of EPC items, it is most
likely to manifest itself with items that have very high average responses to the
obligation form of the item (e. g. , the items making up the two core factors in
question). In regard to those items, the “ceiling” imposed by the scale means that
among participants that trust their employer, there is little room for the predicted
general hierarchical nature (i.e. , expectation > obligation) to be expressed.
However, among those who do not trust their employer, there is the possibility of
responses to the expectation form that are much lower than the obligation form’s high
average, which could cause the expectation form average to be less than the
corresponding obligation form average (depending on the level of distrust among

participants).

 

 

 

 

225

An obvious implication of the findings regarding mean differences across
forms is that if one seeks to understand EPCs through the interpretation of mean item
or factor scores, different conclusions may be drawn depending on the form that is
employed. I would note that the evidence of mean differences across the EPC forms
is not necessarily inconsistent with the LISREL multiple-group CFA findings
supporting congeneric equivalence. Congeneric equivalence requires only equal
factor patterns; scores may vary by an additive or constant.

EPC Fulfillment and Job Satisfaction, Trust, and Turnover Intentions

Obligations versus expectations. The present study is the first to directly
test the claim that one should focus on obligations instead of “mere expectations”
because unmet obligations result in a stronger negative response than unmet
expectations (Rousseau, 1989). This claim did not receive significant support from the
findings of the correlational analyses that were performed to test Hypotheses 7, 8, and
9. There were few significant differences between the correlations obtained using the
obligation fulfillment measures and the corresponding expectation fulfillment
measures.

However, hierarchical regression analyses which investigated the combined
effects the fulfillment measures on job satisfaction, trust, and turnover intentions
detected a nonlinear aspect of the relationships between fulfillment and the criteria
when fulfillment was measured using either the obligation or promise based forms.
This finding indicates that simple correlational analysis is inadequate to assess the

relationships between fulfillment and the criteria, and that the comparison of the R2 s

226

yielded by the hierarchical regression procedure recommended by Cohen and Cohen
(1983) to detect and represent nonlinear relationships would provide a better test of
the relative magnitude of the relationships across forms.

A comparison of the relevant adjusted cumulative R72 3 (reported in Table 18 )
reveals that the obligation fulfillment measures explain 7% more of the variance in
job satisfaction, 12% more of the variance in trust in the employer, and 7% more of
the variance in turnover intentions than the expectation based fulfillment measures.
These findings support the claim that unmet obligations result in a stronger negative
response than unmet expectations, at least when the comparison involves prObabilistic
expectations. The findings also support the essence of the predictions made in
Hypotheses 7, 8, and 9 (i.e., that obligation based fulfillment measures would be
better predictors of job satisfaction, trust, and turnover intentions than expectation
based fulfillment measures).

The results of the hierarchical regression analysis further indicate that in
addition to a difference in the magnitude of the variance explained, there is also a
difference in the nature of the relationships between fulfillment and the criteria across
obligation and expectation forms. Each of the three regressions conducted using the
obligation fulfillment measures yielded a significant change in R2 when the
polynomial terms were entered into the model at Step 4, indicating that each of the
relationships that was being investigated involved a nonlinear aspect (i.e.,
curvilinearity). The slope of the regression line “tails off" or decreases at higher

levels of obligation fulfillment, suggesting that obligation fulfillment has less

227

predictive ability at higher levels of fulfillment (where the employer is perceived as
exceeding obligations) than at lower levels of fulfillment (where the employer is
perceived as failing to meet its obligations). In contrast, of the three regressions
involving the expectation fulfillment measures, only the regression of job satisfaction
on the expectation fulfillment measures provided evidence of curvilinearity.

Obligation versus promise. There were few significant differences between
the correlations between fulfillment and the three criteria (job satisfaction, trust, and
turnover intentions) that were obtained using the obligation fulfillment measures and
the correlations obtained using the corresponding promise fulfillment measures. The
the correlational analyses did not provide support for Hypotheses 10, 11, and 12.
Again, however, hierarchical regression analyses which investigated the combined
effects the fulfillment measures on job satisfaction, trust, and turnover intentions
detected a nonlinear aspect of the relationships between fulfillment and the criteria
when fulfillment was measured using either the obligation or promise based forms.
Under these circumstances, the hierarchical regression procedure recommended by
Cohen and Cohen (1983) provides a better test of the relative magnitude of the
relationships across forms than the use of simple correlations.

A comparison of the relevant adjusted cumulative R2 s reveals that the
obligation fulfillment measures explain 5% more of the variance in job satisfaction,
2% more of the variance in trust in the employer, and 2% more of the variance in
turnover intentions than the promise based fulfillment measures (note: the

“unadjusted” cumulative R2 3 resulting from the regressions involving trust and

228

turnover were essentially equal across obligation and promise forms; see in Table 18).
The findings support the essence of the prediction made in Hypothesis 10 (i.e., that
obligation based fulfillment measures would be better predictors of job satisfaction
than expectation based fulfillment measures).

Summary of Evidence Regarding Equivalence and Implications

The purpose of this study was to obtain empirical evidence that would inform
theoretical thinking regarding the EPC construct. In this section, the extent to which
the present findings support, are consistent with, or contradict theoretical arguments
that I have made are discussed, and the implications of the present findings for future
theorizing regarding the EPC construct are commented on.

Support for the general proposition. While there are similarities in how the
three measures behaved, on balance, the present findings provide support for the
general proposition that obligation, expectation, and promise based EPC measures
may, depending on the measurement purposes, behave in significantly different ways.
The LISREL multiple group CFA results are consistent with that proposition.
Although the CFA results provide evidence of one kind of equivalence, congeneric
equivalence (equal factor patterns), they also refute another kind of equivalence,
parallel equivalence”. Thus, while the CFA findings indicate that the three EPC

forms elicit the same general mental framework, they also provide reason to expect

 

12

A finding of parallel equivalence would have contradicted the general proposition that
obligation, expectation and promised based measures may, depending on the measure-
ment purpose, behave in different ways.

 

229

differences in observed scores. Will these differences make a difference in our
understanding of psychological contracting? Unfortunately, although the CFA
findings indicate that there will be observed differences, the multiple group CFA does
not provide a clear basis for predicting the magnitude or practical significance of
differences.

The investigation of the correlations between the different forms of the EPC
measure and external criteria provides support for the proposition that the different
belief elements may behave in significantly different ways. That investigation revealed
that while there is similarity in the pattern of correlations across forms, when
investigating some relationships that the literature has demonstrated an interest in
(e.g. , work centrality and EPC beliefs; trust and EPC beliefs), the belief element that
was focused on made a significant difference in the findings and the substantive
conclusions that could be logically made based on the findings. The findings indicate
that although the forms appear to be equivalent at the level of congeneric equivalence,
for at least some purposes, there “are differences that make a difference.”

The results showing clear, systematic mean differences across the EPCQ forms
at three levels of item analysis are at least consistent with the general proposition that
the forms may behave differently. Morever, these results may, depending on the use
of the data, constitute affirmative evidence of relevant non-equivalence (e. g. , if one
were attempting to interpret means across studies using different forms, or
investigating whether certain kinds of obligations fall within the scope of given

employees’ EPCs where scope is operationalized as in the present study).

230

Finally, the differences across forms in the magnitude of the variance in job
satisfaction, trust, and turnover intentions that was explained by the EPC fulfillment
measure and differences in the nature of the underlying relationship (linear versus
nonlinear) further support the proposition that the forms may behave differently. The
differences are such that when investigating the effect of EPC fulfillment on trust in
employer and turnover intentions, analyses that are appropriate when using one form,
expectation fulfillment, would not be adequate when assessing the effect of obligation
or promise fulfillment (which require analyses that capture nonlinear aspects of the
relationships). Again, differences that make a difference.

There are theoretical reasons and, now, empirical evidence which indicate that

measures of the EPC that incorporate different belief elements should not be
simply treated as interchangeable measures. The clear implication is that in both
theorizing about the EPC construct and in measuring it, greater attention should be
given to the nature of the belief element that will be the focus of attention. Including
obligations, expectations and promises in one definition risks the inappropriate
aggregation of variables that, we have reason to believe, may relate differently to
other variables of interest. Accordingly, the use of definitions that generally define
EPCs as beliefs about the terms of exchange between employer and employee should
be abandoned in favor of definitions that more carefully specify the nature of the
belief element.

Support for a focus on obligations: Reasoning and empirical evidence.

There is room for obligations, expectations, and promises, as well as other employee

 

231

beliefs, in a comprehensive theory of psychological contracting in organizations. In
fact, in Chapter III, I argued that a comprehensive theory of psychological contracting
should address the relationships among those clearly related beliefs. It might be
argued that all three befiefs should be measured when assessing EPCs. While the use
of combined, multiple measures might be expected to explain more variance in
criteria of interest, there are significant potential tradeoffs. The incremental increase
in explained variance may not warrant the loss of degrees of freedom. There is also
reason to be concerned about possible confusion and order effects. Further, there are
practical limitations on the number of items that could be included in multiple sets
containing the same item stems without creating unreasonable risk of fatigue (or
boredom). Thus, while all three beliefs should be given consideration in a
comprehensive theory of psychological contracting, researchers still nwd to address
the question “which of these beliefs should be incorporated in the focal construct, the
EPC variable itself?” The answer to this question should be based on well reasoned
arguments that are consistent with existing empirical evidence.

I have conceptualized the EPC construct as involving beliefs about obligations.
The reasoning behind my focus on obligations instead of expectations was discussed at
length in Chapters 11 and III, and may be summarized as follows. First and foremost,
a focus on obligations instead of expectations distinguishes the EPC construct from
the met expectations literature, enhancing the construct’s discriminant validity. It was
the “ inherent obligatory quality" of the phenomenon that Levinson et al. (1962, p.

36) observed in their research which caused them to distinguished the phenomenon

232

from the then already extensive literature on expectations and lead to their use of the
term “psychological contract.”

Second, as others have argued (e.g., Rousseau, 1989), it was reasoned that
unmet obligations are likely to result in stronger negative reactions among employees
than unmet expectations.

Third, it was argued that defining the EPC as beliefs about obligations rather
than expectations is likely to sharpen our thinking about the nature of “expectations”
(the terms many meanings), reducing the risk of confusing normative and probabilistic
expectations. For example, by identifying obligations and keeping them distinct from
probabilistic expectations, we will be in a better position to evaluate factors such as
trust, that may relate differently to normative and probabilistic expectations.

Each of these three arguments is at least to some extent supported by the
findings of the present study, reviewed above; none of the arguments are refuted by
the findings. I have also argued that, drawing on the literature regarding the effect of
commitment in the goal setting process, there is a basis for predicting that a focus on
obligations may have a greater potential for predicting behavior than a focus on
expectations. The present findings neither support nor contradict the argument.

Given the historical view of the obligatory quality of the beliefs that make up
the EPC (Levinson et al., 1962; Schein, 1965), the manifest greater potential for the
increased discriminant validity (when contrasted with the met expectations literature)

associated with a focus on obligations instead of expectations, and the recent

233

dominance of Rousseau and colleagues’ work in the PC literature which rejects
expectations as the belief element in the EPC construct and measures beliefs about
obligations, it is likely that the foregoing arguments will fall upon receptive ears. As
a result, they will create little controversy, and quite likely, little discussion. In
contrast, the next matter to be discussed has already created some controversy (as a
result of a conference paper, Roehling, 1996).

In conceptualizing the EPC as involving beliefs about obligations, I have
explicitly rejected Rousseau’s (1989) promise framework. The reasoning behind that
rejection was also discussed at length in Chapters 11 and III, and will now be revisited
in light of the present study’s findings. A foundation for that discussion will be laid
by recalling several factual observations.

First, prior to Rousseau (1989), no one writing in the PC literature purported
to limit the beliefs that they focused on, whether expectations, promises or other
beliefs, to those that are based on promises or perceived promises. On the contrary,
researchers emphasized the role of motives, needs, norms, and societal influences in
addition to the parties’ promises, representation and other interactions as antecedents.

Second, Rousseau’s (1989) application of a promise framework to the PC
construct involved a clear adaptation of traditional contract law. This observation is
based on the fact that traditional contract law terminology was used (“offer,"

“acceptance,” “consideration”; pp. 125-126), citations to sources discussing contract

234

law were provided (e.g. , Famsworth, 198), and there was an absence of any other
explanation of where the promise framework came from or why it was being used.

Third, the one specific argument that has been offered in support of focusing
on promises (or promise based obligations) is that unmet promises will result in
stronger negative employee reactions.

The fundamental reason for rejecting the promise framework is that it focuses
too narrowly on only one kind of obligation that is relevant in exchange or
“contracting” behavior, promissory obligations (Macneil, 1980). This is the main
point of Macneil’s work; it is also why he introduced the concept of relational
contracting in Macneil (1972) and later observed that our knowledge of relational
contracting “is hindered by the immense intellectual barriers we put in the way of its
acquisition, WWW" (Macneil, 1985, p.
525, emphasis added). For a specific example of the kind of employee held
subjective belief in an obligation that would fall outside of the narrow promise
framework, but which represents the kind of phenomenon that, I would argue, most
researchers are interested in capturing as part of the EPC construct, an anecdote that
describes an actual occurrence has been attached as Appendix “B. "

The present study provides two kinds of empirical evidence that is relevant to
the issue of the usefulness of a promise focus. First, the study’s findings regarding
the scope of obligation based EPC measure is greater than the scope of the promise

based EPC measure providing empirical support for the argument that employer-

235

employee exchange obligations are not based solely on perceived promises. Second,
the only specific argument that has been offered in support of focusing on promises
(i.e. , that unmet promises would result in a stronger negative reaction than other
unmet beliefs) is contradicted, at least to some degree, by the findings that the
obligation fulfillment measure explained 5 % more of the variance in job satisfaction,
and slightly more of the variance in trust and turnover intentions, than the promise
fulfillment measure.

Is there a case to be made for the narrower promise focus in conceptualizing
the focal EPC construct? If there is, it has not yet been made. The focus on promise
has its roots in a “literature," traditional contract law, that is itself increasingly
recognizing the limitations of a promise framework and is moving toward greater
recognition of non-promissory obligations in contractual relations. An example is the
duty of good faith and fair dealing imposed on contracting parties under the
provisions of the Uniform Commercial Code. The relational-nansactional/discrete
continuum of contracting behavior that has been widely accepted in the PC literature
as a useful tool for understanding psychological contracting was created by Ian
Macneil as part of a program of research aimed at convincing others of the limitations
of a promise focus. To apply the continuum to a promise based conceptualization of
EPC arguably involves a gross distortion of Macneil’s work. Finally, the findings of
the only study to directly assess the assertion that unmet promises will result in

stronger negative reactions than unmet obligations (i.e. , the present study) did not

236

support the assertion. On the contrary, the findings provide some evidence refuting
the assertion.

Conclusion based on reasoning consistent with existing empirical evidence.
Expectations and perceived promises have roles to play in a theory of psychological
contracting in organizations. However, theory based reasoning that is consistent with
existing empirical evidence leads to the conclusion that a focus on beliefs about
obligations in the definition of the focal EPC construct is likely to maximize the
construct’s potential for making a unique contribution to the understanding of
behavior in organizations. l
Limitations

This study has a number of limitations. The following two limitations are
associated with the scope of the proposed study. First, the questionnaire data was
cross-sectional and suggests nothing about how EPCs change over time. Second, the
correlational analysis used in testing a number of the hypotheses cannot be interpreted
as implying causality. While the foregoing are significant limitations of the study,
they are consistent with the purposes of the study: the study does not set out to assess
the dynamic characteristics of EPCs or establish causal relationships between EPCs
and other variables.

As explained in detail in Chapter V, the survey that was used to assess
employees’ PCs does not purport to assess every belief about obligations that might

possibly be found in EPCs. This limits the generalizability of the conclusions that can

 

237

be made about the factor structure of the EPC based on the present studies findings. It
is possible that specific items would load somewhat differently if a more inclusive or
less inclusive item set was used.

Another limitation of the study is the potential influence of mono-method bias.
The variables are measured using self report data obtained via questionnaires. The
magnitude of the risk posed by the use of self report data must be evaluated in light of
the specific hypotheses to be investigated and the inferences I have sought to make
based on the study's findings. In testing hypotheses 3A, 3B, 4A, 4B, 7, 8, 9, 10, 11,
and 12, the potential effect of mono-method bias is, in a sense, controlled. The main
purpose in investigating these hypotheses is the comparison of relative magnitude of
the correlations between obligation, expectation, and promise based EPC measures,
respectively, and third variables. The investigation of specific substantive
relationships (e.g., between EPC fulfillment and job satisfaction) is subordinate to this
purpose. While mono-method bias may operate to inflate the correlations between
EPC measures and the third variables, I do not have any reason to suspect that mono-
method bias would be more of a problem with the obligation based measure versus
either the expectation based or promise based measures. Thus, when comparisons of
the relative magnitude of the correlations between different forms of the EPC
measures and third variables were made, mono-method bias is essentially held

constant (mono operation bias remains an issue, and will be briefly discussed).

 

238

To some extent, the threat of mono-method bias is likely to be an inherent
limitation on some aspects of EPC research. Measuring EPCs involves measuring
elements of an individual's belief system, and individual's beliefs are best measured
directly using self report measures (Ilgen & Hollenbeck, 1992). There are other
variables involving individual's subjective beliefs (e. g. job satisfaction) whose
relationship to EPC or EPC fulfillment/violation is of theoretical and practical
interest. These variables are also best measured directly by self report measures.
Thus, in investigating some relationships, the risk of mono-method bias is inherent if
the best method of measuring each of the respective variables is to be employed. To
attempt to reduce the amount of common method variance, the questionnaire was
designed to measure the various dependent variables and independent variables in
different sections of the questionnaire, measuring similar constructs as far apart as
possible. These efforts are an attempt to mitigate the threat of mono method bias and
are obviously not a solution to the problem.

Mono operation bias is also a potential limitation of the study. It is possible
that the results of the present study were influenced by the particular
operationalization into which the three belief elements were incorporated.

Summary of contributions

This dissertation study, with the limitations indicated, contributes to the

literature in several ways. The contributions result from both the critical review and

analysis of the PC literature that was the foundation for the study, and from the

239

empirical evidence provided by the study. The areas of contribution, which have been
discussed to varying degrees, may be summarized as follows.

First, and perhaps foremost, through its critical review and the issues chosen
for investigation in the study, the need for the PC literature to focus greater attention
on construct validity issues has been highlighted and brought to the fore.

Second, the investigation of equivalence of the EPC measures provides
empirical evidence regarding similarities and differences associated with obligation,
expectation, and promised based measures. Based on my review, no previous study
has directly compared alternative measures of the EPC. The finding that despite
considerable similarities, the measures may behave in ways that can influence
findings should further alert researchers to the general need to attend to construct
development issues.

Third, the process of assessing the equivalence of the EPC measures resulted
in empirical evidence regarding the antecedents of the beliefs that constitute the
EPC. This includes evidence relevant to hypotheses that have been asserted in the
literature but never empirically assessed. The hypothesized antecedents that were
investigated included a) work centrality, b) equity sensitivity, and c) organizational
tenure.

Fourth, the study is, to my knowledge, the first to directly test the assertion
made by Rousseau (1989) and others, including myself, that an employer’s failure to

meet perceived obligations is likely to result in a stronger negative employee response

240

than the employer’s failure to meet expectations. Support for this assertion was found,
but only after the curvilinear aspects of the relationships were investigated and taken
into account. Previous efforts to assess curvilinearity in the relationship between EPC
fulfillment and criteria are not reported in the literature.

Fifth, the present study, and similar studies that combine the examination of
latent variables using CFA and analysis of correlations with external criteria (e.g.,
Ven de Vijver, 1994) suggests the importance of using multiple approaches to assess
measurement equivalence.

Future Research

There are many areas of future research that would contribute to the
psychological contract literature. Several relate to critical issues that were identified in
Chapter II, but not tested in the dissertation study. For example, in Chapter II it was
observed that although various researchers have conceptualized EPC at a relational or
group level, based on my review, no effort has ever been made to empirically assess the
extent to which there is support for a group level construct. The assumption that there is
such a construct continues (Rousseau, 1995) and should be made the subject of
empirical investigation.

Hypotheses regarding relational contracting continue to proliferate in the PC
literature. Few attempts have been made to empirically test these hypotheses. Those
attempts that have been made have used measures that are not closely linked to
relational contracting theory. The present study resulted in the development of scales

that measure many of the characteristics of relational contracting, as described by the

 

 

241

putative “father” of relational contracting theory, Ian Macneil (1980) . The
psychometric properties of the measures have been assessed and were found to hold up
across three different operationalizations of the EPC construct. The relational measures
include the factor sub-scales: sacrifice, enriched work, extra role, and balance. Future
research should finally put the extensive hypothesizing about relational contracting to the
empirical test.

A link between EPCs and employee behaviors has been often asserted but, based
on my review, never empirically tested. This is an area in obvious need of firture
research. In addition to field settings, the laboratory would provide an appropriate
setting for investigating the relationship between EPCs and behaviors. In particular, a
laboratory setting could be used to firrther investigated the equivalence of obligation,
expectation, and promise based EPCs by comparing the EPCs measured by each form
and how they relate to relevant behaviors in the laboratory.

Finally, although the “grave" individual and organizational consequences said
to be associated with violations of employee psychological contracts have been offered
as a primary reason for employers and researchers to take'an interest in EPCs,
theoretical and operational issues regarding the measurement of contract
violation/ fulfillment have received very little attention and would benefit from further
research. A broader notion of “contract fulfillment" should be considered. The
literature has focused exclusively on perceptions of the employer’s fulfillment of its
obligations. What effect do employees’ perceptions regarding the extent to which

they are fulfilling (or not fulfilling) their obligations have on their attitudes and

242

behaviors? The norm of reciprocity suggests that the relationship between perceived
employer violations and criteria such as job satisfaction, trust, or turnover intentions
may be moderated by individual’s beliefs regarding the extent to which they are

fulfilling their own part of the bargain.

243

APPENDICES

.1
o-/
It.

. {1‘-

 

 

245

MICHIGAN STATE UNIVERSITY
EMPLOYEE BELIEFS AND ATTITUDES SURVEY

GENERAL INSTRUCTIONS

This questionnaire asks questions about your work related beliefs and attitudes. There are no right or Wm!!!
am. The survey is divided into a number of sections. Each section begins with brief instructions. Please

provide all of your responses directly on this questionnaire. e r ' n '

YOUR EMPLOYER
Please answer the following question in the space provided:

Who is your employer?

 

(Please print)
(NOTE: If you have more than one employer, please identify your primary or main employer; in the sections
that follow, when asked questions that relate to "your employer", you should answer the questions with the

employer you have identified above in mind.)

 

246

Please recall: An obligation is a duty or responsibility that is imposed by either law, society. promises that are

made, or morality (“good conscience“)

1 2 3 4 5

Not at all Slightly Moderately Highly Very highly
obligated obligated obligated obligated obligated
Using the scale above, please indicate the extent to which you believe that W is obligated to:
__ 17. provide a caring, supportive work environment
__ 18. provide you a safe work place
_ 19. provide co-workers who are competent in doing their work
__ 20. pay you a bonus based on the organization’s performance
_ 21. allow you be involved in setting your performance goals at work
_ 22. provide you a level of pay based on how well you are performing your job
_ 23. provide you long-term job security
_ 24. provide you a level of pay that is “high” when compared to what most Other employers are paying people
performing your kind of job
__ 25. provide you support with personal problems that you may experience
_ 26. give you an opportunity to improve your performance before firing you for poor job performance
_ 27. provide good opportunities for promorion
_ 28. provide you a chance to tell your side of the story when you believe that you have been treated unfairly by
your immediate boss or supervisor
__ 29. provide you a competitive benefits package
_ 30. have good reasons or "good cause“ for terminating your employment
_ 31. provide explanations that make sense to you when making decisions about your job
_ 32. treat you with respect
_ 33. give you revular pay increases
_ 34. provide you the resources necessary to perform your job
_ 35. accommodate your work scheduling requests when possible
__ 36. provide materials and equipment necessary to perform your job
37.

provide you a level of pay that is competitive with what other employers are paying people performing
your kind ofjob

247

Please recall: An obligation is a duty or responsibility that is imposed by either law, society. promises that are
made, or morality (“good conscience”)

l ' 2 3 4 5

Not at all Slightly Moderately Highly Very highly
obligated obligated obligated obligated obligated

Using the scale above, please indicate the extent to which you believe that W is obligated to:
__ 38. allow you opportunities to talk or socialize with your co-workers

__ 39. provide you the Opportunity to learn new skills

__ 40. provide a comfortable work environment

_ 41. deal honestly with you

_ 42. have good reasons for doing things that are likely to have a negative impacr on you personally
_ 43. provide you the same opportunities given Others with your capabilities and performance history

44. take into account your non-work life (family, relationships, non-work activities) when deciding what
demands can be reasonably made of you at work

Please list any other things which you believe vgg; gmplovgr is obligated to provide or do for mg:

 

 

 

 

248

SECTION B: YOUR OBLIGATIONS TO YOUR EMPLOYER

This secrion assesses your beliefs regarding what mu are obligated to provide or do for yogr gmp oner.

For each item presented below, you should indicate your belief regarding the extent to which either the law, society,
promises you have made, or morality obligates you to do or provide the thing that is described. Again, write your
response to each item directly on this questionnaire, on the short line next to the item number.

Write in a '1' if you believe that you are not at all obligated

Write in a '2' if you believe that you are slightly obligated

Write in a '3' if you believe that you are moderately obligated

Write in a '4' if you believe that you are highly obligated

Write in a '5' if you believe that you are very highly obligated

l 2 3 4 5

Not at all Slightly Moderately Highly Very highly
obligated obligated obligated obligated obligated

Using the scale above, please indicate your belief regarding the extent to which you are obligated to:

«New

1. volunteer to do things at work that are n0t formally required by your job
come up with ideas or suggestions for improving your employer's operations

help your co-workers who are having difficulty performing their job

2
3
4. refuse to directly or indirectly support your employer‘s competitors
5

accept a temporary cut in pay if your organization is experiencing financial difficulties and management
indicates that pay cuts are necessary

.0‘

work extra hours even if it does not involve extra pay
7. help solve problems relating to your job

8. make adjustments in your personal life (family, friendships, social activities) to meet changing demands
on your job

9. not seek a new job with another employer so long as your present employer is meeting
its obligations to you

__ 10. have good reasons for leaving your employer

__ ll. get along well with your co-workers

_ -. place company goals ahead of your personal goals on some occasions

_ l3. conduct your personal life so as not to hurt the public image of your employer

14. protect your organization’s pr0prietary information (e.g. trade secrets, information about products, etc.)

15. be loyal to your employer
16. participate in ongoing training to enhance you ability to contribute

17. show up for work every day that you are scheduled to work, so long you are physically able to do so

..
i:
7", ?
We. . ,
[.tl.’ '
.
I
M442: 2": ~ . 4‘ .
ll t

rim: Vito gust!!!“ e: . .: l. -,,.

:;,u.—: " " 4

EUROS... L

taken/qua 1‘ my ‘ , 2;". .

(an mum: tum. nuxercyv r‘ , v.2: -..»- -
32:69qu

or (‘5 at alt wta‘qria :u not pm! :2 .mr

. » > u .' 't 1.0113".
iv 52' La“: nub-d
.u nvf no“
, _ asst.‘ 5 ‘E’Bm w
.- . Alter. at; e". l
. r . ~ i." Ant-”Jain.

     
  
   
    
  
  
   
   
   
  

'.‘.\.N'iyin"

. v—i ”Hm“

.. , .__ V;_--..I-!oo“
. ' m; '-:.td'w”.".¢,‘

uyd‘JLC
‘:.,'m
x 223.1,“; MRNQ
igwttobu .~ ,
'52-; .9 MN” . '.
--r ".‘N 0911mm“ ' __

,, _.A,,"

~.; .. ,_ ma >0 (hm. a

.‘ . . , .'. u: it 'u; rim".
.,,»,. :4)!qu

l.’ 'L
’v-

‘~.
' 4.. ‘.. a": {.394 nu

‘5

”4 mt (my; :1 ram.“

we: 1y” .‘3 (WM 4

1, ‘ "‘.-Lnrfi«-fo‘wrul _ {
ma :2: arm i
I . .
' ‘.“ m ‘ 1 11W] but
. , . I n-
. , l:;n.1_'l"¥ti~“ ’r’
J.‘[‘_ If:
: .r 3 2': ;.;nf"'- 21.-,1: [Ltd] (m - ‘
l 1 vi .141.th :..:l.l¢.-.c:nqwog
‘21.“. -u . Jun: ..1; nLHi‘urto M!
iefuiqris M.
V a t g. a Li} gitrvr...‘ gntfitflofl -
,.
u Mute-:4 .~ .'f’.l \t‘: 1'th aicvaflp ~ ,

249

Please recall: An obligation is a duty or responsibility that is imposed by either law. society, promises that are
made, or morality ('good conscience”)

l 2 3 4 5

Not at all Slightly Moderately Highly Very highly
obligated obligated obligated obligated obligated

Using the scale above, please indicate your belief regarding the extent to which you are obligated to:
__ 18. work as a team player

_ 19. help promote a positive work atmosphere

_ 20. show up for work on time

_ 21. be flexible in your dealings with your employer

_ 22. perform all assigned job tasks to the best of your ability

_ 23. provide a high quality product or service

_ 24. accept a transfer to anOther geographic location

__ 25. make personal sacrifices when the company is experiencing hard times

_ 26. deal honestly with your employer

_ 27. provide advance notice of taking a job elsewhere

_ 28. let your employer know if you have complaints about the way you are being treated
_ 29. look for new ways to contribute to the organization's success

_ 30. follow all company rules and policies

__ 31. accept a transfer to another position if upper management believes that it is necessary
__ 32. spend a minmum of two years in the organization

_ 33. treat your immediate supervisor or boss with respect

34. Please list any other things which you believe you are obligated to provide or do for mu; gmolgvgr:

a.

 

b.

 

C.

 

250

SECTION C: BELIEFS ABOUT WORK IN GENERAL

Below are a number of statements each of which you may agree or disagree with depending on your own personal
evaluation of )3ng in ggngrgl (without reference to any specific job). Please indicate the extent to which you agree or
disagree with each statement by writing the appropriate number from the scale below (1 to 7) in the blank next to each

statement. To what extent do you agree or disagree with each of the following statements?

1 2 3 4 5 6 7
Strongly Disagree Slightly Neither agree Slightly Agree Strongly
disagree disagree nor disagree agree agree

__ l. The most important things that happen in life involve work.

_ 2. Work is something people should get involved in most of the time.
__ 3. Work should only be a small part of one's life.

__ 4. Work should be considered central to life.

_ 5. In my view, an individual's personal life goals should be work-oriented.
_ 6. Life is worth living only when peeple get absorbed in work.

The statements below ask what you would like for your relationship to be with any organization for which you might
work. On each question, divide 10 points between the two choices (choice A and choice B) by giving the most points
to the choice that is most like you and the fewest points to the choice that is least like you. You can, if you like, give

the same number of points to both choices (for example, 5 points to choice A and 5 points to choice B). And you can

use zeros if your like. Just be sure to allocate all Wm between each pair of possible responses.
In any organization I might work for:

i. It would be more important for me to: (allocate 10 points)
A. Get from the organization

B. Give to the organization

2. It would be more important for me to: (allocate 10 points)
A. Help others

B. Watch out for my own good

3. I would be more concerned about: (allocate 10 points)

A. What I received from the organization

 

B. What I contributed to the organization

 

4. The hard work I would do should: (allocate 10 points)

A. Benefit the organization

 

B. Benefit me

 

5. My personal philosophy in dealing with the organization would be: (allocate 10 points)

A. Ifl don't look out for myself, nobody else will

251

SECTION D: THE EXTENT TO WHICH YOUR EMPLOYER HAS FULFILLED ITS OBLIGATIONS

‘Ihis section asks you to indicate your beliefs regarding the extent to which your current employer has met, failed to

meet, or exceeded its obligations to you. Please carefully read each item and cirglg the appropriate answer (to the
right of each question).

10.

ll.

13.
H.
15.
l6.

17.

Circle '1' if what your employer has provided or done is much less than obligated
Circle '2' if what your employer has provided or done is less than obligated

Circle '3‘ if what your employer has provided or done is about the same as obligated
Circle '4' if what your employer has provided or done is more than obligated

Circle '5' if what your employer has provided or done is much more than obligated

Much
Lean

provide you work that is challenging 1
allow you time off from work to participate in non-work activities that are important 1
to you
provide you training to prepare you to assume a better position in the organization I
provide you work that is interesting 1
give you recognition for the contributions that you make to the organization 1
help you develop your career 1
(item #7 as been removed; please go on to item #3)
provide you opportunities to advance and grow 1
allow you to participate in decisions that may have an important impact on your work 1
provide you work that gives you a sense of purpose I
keep you informed regarding the organization's financial condition . l

. be flexible in its practices in order to allow you to try to balance your work 1
and personal lives
provide you work that gives you a great deal of responsibility I
provide a friendly work atmosphere 1
treat you as a partner 1
provide you regular feedback about how you are performing your job 1
provide a caring, supportive work environment 1

Id

'0

I.)

Some

Le)

(4)

Much
Adore

Us

'To what extent has your current employer met, failed to meet, or exceeded its obligations to you?

18.
19.
20.
21.
22.
23.

24.

25.

37.

33.
39.

40.

 

2s2

provide you a safe work place

provide you co-workers who are competent in doing their work

pay you a bonus based on the organization's performance

allow you to be involved in setting your performance goals at work
provide you a level of pay based on how well you are performing your job
provide you long-term job security

provide you a level of pay that is “high“ when compared to what most other
employers are paying people performing your kind of job

provide you support with personal problems you may experience

. give you an opportunity to improve your performance before firing you for poor

job performance

. provide good Opportunities for promotion

. provide you a chance to tell your side of the story when you believe that you have

been treated unfairly by your immediate boss or supervisor

. provide you a competitive benefits package

. (item #30 has been removed; please go on to item #31)

. provide explanations that make sense to you when making decisions about your job
. treat you with respect

. give you regular pay increases

. provide you the resources necessary to perform your job

. accommodate your work scheduling requests when possible

. provide materials and equipment necessary to perform your job

provide you a level of pay that is competitive with what other
employers are paying people performing your kind of job

allow you Opportunities to talk or socialize with your co-workers
provide you the opportunity to learn new skills

provide a comfortable work environment

Much
Leu

1

l

[J

IQ

DJ

DJ

Le)

b)

.3.

Much
More

Us

 

253

'To what extent has your current employer met, failed to meet, or exceeded its obligations to you?

Much Much

Lou Some More
41. deal honestly with you 1 2 3 4 5
42. have good reasons for doing things that are likely to have a negative impact 1 2 3 4 5

on you personally

43. provide you the same Opportunities given Others with your capabilities and 1 2 3 4 5
performance history
44. take into account your non-work life (family, relationships, non-work activities) 1 2 3 4 5

when deciding what demands can be reasonably made of you at work

 

SECTION E

Please answer each of the following six questions by finding the appropriate response (from I to 7) on the scale
presented immediately below each question.

1. How likely is it that yOu will actively look for work in a different organization within the next year?

I 2 3 4 5 6 > 7
Not at all Somewhat Quite likely Extremely likely
likely likely

2. How likely is it that you will leave your organization in the next year?

i 2 3 4 5 6 7
Not at all Somewhat Quite likely Extremely likely
likely likely

3. To what extent do you agree or disagree with the following Statement: "I often think about quitting my job.‘'

1 2 3 4 5 6 7
Strongly Disagree Slightly Neither agree Slightly Agree Strongly
disagree disagree nor disagree agree agree

4. All things considered, how satisfied are you with your job?

1 2 3 4 S 6 7
Extremely Dissatisfied Slightly Neutral Slightly Satisfied Extremely
dissatisfied dissatisfied satisfied satisfied

5. Considering all of the jobs you might realistically have at this point in your career, how does your current job
compare?

1 2 3 4 5 6 7
One of Worse Slightly About Slightly Better One of
the worst than most worse as erage better than most the best

than. ms: than most

 

 

 

254

' 6. Considering work in general (without reference to any specific job), Overall, how important is work in your life?

1 2 3 4 5 6 7
Not at all Somewhat Quite Extremely
important important important important

 

SECTION F: DEMOGRAPlllCS
GENERAL INSTRUCTIONS: Sometimes results of studies, such as these, will differ based on some aspect of
demography (e.g. age or gender). I am asking the following quesfions for this reason. Please write in, or circle, the
apprOpriate answer for each question. '
I. What is your gender: a. Male b. Female
2. What is your age?
3. What is your current martial status? a. married b. single c. divorced d. separated e. widowed

4. Do you have minor children (l7 years old or younger) ? a. Yes b. No

If so, what are you minor children’s ages?
if so, do you consider yourself your children's primary care provider?

 

a. Yes
D. l share child care responsibilities approximately equally with another person (e.g. a spouse)
c. No

5. What is your race or ethnic origin: a. White b. African American c. Asian/Pacific lslander

d. Hispanic e. Other

6. Education level - Please circle the highest level of formal education that you have obtained:
some high school, but did nor graduate

. graduated from high school

. two years or less of college

. more than two years of college. but did not graduate

. four year college degree

masters degree

. degree work beyond a masrers degree

«manner.»

7. What is your department or functional area:
a. bakery
b. carry-out
c. cashier
(1. dairy 6t frozen foods
e. deli

f. meat

g. produce

h. office

i. stock - day

j. night stock

it. “other" (if I of the above areas does not apply)

 

(your department or functional area

7A. Are you a manager or a supervisor? a. Yes b. No

 

255

' 8. How long have you worked for your current employer: years months
9. On average, how many hours a week do you work for your employer? hours per week

l0. Which of the following statements best describes your preference for full-time versus part-time employment?
(consider “part-time employment" as involving less than 35 hours of work per week)
a. 1 strongly prefer being employed pan-time
b. I prefer being employed part-time
c. l somewhat prefer being employed part-time
d. i do nOt have any preference for either full-time or part-time employment
e. l somewhat prefer full-time employment
f. I prefer full-time employment
g. 1 strongly prefer full-time employment

ll. fl hink back to when you first began working for your present employer. At tha t timg, how long did you plan
to work for your present employer:

a. less than 1 year d. 7 to 9 years
b. l to 3 years c. l0 years or more, but less than your entire working career
c. 4 to 6 years f. your entire working career

12. Approximately how many years of work experience do you have?
(include your current job and your experience in all previous jobs that you have had)

 

SECTION G

Using the scale below, please indicate the extent to which you agree with the following statements by circling the
appropriate response.

1 = Strongly disagree 5 == Agree somewhat
2 = Disagree 6 = Agree
3 = Disagree somewhat 7 = Strongly agree
4 = Neutral (neither agree or disagree)
Strongly Strongly
Disagree Neutral Agree
1. I am not sure that I fully trust my employer. l 2 3 4 5 6 7
2. Given the opportunity, my employer would take unfair l 2 3 4 5 6 7

advantage of me.

3. My employer is open and up~front with me. i 2 3 4 5 6 7
4. I believe my employer has high integrity. l 2 3 4 S 6 7
5. In general. i believe my employer's intentions and l 2 3 4 5 6 7

mOtives are good.

6. My employer is not always honest and truthful. l 2 3 4 5 6 7
7. I don't think that my employer treats me fairly. l 2 3 4 5 6 7
8. l can count on my employer to treat me in a consistent and l 2 3 4 5 6 7

predictable way.

256

Please recall: An obligation is a duty or responsibility that is imposed by either law, society, promises that are

made, or morality (“good conscience”) ‘

l - 2 3 4 5

Not at all Slightly Moderately Highly . Very highly
obligated obligated obligated obligated obligated
Using the scale above, please indicate the extent to which you believe that W is obligated to:
__ 17 provide a caring, supportive work environment
_ 18. provide you a safe work place
_ 19. provide co-workers who are competent in doing their work
_ 20. pay you a bonus based on the organization’s performance
__ 2!. allow you be involved in setting your performance goals at work
_ 22. provide you a level of pay based on how well you are performing your job
_ 23. provide you long-term job security
_ 24. provide you a level of pay that is 'high' when compared to what most other employers are paying people
performing your kind of job
_ 25. provide you support with personal problems that you may experience
__ 26. give you an opportunity to improve your performance before firing you for poor job performance
__ 27. provide good Opportunities for promotion
__ 28. provide you a chance to tell your side of the story when you believe that you have been treated unfairly by
your immediate boss or supervisor
__ 29. provide you a competitive benefits package
__ 30. have good reasons or ”good cause“ for terminating your employment
__ 31. provide explanations that make sense to you when making decisions about your job
_ 32. treat you with respect
_ 33. give you regular pay increases
_ 34. provide you the resources necessary to perform your job
__ 35. accommodate your work scheduling requests when possible
__ 36. provide materials and equipment necessary to perform your job
37

. provide you a level of pay that is competitive with what other employers are paying people performing

your kind ofjob

257

SECTION A: EMPLOYER’S OBLIGATIONS TO YOU
This section assesses your beliefs regarding what W is obligated to provide or do for you. An obligation is
a duty or responsibility that is imposed by either law, society, promises that are made, or morality (“good con.
science'). For each item presented below, you should indicate your belief regarding the extent to which either the law,
society, promises your employer has made, or morality obligates your employer to do or provide the thing that is
described. Write your response to each item directly on this questionnaire, on the blank next to the item number.
L Write in a '1' if you believe that your employer is not at all obligated
_J_~__ Write in a '2' if you believe that your employer is slightly obligated
1 Write in a '3' if you believe that your employer is moderately obligated
l Write in a '4' if you believe that your employer is highly obligated
I Write in a '5‘ if you believe that your employer is very highly obligated

1 2 3 4 5

Not at all Slightly Moderately Highly Very highly
obligated obligated obligated obligated obligated

Using the scale above, please indicate the extent to which you believe that W is obligated to:

h

__ i. provide you work that is challenging

_ 2. allow you time off from work to participate in non-work activities that are important to you
_ 3. provide you training to prepare you to assume a better position in the organization

_ 4. provide you work that is interesting

__ 5. give you recognition for the contributions that you make to the organization

__ 6. help you develop your career

__ 7. continue employing you so long as you are performing your job reasonably well

__ 8. provide you Opportunities to advance and grow

__ 9. allow you to participate in decisions that may have an important impact on your work

_ 10. provide you work that gives you a sense of purpose

__ ll. keep you informed regarding the organization‘s financial condition

_ 12. be flexible in its practices in order to allow you to try to balance your work and personal lives
__ 13. provide you work that gives you a great deal of resporuibility

__ 14. provide a friendly work atmosphere

__ 15. treat you as a partner

16. provide you regular feedback about how you are performing your job

 

258

APPENDIX B

 

 

259

Anecdotal Illustration of the Limitations of a Promise Framework

A man who was discharged by his employer of 13 years goes to an attorney.
His anger at being discharge without, in his mind, good reasons borders on outrage.
He expresses his belief that he was “owed better treatment than that." The man
produces a paperback employee manual from a brown paper bag. It contains the
following express disclaimer: “EMPLOYMENT IS AT-WILL. YOUR
EMPLOYMENT MAY BE TERMINATED AT ANY TIME, WITH OR WITHOUT
NOTICE, AND WITH OR WITHOUT CAUSE.” The man acknowledges reading
the employee manual but doesn’t recall what he made of the quoted statement. In I
response to the attorney’s question, the man concedes that “No, they never promised
me that they would only fire me for good reasons." He then volunteers “But they
owed me better treatment than that.”

The man’s subjective belief that his employer owed him better treatment than
he had received is, I would argue, part of the phenomenon most researchers would
like to capture with the EPC construct. However, the man’s subjective belief would
not be included in conceptualization of the EPC construct that truly limits the
construct to promise based obligations , i.e., one that does not employ the term

“promise” as a fiction.

260

LIST OF REFERENCES

261

LIST OF REFERENCES

Ackerman, P. L. & Humphreys, L. G. (1990). Individual differences theory
in industrial and organizational psychology. In M. D. Dunette & M. L. Hough

(EdS) WWWQM ed) Vol 2.
165- 207. Palo Alto, CA. Consulting Psychologist Press.

Argyris, C. (1960). Wm. Homewood Ill.: The

Dorsey Press, Inc.

Baker H. G. (1986). The unwritten contract: Job perceptions. W.
37—41.

Barnard, C. I. (1938). W. Cambridge, MA: Harvard.

Barnett, R. E. (1992). Conflicting visions: A critique of Ian Macneil's relational

theory of contract. MW, 78, 1175-1206.

Black, H. C. (1968). W. New York: West Publishing.

Blau, P. M. (1964). W. New York: Wiley.

Boss, R. (1985). The psychological contract: The key to effective organizational
development consultation. WM, 4, 284-304.

Bowditch. J. L.. & Bruno, A. F. (1982). MW
W. Boston, MA: Auburn House.

Carroll, L. (1923). W55. New York: MacMillan.

Cohen J &Cohen P (1983) W
W. Hillsdale, NJ: Lawrence Erlbaum Associates.

Cole. D. W. (1981). W New
York: McGraw-Hill.

 

 

262

Collella, A., & DeNisi, A. S., & Wanous, J. P. (1994). On the efficacy of using
newcomer expectations to predict subsequent job reactions. Paper presented at the
Ninth Annual Society for Industrial and Organizational Psychology, Nashville.

Commons, J. R. (1924). W. New York: Macmillan.

Cook T D &Carnpbell D T (1979) WWW
analysisjqn fielisettings. Boston: Houghton Mifflin Company.

Covey, S. R. (1988). Six conditions of effectiveness. WW, 5(11),
8-9.

Dunahee, M. H. & Wangler, L. A. (1974). The psychological contract: A conceptual
structure for management/employee relations. BermnneLanmal, 518-526.

Dwyer, F. R., Schurr, P. H., & Oh, S. (1987). Developing buyer-seller
relationships. Marketing, 52, 11-27.

Farnsworth, E. A. (1982). antmcts. Boston: Little, Brown

Fleishman, E. A. (1953a). A description of supervisory behavior. ,LQumaLQLApnnen,
12511911919211. 37. 1-9.

Fleishman, E. A. (1953b). The measurement of leadership attitudes 1n industry.
JoumaLQLAnnlicifimhology, 37, 153-158.

Folger, R. (1993). Reactions to mistreatment at work. In J. Murnighan (Ed. )_S_QciaL

WW Prentice Hall:
Englewood Cliffs, NJ.

Ford, J. K., MacCallum, R. C., & Tait, M. (1986). The application of exploratory
factor analysis in applied psychology: A critical review and analysis. Rename}

125mm, 39, 291-314.
Fox, A. (1974). WWW London: Faber.

Forbes, F. S., & Jones, I. M. (1986). A comparative, attitudinal, and analytical
study of dismissal of at-will employees without cause, WM, 157-
166.

 

263

French, W. L. Kast, F. E., &Rosenzweigh, J. E. (1985). WI;
behaxinunmganiznnnns. New York: Harper & Row.

GabarrO. J. 1.. & AthOS. J. (1976). W. New
York: Prentice Hall.

Gibson, R. O. (1966). Toward a conceptualization of absence behavior of personnel

inorganizations AdministratiXLScicnccfluarterly. 11. 107-133.

Glass, G. V., & Hopkins, K. D. (1984). Stan'sn'ga] 31;;th in education and
12519115219231. Englewood Cliffs, NJ: Prentice Hall Inc.

Goddard, R. W. (1988). Inside the psychological employee. Managememflqfld,
17. 24-26.

George, J. M. (1990). Personality, affect, and behavior 1n groups. 11mm
Appliedflxchclm 75 107-116

Goetz, C. J ., & Scott, R. E. (1981). Principles of relational contracts. yingjm'a lay!
Rel/icy, 67. 1089-1150.

Goldberg, V. P. (1980). Relational exchange. MW, 23,
337-352.

Gorsuch, R. L. (1974). W5. Philadelphia, PA: W.B. Saunders.
Cough (1957. 1978). IhLSmiaLQQmm.

Gould, S. ( 1979). An equity-exchange model of organizational involvement.
AcademLQLManagemanRuim. 4. 53-62.

Gouldner, A. W. (1960). The norm of reciprocity: A preliminary statement.
MW. 25, 161-178.

Gulliksen, H. (1968). Methods for determining equivalence of measures.
Whit]. 70. 534-544

Guzzo, R. A., Noonan, K. A. & Elron, E. (1994). Expatriate managers and the
psychological contract. MW, 79, 617- 626.

Herriot, P. ( 1988) Graduate recruitment. Psychological contracts and the balance of

power.Britishqumal_QLQ1udanceand_Ccunsclling 16 228- 241

264

Herzberg, F., Mausner, B., & Snyderman, B. (1959). W.
New York: John Wiley and Sons.

Higgins, E. T. (1992). Social cognition as a 5 social science: How social action
creates meaning. In D. Ruble, P Costanzo, & M. Oliveri (Eds.) Mal

W. New York: The Guilford Press.

Holtz, J. L. (1978). Exploring the psychological contract over the life cycle.
Unpublished dissertation.

Huse, E., & Bowditch, J. (1977). WW9.
managing. Reading, MA: Addison-Wesley.

Huseman, R. C., Hatfield, J. D., & Miles, E. W. (1987). Test for individual

perceptions of job equity: Some preliminary findings. W,
61, 1055-1064.

Ilgen, D. R., & Hollenbeck, J. R. (1991). The structure of work. Job design and

roles. In M. D. Dunette & M. L. Hough (Eds.,) W

W (2nd ed..) Vol. 2,165-207.Palo Alto, CA: Consulting
Psychologist Press.

Ironsen, G. H., Smith, P. C, Brannick, M. T. Gibson, W. M., & Paul, K. B.
(1989). Construction of a job 1n general scale: A comparison of global, composite,

and specific measures. WM, 74, 193- 200.

Jackson, S. E., Brett, J. F., Sessa, V., Cooper, D.M., Julin, J., & Peyronnin (1991).
Some differences make a difference: Individual dissimilarity and group
heterogeneity as correlates of recruitment, promotions, and turnover. LonmaLQf

Applicdflxchclozx 76 675-689

James, L. R. Demaree, R. G. & Wolf, G. (1984). Estimating within- -group
interrater reliablity with and without response bias. lQumaLQLApnlied

13thle 69 35 89

James, L. R., Demaree, R. G, & Wolf, G. (1993). rwg: An assessment of within-
group interrater agreement. MW, 78, 306- 309.

Jurek, W. R. (1968). A study of the relationship between the psychological contract
of routemen and their sales performance in eight selected laundry and dry cleaning
companies. Unpublished dissertation.

 

265

KM. D.. & Kahn. RL. (1978). W (2nd ed.)-
New York: Wiley.

Kaufman, P. J. & Stern, L. W. (1988). Relational exchange norms, perceptions of
unfairness, and retained hostility 1n commercial litigation. ,LQnmaLQLQQnfljg
reschttim. 32, 534- 552.

Kazdin, A. E. (1982). Sinlfe case experimental designs. In P. C. Kendall & J. N.

Butcher (Eds). Handeancarcleethodsmchmcaanholm (PP 416-
490). New York: Wiley.

Kerlinger, F. N. (1986). W. Chicago: Holt,
Rinehart, & Winston, Inc.

Klein, K. J ., Dansereau, F ., & Hall, R. J. (1994). Level issues in theory

development, data collection, and analysis. MW, 19.
195-229. A

Kolb, D. A. Rubin, I. M., & McIntyre, J. M. W

W. Englewood Cliffs NJ Prentice-
Hall, Inc.

Kotter, J. P. (1970). The psychological contract: Expectations and the joining up
process (unpublished master's thesis, Sloan).

Kotter, J. P. (1973). The psychological contract. Managing the joining up process.
MW. 15. 91-99.

Kozlowski, S. W. J ., & Hults, B. M. (1987). An exploration of climates for
technical updating and performance. W, 40, 539-563.

Kozlowski, S. W. J., & Hattrup, K. (1992). A disagreemnet about within- -group
agreement. Disentangling 1ssues of consistency versus consensus. Lonmanf

AW 77 161-167

LawlCSS. D. J. (1972).Effcctme_managcmcnt:_5.nc1aLpsxchologicaLannmach
Englewood Cliffs, N.J.: Prentice-Hall, Inc.

Levin, J ., & McDowell, B. (1983). Balance theory of contracts. 1219;11le
_J_Qnma1, 29, 24-87.

 

 

266

Levinson, H., Price, C. R., Munden, K. J., & Solley, C. M. (1962). Men,
managernenLanimentaLhealth. Cambridge, MA: Harvard University Press.

Lucero, M. A., & Allen, R. E. (1994). Employee benefits: A growing source of
psychological contract violations. EmmanReseuLceManagemem, 33, 425-446.

Macneil, I. R. (1974). The many futures of contracts. SeuthernfialifemiaLaw
89119.81. 47, 691-816.

Macneil, I. R. (1978). Contracts: Adjustment of long-term economic relations under

classical, neoclassical, and relational contract law. Wm, 72,
854-905.

Macneil, I. R. (1983). Values in contract: Internal and external. WW
Reliew, 78, 340-418.

Macneil, I. R. (1980). Wm. New Haven: Yale University
Press.

Macneil, I. R. (1985). Relational contract: What we do and do not know. ansin
Law. 483-525.

Macneil, R. W. ( 1986). Contract in China: Law, practice, and dispute resolution.
SEW. 38, 303-397.

Major, D. A., Kozlowski, W. J., Chao, G. T., & Gardner, P. D. (1995). A
longitudinal investigation of newcomer expectations, early socialization outcomes,
and the moderating effects of role development factors. LomnaLeLAnflied

Esyehelegy, 80, 418-431.

March, J. E., & Simon, H. A. (1958). W. New York: John Wiley &
Sons, Inc.

Martocchio, J. J, & Judge, T. A. (1994). A policy capturing approach to

individual' 3 decisions to be absent. WWW
Processes, 57, 358- 386.

McFarlane Shore, L., & Tetrick, L. E. (1994). The psychological contract as an
explanatory framework 1n the employment relationship. In C. L. Cooper & D. M.

Rousseau (Eds..) WW. 1. 91- 109

267

Menninger, K. (1958). W. New York: Basic
Books.

Menninger, K., & Holzman, P. W (2nd ed.). New
York: Basic Books.

Miller, J .G.(1978). W. New York: McGraw-Hill.

Mitchell, T. R. (1985). An evaluation of the validity of correlational research
conducted in organizations. Wm, 10, 192-205.

Nelson, D. L., Quick, J. C, & Joplin (1991). Psychological contracting and
newcomer socialization: An attachment theory foundation. Iheleumalef

Seeialflehaxierandlzersonalim, 6, 55-.72

Nicholson, N., & Johns (1985). The absence culture and the psychological contract -
Who's in control of absence? AeademLQLManagemenLReyiey, 10, 397-407.

Noe, R. A, Hollenbeck, J. R., Gerhart, B., &Wright, P. (1993). Humanresenree
managememfiaininaaeemmtitixudxantage Boston Irwin

Nunnally, J. C. (1978). Esyehemetdetheery. New York: McGraw-Hill.

Ofori-Dankwa, J. (1991). Unpublished dissertation.

Olsen, J. M., Roese, N. J., & Zanna, M. P. (1994). Expectancies. In E. T. Higgins

&A. W. Kruglanski (Eds) SeciaLnsxchelegxzflandbeekeLbasieprineinles.
New York: Guilford Press.

Portwood, J. D., & Miller, E. L. (1976). Evaluating the psychological contract: Its
implications for employee satisfaction and work behavior. Proceedings of the
Annual Academy of Management, 109-113.

Provan, K. G., & Gassenheimer, J. B. (1994). Supplier commitment in relational
contract exchanges: A study of interorganizational dependence and exercised

power. MW. 31. 55-68.

Robinson, S. L. (1994). The impact of psychological contract violation on
employment relations. Annual Academy of Management Meeting, Dallas.

Robinson, S. L., & Morrison (1995). Developing a standard measure of the
psychological contract. Annual Academy of Management Meeting, Vancouver.

 

268

Robinson, S. L., Kraatz, M. S., & Rousseau, D. M. (1994). Changing obligations
and the psychological contract. AeademLQLManagemenLLQnma], 37, 437-452.

Robinson, S. L. & Rousseau, D. M. (1994). Violating the psychological contract.
Not the exception but the norm. leumaLeerganizanenalflehamr, 15, 245- 259.

Roehling, M.V. (1996). Conceptualizing the psychological contract construct:
Critical issues. Paper presented at the Annual Meeting of the Academy
of Management, Cincinnati, Ohio.

Rosen, B., & Schwoerer, C. (1990, February). Balanced protection policies.
HRMasazine. 59-64.

Rousseau, D. M. (1985). Issues of level in organizational research. Multi- level and
cross- -level perspectives. In B. Staw & L. Cummings (Eds. ), Researelun

ergan'uanenauzehayjer (Volume 7). Greenwich CT: JAI Press.

Rousseau, D. M. (1989). Psychological and implied contracts in organizations.

ErnnloxeLResmnsihilitiesandBightslaumal.2.121-139

Rousseau, D. M. (1990). New hire perceptions of their own and employer' 5

obligations: A study of psychological contracts. lenmalefiflrganizanenal
Behayier, 11,389-400.

Rousseau, D. M., & McLean Parks (1993). The contracts of individuals and

organizations. In B. Staw & L. Cummings (Eds), Reseanenjnmganizanenal
behaxier, 15, 143. Greenwich CT. JAI Press.

Rousseau, D.M., & Greller, M. M. (1994) Human resource practices:
Administrative contract makers HnmanResQnreeManagement, 33, 385-401

Rousseau, D.M., & Wade-Benzoni, K.A. (1994). Linking strategy and human
resource practices: How employee and customer contracts are created. Human

KW. 33, 463-489.

Schein, E. H. (1965, 1970, 1980). QuanizatjenaLnsyehelegy. Engelwood Cliffs,
NJ :Prentice Hall.

Schneider, B. (1983). An interactionist perspective on organizational effectiveness.

In L L Cummings & B M Staw (Eds ) ResearehjnerganizatienaLbehaxier (Vol.
7, pp. 1-38). Greenwich, CT. JAI Press.

269

Schneider, B. (1987). The people make the place. Persennelflsyehelegy, 40, 437-
453.

Schuler, R. S., Beutell, N. J ., & Youngblood, S. A. (1989). Effeetixefiensennej
Management. New York: West Publishing Company.

Simon, H. A. (1991). Organizations and markets. Journaleflfieenemie
Berspeetises. 5. 25- 44

Sims, R. R. (1995). Human resource management's role in clarifying the new

psychological contract. HemanResenmeManagement, 33, 373-382.

Smith, P. C, Kendall, L. M., &Hulin, C. L. (1969). Ihemeasnmflsafisfaefienjn

WWW. Chicagoz Rand McNally
& Company.

Spindler, G. S. (1994). Psychological contracts in the work place: A lawyer's view.
HumanBeseureeManagemem. 33, 325-334.

Tomow, W. W. (1988). Contract redesign. BersenneLAdministrater, 33(10),
97-101.

Van Dyne, l., Grahm, J. W., & Dienesch, R. M. (1994). Organizational citizenship
behavior: Construct redefinition, measurement, and validation. leurnalgf

W. 37, 765-802.

Wayland, R. F. Clay, J. M., & Payne, P. L. (1993). Employment at-will

statements: Perceptions of job applicants. InternatienallenmaLQLManpem, 14,
22 33.

Weick.K. E. (1979). WWW MAI Addison-
Wesley.

Whitford, W. C. (1985). Ian Macneil' 3 contribution to contracts scholarship.
W. 545- 560.

Yates, R. E. (1993). Downsizing's bitter pill. theagflfibnneMagazine, Section
10, November 21.

"11111111117111.111111111115