THE EFFECTS OF ORGANIZATIONAL COMMITMENT ... - TDL

THE EFFECTS OF ORGANIZATIONAL COMMITMENT, PROFESSIONAL

COMMITMENT, LIFE-SPAN CAREER DEVELOPMENT, AND

SELF-MONITORING ON JOB SATISFACTION AND JOB

PERFORMANCE AMONG STAFF ACCOUNTANTS

by

PETER JOHN POZNANSKI, B.S., M.S.

A DISSERTATION

IN

BUSINESS ADMINISTRATION

Submitted to the Graduate Faculty of Texas Tech University in

Partial Fulfillment of the Requirements for

the Degree of

DOCTOR OF PHILOSOPHY

Approved

May, 1991

So/ rz

e<

Copyright, 1991, Peter John Poznanski

ACKNOWLEDGMENTS

The participation of the professional accountants at

the office of the public accounting firm which

participated in this study is appreciated. Comments on

early drafts of this study from the accounting doctoral

students at Texas Tech University, notably Michael Shaub,

Randy Stitts, and Gary Caplan, are also appreciated.

Dr. Donald Clancy and Dr. James Wilcox were of great

assistance in the initial development of this study and

the data analysis, even though they were not committee

members.

A study such as this could not be successfully

completed without the assistance and encouragement of an

advisory committee, and the leadership of a committee

chairman. Dr. Don Finn's contributions as chairman have

made the completion of this study possible.

Finally, the continual support from family and

friends throughout the process is greatly appreciated.

Thank you all very much.

11

CONTENTS

ACKNOWLEDGMENTS ii

ABSTRACT v

LIST OF TABLES vii

LIST OF FIGURES ix

I. INTRODUCTION AND BACKGROUND 1 Need for the Study 3 Purpose of the Study 4 Organization of the Study 5

II. LITERATURE REVIEW 7 Job Satisfaction and Job Performance 8

Job Satisfaction in the Accounting Profession 9

Job Performance in the Accounting Profession 14

Organizational and Professional Commitment, Life-Span Career Development, and Self-Monitoring 21 Organizational and Professional Commitment in the Accounting Profession 21

Life-span Career Development 26 Self-Monitoring 32

Chapter Summary 42

III. THE THEORETICAL MODEL, RESEARCH MODEL, AND FORMULATION OF HYPOTHESES 45 The Theoretical Model 45 The Research Model 46 Formulation of the Hypotheses 48 Chapter Summary 54

IV. RESEARCH METHODOLOGY 56 Research Design 56

The Population for the Study 57 Development of the Research Instrument 58 Job Satisfaction Measurement 59 Job Performance Measurement 60 Organizational and Professional Commitment Measurement 62

111

Life-Span Career Development Measurement 63 Self-Monitoring Measurement 64 Research Instrument Distribution and Collection Technique 65

Pretesting the Research Instrument 66 Statistical Technique 71

LISREL Model Specification 72 Chapter Summary 79

V. EMPIRICAL ANALYSIS OF THE DATA 81 Response Information 81

Response Rates 81 Test for Nonresponse Bias 82

Statistical Analysis 84 Testing for Identification 84 Testing the Measurement Model 86 Hypothesis Testing 102 The Structural Model 103 Sequential Chi-square Difference Tests 106 Evaluating the Theoretical Model 114

Summary of the Data Analysis 123

VI. SUMMARY AND CONCLUSIONS 127 Summary of the Findings and Implications 127

Research Objective One 128 Research Objective Two 131 Research Objective Three 132

Contributions 134 Limitations 135 Future Research 136

REFERENCES 138

APPENDICES

A. RESEARCH INSTRUMENT 150

B. RESEARCH INSTRUMENT APPROVAL , 162

C. DEMOGRAPHIC DATA AND DESCRIPTIVE STATISTICS 164

D. LISREL RESULTS FOR INDIVIDUAL HYPOTHESES AND PARAMETERS 167

IV

ABSTRACT

Increasing the level of employee job performance and

job satisfaction is a goal of most organizations. No one

theory or variable has been found to explain job

performance and job satisfaction; however, research

evidence has indicated some correlation between these two

constructs (for example, see Vroom, 1964; Lawler and

Porter, 1967; Ross and Bomeli, 1971; Benke and Rhode,

1980, and Bullen and Flamholtz, 1985).

The importance of maximizing job satisfaction and job

performance among employees within organizations is

continually sought by managers. Maximizing job

satisfaction and job performance may lead to higher

productivity, lower overall costs for the firm, and

reduced turnover of employees.

This study investigated constructs that can affect

job satisfaction and job performance, as well as the

relationship between these two constructs. The research

objectives of this study were to: (1) investigate the

effects of organizational commitment, professional

commitment, life-span career development, and self-

monitoring on job satisfaction and job performance for

staff accountants in public accounting firms; (2)

investigate the relationship between job satisfaction and

job performance of staff accountants in public accounting

firms; and, (3) investigate the relationship among

organizational commitment, professional

commitment, and life-span career development for staff

accountants in public accounting firms.

VI

LIST OF TABLES

2.1 LIFE-SPAN CAREER DEVELOPMENT STAGES AND AGES 28

4.1 SUMMARY OF QUESTIONNAIRE RESPONSE RATES FROM THE PRETEST FOR ALL SUBJECTS 68

5.1 SUMMARY OF QUESTIONNAIRE RESPONSES AND ELIMINATIONS 83

5 . 2 NONRESPONSE BIAS TEST RESULTS 85

5.3 MEASUREMENT STATISTICS FOR THE TWENTY-ONE COMPOSITES 90

5.4 MEASUREMENT STATISTICS FOR THE SIXTEEN COMPOSITES 97

5 . 5 PEARSON CORRELATION COEFFICIENTS 99

5 . 6 POLYSERIAL CORRELATION COEFFICIENTS 100

5.7 PSEUDO TEST STATISTIC FOR THE NULL AND SATURATED MODELS 105

5.8 COMPARISON OF THE NULL -AND SATURATED MODELS 107

5.9 COMPARISON OF THE THEORETICAL AND SATURATED MODELS 110

5.10 COMPARISON OF THE THEORETICAL AND CONSTRAINED MODELS 113

5.11 COMPARISON OF THE THEORETICAL AND UNCONSTRAINED MODELS 115

5.12 GOODNESS-OF-FIT INDICES AND ROOT MEAN SQUARE RESIDUiALS 117

5.13 LISREL VALUES—NONSIGNIFICANT RELATIONSHIPS 119

5 .14 LISREL VALUES—ALL HYPOTHESES 124

C . 1 DEMOGRAPHIC DATA 165

C.2 DESCRIPTIVE STATISTICS 166

Vll

D.l TESTING INDIVIDUAL PHI 168

D.2 TESTING INDIVIDUAL BETAS—WITH PHI 169

D. 3 TESTING INDIVIDUAL GAMMAS—WITH PHI 170

D.4 LISREL VALUES FOR THE HYPOTHESES 171

D.5 LISREL VALUES FOR LAMBDA Y AND LAMBDA X 172

D.6 STANDARDIZED SOLUTIONS 173

Vlll

LIST OF FIGURES

3 .1 Theoretical Model 46

3 . 2 The Research Model For H1-H8 47

3. 3 The Research Model For H9 47

3.4 The Research Model For H10-H12 48

3 . 5 The Research Model For All Hypotheses 55

4.1 LISREL Model 73

4 .2 LISREL Notation And Terminology 75

4 . 3 The Structural Equation 76

4 .4 The Measurement Model For Y Indicants 77

4 .5 The Measurement Model For X Indicants 78

5.1 Respecified Measurement Model For Y Indicants 92

5.2 Respecified Measurement Model For X Indicants 93

5.3 Respecified Measurement Model For Composite Y Indicants 95

5.4 Respecified Measurement Model For Composite X Indicants 96

5 .5 Decision-Tree Framework For SCDTs 108

5.6 Research Model—Results Of Hypothesis Testing.... 126

IX

CHAPTER I

INTRODUCTION AND BACKGROUND

Increasing the level of employee job performance is a

goal of most organizations. No one theory or variable has

been found to explain job performance; however, research

evidence has indicated some correlation between job

performance and job satisfaction (for example, see Vroom,

1964; Lawler and Porter, 1967; Ross and Bomeli, 1971;

Benke and Rhode, 1980, and Bullen and Flamholtz, 1985).

Job satisfaction is a concept that differs across

individuals. Research evidence on job satisfaction can be

divided into categories such as leadership, psychological

needs, effort-reward, management ideology and values, and

job design factors and the content of work. Vroom [1964]

describes job satisfaction as having positive attitudes

towards one's job.

Linking the relationship between job performance and

job satisfaction is difficult because of the numerous

factors that affect the direction and magnitude of the

relationship [Vroom, 1964]. In their study on job

satisfaction and employee turnover, Bullen and Flamholtz

[1985] stated:

Other factors which may affect job satisfaction and turnover that have not been tested by the researchers include performance, ability, personality, and certain other characteristics.

geographical location, and general economic conditions, [p. 289]

A job satisfaction study concluded that job

performance causes job satisfaction [Vroom, 1964, p.

182]. Lawler and Porter [1967] took a position contrary

to Vroom [1964] and hypothesized that job satisfaction

directly affects job performance. In a subsequent study.

Porter and Lawler [1968] then agreed with Vroom [1964]

that job performance causes job satisfaction. Finally,

after offering conflicting theories on the causality

between the two constructs, Lawler [1971] discussed the

difficulty of testing and developing theories of job

satisfaction and job performance.

There have been few studies on the causality between

job satisfaction and job performance among accountants.

Using accountants as subjects for a study on job

satisfaction and job performance, Ross and Bomeli [1971]

stated:

. it is not clear whether high satisfaction with autonomy needs results in high performance and quality or whether high performance and quality result in high satisfaction. [p. 384]

In another study on job satisfaction among accountants,

Benke and Rhode [1980] stated:

. job satisfaction and performance are not necessarily positively related. . . . Consequently, job satisfaction [and] job performance . . . may be interrelated in a complex manner, [pp. 187-88]

While an investigation of causality for the job

satisfaction and job performance question is an important

one, it is an empirically unresolved issue. As a result

of the conflicting evidence the question of causality

between job satisfaction and job performance was not

resolved by this research study. Rather, the relationship

between these two constructs was investigated.

Need for the Study

The importance of maximizing job satisfaction and job

performance among employees within organizations is

continually sought by managers. Maximizing job

satisfaction and job performance may lead to higher

productivity, lower overall costs for the firm, and

reduced turnover of employees. Public accounting firms

experience annual non-partner turnover rates as high as

45%, primarily at the entry-level staff accountant

position [Bao, Bao, and Vasarhelyi, 1986]. Rhode,

Sorensen, and Lawler [1976] reported turnover rates of 85%

for initial hires over a ten year period. Lcimpe and

Earnest [1984] reported an annual turnover rate of 23.9%

for staff accountants who had one to three years'

experience with a firm. Their reported annual turnover

rate was higher than any of the prior five year annual

turnover rates. This indicates that accountants are

leaving firms at earlier stages in their career. The cost

of replacing an entry-level staff accountant has been

estimated at between $10,000 and $15,000 [Bullen and

Flamholtz, 1985]. While attrition may be necessary in the

public accounting profession due to the limited number of

supervisory and managerial positions, early detection of

job dissatisfaction and low levels of job performance,

which can be caused by other factors, can diminish the

premature or unexpected attrition of employees whom the

firm would like to retain. In addition, better

understanding of the relationship between job satisfaction

and job performance may lead to minimizing expenditures to

increase job performance, job satisfaction, and costs

associated with employee turnover.

Purpose of the Study

This purpose of this study was to contribute to the

academic literature in the areas of job satisfaction and

job performance, and provide useful information to

practitioners in public accounting firms. This will be

accomplished by investigating constructs that can affect

job satisfaction [Bullen and Flamholtz, 1985] and job

performance cimong staff accountants in a public accounting

firm. The effects on job satisfaction and job performance

of organizational commitment, professional commitment,

life-span career development, and self-monitoring were

examined. These constructs are discussed in depth in

Chapter II.

The relationships among these constructs are

presented in three research objectives: (1) investigate

the effects of organizational commitment, professional

commitment, life-span career development, and self-

monitoring on job satisfaction and job performance for

staff accountants in public accounting firms; (2)

investigate the relationship between job satisfaction and

job performance of staff accountants in public accounting

firms; and (3) investigate the relationship among

organizational commitment, professional commitment, and

life-span career development for staff accountants in

public accounting firms.

Organization of the Study

This study is further divided into five chapters.

Chapter II is a review of the relevant works for this

study relating to job satisfaction, job performance,

professional and organizational commitment, life-span

career development, and self-monitoring. Chapter III

contains the theoretical model, the research model, and

the development of the hypotheses. In Chapter IV, the

research design, the pretest results of the research

instrument, and statistical tools for this study will be

presented. Chapter V contains the statistical analysis of

this study. In Chapter VI the summary and conclusions are

presented.

CHAPTER II

LITERATURE REVIEW

Much of the research on job satisfaction and job

performance use these constructs as dependent variables

and investigate the impact of numerous independent 1

variables. As for the research evidence, subjects were

chosen typically based on the field or profession under

study. For example, accounting researchers used

accountants, psychologists tended to use students or the

general public, and management researchers used managers

as the subjects for their studies.

Except for self-monitoring, which has been used

primarily in psychological research, the constructs for

this study were selected from existing accounting research

evidence. For example, to determine the effects of

budgetary participation on attitudes and behavior,

Chenhall and Brownell [1988] examined the intervening

effect of role ambiguity on job satisfaction and job

performance. In a study where auditors were the subjects,

Choo [1986] used personality variables to examine their

1 Edwin Locke estimated a conservative figure of 3,350

articles and dissertations on job satisfaction from the period 1957-1972 [Dunnette, 1983]. This indicates the lack of conclusive evidence and continued interest in the area of job satisfaction.

relationship with role stress, and the effect of role

stress on job performance. Rasch and Harrell [1990]

investigated the impact of personal characteristics such

as job satisfaction, on determinants of job turnover.

Ferris [1981] investigated the effect of organizational

commitment on job performance eimong accountants in public

accounting firms. Harrell and Eickhoff [1988] researched

the effect of commitment on the level of job satisfaction

among auditors. The emphasis on accounting related

literature is because the sample under investigation for

this study were staff accountants in a public accounting

firm, and the relationship between this study and prior

research evidence was on constructs that may affect job

satisfaction and job performance.

This literature review is presented in the following

order: (1) job satisfaction and job performance;

(2) professional and organizational commitment, life-span

career development, and self-monitoring; and (3) a

discussion of how new empirical research evidence may

contribute to the accounting body of knowledge and the

accounting profession.

Job Satisfaction and Job Performance

Lawler and Porter [1967] recognized a consistent

relationship between job satisfaction and job performance

This relationship has been explained by the introduction

8

of such variables as intrinsic and extrinsic rewards

[Vroom, 1964], positive and punitive reward behavior [Sims

and Szilagyi, 1975], situational performance constraints

[Phillips and Freedman, 1984], and leadership style and

budgetary participation [Brownell, 1983; Collins, Munter,

and Finn, 1987; Chenhall and Brownell, 1988].

Job Satisfaction in the Accounting Profession

In general, accountants appear to be more satisfied

than other types of employees [Seller and Sapp, 1979]. A

study of job satisfaction among certified public

accountants used the needs hierarchy framework developed

by Maslow [1954] as a test of job satisfaction [Strawser,

Ivancevich, and Lyon, 1969]. The results of the study

were that accountants' needs satisfaction should be

investigated before developing motivational programs.

However, their study utilized a sample of middle- and top-

level accountants in certified public accounting firms.

Therefore, the results may not be considered generalizable

to staff accountants. Ross and Bomeli [1971] gave a

caveat on the Strawser et al. study of the complexities

and ambiguities of the variables job satisfaction and

motivation, which were not covered by Strawser et al. In

a review of previous studies that used Maslow's need

hierarchy, Wahba and Bridwell [1976] found little support

or consistency in the reported results. In fact, some of

the studies rejected Maslow's propositions [Huizinga,

1970; Alderfer, 1972].

Brenner, Carmack, and Weinstein [1971] applied the

Motivation-Hygiene Theory proposed by Herzberg, Mausner,

and Snyderman [1959] to determine which factors influence

job satisfaction among certified public accountants.

Motivation factors (recognition, achievement, work itself,

advancement, and responsibility) affect job satisfaction,

and hygiene factors (salary, company policies, technical

competence, interpersonal relations, and working

conditions) were found to cause job dissatisfaction.

Subjects in the study were randomly selected from the

population of certified public accountants. This

population would include accountants employed in public

accounting firms at different levels in the organization;

thus, the results are not limited only to staff

accountants.

Benke and Rhode [1980] studied personal

characteristics as indicators of job satisfaction. These

personal characteristics were cautiousness, original

thinking, personal relations, vigor, ascendancy,

responsibility, emotional stability, and sociability.

Subjects in the study were at the senior level or higher

in their accounting firms, perhaps providing some

limitation for generalizability to staff accountants.

10

However, their results indicated that personal

characteristics were statistically significant predictors

of job satisfaction for audit and tax specialists in

public accounting firms.

Ferris [1977a] tested perceived environmental

uncertainty and its effect on job satisfaction among staff

accountants (environment is defined as a set of stimuli to

which an organization is exposed and its response). He

hypothesized that environmental uncertainty will inversely

affect job satisfaction. Ferris found that as perceived

environmental uncertainty increased, job satisfaction

decreased.

Jiambalvo and Pratt [1980] hypothesized that the

leader behavior of in-charge auditors can affect the

satisfaction with supervision of their staff assistants.

The considerate style of leadership (mutual trust,

respect, and consideration) had a significant and positive

effect on staff assistants' satisfaction with supervision.

Senatra [1980] investigated the effects of role

conflict and role ambiguity on job satisfaction using

audit seniors in public accounting firms. He hypothesized

that role conflict and role ambiguity are negatively

related to job satisfaction. Senatra reported that only

the relationship between job satisfaction and role

ambiguity were significant and negatively related.

11

Senatra's results conflict with those of other researchers

using the same scale. Teas [1983] and Tracy and Johnson

[1981] reported that role conflict and role ambiguity are

a single measure of stress. Tracy and Johnson noted that

all of the items for role conflict were negatively worded,

while the items for role ambiguity were positively worded,

which may have contributed to the results. If the scales

were unidimensional, all of the items in each scale should

have related similarly to other items not in the scale.

Aranya, Lachman, and Amernic [1982] analyzed the

effects of organizational and professional commitment on

job satisfaction for non-partner chartered accountants.

They reported a statistically significant correlation

between organizational commitment and job satisfaction.

Professional commitment affected job satisfaction

indirectly through organizational commitment.

Harrell, Chewing, and Taylor [1986] examined the

effects of organizational and professional commitment on

job satisfaction. Their model differed from the model

used by Aranya et al. [1982] in that the effects of both

organizational and professional commitment on job

satisfaction were indirect through organizational-

professional conflict. The results of the study were

statistically significant for the multiplicative

interaction of the three variables—organizational

12

commitment, professional commitment, and organizational-

professional conflict.

Harrell and Stahl [1984] utilized McClelland's [1961]

trichotomy of needs theory (need for affiliation, need for

power, need for achievement) to explain job satisfaction

for certified public accountants employed in public

accounting firms. They found statistically significant

positive relationships between job satisfaction, need for

achievement, and need for power among junior-level

accountants in the audit/tax areas, but not among junior-

level accountants in the management consultant area. At

the partner and manager level, the results were also

statistically significant, except for need for achievement

as it related to the number of hours worked per week.

Bullen and Flamholtz [1985] attempted to predict

potential job turnover of non-partners in a certified

public accounting firm. They measured the level of job

satisfaction among non-partners, and then predicted

potential turnover based on the level of job satisfaction.

Bullen and Flamholtz reported statistically significant

results for predicting the probability of turnover based

on the level of job satisfaction. In their research

model, Bullen and Flamholtz theorized that personal and

psychological factors may affect job satisfaction.

However, they did not test these factors.

13

Reed and Kratchman [1987] reported that job

satisfaction was related to job attributes that fulfilled

individual needs, and one of the most important needs,

^^^^-f^lfillinent, can be satisfied through an accounting

position because the position allows for self-expression.

This may suggest a statistically significant correlation

between the constructs life-span career development and

job satisfaction.

While there have been studies on job satisfaction

using public accountants as subjects, these studies are

not all applicable to staff accountants, nor were all the

individual characteristics that can affect job

satisfaction considered. These limitations in empirical

accounting research can be attributed to the all-

encompassing nature of job satisfaction with many

variables, and the different levels and types of

accountants. The results of these empirical studies do

not necessarily conflict, but differ due to the employee

level of the subjects used in the studies (e.g., junior-

level accountants versus managers/partners), and the

variables employed to explain job satisfaction.

Job Performance in the Accounting Profession

Early studies on job performance by accounting

researchers were primarily laboratory experiments. Cook

[1967] found the frequency of feedback (communication)

14

affected the performance of students. Sorensen and Franks

[1972] used student subjects to correlate psychological

factors to performance. They tested personal

characteristics, ability and self-esteem, and how these

characteristics contributed to performance. They found

increasing self-esteem through supportive feedback could

lead to better performance. Decreasing self-esteem, which

occurred when neutral feedback was given, was found to

result in decreased performance.

Studies conducted after the laboratory experiments on

job performance among accountants were quasi-experimental.

Todd, Thompson, and Dalton [1974] reported that a strong

relationship between performance and rewards is necessary

to manage accountants effectively. However, Todd et al.

[1974] did not investigate which performance evaluation

dimensions that are considered important to supervisors

and subordinates.

Ramanathan, Peterson, and Maher [1976] reported that

better communication between supervisors and subordinates

is needed on job performance criteria. Juniors/seniors,

managers/supervisors, and managing partners had different

perceptions of the relative importance of the criteria

used to evaluate performance. Jiambalvo [1982] researched

audit seniors' (subordinates) perceptions of the

importance of performance evaluation dimensions with audit

15

partners' (supervisors) ratings of the importance of these

same dimensions. He reported that there was a difference

in what the subordinates and the supervisors considered

important. Audit seniors "overestimated" the importance

of working with the client's personnel, demonstrating an

area of special competence, and obtaining the respect and

cooperation of the client's personnel. Audit seniors

"underestimated" the importance of revising audit programs

and reviewing the work of the assistants on the audit.

Jiambalvo, Watson, and Baumler [1983] attempted to

determine what comprises a good evaluation of job

performance in a certified public accounting firm. Using

accountants who worked in the audit, tax, and management

services departments as subjects, Jiambalvo et al.

reported significant individual differences in the

contents of a good evaluation of job performance. These

differences were attributed to the subjects' placing too

much importance on evaluation categories that were not

considered important by their supervisors, and too little

importance on areas considered important by their

supervisors. They also reported that there were

differences in performance evaluation decision making

among the audit, tax, and management service areas of the

firm. Even though these studies indicated that

performance evaluations are an important factor in an

16

accountant's behavior, Wright [1980] reported, on average,

only five hours were used to train seniors on job

performance evaluations of subordinates.

In another study, Wright [1982] reported that

technical ability was a primary focus for appraising staff

personnel's job performance. One reason could be the

difficulty in evaluating other factors that are important

in job performance. Arrington, Bailey, and Hopwood [1985]

reported that audit behavior, audit procedures, and audit

tasks were used as judgment factors for evaluating

performance. These studies indicated that there are

differences in the perceived importance of criteria used

in performance evaluations between subordinates and

supervisors, and between areas in public accounting firms.

Ferris [1982] reported that the accounting

organization's ability to adapt to the outside environment

(environment is defined as a set of stimuli to which an

organization is exposed and its response) was directly

related to employee job performance. As the level of

coping with the outside environment increased for the

organization, employee job performance also increased.

However, Ferris used employee performance as a surrogate

measure for the coping effort. He commented that there

could be a potential problem using job performance as a

surrogate measure for the coping effort. This potential

17

problem is because of the effects of other variables that

can affect employees performance not included in the

study, but not the coping effort.

Chow [1983] explored the effects of job standard

tightness and the type of compensation scheme on job

performance. He hypothesized that job performance will be

directly and interactively affected by the compensation

scheme and job standard tightness. There were

statistically significant results for the direct effects

of compensation schemes and job standards tightness on job

performance, in groups where subjects were assigned to

compensation schemes. There were no significant

interactive effects of compensation scheme and job

standard tightness on job performance. Job performance

was enhanced when subjects chose their own compensation

schemes. Chow felt that the results indicated that job

standards and compensation schemes may play significant

screening and motivational roles.

Ferris and Larcker [1983] reported that rated

performance was a function of motivation and

organizational commitment, while rewarded performance

(compensation) was a function of both the interpersonal

attraction between the auditor and his/her supervisor, and

the task-related ability of the individual. This is an

important study because it distinguished between some

18

variables that can affect rated job performance and

rewarded job performance. Rewarded job performance can be

biased based on the interpersonal attraction between the

auditor and supervisor.

A primary difference between the Chow [1983] study

and the Ferris and Larcker [1983] study was the way

compensation schemes were distributed and the experimental

design of the studies. The Chow study was a laboratory

study in which subjects had their choice of compensation

schemes, and these schemes were not based on a

supervisor's recommendation. Ferris and Larcker used a

public accounting firm in which compensation was assigned

based on a number of different factors, one of the primary

factors being a supervisor's recommendation.

There have been studies directly relating the level

of job performance among staff accountants to other

variables. Ferris [1977b] applied expectancy theory to

staff accountants in two public accounting firms and

reported results that showed expectancy theory to be a

poor predictor of job performance. Ferris [1981] studied

the effects of organizational commitment on job

performance. This relationship was hypothesized to be

moderated by an individual's task-related ability, which

is consistent with prior Expectancy Theory research

[Lawler, 1966]. Subjects were junior-level and senior-

19

level auditors who were employed for at least one year by

a large public accounting firm. Ferris reported that

there was no statistically significant relationship for

the moderating effect of task-related ability. Ferris

attributed this result to the limitations in the

measurement of task-related ability. However, junior-

level auditors' performance was influenced by a

willingness to exert effort on behalf of the organization

(organizational commitment).

Jiambalvo [1979] departed from the traditional

expectancy model to investigate performance evaluation and

job effort cimong auditors. He departed from the

traditional expectancy model by dividing it into two

components, and adding sequential steps in the first

component. The first component was the linkage between

efforts and rewards in a sequential evaluation context.

The sequential steps were effort, performance, evaluated

performance, overall evaluation, and rewards. By adding

the three intermediate steps between effort and reward,

Jiambalvo wanted to identify where the effort—>reward

process may break down. The second component was the

decomposition of job effort into performance evaluation

dimensions. Using the sequential evaluation steps,

Jiambalvo reported a large difference between self- and

manager-rated performance. He stated this was

20

"unanticipated theoretically, but is a typical finding in

expectancy studies" (p. 449). Jiambalvo reported that

senior auditors directed their job effort on activities

that were used to evaluate their performance.

Choo [1986] researched the relationship between job

performance and stress among practicing chartered

accountants. Choo found the relationship between job

performance and job stress to be an inverted U-shaped

function. As job stress increased, so did job

performance, to a maximum point; after that point, job

performance decreased as job stress increased.

Like job satisfaction, job performance is an area in

which there have not been many empirical studies that used

staff accountants as subjects. However, there has been a

recent trend to use non-technical individualistic factors

in studies concerning accountants' job performance (for

example, Ferris [1982]; Ferris and Larcker, [1983]; Choo

[1986]; Rasch and Harrell [1990]). This trend indicates a

continued interest in job performance research.

Organizational and Professional Commitment> Life-Span Career Development,

and Self-Monitoring

Organizational and Professional Commitment in the Accounting Profession

Public accounting was accepted as a profession in

1890, when certification of accountants began [Freidson,

21

1973]. The label "professional" has implications for an

individual at the organizational and occupational level.

A level of behavior is expected by the organization

employing the professional, as well as by the external

peer group that makes up the profession [Harrell, Chewing,

and Taylor, 1986]. The extent to which individuals behave

in the expected manner can be reflected in their

commitment to the organization and profession. Commitment

can be defined as: (1) a belief in and acceptance of the

goals and values of the organization and/or profession;

(2) a willingness to exert considerable effort on behalf

of the organization and/or profession; and (3) a desire to

maintain membership in the organization and/or profession

[Aranya, Pollock, and Amernic, 1981a]. The following

studies had results that showed significant relationships

among the variables organizational commitment and

professional commitment, with job satisfaction, job

performance, and career intentions.

Aranya et al. [1981a] reported organizational and

professional commitment to be compatible due to the

positive relationship between the two variables. They

modeled organizational commitment as an independent

variable used to predict professional commitment in

chartered accountants employed at public accounting firms

in Canada. At the level of partner, organizational

22

commitment was reported to be greater than professional

commitment. At firm levels below partner, Aranya et al.

reported that there is lower organizational commitment

than professional commitment. This may indicate that

staff accountants have yet to reach the stage in their

career development to commit to an organization as a

longtime employee, but are committed to a career in public

accounting. Aranya, Lachman, and Amernic [1982] used

professional and organizational commitment as predictors

of job satisfaction. One of the differences between the

Aranya et al. [1982] study and the Aranya et al. [1981a]

study, was the reversal of organizational commitment from

an independent to a dependent variable, and professional

commitment from a dependent to an independent variable.

The results in that study were that professional

commitment precedes organizational commitment, and both

exhibit significant influences on job satisfaction.

Lachman and Aranya [1986] provide further empirical

support that professional commitment precedes

organizational commitment, and that the two constructs are

positively correlated, as previously reported by Aranya et

al. [1982],

Norris and Niebuhr [1983] also supported the

conclusions of Aranya et al. [1981a and 1982], that

organizational and professional commitment are correlated.

23

and that there are statistically significant correlations

between job satisfaction and both professional and

organizational commitment. They reported organizational

commitment to be greatest at the partner level, and

professional commitment to be greater than organizational

commitment at most levels below partner. However, Norris

and Niebuhr used subjects at all levels of different

public accounting firms. Their results merged the

different levels and firms, which makes generalizing the

results to only staff accountants difficult.

Harrell et al. [1986], using internal auditors in

their sample, reported that organizational commitment

positively affects job satisfaction. For individuals who

were members of the Institute of Internal Auditors, a

professional organization, professional commitment was

more positive than for internal auditors who were not

members. As with the Norris and Niebuhr [1983] study, the

generalizability of the results reported by Harrell et al.

is limited because the sample consisted solely of internal

auditors, and not staff accountants in public accounting

firms.

Harrell and Eickhoff [1988] reported higher levels of

organizational commitment and job satisfaction among

auditors who were influence-oriented, or attracted to

working in a competitive environment in which leadership

24

status can be achieved, than those who were not influence-

oriented. They concluded that the auditors who were

influence-oriented had more positive "Big Eight" career

intentions than the auditors who were not influence-

oriented.

Ferris [1981] reported that at junior-level positions

(e.g., staff accountants), organizational commitment

influenced job performance by the individual's willingness

to exert effort on behalf of the organization—that is, to

perform well on the job. At both the senior and junior

staff accountant levels, organizational commitment is

positively related to job performance. For employees

beyond the level of staff accountant, high levels of

organizational commitment were due to a desire to remain

with the firm. Ferris states that these differences are

because of the nature of organizational commitment, which

changes over time. Hence, as an individual becomes more

positively entrenched in his or her career development,

the level of organizational commitment should increase.

Aranya and Ferris [1984] reported that certified public

accountants employed by public accounting firms had higher

levels of organizational and professional commitment,

compared to public accountants working in nonprofessional

organizations. As with the Ferris [1981] study, Aranya

and Ferris reported lower levels of organizational and

25

professional commitment at the staff accountant level than

at the manager and partner levels.

The results of these studies provide an indication

that a relationship exists among the variables job

performance, job satisfaction, career development,

organizational commitment, and professional commitment.

Professional commitment precedes organizational commitment

in an individual's career [Aranya, Lachman, and Amernic,

1982; Lachman and Aranya, 1986], organizational commitment

is an indicator of career intentions [Harrell and

Eickhoff, 1988], and organizational commitment and

professional commitment can affect the level of job

satisfaction [Sorensen and Sorensen, 1974; Aranya,

Lachman, and Amernic, 1982; Harrell and Eickhoff, 1988]

and job performance [Ferris, 1981].

Life-span Career Development

Life-span career development is a concept that

considers an individual's career objectives to be an

ongoing process. As congruence grows between individuals

and their work environment, there is greater satisfaction

on the job and with the chosen career. Super [1963] has

identified attitudes and behaviors that comprise five

stages of vocational development. These five stages are

crystallization, specification, implementation,

stabilization, and status and advancement.

26

Crystallization is the stage in which an individual

formulates thoughts about the type of suitable work for

himself/herself. This stage will usually occur between

the ages of fourteen to eighteen. The second stage is

the specification of a career. This stage requires an

individual to make a specific choice of a career and take

the steps necessary to implement the choice.

Specification will usually occur between the ages of

eighteen to twenty-one. The third stage, implementation,

will usually occur between the ages of twenty-one to

twenty-five. It is during this time that an individual

will enter employment based on the career choice and

complete some training in his/her chosen career. The

fourth stage is stabilization. This stage will usually

occur between the ages of twenty-five to thirty-five. An

individual will settle down within the field of work

he/she has chosen. The last stage is status and

advancement. This will usually occur between the late

thirties to the mid-forties. During this stage, an

individual will attempt to achieve a comfortable and

secure position in the field of work he/she has chosen.

The rate and level of an individual's development in

his/her career is known as vocational maturity. Table 2.1

summarizes these stages.

27

TABLE 2.1 LIFE-SPAN CAREER DEVELOPMENT STAGES AND AGES

Vocational Development Stages Ages

Crystallization 14-18

Specification 18-21

Implementation 21-25

Stabilization 25-35

Status & Advancement 35-49

The age levels for these different stages were

questioned by Super [1981] after further research on

vocational development. He believed that vocational

maturity is not necessarily something that will increase

with age, because of two factors. The first factor is the

changing labor market. The advances in technology make

for uncertain jobs and professions in the future. While

technology has changed the way accountants perform their

jobs (for example, paperless audits), it does not seem

likely that technology will make accounting, an uncertain

profession. Second, there is a lack of empirical evidence

that shows an increasing ability to cope with career

developmental tasks as age increases.

Life-span career development is not specifically

referred to in the accounting literature. However,

researchers in psychology have studied the effects

of career development on job satisfaction and found that

28

vocational maturity generally predicts occupational

satisfaction [Osipow, 1983]. Studies on careers using

accountants as subjects have focused on career intentions

rather than the stage or level of life-span career

development.

Sorensen [1970] hypothesized that there was conflict

between the educational subculture of an accounting

graduate and the professional subculture of the graduate.

Without an understanding of the differences, the graduates

have less chance of fulfilling their career objectives.

This conflict is particularly applicable to staff

accountants, especially those who enter public accounting

upon graduation. This conflict occurs in what would be

considered the implementation stage of vocational

development.

Lengermann [1971] studied professional autonomy among

accountants. Professional autonomy is the freedom allowed

professionals to do their work in the manner they feel is

best, according to the accepted standards of their

profession. Lengermann [1971] reported that professional

autonomy was a positive resource, especially among lower-

level certified public accountants, such as staff

accountants in public accounting firms.

DeCoster and Rhode [1972] reported that younger

accountants were less conservative, more aggressive, and

29

less conforming than older accountants. The variation in

these traits over time could be attributable to Super's

stages of vocational development. Using Holland's [1973]

theory of vocational choice, Aranya, Barak, and Amernic

[1981b] reported accountants to be conventional,

enterprising, and social. Mossholder, Bedeian, Touliatos,

and Barlonan [1985] reported that accountants are

distinguishable in terms of personality, outcome

preferences, and perceived work climate. Zytowski and Hay

[1984] were hesitant to generalize about similarities

found within occupations, including accountants, because

they found differences cimong members of the same

occupation. One of the sources of these differences is

the stage of vocational development.

Lawler, Kuleck, Rhode, and Sorensen [1975] studied

job choice and post-decision dissonance among college

graduates. The study reported negative attitudes that

recent graduates had towards the firms during the initial

employment period. The attractiveness in working at any

public accounting firm decreased after about one year.

There are many reasons for these results, two of which

could be dissatisfaction with the organization or

profession. This would occur during the implementation

stage of vocational development.

30

Dillard [1979] used the variables goal and expectancy

to investigate the occupation-position choice decision

among public accountants. Occupation and position choices

are derived from an individual comparing alternative

occupations and positions to his/her current work

environment. These comparisons lead to an individual

determining his/her goals, and ultimately behavior.

Expectancy is the subjective probability that an

individual could attain the occupation-position he/she

desires. Expectancy is multiplied by the utility an

individual believes the occupation-position will provide,

which gives the expected utility position. Dillard

reported that individuals' occupation-position goals are

significantly related to the position with the highest

expected utility. Individuals will attempt to achieve the

occupation-position that will give them the highest

utility, based on the probability of achieving the

occupation-position.

Reed and Kratchman [1987] reported that job

satisfaction was related to the attributes of the job that

fulfilled individual needs. They further stated that one

of the most important needs, self-fulfillment, can be

satisfied through an accounting position because the

position allows for self-expression. Harrell and Eickhoff

[1988] reported that influence-oriented auditors have more

31

positive feelings toward "Big Eight" career intentions

than other auditors.

There is evidence in the psychological literature

that indicates a correlation between career stage and job

satisfaction [Osipow, 1983]. However, the accounting

literature is lacking empirical research on the stages of

career development and its effect on job satisfaction and

job performance. The first career stage in which job

satisfaction or job dissatisfaction should appear is

during the implementation stage. It is during this stage

that individuals will usually first encounter the work

place that was their career choice.

Self-Monitoring

Self-monitoring is the process by which individuals

observe and control how others perceive them. This is

done by attempting to create images that seem appropriate

for a given set of circumstances [Snyder, 1979]. Self-

monitoring is divided into two conditions, high and low.

A high self-monitor attempts to determine what type of

characteristics a situation calls for, adopts those

characteristics, and then acts accordingly. Low self-

monitors do not attempt to change their characteristics to

adapt to a situation, but act in a manner in which they

are most comfortable. Researchers have shown through

empirical studies that self-monitoring is a good

32

explanatory variable of attitudes and behavior in

different situations.

Because self-monitoring has been shown to explain

attitudes and behavior in different situations, it is a

variable of interest because of the nature of a staff

accountant's work. Staff accountants freguently change

job locations and job supervisors. Their perceived

ability to adapt may influence their level of job

satisfaction. Staff accountants' feelings that they are

adapting sufficiently may lead to greater levels of job

satisfaction. Self-monitoring may also affect perceived

job performance in the same manner as job satisfaction.

Self-monitoring is a behavioral concept that has been

investigated extensively in psychology (e.g., see Snyder

1974, 1979, 1983; Snyder, Berscheid, and Click, 1985;

Snyder and Cantor, 1980; Snyder and Gangestad, 1982;

Snyder, Gangestad, and Simpson, 1983; Snyder and Monson,

1975; Snyder and Simpson, 1984; Snyder and Swann, 1976;

Snyder and Tanke, 1976; Snyder, Tanke, and Berscheid,

1977; Ajzen, Timko, and White, 1982; Briggs, Cheek, and

Buss, 1980). A thorough review of the accounting

literature did not reveal any references to this concept;

therefore, the following literature review on self-

monitoring presents the results from selected

psychological studies.

33

A study of attitude and behavior using self-

monitoring was done by Snyder and Swann [1976]. They

tested whether individuals' behaviors were indicative of

their attitude after they were given an opportunity to

express their opinion on a social issue. There was a

higher covariance between low self-monitors and their

attitude and behavior than high self-monitors.

Snyder and Tanke [1976] further tested consistency

between attitude and behavior among low and high self-

monitors using essay writing. In the first phase of the

study, subjects wrote counterattitudinal essays under two

conditions, choice and no choice. As compared to high

self-monitors, low self-monitors who chose to write

counterattitudinal essays had final attitudes that were

more like their behavior. In the next phase of the study,

subjects were given more choice in their topic so there

would be a wider range of proattitudinal and

counterattitudinal essays. By allowing subjects the

choice of writing counterattitudinal essays, Snyder and

Tanke hypothesized that low self-monitors would be more

likely to change their attitudes than high self-monitors.

Nonparametric tests were used to determine the proportion

of individuals who changed their attitude. Approximately

eighty-three percent of low self-monitors showed a change

in their initial and final attitudes, while only forty-

34

five percent of high self-monitors showed an attitude

change. The results were possibly due to high self-

monitors attributing their essays to the experiment

(situation). Low self-monitors would attribute their

responses to their character (disposition).

Snyder [1983] researched situational and

dispositional individuals and social behavior. Low self-

monitors are likely to be less responsive to situations,

while high self-monitors are likely to be more responsive.

Low self-monitors are characterized by attitude-behavior

consistencies, and seek out situations that will conform

with their current behavior.

Ajzen et al. [1982] also tested differences in

attitudes and behaviors between low and high self-

monitors. One of the results of their experiment showed a

difference in intention-behavior between the two groups.

As compared to that of high self-monitors, the behavior of

low self-monitors was more accurately predicted based on

their intentions. One reason suggested for this finding

is that low self-monitors are not as affected by external

events; therefore, they are more stable and more

predictable.

Kulik and Taylor [1981] investigated how high and low

self-monitors use desirable and undesirable consensus

information to make predictions about their own and

35

others' behavior. Overall, high self-monitors used

consensus information more than low self-monitors.

However, this was significant only for undesirable

information. Rather than use consensus information, low

self-monitors will consider their own opinions to predict

behavior.

Snyder and Gangestad [1982] tested the attitude of

high and low self-monitors who were entering different

social situations. The first phase of their study gave

individuals, both low and high self-monitors, the choice

to enter a situation that required being sociable. In

some cases the type of sociability was prescribed; in

other cases it was not prescribed. It was expected that

high self-monitors would be more willing to enter the

well-defined situation than the less defined situation.

Snyder and Gangestad also expected low self-monitors to

base their decision on their introverted or extraverted

characteristics, not the amount of information given about

the social situation. The evidence supported expectations

regarding high self-monitors. They reported that both

extraverted and introverted high self-monitors were

willing to spend time in the social situations. Neither

extraverted not introverted low self-monitors as a group

were inclined to spend time in the social situations. Of

those individuals who were willing, extraverts were more

36

willing than introverts to spend time in social

situations.

The second phase of their study asked high and low

self-monitors how a situation should be changed so they

would participate in it. High self-monitors more clearly

defined the situation they would be willing to enter than

low self-monitors. Low self-monitors defined the

situation in terms of their introverted and extraverted

characteristics. Snyder and Gangestad concluded that

social behavior can also be a function of the " . . .

processes by which the individual finds himself or herself

in that situation" (p. 134). Snyder [1983] discussed this

in more detail. He felt the situation appeared to

influence the individual, but the more important influence

is the activity of the individual.

In another study, low self-monitors were more

committed in dating relationships than high self-monitors

[Snyder and Simpson, 1984]. Low self-monitors were more

reluctant to change partners, they dated fewer individuals

in the same period of time than high self-monitors, and

the length of and intimacy in a relationship was greater

among low self-monitors. The results were shown to have

two potentially significant implications. First, low

self-monitoring individuals find romantic relationships to

be an important part of what makes their lives meaningful.

37

Second, the results can help one understand other

relationships that go beyond dating, such as marriage.

Snyder and Cantor [1980] tested self-monitoring and

social knowledge. Theoretically, low self-monitors should

be knowledgeable about their own traits, while high self-

monitors would be knowledgeable about the traits of

prototypical individuals. The experiment supported the

theory. Low self-monitors were able to give more

information about their own traits, while high self-

monitors provided more information about the traits of

others. The result suggests that when a certain social

behavior is called for, low self-monitors have more

ability to look to themselves for guidance on how to act.

High self-monitors would consider the person (real or

imagined) whose behavior is appropriate for the situation

as the proper model for their own behavior.

Self-monitoring has been found to be a significant

factor during initial interactions between individuals.

During initial interaction, conversation tends to be

highly structured [Douglas, 1983]. Douglas tested high

and low self-monitors' responses to interaction in an

initial encounter between a male and a female. The

encounters could be typical—a normal conversation—or

atypical, where the individuals were abrasive and previous

mental problems were revealed. High self-monitors

38

generated more responses to the typical conversation than

the atypical conversation. Individuals categorized as

high self-monitors also generated more responses than low

self-monitors who viewed the typical initial encounter.

Low self-monitors' responses to the typical and atypical

conversations were about the same. Overall, responses to

the atypical conversation were similar between the low and

high self-monitors.

Douglas [1983] believed that high self-monitors may

not process more information, but use more information

than low self-monitors. This could lead to friendlier

interaction by high self-monitors than low self-monitors.

To test this hypothesis, he had independent judges rate

the replies of the subjects. High self-monitors were

found to be more likely to have friendly interactions than

low self-monitors. He continued his research in initial

interaction by testing whether cognitive differences are

associated with behavioral differences at different levels

of self-monitoring. Summaries of high self-monitors'

conversations and behavior were different than those of

low self-monitors. High self-monitors tended to view

their conversation as strategic, whereas low self-monitors

saw their conversations as a way of achieving interaction

objectives.

39

Snyder et al. [1985] investigated how the selection

process of one type of initial interaction, dating,

differed between low and high self-monitors. They

hypothesized there would be different approaches between

low and high self-monitors in choosing a dating partner.

Males were given portfolios containing a picture and

information about females they could choose for a date.

The males' process for choosing a partner was studied and

compared to their level of self-monitoring. As

hypothesized, low self-monitors spent more time looking at

the information about the attributes of the female rather

than the pictures. High self-monitors spent more time

looking at the photographs rather than the information.

This could be because the high self-monitors believe that

this was what was expected of them, and then conform their

behavior accordingly. Seventy-five percent of low self-

monitors reported that interior attribution information

was the most important reason for choosing a partner.

Seventy-seven percent of high self-monitors reported the

pictures were the most important factor in making their

selection. Overall, low self-monitors spent more time in

the selection process than high self-monitors, though both

groups looked at about the same number of files.

The results of the Douglas [1983] and Snyder et al.

[1985] studies appear to be contradictory. In the Douglas

40

study, high self-monitors used more information, while in

the Snyder et al. study low self-monitors spent more time

looking at information. However, in the Douglas study an

actual oral exchange of information took place, whereas

the Snyder et al. study had subjects only look at

information.

Interpersonal skills needed to initiate and preserve

a relationship can be discussed in terms of humor

production [Turner, 1980]. High self-monitors produced

more humorous material in terms of both quantity and

quality. This was measured by writing cartoon captions

and group discussions. When a situation calls for a

humorous remark, as can be required when initiating or

maintaining a relationship, the high self-monitor is

better prepared to deliver the remark.

Ickes and Barnes [1977] tested behavior during

initial interactions in same-sex dyadic combinations of

three levels of self-monitoring—low, medium, and high.

The results of any pair who had previously met were

excluded from the data to ensure only initial interactions

were being tested. In both male and female pairs in which

talking occurred, the high self-monitor initiated the

conversation significantly more often than the low self-

monitor. During the initial interaction, dyads in which

low and high self-monitors were paired showed more periods

41

of silence than any other combination of self-monitors.

Female dyads showed more expressive behavior during the

initial interaction than male dyads. Expressive behavior

was measured by gazes, verbalization, and expressive

gestures. Ickes and Barnes felt dyads of high and low

self-monitors had more interaction difficulties than other

pairs. Their results supported the construct of self-

monitoring being concerned with appropriate behavior in

social interaction.

Researchers have shown through empirical studies that

self-monitoring is a good explanatory variable of

attitudes and behavior in different situations. Results

of empirical studies using self-monitoring as a variable

have been supported by the theory of self-monitoring.

However, there has not been published empirical research

using self-monitoring as an explanatory variable for an

individual's satisfaction and performance at work.

Chapter Summary

The large amount of literature on job satisfaction

and job performance shows a continuing interest and a need

for understanding these two constructs (see Footnote 1,

Chapter II). Accounting literature on job satisfaction

and job performance is not as prevalent nor is it focused

on specific accounting occupations or positions in public

accounting firms, as evidenced by this literature review.

42

However, there have been some studies published recently

in the accounting literature on job satisfaction and job

performance [for example, Bullen and Flamholtz, 1985;

Kida, 1985; Arrington et al., 1985; Ferris and Larcker,

1983; Harrell et al., 1986; Choo, 1986; Reed and

Kratchman, 1987].

A research trend in the areas of job satisfaction and

job performance among accountants is to investigate the

effects of personality variables. In their study on job

satisfaction, Bullen and Flamholtz [1985] identified a

large number of factors, including personality variables,

that can be used to study job satisfaction. Choo [1986]

stated that previous studies on auditors tended to focus

on variables other than personality variables that could

be significant in understanding behavior. Rasch and

Harrell [1990] investigated the impact of several personal

characteristics of accounting professionals on

determinants of personnel turnover, including job

satisfaction.

An interesting area for research in accounting is to

study the effects of personality variables on job

satisfaction and job performance among staff accountants.

Chapter III contains the theoretical and research models,

and the hypotheses and justification for testing

43

personality variables that may affect the level of job

satisfaction and job performance among staff accountants

44

CHAPTER III

THE THEORETICAL MODEL, RESEARCH MODEL,

AND FORMULATION OF HYPOTHESES

The Theoretical Model

Gruneberg [1976] broadly categorized job satisfaction

to be associated with the job itself, environmental

factors, and individual factors. Landy, Zedeck, and

Cleveland [1983] used organizational considerations,

individual considerations, and sociopolitical

considerations as the primary factors affecting job

performance. Different theories behind the relationship

between job satisfaction and job performance have been

well-documented (for example, Vroom, 1964; Lawler and

Porter, 1967; Porter and Lawler, 1968; Lawler 1971; Ross

and Bomeli, 1971; Benke and Rhode, 1980; Harrell and

Stahl, 1984; Bullen and Flamholtz, 1985). There is no one

theory to explain this relationship because of the large

number of variables that affect job satisfaction and job

performance.

Vroom [1964] discussed the difficulty of describing

the relationship between job performance and job

satisfaction because of the numerous factors that affect

the direction and magnitude of the relationship. In this

study, the factors of interest are the constructs

45

organizational commitment, professional commitment, life­

span career development, and self-monitoring.

The model developed by Bullen and Flamholtz [1985]

shows the complexity of attempting to explain the causes

of job satisfaction because of the numerous variables

affecting this construct. Figure 3.1 is a portion of the

theoretical model developed by Bullen and Flamholtz.

Overall Assessment of

Job Satisfaction /

\

\

/

Area Level

Career Goals Demographic s Other Factors

Figure 3.1. The Theoretical Model

"Other Factors" include performance, personality, and

certain other personal characteristics [Bullen and

Flamholtz, 1985]. Investigating the interrelationships

among the constructs is the primary purpose of this study.

The Research Model

Researchers have theorized and empirically tested the

relationship among job satisfaction, job performance, and

other variables [for example, Bagozzi, 1980a]. The

research model used in this study was adapted from the

Bagozzi [1980a] research model, and based on the theory

proposed by Bullen and Flamholtz [1985]. The first

46

research objective of this study was to investigate the

hypothesized (H) effects of organizational commitment,

professional commitment, life-span career development, and

self-monitoring on job satisfaction and job performance

for staff accountants in public accounting firms. Figure

3.2 is the research model for these relationships.

Professional Commitment Organizational Commitment Life-Span Career Development

Self-Monitoring

Job Satisfaction

Job Performance

Figure 3.2. The Research Model for H1-H8

The second research objective was to investigate the

relationship between job satisfaction and job performance

of staff accountants in public accounting firms. This

relationship is illustrated in Figure 3.3.

Figure 3.3. The Research Model for H9

47

Research objective three was to investigate the

hypothesized relationship among organizational commitment,

professional commitment, and life-span career development

for staff accountants in public accounting firms. These

relationships are illustrated in Figure 3.4.

Professional Commitment

/

\

\

/

Life-Span Career

Development

\ /

\ /

Organizational Commitment

Figure 3.4. The Research Model for H10-H12

Formulation of the Hypotheses

The research objectives stated in Chapter I were to:

(1) investigate the effects of organizational commitment,

professional commitment, life-span career development, and

self-monitoring on job satisfaction and job performance

for staff accountants in public accounting firms; (2)

investigate the relationship between job satisfaction and

job performance of staff accountants in public accounting

firms; and (3) investigate the relationship among

organizational commitment, professional commitment, and

48

life-span career development for staff accountants in

public accounting firms.

The hypotheses for the first research objective are

predicated on the basic theory that there are many

variables such as personality and certain other personal

characteristics that can affect job satisfaction and job

performance. The first four hypotheses are based on the

effects of organizational commitment and professional

commitment on job satisfaction and job performance.

Models illustrating commitment have antecedents that

precede commitment, and job-related outcomes resulting

from the level of commitment individuals possess [Hunt,

Chonko, and Wood, 1985]. These job-related outcomes

include performance and satisfaction [Hunt, Chonko, and

Wood, 1985]. Ferris [1981] hypothesized that

organizational commitment preceded job performance among

auditors. He reported that junior-level auditors'

performance was influenced by a willingness to exert

effort on behalf of the organization. Few empirical

accounting studies have examined the effects of both

organizational and professional commitment on job

satisfaction. Norris and Niebuhr [1983] reported that

organizational and professional commitment were strongly

related to job satisfaction for individuals employed in

public accounting firms. However, the results reported by

49

Norris and Niebuhr [1983] were for all levels of employees

in all functional areas of the firm. The hypotheses, in

alternate form, for the relationship between

organizational and professional commitment with job

satisfaction and job performance are:

HI: Job satisfaction and organizational commitment will be positively related for a staff accountant.

H2: Job performance and organizational commitment will be positively related for a staff accountant.

H3: Job satisfaction and professional commitment will be positively related for a staff accountant.

H4: Job performance and professional commitment will be positively related for a staff accountant.

Super [1981] stated that there is a lack of empirical

studies on vocational maturity in adulthood. To

investigate vocational maturity in adulthood, these

hypotheses are based on the effects of life-span career

development on job satisfaction and job performance. The

rate and level of an individual's development in his/her

career—vocational maturity—has an effect on satisfaction

[Osipow, 1983]. The relationship between job performance

and life-span career development was proposed by Super

[1957] when he stated that,

. the learning of job or position performance is necessarily much more specific than in the learning of occupational performance (and) virtually all workers have some problems of induction and

50

transition when they begin a new job, and in the inexperienced worker these are often of some significance, [p. 119]

The implementation stage of life-span career

development was investigated. Theoretically, the

implementation stage will occur between the ages of

twenty-one to twenty-five. This is the usual age in which

an individual will enter the accounting profession as a

staff accountant, and go through the induction and

transition processes. Super [1957] theorized a

relationship between life-span career development and both

job satisfaction and job performance during the

implementation stage when he stated that

. [recent college graduates] change in search of something more satisfying to themselves [job satisfaction] and in which they are more satisfactory to employers [job performance], [p. 129]

The hypothesized relationships, in alternate form, among

life-span career development, job satisfaction, and job

performance are:

H5: A positive relation exists between job satisfaction and the level of the implementation stage in a staff accountant's life-span career development.

H6: A positive relation exists between job performance and the level of the implementation stage in a staff accountant's life-span career development.

Hypotheses seven and eight are based on the

relationship between job satisfaction and self-monitoring.

51

and job performance and self-monitoring. Self-monitoring

is used in psychology to explain different attitudes and

behaviors involving interactions.

Because of staff accountants' numerous interactions

with others, self-monitoring should be a construct for

inclusion in this study on job satisfaction and job

performance. If staff accountants who are low self-

monitors do not feel that they fit in a continually

changing work environment, they may be less satisfied with

their job. Their perceived level of job performance can

also be lessened by their perceived lack of adaptability.

Douglas [1983] reported that high self-monitors view

conversations (interactions) as being strategic. Staff

accountants who are high self-monitors should be more

strategic and adaptable than low self-monitors, which can

lead to greater levels of job satisfaction and higher

perceived job performance. Ferris [1982] reported that an

accounting organization's ability to adapt to the outside

environment was directly related to employee job

performance. The ability of staff accountants to adapt to

their environment may also affect their level of job

performance and job satisfaction. Hypotheses seven and

eight are presented below in alternate form.

H7: A positive relation exists between the level of self-monitoring and job satisfaction for a staff accountant.

52

H8: A positive relation exists between the level of self-monitoring and job performance for a staff accountant.

The results of previous empirical research studies

show that job satisfaction and job performance tend to

correlate significantly [Bagozzi, 1980]. This hypothesis

results in theorizing a nonrecursive model in which the

interrelationship between the two constructs is

investigated. This differs from a recursive model in

which one of the constructs precedes or "causes" the

second construct. A nonrecursive model is used because

the theories regarding the causal relationship between job

satisfaction and job per*formance are uncertain and

conflicting [Vroom, 1964; Lawler and Porter, 1967; Porter

and Lawler, 1968; Lawler, 1971; Ross and Bomeli, 1971;

Benke and Rhode, 1980]. The second research objective

hypothesizes, in alternate form, the following:

H9: There is a positive nonrecursive relation between a staff accountant's job satisfaction and job performance.

As stated previously, professional commitment has

been found to precede organizational commitment,

supporting hypothesis ten [Aranya, Lachman, and Amernic,

1982; Lachman and Aranya, 1986]. Hypotheses eleven and

twelve are based on the level of satisfaction an

individual has with the career choice of becoming a

certified public accountant. If an individual is

53

satisfied with his/her career choice, professional

commitment and organizational commitment should be

significantly related in a positive manner with life-span

career development. The last research objective

hypothesizes, in alternate form, the following:

HIO: A staff accountant's organizational commitment will be directly predicted by professional commitment.

Hll: There is a positive relation between professional commitment and life-span career development for a staff accountant.

H12: There is a positive relation between life-span career development and organizational commitment for a staff accountant.

Chapter Summary

Figure 3.5 is the research model for the twelve

hypotheses presented in this chapter. These hypotheses

are based on the theoretical model developed by Bullen and

Flamholtz [1985]. In addition to the twelve hypotheses

under investigation, demographic data such as sex, age,

and years as a staff accountant were collected. While

there are no hypotheses presented for the demographic

data, analyses of the data will be conducted. Chapter IV

contains the research design, the results of pretesting

the research instrument, and the statistical analytical

technique for this study.

54

Figure 3.5. The Research Model for all Hypotheses

55

CHAPTER IV

RESEARCH METHODOLOGY

Having discussed the continuing interest in the

relationship between job satisfaction and job performance,

and the constructs that affect job satisfaction and job

performance, this chapter describes the development of the

approach to test the research model. The analyses of the

data are presented in Chapter V.

The three purposes of this chapter are to: (1)

explain the research design; (2) discuss the pretesting of

the research instrument; and (3) introduce the statistical

techniques that were used to analyze the data.

Research Design

This research study is a field study. Kerlinger

[1986] defines field studies as "nonexperimental

scientific inquiries aimed at discovering the relations

and interactions among sociological, psychological, and

educational variables in real social structures" (p. 372).

He further refines his definition to include studies that

pursue relations and test hypotheses that are done in life

situations such as businesses. Field studies can be

either exploratory, seeking what is occurring, or

hypothesis-testing. This study comes under the latter

56

category and its aim is to uncover relationships among the

constructs.

Field studies have several strengths such as realism,

significance, strength of variables, theory orientation,

and heuristic quality. Weaknesses of field studies

include the large number of variables and variance, which

can influence the study and the lack of precision in

measurement [Kerlinger, 1986]. An attempt was made to

control for extraneous variables that may affect the study

by using a single office of a large public accounting

firm. Using only one office of the firm controlled for

the variation that could occur due to spatial

autocorrelation, a problem common to studies that sample

firms in different geographical locations. Types of

variation that are controlled for by using only one firm

include personnel differences, the leadership style of the

managing partner, and the office environment. Measurement

instruments were refined to maximize the precision of the

instruments (see Chapter V).

The Population for the Study

The subjects used in this study are employed at a

national public accounting firm located in a large

midwestern city. The firm categorizes levels within the

organization as partner, manager, or staff. Within the

level of staff are the categories senior and assistant.

57

To test the hypotheses of this study, staff

accountants in a public accounting firm who had at least

six months of public accounting experience were selected.

This limitation on subject selection is imposed because

some training and relevant employment—employment specific

to the chosen career—should occur during the

implementation stage of career development [Osipow, 1983].

Development of the Research Instrument

A research instrument was used in this study to

ensure quantifiable data, maintain equivalency of

responses, and because of the sensitive nature of the

material. In addition to the measurement scales used for

the constructs under study, the research instrument also

included requests for demographic information such as age,

sex, position with the firm, years as a staff accountant,

years with the firm, professional certifications, and

education. The demographic data were used as a screening

tool to ensure that the respondents held both the staff

position in the firm and had at least six months of public

accounting experience, as required for this study.

Participation in the study was voluntary, and no attempt

was made to identify respondents by the demographic data

supplied. A copy of the research instrument, a letter to

the subjects informing them of the study, and a letter

58

from a partner in the firm encouraging them to participate

in the study are presented in Appendix A.

Federal regulations require material that will be

used on human subjects to be reviewed. It is a Texas Tech

University policy to review all research involved with

human subjects. The research instrument was submitted to

the Texas Tech University Committee for the Protection of

Human Subjects for review and approval was received. A

copy of the letter granting approval is presented in

Appendix B.

Job Satisfaction Measurement

The Minnesota Satisfaction Questionnaire (MSQ)

[Weiss, Dawis, England, and Lofquist, 1967] was used to

measure job satisfaction. Price and Mueller [1986] stated

that the MSQ was developed to link both measurement and

theory. This scale has been used in accounting studies to

measure job satisfaction (for example, Brownell, 1983).

Weiss et al. [1967] report construct validity based on

the results of testing the instrument within occupations,

and between different occupations. A general internal

reliability coefficient of .90 is reported [Weiss et al.,

1967, p. 23].

59

Job Performance Measurement

Employee performance can be actual or perceived. It

can be further subdivided into historical, expected, or

aspired performance. In public accounting firms, employee

performance is usually based on a supervisor's judgments

and is supported by the employee's (subordinate)

historical actions.

National public accounting firms are reluctant to let

confidential information such as performance evaluations

be seen or used by individuals outside of the firm. An

alternative is to use self-reported performance

evaluations. Self-reported performance evaluations have

been used in accounting studies (for example, Brownell and

Mclnnes, 1986; Choo, 1986; Chenhall and Brownell, 1988).

A potential bias in the use of self-reported performance

evaluations is the inflation of one's performance or a

leniency in self-evaluation [Prien and Liske, 1962;

Lawler, 1968; Thornton, 1980]. Other studies have shown

the opposite effect in self-rating performance evaluations

[Heneman, 1974; Teel, 1978].

There is also the possibility of bias when

supervisors' ratings are used to measure job performance.

Ramanathan et al. [1976] and Jiambalvo [1982] have

reported that there are differences in the perceived

importance of the criteria used in performance

60

evaluations. This may indicate that in studies where

individualistic factors are related to job performance, an

individual's perceived job performance may better relate

to these factors than a superior's job performance rating.

Ferris and Larcker [1983] reported that rewarded

performance (compensation) was a function of both the

interpersonal attraction between the auditor and his/her

supervisor and the task-related ability of the individual.

Therefore, rewarded job performance may be biased based on

the interpersonal attraction between the auditor and

supervisor. It appears that there is potential for bias

using either supervisor or self-rating performance

evaluations [Landy and Farr, 1983]. Since supervisory

performance evaluations were not made available, and there

is a potential for bias in either self-reported or

supervisory performance evaluations, self-reported

performance scales were used in this study.

The self-reporting performance scale used in this

study was developed by Choo [1986], specifically to

measure accountants' job performance. There was no

statistically significant difference found between self-

reported performance evaluations and supervisor

evaluations. The mean score for self-reported performance

evaluations was .50 less than the mean score reported by

61

superiors [Choo, 1986, p. 26]. This indicates a lack of

leniency bias in the instrument and reported scores.

Organizational and Professional Commitment Measurement

Organizational commitment was measured using the

Organizational Commitment Questionnaire (OCQ) developed by

Porter, Steers, Mowday, and Boulian [1974]. Mowday,

Steers, and Porter [1979] were critical of organizational

commitment scales because they believed most were created

on an a priori basis, with little, if any, reliability and

validity data presented. Using the OCQ, a variety of

analyses were carried out using different samples [Mowday

et al., 1979]. Acceptable levels of reliability and

convergent, discriminant, and predictive validity were

evident for the OCQ. This scale has been used in the

accounting literature for studies measuring organizational

commitment (for example, Aranya et al., 1981a; Aranya et

al., 1982; Norris and Niebuhr, 1983; Aranya and Ferris,

1984).

The Professional Commitment Scale (PCS) developed by

Aranya et al. [1981a] was used in this study. The PCS was

primarily adapted by taking the OCQ and substituting the

word profession for organization. In their study, Aranya

et al. [1981a] reported a coefficient alpha of .87.

62

Nunnally [1978] recommends alpha values of .80 or above to

be good indicators of reliability.

Since the organizational and professional commitment

scales are similar, there is the possible effect of common

method variance on the correlation between the two

measures. Aranya and Ferris [1984] used these two scales

in a study and found the correlation between them did not

indicate a strong response bias.

Bline, Duchon, and Meixner [1991] investigated the

psychometric properties of the professional and

organizational commitment scales used in this study.

Their results provide support that the scales do measure

two different constructs.

Life-Span Career Development Measurement

An individual can be satisfied or dissatisfied with

his or her career choice. Osipow [1983] reports that

dissatisfied individuals would be more likely to consider

a career change than satisfied individuals (p. 220).

Personal interviews or longitudinal studies have been used

to determine an individual's life-span career development

[Osipow, 1983].

An alternative to the interviewing process is to

develop a scale that would measure the level of an

individual's career stage. This study used subjects who

should be in the implementation stage of their career

63

development. Therefore, the scale must contain questions

that will allow the researcher to ascertain whether the

individual is in the implementation stage. The scale

developed for this study is based on Super's [1957] six

descriptors for adjusting to work requirements in the

implementation stage. The six descriptors are technical

competence, the routine or tempo of the work, the work

load, work attitudes and roles, security, and time-in-job

before advancement. The questions were derived from the

characterizations given by Super [1957] for the six

descriptors. The characterizations used for the

questionnaire were limited to those that dealt with career

development, to avoid questions that may measure job

satisfaction.

Self-Monitoring Measurement

Snyder's [1974] Self-Monitoring Scale was used to

measure an individual's level of self-monitoring. This

scale requests the subjects to answer twenty-five true-

false self-descriptive statements. Subjects are then

categorized into a low group or high group based on their

responses to the questions. The technique used to

distinguish between high and low self-monitors is to use a

median cutoff (for example, Snyder, 1974; Douglas, 1983).

The upper fifty-percent are considered high self-monitors

64

of the sample, the lower fifty-percent are considered low

self-monitors of the sample.

Research Instrument Distribution and Collection Technique

The researcher was not allowed the time needed to

meet with all of the subjects at the branch and home

offices, as many of them went to the offices only once a

month because they were at remote areas working on audits.

Therefore, for both the pretest and study, a

representative of the firm distributed the research

instrument, which included a letter from the researcher

describing the study, a letter from a partner in the firm

encouraging subjects to participate in the study, and a

self-addressed stamped envelope to return the research

instrument. This was done through the office manager at

the branch office, and a partner at the home office. For

subjects who did not pick up their mail within-one week of

the distribution of the research instrument, the firm

mailed the research instrument and self-addressed stamped

envelope to their home address. The researcher verified

with the firm's office manager and partner assigned to the

study that this procedure was done. The office manager

and partner sent a follow-up memo requesting completing

and mailing the research instrument to all subjects in

65

both the pretest and study. All completed questionnaires

were mailed directly to the researcher's home.

The distribution of the questionnaires by a

representative of the firm, rather than by both the

researcher and a representative of the firm, controls for

potential experimenter effects. These effects can occur

if the researcher distributes and collects some of the

questionnaires, and a representative of the firm

distributes and collects some of the questionnaires.

Subjects may respond differently to questionnaires based

on the individual who distributes the questionnaires.

This distribution and collection technique has been used

in previous accounting research at public accounting firms

[Sutton and Lampe, 1988].

Pretesting the Research Instrument

Hunt, Sparkman, and Wilcox [1982] list three

categories of items that should be pretested: "(1) items

about the questionnaire itself, (2) items about specific

questions, and (3) items about data analysis" (p. 269).

Converse and Presser [1986] expanded these three

categories into ten specific purposes for pretesting a

questionnaire. The first four purposes are tests for

specific questions: variation (data analysis), meaning,

task difficulty, and respondent interest and attention.

The last six purposes bear on the questionnaire as a

66

whole: flow and naturalness of the sections, the order of

questions, skip patterns, timing, respondent interest and

attention (overall), and respondent well-being.

The research instrument was pretested in a branch

office of the national firm used in this study. The

branch office was located approximately ninety miles from

the home office where this study was conducted. The close

proximity between the offices decreased the possibility of

spatial autocorrelation, which could result from using two

or more offices or firms geographically far apart.

Table 4.1 is the summary of the research instrument

response rates from the pretest. Twenty research

instruments were distributed by the branch office manager

to the staff accountants. Initially, the subjects were

not made aware that they were participating in a pretest;

however, at the end of completing the research instrument,

subjects were requested to list any comments they had

regarding the questions and the overall research

instrument. Subjects were instructed to complete the

research instrument and mail their responses in self-

addressed stamped envelopes, which were provided. These

procedures were used at the branch office to simulate the

conditions that would be employed at the home office for

this study. Within nine days, sixteen of the

questionnaires were returned, of which fourteen were

67

TABLE 4.1 SUMMARY OF QUESTIONNAIRE RESPONSE RATES FROM THE

PRETEST FOR ALL SUBJECTS

1 1 Staff Managers Partners Total

Questionnaires distributed

No Response

Responses

Responses not usable for data analysis

Responses usable for data analysis

20

4

16

2

14

6

1

5

1

4

3

1

2

0

2

29

6

23

3

20

Response rate based on questionnaires distributed 80% 83% 67% 80%

Research instruments were distributed to managers and partners to pretest the life-span career development measurement instrument.

68

usable. Two questionnaires were not usable because the

subjects were employed at the firm for substantially less

than six months. Four managers and two partners completed

questionnaires as part of the pretest. This was to

compare mean responses on the measurement instrument for

life-span career development. Hunt et al. [1982, p. 270]

reported that the size of the pretest sample is not fixed,

and that sample sizes of twelve, twenty, and thirty have

been reported to be satisfactory.

The first purpose of pretesting the specific

questions, variation (data analysis), was met by examining

Pearson correlation coefficients for the constructs

professional commitment, organizational commitment, life­

span career development, job satisfaction, and job

performance. No changes were made to the research

instrument based on this analysis. Since self-monitoring

is measured on a dichotomous scale, high or low, it was

excluded, from this analysis. PRELIS and LISREL (linear

structural relations) VII, the statistical packages used

in this study, allow the use of continuous and dichotomous

constructs in the Scime model. However, the number of

responses in the pretest was too small to use LISREL VII.

This posed no problem for this study, since the number of

responses from the home office was large enough to use

LISREL VII.

69

The remaining three purposes of pretesting the

specific questions—meaning, task difficulty, and

respondent interest and attention—were analyzed by

reading the comments provided by the subjects. There were

no comments regarding these three areas of the pretest.

The last six purposes of the pretest—flow and

naturalness of the sections, the order of questions, skip

patterns, timing, respondent interest and attention

(overall), and respondent well-being—were also analyzed

by reading the comments provided by the subjects. Based

on the subjects' responses, two areas that rec[uired

changes in the research instrument were the ordering of

the questions, and the form requesting demographic data

and job titles.

Initially, the questions for organizational and

professional commitment were sequential. Some of the

subjects assumed they were the same questions because of

the similarities between the questions. This was changed

for the research instrument distributed to the home office

by separating the two sets of questions, and inserting a

statement that the questions were not the same as the ones

asked previously. A second change was made regarding the

ordering of the demographic data, and the wording of the

job titles.

70

No other changes were made to the research instrument

after the pretest. On average, it took the subjects

fifteen minutes to complete the research instrument, with

a range of nine minutes to twenty minutes.

Statistical Technique

The constructs under study are latent or unobserved.

In their purest form they are abstract unidimensional

concepts, making them only indirectly measurable [Bollen,

1989]. The latent variable model consists of the

structural equations that summarize the relationships

among the constructs, and the measurement model. The

totality of relationships among the constructs is usually

referred to as the structural equation, and their

measurement as the measurement model. This is misleading

because it indicates that the measurement model is not

structural. While this is not the case, to differentiate

between the relationship among the constructs and the

measurement model, this study will use the traditional

approach of referring to the relationships among the

constructs as the structural model, and the measurement of

these constructs as the measurement model.

The primary statistical tools used to analyze the

data in this study are PRELIS and LISREL VII. PRELIS and

LISREL VII were developed by Joreskog and Sorbom [1988,

1989a, 1989b]. PRELIS is data screening software that can

71

read grouped and patterned data. In addition to data

screening, PRELIS computes correlations and matrices which

can be used as input files for LISREL VII. The necessity

of using PRELIS in this study is that it can compute

correlations and other product moments for continuous and

dichotomous variables that other statistical programs

cannot produce [Joreskog and Sorbom, 1988]. This will

allow the dichotomous construct self-monitoring to be

analyzed with the continuous constructs in this study.

LISREL VII is used to analyze the structural and

measurement models. LISREL VII will minimize the

difference between the sample covariances and the

covariances predicted by the model. An advantage to using

LISREL VII with the data for this study is that it will

integrate measurement concerns with the structural

equation model. This is done by incorporating both the

measured indicator variables with the latent theoretical

concepts into a single structural equation model [Hayduk,

1987, pp. 87-88]. LISREL thus allows for integrating

model development, estimation, evaluation, and

interpretation with measurement concerns [Bohrnstedt,

1983].

LISREL Model Specification

The LISREL model is illustrated in Figure 4.1, and

the LISREL terminology and notations for this study are

72

*l-'n

^t- ^n

^ - ^ i '.6/n

ib.i^yi.i

^)6-'-7

^.6- ^7

Figure 4.1. The LISREL Model

73

presented in Figure 4.2. The model consists of three

latent exogenous constructs (xis)—professional

commitment, life-span career development, and self-

monitoring—which will be predicted outside of the model,

and three endogenous constructs (etas)—organizational

commitment, job satisfaction, and job performance—which

will be predicted within the model. This is done by the

three ecjuations which make up the general structural

equation model. (See Figures 4.3 - 4.5 for the specific

research equations and the general form of these

equations.)

The first equation (Figure 4.3) contains the

hypothesized direct effects among the concepts. These

were estimated by the matrices of structural coefficients

(betas and gammas), the vectors of the endogenous and

exogenous concepts, and a vector of "errors" in the

conceptual model (zeta). The covariances among the zeta

terms will make up matrix psi. Xis will generate a

matrix, phi, which is the covariance among the xis. The

second and third equations (Figures 4.4 and 4.5) link the

conceptual variables to their observed indicators [Hayduk,

1987]. The second equation is the measurement model for

the endogenous indicators, consisting of a matrix of

structural coefficients (lambdas), the vector of the

endogenous concepts, and a vector of "errors" in the

74

S p e c i f i c Terms I n c l u d e d in I h e LISREL Model

^15

><1

P r o f e s s i o n a l Commi tiuen t L i f e S p a n C a r e e r D e v c 1 o p m c n I

S e 1 f - M o n i l o r i ng Orga i i i 2a t i ona i Commi Uiien t J o b S a l i s f a c l i on J o b P e r f o r m a n c e i n d i c a n t s of Piof ess i ona 1 Commi linen I i n d i c a n l s of Life S p a n C a r e e r D e v e l o p m e n t i n d i c a n t s of Se 1 f-Mon i t or i ni; i n d i c a n t s of O r g a n i z a t i o n a l Coimiii tuien t i n d i c a n t s of Job S a t i s f a c t i o n i n d i c a n l s of Job P e r f o r m a n c e

D e f i n i t i o n s of L I S R E L ' S G e n e r a l T e r m s

I Tl X y p q (U n

(xi) (ela)

an unobservable (latent) cxog^enous variable an unobservable (latent) cndogejtous variable an observed indicant of an exogenous variable an observed indicant of an endogenous variable the number of y-variables the number of x-variables the number of T^-variables the number of ^-variables

Parameter Matrices

Ay (Larabda-y)

Ay (Lambda-x)

B (Beta)

r (Gamma)

i (Phi)

Y (Psi)

Sj ( T h e t a - E p s i I o n )

e, (Theta-Dclta)

a pxm among a qxn among a mxm among a mxn among a nxn among a mxm among a pxp among a qxq among

matrix of y ' s and T

matrix of x's and ; matrix of

n matrix of ; ' s and T\ ma t r i X of

the I m a t r i x of

I he

the

the

the

the

the

the T)' s e r r o r m a t r i x of the the y's er r o r m a t r i x of the

the x's er r o r

r e 1 u I i oiisii i ps

re 1 a t i o n s h i p s

r e l a t i o n s h i p s

r e l a t i o n s h i p s

r e i a l i o n s h i p s

r e l a t i o n s h i p s Ierms (C's ) re1ali o n s h i p s lerms ( € ' s ) r e l a l i o n s h i p s Ierms ( 6 ' s )

Figure 4.2. LISREL Notation and Terminology

75

G e n e r a l . Forxc: T] = BT) -K pf * C

S p e c i f i c Foruj

T)

Tl

T)

B. , B:;

0 0 "D

TJ "^ 1 Tl .

Tl • , 0

* '^ 11 • 21 ^ J l

r M V , ,

1

0

1 ' . .

i • 1 ' • I - •

1 ; .

^ : 1 :1 c ;

Figure 4.3. The Structural Equation

76

G e n e r a l :

Spec 1f i c :

Y = Ayt,

> • ' :

yi >i

>5

>-ll

>li >-15 >1T >!l ^15 y V

20 21

yu

>25

> 3 0

>-3l > -3 : > 3 3 >-3< >'3i v.-* n yyi V V

3S 35

•Nl

yr, > • «

• 4i

.1.1 A , ,

^ t , l A., . • • < 1 A . ,

.10 ,1

^11,1

^ i : . i A , , ,

0 u 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

li.l

0 0 0 0 u u 0 0 u 0 u 0 0 0 0

A l , . . t »li

^\', ^ « * . t . A»(( . . . . • . . • « • . A . « ^ . • « • . A . . ^

A l ( 1

A i ; <

A * ' 7 . •^•• A l - 1 . •• •' A i r 1

• . • * A1C 1

• • • •

f3 t . : A * ' ^ , « • • • . • - • • /»• • ^ • •« • *

A , : ,

0*"* 0 0 0 0 0 0 0 0 0 0 0

0 u u 0 0 u 0 u 0 u 0 u 0 0 0 0 0 0 u 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

^1

^ 2

Tl;

* A - -. - ( l .

/..• • • ' • A . ,

A v .

. • • • A | r

A , .

. '•• ^r.. ^ < : .

^u. A , : . '•• A , . . U. ' T .

3 3

! ' i e. «i ^i ^ s ^ j € i

^11 ^ 1 !

^15

•11

^\)

«!1

^33

^2<

^!J

^:i

€„ ^ 3 1 ^31 € . , ^33

^ 3 :

3i

' 3 1 «3?

^ 4 1 ^41 ^ 0 e<3 ^ H

Figure 4.4. The Measurement Model for Y Indicants

77

General:

Specifie:

X = A , ^ ^ 6

10

X ,

x« X .

^ I Xj

^ { X i

^1 X ;

X

X . ^

x,i

^11 ^IJ >^:o >-:i X . .

X.I . . I

^21 X i ;

xi: • «

X . .

^31 X . t

X j j

X i , 4 i

X . I x -

* 34

^3J X i l

>^3I

>^3:

X n

^13

A , . > • A . • . I A . . • •

A( • * * « * * A . • ' t

A ,

A c '

X" .\: A . . . ' ' A . . . » • A . . . » .

A , t

* • u 0 0 0 u 0 0 0 0 0 0 0 0 0 0 0 u 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 u u 0 0 0 u u 0 0 u u 0 u

A l . ^ . i • I .

"11.2 ^S.2

• fc^ t •

A 4 « « • m 9 9 i

A«« ^

u 0 0 0 0 u 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

u 0 u u 0 u u u 0 0 u u 0 u 0 u u 0 0 0 0 0 0

>-u.: ^3:.3 ^•2i.3 / • : : . ; ^•2t.3 A n :

l i t I t

/••• -

C34.3 A i ; . . - - - J A . • . . . b • . A . • • . . . . . A . r •

A . 5 •

A i l , • . *••« A 11 • . ' i • • / • I • . , . . . . A I . t

. «... A I • .

/ • I • •

. ' • •• / • I r •

4 • • rf

f

f

f

^( t T

^ E T

r

r

6 ' 0

"11 f U t «

"n

"IE

"11 "1?

"20

" I I

6=' 24

?31

^31 f

6 : :

"34 ^33 ^ ; ; 6 , . hi "3" 'ii "«3

6.; H

" U f

"4 i

Figure 4.5. The Measurement Model for X Indicants

78

measurement model (epsilon). The covariance among the

epsilon terms constitute matrix theta epsilon. The third

equation is the measurement model for the exogenous

indicators, consisting of a matrix of structural

coefficients (lambdas), the vector of exogenous concepts,

and a vector of "errors" in the measurement model (delta).

The covariance among the delta terms constitute matrix

theta delta.

To test the model, a two-step approach is used

[Anderson and Gerbing, 1988]. This approach evaluates the

measurement model separately from the full structural

equation model. The first step is to evaluate the

measurement model for unidimensionality of the measurement

constructs, and respecify the model if necessary [Anderson

and Gerbing, 1988]. The second step is to simultaneously

test the measurement and structural model after

respecification. These steps are presented in Chapter V.

Chapter Summary

The research design, pretesting of the research

instrument, and the statistical techniques used in this

study were presented in this chapter. The research design

is based on field study research [Kerlinger, 1986].

Pretesting the research instrument was conducted as

closely as possible to the distribution and collection

techniques used in this study. There were no significant

79

changes made to the research instrument or planned

distribution and collection techniques based on the

pretest. Based on the constructs being investigated,

structural equation modeling with a two-step approach was

selected to analyze the data, using PRELIS and LISREL VII

Chapter V contains the analyses for this study, including

response rates, testing for nonresponse bias, and

statistical testing of the measurement models and

structural equations.

80

CHAPTER V

EMPIRICAL ANALYSIS OF THE DATA

The study to this point has used the theories of

Bullen and Flamholtz [1985] and Bagozzi [1980] to

construct a research model to investigate the

interrelationships among variables that can affect job

satisfaction and job performance, as well as the

relationship between job satisfaction and job performance

The research design, population, research instrument,

pretest results, and statistical procedures used in this

study were presented in Chapter IV. In the first part of

this chapter the response rates and the test for

nonresponse bias are presented. The second part presents

the statistical analyses used in the study to test the

research model (see Figure 3.5). This analysis includes

testing and refining the measurement models, simultaneous

tests of the measurement and structural models, and tests

of the individual structural coefficients.

Response Information

Response Rates

The study was conducted at a national public

accounting firm located in a large midwestern city. The

distribution and collection of the research instrument

were described in Chapter IV. The firm employed 421

81

accountants classified as staff at the time of this study.

Of these, 312 responded, for a 74% response rate.

Listwise deletion is used in this study, so any

respondents not answering all of the questions, or meeting

the experience requirements for this study were eliminated

(see Statistical Analysis, p. 84). This reduces the data

matrix to a matrix without any missing observations. This

resulted in 31 responses being eliminated for a usable

response rate of 67%. Table 5.1 presents the response and

elimination rates.

Test for Nonresponse Bias

Armstrong and Overton [1977] recommend the best

protection against nonresponse bias is the reduction of

nonresponse itself. However, when nonresponse is present,

subjects who respond late or less readily are assumed to

be more like nonrespondents [Armstrong and Overton, 1977].

In this study, the nonresponse rate was 26%, possibly due

to the distribution and collection techniques. To test

for nonresponse bias, a comparison of the means of early

and late respondents was undertaken. This testing was

done by using multivariate analysis of variance.

Multivariate analysis of variance is used when means on

several different variables are being compared

simultaneously, as in this study [Bray and Maxwell, 1985].

82

TABLE 5.1 SUMMARY OF QUESTIONNAIRE RESPONSES AND ELIMINATIONS

Number Percent

Total Questionnaires Distributed

Questionnaires Returned 1

Questionnaires Eliminated

Usable Responses

421

312

31

281

100%

74%

7%

67%

1 Questionnaires were eliminated for candidates who were

not employed at the firm for at least six months, or who did not answer all of the questions pertaining to the measurement of the constructs. This latter requirement is because of the use of listwise deletion.

83

Table 5.2 presents the nonresponse bias test result.

The null hypothesis for testing nonresponse bias is that

there is no difference in the mean responses between the

early and late respondents. Wilks' Lambda results in not

rejecting the null hypothesis at the .05 level. No

significant nonresponse bias is indicated by the results

of the test statistic.

The demographic data and descriptive statistics are

presented in Appendix C. Statistical tests were conducted

for differences in mean responses for the demographic

data. There were no significant differences between mean

responses based on the demographic data at the .05 level.

Statistical Analysis

Testing for Identification

Equations in a model must be identified if meaningful

results are to be obtained [Fisher, 1966]. Berry [1989]

stated:

An equation is identified if the acquisition of data for a properly chosen infinite sample would be sufficient to determine a unique set of parameters for the equations; if the equation is identified, only one explanation (i.e., one set of parameters) would be consistent with both the data from the infinite sample and the model's restrictions, [p. 25]

LISREL VII checks for the identification of the model by

making tests in the following way [Joreskog and Sorbom,

1989b]:

84

TABLE 5.2 NONRESPONSE BIAS TEST RESULT

Test F-Test PR > F

Wilks' Lambda .9434 .4548

85

The program checks the positive definiteness of the information matrix. (The information matrix is the probability limit of the matrix of second order derivatives of the fit function used to estimate the model.) If the model is identified, the information matrix is almost certainly positive definite. If the information matrix is singular, the model is not identified and the rank of the information matrix indicates how many parameters are identified, [p. 17]

There were no instances of lack of identification in the

model during the statistical analysis using LISREL VII.

Testing the Measurement Model

Anderson and Gerbing [1988] recommend a two-step

approach to be used for structural equation modeling.

This is accomplished by first testing the measurement

model prior to testing the measurement and structural

submodels simultaneously. The reliability and validity of

the measurement instrument can be assessed by testing the

measurement model in a two-step approach.

Because of the large number of indicants used in this

study, composites of the 95 questions used in the

measurement instrument were formed. The use of composites

and single indicators has precedence when using LISREL as

the statistical tool for data analysis (for example, see

Bagozzi, 1980a and 1980b; Gaski, 1986; Howell, 1987).

Questions 1-70 measured the constructs professional

commitment, life-span career development, organizational

commitment, job satisfaction, and job performance. These

86

seventy questions were grouped into twenty composites,

four for each construct. Self-monitoring was measured by

questions 71-95. Based on the responses to these

questions, subjects were numerically categorized as a low

(1) or high (2) self-monitor. Therefore, a total of

twenty-one composite measurement indicants were used for

the six constructs.

Next, the input files were created for analysis using

LISREL. This was accomplished by using PRELIS, the

predecessor analytical software for LISREL. Since self-

monitoring is a dichotomous construct, and the other

constructs are continuous, Joreskog and Sorbom [1988 and

1989a] recommend using the weighted least squares method

for data analysis. This entails the use of listwise

deletion because there are a different number of cases

(observations) involved in the computations if pairwise

deletion is used [Joreskog and Sorbom, 1989a]. PRELIS

computed Pearson product moment correlations among the

constructs professional commitment, life-span career

development, organizational commitment, job satisfaction,

and job performance. Polyserial correlations were

computed between self-monitoring and professional

commitment, life-span career development, organizational

commitment, job satisfaction, and job performance. In

addition to the correlation matrix, PRELIS created the

87

asymptotic covariance matrix to be used as an input file

in the LISREL analyses. This input file is used with the

weighted least squares method so that "an estimate of the

asymptotic covariance matrix of estimated correlations can

also be obtained under the same general assumptions of

non-normality" [Joreskog and Sorbom, 1989a, p. 21].

After the input files were created, LISREL was used

to measure the reliability of the composites. This is

done by constraining to zero all of the parameters that

relate the constructs. Squared multiple correlations are

used to determine the reliability, where values greater

than .50 are considered desirable [Bagozzi and Yi, 1988].

Measurement error for self-monitoring was specified at

.20, producing the squared multiple correlation of .80.

Hayduk [1987] recommends measurement reliabilities to be

fixed rather than free when the researcher is familiar

with the data and measurement quality. As discussed in

Chapter IV, subjects are categorized as high or low self-

monitors based on a median split. Since the median can

differ based on the subjects responding to the research

instrument, it was felt that measurement error was present

and .20 would represent an acceptable amount of

measurement error.

The results are a goodness-of-fit index of .957 and

root mean square residual of .049, both indicating a very

88

good overall fit [Bagozzi and Yi, 1988]. However, job

satisfaction 1 and 2, job performance 4 and life-span

career development 1 and 3 composites did not meet the

desired squared multiple correlation level of .50, so the

measurement model was respecified. The squared multiple

correlations for the twenty-one composites are presented

in Table 5.3.

Respecification was done by using factor analysis to

examine the indicants for low loadings to obtain

candidates for deletion. Indicants that did not have a

factor loading of .30, or had loadings on a factor other

than the common factor for a measurement scale, are

candidates for deletion [Kim and Mueller, 1986]. The

measurement scales for organizational commitment and job

performance did not have any candidates for deletion.

Two indicants for professional commitment, x7 and

xl2, were deleted, since they did not have factor loadings

large enough to warrant inclusion based on the criteria

for deletion. The measurement scale for life-span career

development had four candidates for deletion, xl6, xl7,

x20, and x21. These were deleted because they did not

load on the factor for the construct life-span career

development. These variables each loaded on a different

factor, therefore they did not, as a group, represent a

different construct. Possible reasons for this are that

89

TABLE 5.3 MEASUREMENT STATISTICS FOR THE TWENTY-ONE COMPOSITES

Item Squared Multiple Correlation

Organizational Commitment 1 .717

Organizational Commitment 2 .645

Organizational Commitment 3 .750

Organizational Commitment 4 .755

Job Satisfaction 1 .490

Job Satisfaction 2 .468

Job Satisfaction 3 .721

Job Satisfaction 4 .606

Job Performance 1 .694

Job Performance 2 .766

Job Performance 3 .558

Job Performance 4 .430

Professional Commitment 1 .611

Professional Commitment 2 .590

Professional Commitment 3 .734

Professional Commitment 4 .698

Life-Span Career Development 1 .022

Life-Span Career Development 2 .578

Life-span Career Development 3 .255

Life-Span Career Development 4 .517

Self-Monitoring (error specified at .20) .800

90

the scale, which was created for this study, measured

other stages of life-span career development, or a

construct other than life-span career development. Job

satisfaction had the most indicants deleted, eight. The

squared multiple correlations for two of the composites

were less than .50, which may indicate measurement

problems for this construct. The job satisfaction scale

consists of twenty questions comprising many aspects of an

individual's job. It has a reported internal reliability

coefficient of .90 [Weiss et al., 1967, p. 23]. In this

study, the indicants yl8, yl9, y23, y24, y25, y27, y28,

and y29 did not load on the common factor, nor did they

load on the same factor, which may have indicated some

construct other than job satisfaction being measured. A

possible reason for the measurement problem that occurred

in this study is that the questions were not applicable to

staff accountants. For example, some of the questions

asked about supervising individuals, and subjects trying

their own methods of doing the job. Staff accountants

usually do not supervise, and they are told how to go

about doing the job. The respecified measurement models

for the y and x indicants are presented in Figures 5.1 and

5.2.

The respecified measurement model consists of sixteen

composite measurement indicants for the six constructs.

91

c i e n e r a ; :

S p e c i f i c

A .Ti - e

>-i

y .

V . ' » > 1

> l >5

; . i ' 1

>2t I

vji v . . - i t

>:o >: i I

\ . •

>'3I

-•"4; *' 4!

- i 4 !

>-4: !

U:

A .

A( , . • • • A . ,

A . , . • . » Aj .

At . . • • • ' • ic . A t • . I . .

A i . . » • • A . • . . . . ?u. A f .

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

, 0

0 0 u 0 u 0 u u 0 u 0 0 u 0 u A i . . , I k , i A ( 1 I)

. 1 . . »

. . . . . A . i •

. . . . . A . . 1

. . . . . A . . . . . . . . A ^ » 1 . *» ' • A . . 1

A 3 3 . 3 A3:.3 Aji .3 A i i . : 0 0 0 c 0 0 0 0 0 0 0 0

u 0

u 1 u U 1

0 1 u ! u ! 0 u u u u u 1)

0 0 0 0 u 0 0 0 0

^1 e.

£4

5 « { e. i

£ !

^10 ^11 ^ 1 : ^t: 14

^ l i ^It 13

1

^ !

Tl.

^21 ^31

• •

^ i ; " 1 ^30

\ «31

T13 ' ! ^33 0 I ; £ : : 0 ; 1 £ j i

0 1 i ^3: A : : . 3 *^t ; • • • ^ • * A . K . 1 . "

• t ( c

• • • • • 1 1 c "

• 1 ( •

'•41.3 *< : A , . . 1 1 ^13

• ' 1 \ c ^K.3 ; J ^«< ^4i.3 1 ; ^l i A I : • 1 • ~4'i

/ . 1 ^ • 1 . t f -

Figure 5 . 1 . Respecified Measurement Model Indicants

for Y

92

General:

Speci fie:

X = A , $ + C

X ,

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

» I

* I

X. . ' • A , . • » A .

A l . .. Art . t . A f . •. A ,

; . 1 4

0 0 0 u 0 0 0 0 0 U 0 0 0 0 0 0 0 0 C 0 0 0 0 0 0 0 0 0 0

U 0 0 0 u 0 0 u 0 u 0 u u A , ,

. '•• A l » . ' • . A . . . . . • A . .

0 " ' 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 u u u u 0 n 0 u u 0 0 0 0

^24.3 A . t • . . . . J A . . . , • • • w A * . • . . . . J A . r •

. . . . Xii.: A3:.3 A . , .

A . i .

A j j ,

A34. •; i r

33. . J A . ? 1

. . . . A ^ : 1 . . . . J

A . r 1

^ ^ * • • A . .

. <J.

l-r.t

'"4J.

r *3

I i*

t " I

^1

" {

I

"10

I #

"Ij ^11 "!!

«-««i

^24 ^3i "2t "33 "21

^30

'34

.c W . "

trf-i . t f

"3! "4C T

^41 9

f43 ^ | ? > e

" 4 i

hi ?'f "13 r

"41

F i g u r e 5 . 2 . Respecified Measurement Model for X Indicants

93

There are three composites for each of the constructs

professional commitment, life-span career development,

organizational commitment, job satisfaction, and job

performance. Each of the three composites are created

from a common factor for each construct. There is one

composite measurement indicant for self-monitoring.

This study minimized the effects of deleting these

indicants by using composites. If the questions that were

deleted did not contribute to the reliability of the

measurement scales, then the new composites should show

increased reliabilities through the recalculated sguare

multiple correlations. By creating new composites based

on the results of the factor analysis, it also makes it

unnecessary to correlate error terms, as these will be

minimized by grouping the indicants into the composite

scales. The respecified measurement models for the

composite y and x indicants are presented in Figures 5.3

and 5.4.

The squared multiple correlations for the respecified

measurement model are presented in Table 5.4. life-span

career development composites 2 and 3 are marginally short

of the desired .50 level, while all of the other

composites exceed .50. The goodness-of-fit index improved

increasing from .957 to .990, and the root mean square

residual has improved by decreasing from .049 to .041.

94

General:

Spec:fie:

Y = A,.T]

1 >> V .

>-4 y^ yi V . • 1 y\

X. x: m •

A..

0' 0 0 0 0 0

.1

.1 0 0 0 A( 1

Al , . *.* '^i ' . • » .

A

0 0

0 0 0 0 0 0 X. 1 . ' .< A . , . *«• A j ^

^1

T i l

1 «

e,

Figure 5.3. Respecified Measurement Model For Composite Y Indicants

95

General:

Speci fie:

1 X,

^ 3 ^ <

^ :

1

X = A P

x: l.l ;i: :.i 3.1

0 0 0 X ^ . XI':

0 0 0 0 0 0 A, ,3

5 , I

tl t

f 0(

>7

Figure 5.4. Respecified Measurement Model For Composite X Indicants

96

TABLE 5.4 MEASUREMENT STATISTICS FOR THE SIXTEEN COMPOSITES

l ni Squared Multiple Correlations

Organizational Commitment 1 .740

Organizational Commitment 2 .836

Organizational Commitment 3 .771

Job Satisfaction 1 .661

Job Satisfaction 2 .689

Job Satisfaction 3 .588

Job Performance 1 .708

Job Performance 2 .717

Job Performance 3 .749

Professional Commitment 1 .739

Professional Commitment 2 .700

Professional Commitment 3 .776

Life-span Career Development 1 .577

Life-span Career Development 2 .452

Life-span Career Development 3 .493

Self-Monitoring (error specified at .20) .800

97

These results indicate an overall improvement in the

measurement model.

Construct validity was examined by the correlations

among the constructs produced with PRELIS, with their

hypothesized relationships. The hypothesized

relationships cimong the constructs were for significant

correlations. Table 5.5 presents the correlations among

five of the constructs. The correlations between job

performance, and professional and organizational

commitment are positive and significant at the .05 level.

The correlation between job performance and life-span

career development is significant at the .01 level. All

of the other correlations are positive and significant at

the .0001 level. This gives evidence for construct

validity.

Table 5.6 presents the polyserial correlations

between self-monitoring, and the composites in the

respecified measurement model for job satisfaction, job

performance, organizational commitment, professional

commitment, and life-span career development. The

composites rather than the individual indicants were used

to attempt to better respecify the measurement model. The

polyserial correlations are not significant for the

hypothesized relationships. Correlations between self-

monitoring and professional commitment, life-span career

98

TABLE 5.5 PEARSON CORRELATION COEFFICIENTS/

PROB > |R| UNDER HO:RHO==0

OC JS JP PC LSCD

Organizational Commitment (OC) 1.000

0

Job Satisfaction (JS) .6216 1.000 .0001 0

Job Performance (JP) 1476 .3244 1.000 .0132 .0001 0

Professional Commitment (PC) 7680 .5671 .1252 1.000

.0001 .0001 .0359 0

Life-Span Career Development (LSCD) 6099 .5630 .1787 .6826 1.000

.0001 .0001 .0026 .0001 0

99

TABLE 5.6 POLYSERIAL CORRELATION COEFFICIENTS/

PROB > |R| UNDER HO:RHO=0

Self-Monitoring

Job Satisfaction 1 .092 / .122

Job Satisfaction 2 .115 / .053

Job Satisfaction 3 -.030 / .615

Job Performance 1 .011 / .856

Job Performance 2 .006 / .924

Job Performance 3 .059 / .325

Organizational Commitment 1 .074 / .215

Organizational Commitment 2 .112 / .056

Organizational Commitment 3 .014 / .815

Professional Commitment 1 .054 / .370

Professional Commitment 2 .144 / .016

Professional Commitment 3 .112 / .061

Life-Span Career Development 1 .009 / .887

Life-Span Career Development 2 .037 / .542

Life-span Career Development 3 -.002 / .967

Self-Monitoring 1.000 / 0

100

development, and organizational commitment are presented

for informational purposes. There are no hypothesized

relationships among these constructs.

Criterion-related validity can be assessed by

examining concurrent validity. Concurrent validity is

relevant in psychological tests for the diagnosis of an

existing status, rather than the prediction of future

outcomes [Anastasi, 1976]. Concurrent validity is used as

a substitute for predictive validity when it is

impractical to extend the validation procedures over time,

as in this study. Questions must be phrased so that they

do not attempt to predict an outcome- The questions used

in this study are not phrased to be predictive.

Evidence of concurrent validity can be obtained by

examining the relationships among the constructs based on

some criteria. For example, professional and

organizational commitment are hypothesized to be

significantly related in a positive direction. The

Pearson correlation coefficient for this relationship is

.768, significant at the .0001 level. The correlations

reported in Table 5.5 give evidence of concurrent validity

among the constructs. The correlations among the

constructs are statistically significant in a positive

direction as hypothesized in this study.

101

Hypothesis Testing

Rather than freeing the individual hypothesized

causal links in the model, referred to as a one-step

approach [Anderson and Gerbing, 1988] (see Appendix D for

the results of this approach), a two-step approach is used

to test the hypotheses. With a two-step approach the

structural model is tested first, and second, a series of

nested structural models are estimated and sequential chi-

square difference tests (SCDTs) are performed [Anderson

and Gerbing, 1988]. This is accomplished by estimating a

null model in which the beta, gamma, and phi parameters

are fixed at zero. This model gives the largest number of

degrees of freedom for any structural model. Next, a

saturated model in which the beta, gamma, and phi

parcimeters are free is then estimated. This model gives

the smallest number of degrees of freedom for any of the

models under consideration. The theoretical (research)

model under investigation is then estimated. Finally, two

additional structural submodels are estimated, which are

the "next most likely constrained and unconstrained

alternatives from a theoretical perspective to the

substantive model of interest" [Anderson and Gerbing,

1988, p. 418]. These five models are referred to as the

null, saturated, theoretical, constrained, and

unconstrained models. These five structural submodels are

102

nested in a sequence such that, for the number of

parameters to be free, the null < constrained

< theoretical < unconstrained < saturated. Sequential

chi-square difference tests for the five nested models are

performed after it is first assessed whether a structural

model exists that would have an acceptable goodness-of-

fit.

Testing a nested sequence of models using a chi-

square difference test rather than testing individual

paths is preferred [Anderson and Gerbing, 1988] and is

recommended in research literature (for example, Bentler

and Bonett, 1980; James, Mulaik, and Brett, 1987; Bollen,

1989). Anderson and Gerbing [1988] cite four strengths

that the two-step approach has over a one-step approach.

These strengths are:

First, it allows tests of the significance for all pattern coefficients. Second, the two-step approach allows an assessment of whether any structural model would give acceptable fit. Third, one can make an asymptotically independent test of the substantive or theoretical model of interest. Finally, the two-step approach provides a particularly useful framework for formal comparisons of the substantive model of interest with next most likely theoretical alternatives, [p. 422]

The Structural Model

To determine if any structural model has an

acceptable goodness-of-fit, a "pseudo" chi-square test

statistic is calculated (Anderson and Gerbing, 1988,

103

p. 418). The pseudo chi-square test statistic is

comprised of the results from two different models, the

saturated and null models. The chi-square value of the

saturated model and the degrees of freedom from the null

model are used to test for significance. The saturated

model has the smallest value for any structural model

since all of the parameters are free, while the null model

has the largest number of degrees of freedom since all of

the parameters are fixed. If this test statistic is

significant, then there is no structural model that would

give an acceptable fit. This is because any alternative

models would have fewer degrees of freedom than the null

model, with a chi-square value greater than or equal to

the saturated model.

Table 5.7 presents the results of the pseudo chi-

square test statistic. The test statistic is significant

which could indicate a misspecification of the measurement

model [Anderson and Gerbing, 1988]. The measurement model

was reexamined by investigating modification indices

obtained in the LISREL output, and attempting to further

refine the composite indices. No further respecification

could be done to obtain a measurement model that would

produce an acceptable pseudo chi-square test statistic.

Anderson and Gerbing [1988] stated:

. . . a researcher can build a measurement model that has the best fit from a content and

104

TABLE 5.7 PSEUDO TEST STATISTIC FOR THE NULL AND SATURATED MODELS

Saturated Model Chi-square Test Value 436.74

Null Model Degrees of Freedom 106

Significance Level < .001

105

statistical standpoint, where respecification may have been employed to accomplish this, and still provide a statistical assessment of the adequacy of the theoretical model of interest, [p. 419]

In this study the measurement model is deemed to have the

"best fit." The improvement in fit of the null versus the

saturated model is highly significant as illustrated in

Table 5.8 [Bentler and Bonett, 1980]. Also, the goodness-

of-fit index of .990 demonstrates that the remaining

improvement of .01 (1.00-.990) that might be obtained if

the measurement model is improved is not significant from

a practical viewpoint [Bentler and Bonett, 1980]. The

next step after the structural model has been tested for

goodness-of-fit is to compare alternative models using

sequential chi-square difference tests (SCDTs).

Sequential Chi-sguare Difference Tests

After the saturated model is evaluated, a decision-

tree approach is used to test sequential chi-square

differences for the models. Figure 5.5 illustrates the

order of testing the nested models for SCDTs [Anderson and

Gerbing, 1988]. The null hypothesis to compare a set of

SCDTs is that there is no significant difference between

two nested structural models based on the difference in

chi-square values and the difference in degrees of

freedom. For each SCDT in which the null hypothesis is

not rejected, the model that is more constrained (i.e.,

106

TABLE 5.8 COMPARISON OF THE NULL AND SATURATED MODELS

Chi Significance Model Square Difference D.F Level

Null 10693.69 — 106 1

Saturated 436.74 10256.95 88 .001

1 Difference in degrees of freedom, 18, with difference

in the chi-square value, 10256.95.

107

MC - MS

sign

accept MC

accept MI

acceoc MU

MC - Ms — ^ accept He

sign

V ns MI - .MU ;^—^ Respecify MU as alternaie

mcd<l Hu'; then Ni - Mu'

sicn .-.s Mu - MS

sicn

^ accept M;

Relax ccnstrai.-t m MU t. a: is next jicst-lixeiy model Mu2; then Mu: - MS.

ns - not significant at specified crctability i*v€2 Sign - significant at specified profcatiiity level Mt - theoretical model MS - saturated acdel MC - constrained model MU - unconstrained model

Figure 5.5. Decision-Tree Framework for SCDTs

108

more fixed parameters) would be accepted. To test the

null hypothesis, a chi-square test statistic value is

calculated for the saturated, constrained, theoretical,

and unconstrained models.

The first SCDT test is between the theoretical and

saturated model. The chi-square values and associated

degrees of freedom are presented in Table 5.9. The five

relationships not hypothesized in the theoretical model

(SM->PC, SM->LSCD, JS->OC, JP->OC, and SM->OC) represent

the five degrees of freedom difference. Since

constraining these causal links does not make a

statistically significant difference at the .001 level,

the hypothesis that there is no significant difference

between the two models is not rejected. Therefore, the

second SCDT is between the theoretical and constrained

models.

Constraining additional parameters should be done

from a theoretical perspective by seeking a "weak-link" in

the theoretical model. One method of constraining or

unconstraining paths is to examine the modification

indices provided by LISREL. The modification index is "a

measure of predicted decrease in chi-square if a single

constraint is relaxed and the model is reestimated"

[Joreskog and Sorbom, 1989b]. Theoretical justification

should be the prime criterion for constraining or

109

TABLE 5.9 COMPARISON OF THE THEORETICAL AND SATURATED MODELS

Chi Significance Model Square Difference D.F Level

Theoretical 447.94 93 1

Saturated 436.74 11.20 88 < .05

1 Difference in degrees of freedom, 5, with difference in

the chi-square value, 11.20.

110

unconstraining paths between constructs. Bagozzi and Yi

[1988] stated:

Changing a model by indiscriminatly freeing parameters suggested by modification indices is of course an example of capitalizing on chance. One should avoid relying on such a practice except perhaps in the most exploratory of studies. As a matter of fact, models should not be modified unless one has theoretical and/or methodological reasons justifying any modification, [p. 81]

An examination of the modification indices resulted in a

decision not to constrain or free any of the paths based

solely on these indices.

Based on prior research that also used professionals

as subjects, the path job satisfaction—>job performance

will be constrained. Bagozzi [1980], using professional

salespeople as subjects, hypothesized that job performance

will influence job satisfaction, but that job satisfaction

will not significantly affect job performance. His

theoretical rationale for the job performance—>job

satisfaction sequence is that an individual will form

positive or negative feelings (job satisfaction) after

outcomes on the job (job performance) are obtained, and

compared to the expected and obtained rewards. Bagozzi

[1980] believed that the job satisfaction—>job

performance sequence was more difficult to justify for two

reasons. First, a person must be aware of his/her

feelings (job satisfaction) and attribute these feelings

to specific aspects of the job (job performance). Second,

111

feelings have many antecedents that may not necessarily be

related to job performance. Bagozzi [1980] reported that

job satisfaction did not have a statistically significant

impact on job performance, and job performance had a

statistically significant impact on job performance.

Table 5.10 shows the result of constraining the job

satisfaction—>job performance path. There is a

statistically significant result that indicates that this

path is significant in the theoretical model and should

not be constrained. The null hypothesis that there is no

significant difference between the two nested structural

models is rejected. These results differ from those

reported by Bagozzi [1980]. An explanation for this could

be the difficulty in explaining causality between these

two constructs (for example, Lawler and Porter, 1967;

Porter and Lawler, 1968; Lawler, 1971; Ross and Bomeli,

1971; Benke and Rhode, 1980).

The next SCDT is between the theoretical and

unconstrained model. As with the constrained model,

parameters that are freed in the unconstrained model

should be done in a theoretical perspective. The

parameter that is freed in the unconstrained model is the

sequence job satisfaction—>organizational commitment. As

in this study, other studies have used organizational

commitment as an antecedent to job satisfaction (for

112

TABLE 5.10 COMPARISON OF THE THEORETICAL AND CONSTRAINED MODELS

JOB SATISFACTION—>JOB PERFORMANCE CONSTRAINED

Chi Significance Model Square Difference D.F Level

Constrained 1008.85 94 1

Theoretical 447.94 560.91 93 < .001

1 Difference in degrees of freedom, 1, with difference in

the chi-scjuare value, 560.91.

113

example, Aranya, Lachman, and Amernic, 1982; Hunt, Chonko,

and Wood, 1985; Harrell, Chewning, and Taylor, 1986).

Still other studies have used job satisfaction as an

antecedent to organizational commitment (for example.

Steers, 1977; Zahra, 1984). In these latter studies, job

satisfaction was classified as an attitudinal variable

that precedes organizational commitment.

Table 5.11 presents the results of the SCDT between

the theoretical and unconstrained models. Freeing the

parameter between job satisfaction and organizational

commitment did not result in a statistically significant

change in the chi-square value based on the change in the

degrees of freedom. The null hypothesis that there is no

significant difference between the two nested structural

models is not rejected. Since this path is not

significant and does not contribute to the explanation of

the construct covariances, the theoretical model would be

tentatively accepted.

Evaluating the Theoretical Model

Even though the theoretical model has been

tentatively accepted using the decision-tree framework for

SCDTs, further evaluation is needed to justify accepting

the theoretical model. Anderson and Gerbing [1988]

recommend practical in addition to statistical

considerations. Also, relationships that were

114

TABLE 5.11 COMPARISON OF THE THEORETICAL AND UNCONSTRAINED MODELS

JOB SATISFACTION—>0R(3ANIZATI0NAL COMMITMENT UNCONSTRAINED

Chi Significance Model Square Difference D.F Level

Theoretical 447.94 93 1

Unconstrained 445.49 2.45 92 > .10

1 Difference in degrees of freedom, 1, with difference in

the chi-square value, 2.45.

115

hypothesized to be statistically significant but were not

should be examined.

Practical considerations involve assessing the

goodness-of-fit for the models, in addition to chi-square

test statistics. This is because the value of the chi-

square test statistic is dependent on the size of the

sample. If the sample is large, significant values can be

obtained even if there are only small discrepancies

between the data and model [Anderson and Gerbing, 1988].

Therefore, goodness-of-fit from a practical standpoint

should be considered along with SCDTs. Table 5.12

presents the results of the goodness-of-fit indices,

adjusted goodness-of-fit indices, and root mean square

residuals for each of the models under consideration.

These three assessments are measures of the overall fit of

the model to the data [Joreskog and Sorbom, 1989a].

Goodness-of-fit indices fall between zero and one,

with a desired value of .90 or greater [Bagozzi and Yi,

1988]. The adjusted goodness-of-fit index adjusts the

goodness-of-fit index for the degrees of freedom in the

model. For all of the models, both the goodness-of-fit

indices and the adjusted goodness-of-fit indices surpass

the desired value of .90. The root mean square residual

is the measure of the average of the fitted residuals

[Joreskog and Sorbom, 1989a]. A root mean square residual

116

TABLE 5.12 GOODNESS-OF-FIT INDICES AND ROOT MEAN SQUARE RESIDUALS

FOR THE THEORETICAL, SATURATED, CONSTRAINED, AND UNCONSTRAINED MODELS

Model Goodness-of • Fit Index

Adjusted Goodness-of- Root Mean Fit Index Square Residual

Theoretical .990

Saturated .990

Constrained .978 (JS—>JP)

Unconstrained .990 (JS—>0C)

986

985

968

986

.044

.041

.047

.043

117

of .05 or less is considered an acceptable value to

indicate a good fit [Bagozzi and Yi, 1988]. For all of

the models, the root mean square residual is less than the

.05 criterion. Considering these additional assessments

of the models, all are acceptable. The differences in the

goodness-of-fit indices, adjusted goodness-of-fit indices,

and root mean square residuals are marginal. Since the

SCDTs favor the theoretical model, and the additional

assessments of the theoretical model are acceptable, the

theoretical model is still chosen over the other models.

Inferences about the relationships between constructs

should be made only if they are based on established

principles of scientific inference [Anderson and Gerbing,

1988]. Theory for the theoretical model was established

in the first four chapters of this study. The values

reported in Appendix D support seven of the twelve

hypotheses posited in this study. The five hypotheses (in

alternate form) that were rejected based on statistical

significance should be investigated for other

possibilities. These hypothesized relationships did not

significantly contribute to the theoretical model. These

five hypotheses and associated values are reported in

Table 5.13.

The first hypothesis not accepted is H4, which

posited a positive relationship between professional

118

TABLE 5.13 LISREL VALUES—NONSIGNIFICANT RELATIONSHIPS

1 Hypothesis Value Significance Level

PC->JP/H4 -.901 > .10

LSCD->JP/H6 -1.008 > .10

SM->JS/H7 1.559 > .05

SM->JP/H8 -.962 > .10

LSCD->0C/H12 -.722 > .10

1 The significance of the values are based on a

comparison to a normal distribution table.

119

commitment and job performance. Ferris [1981] described

conflicting findings on the relationship between

commitment and performance, in which highly committed

employees outperform less committed employees, and other

situations in which these two constructs are unrelated [p.

317], In this study, the lack of relationship may be

attributable to the lack of importance placed on

professional commitment for job performance evaluation.

Ramanathan et al. [1976] reported that managing partners

rated personal professional development as the least

important factor for public accounting firms' success,

while staff accountants would like to see additional

professional-oriented factors used in performance

evaluation. This may indicate that staff accountants do

not perceive professional commitment as being an important

factor in job performance evaluations, even though they

would like to see this considered as an important element

in performance evaluation.

Hypotheses six and twelve theorized a significant

positive relationship between life-span career

development, and job performance and organizational

commitment. The results of this study show that both of

these relationships are not significant. These results

may be explained by prior research.

120

Earnest and Lampe [1987] investigated behavioral

choice processes and turnover, reporting that lower level

auditors compared auditing unfavorably with job

alternatives on two intrinsic and two extrinsic factors.

The intrinsic factors are "the feelings experienced as a

result of working on the types of tasks assigned and as a

result of attainment of the task-related goals" [p. 236].

The unfavorable extrinsic factors are performance-

contingent financial compensation and rewards of adequate

leisure time. The unfavorable comparison of auditing with

alternative jobs may explain the negative relationship

between life-span career development and organizational

commitment. Extrinsic rewards such as those reported by

Earnest and Lampe [1987] are associated with the

organization as a whole. If staff auditors feel that

there are alternative jobs that have better extrinsic

rewards, then there may be dissatisfaction with public

accounting as a career choice, and thus lower

organizational commitment. An additional explanation may

be that some individuals enter the public accounting

profession to meet state experience requirements to be a

licensed certified public accountant.

Harrell and Eickhoff [1988] reported that when an

individual's intrinsic needs are not congruent with the

demands of a work environment, he/she is likely to react

121

negatively to that environment (such as job performance),

and seek a more congruent alternate work environment.

These individuals are more likely to express negative

career intentions than individuals who are influence-

oriented, that is, individuals with relatively high power

needs.

It appears that prior research indicates a

relationship between career intentions, organizational

commitment, and job performance based on intrinsic and

extrinsic factors. In this study, these factors were not

considered and may influence the results of hypotheses six

and twelve.

Hypotheses seven and eight posited a positive

relationship between the level of self-monitoring, and job

satisfaction and job performance. The results of this

study are that there is not a statistically significant

relationship between the level of self-monitoring, and job

satisfaction and job performance. Prior research has not

investigated the relationship between these constructs,

making these hypotheses exploratory.

It appears that accountants, as professionals, have

the ability to adapt, or the appearance of adapting, and

thus there is no effect on their perceptions of how they

perform on the job. The relationship between self-

monitoring and job satisfaction was marginally not

122

significant (<.10) at the .05 level. The lack of

statistical support for this hypothesis indicates that

self-monitoring does not make a contribution as an

attitude towards one's level of job satisfaction.

Summary of the Data Analysis

The results of the sequential chi-scjuare difference

tests support the acceptance of the theoretical model over

alternative models. This series of testing "provides

further understanding of the explanatory ability afforded

by the theoretical model of interest" [Anderson and

Gerbing, 1988, p. 419]. The goodness-of-fit index,

adjusted goodness-of-fit index, and root mean square

residual are very good for the theoretical model. In

addition, Bagozzi [1984] stated that rival hypotheses

should be considered in theory construction, and that

these hypotheses should be tested within the same study.

This study considered rival hypotheses by testing and

comparing the constrained and unconstrained models with

the theoretical model.

Significant relationships were evidenced by the

values obtained from the LISREL output. The alternate

hypotheses for the relationships OC->JS, OC->JP, PC->JS,

LSCD->JS, JS->JP, JP->JS, PC->OC, and PC<->LSCD are

supported by the data. Table 5.14 summarizes the

significant and nonsignificant relationships hypothesized

123

TABLE 5.14 LISREL VALUES—ALL HYPOTHESES

Hypothesis

0C->JS/H1

0C->JP/H2

PC->JS/H3

PC->JP/H4

LSCD->JS/H5

LSCD->JP/H6

SM->JS/H7

SM->JP/H8

JP->JS/H9

JS->JP/H9

PC->OC/H10

PC<->LSCD/H11

LSCD->0C/H12

Value

8.430

-6.253

-2.974

-.901

4.565

-1.008

1.559

-.962

-14.212

37.076

12.024

51.829

-.722

1 Significance Level

.001

.001

.001

> .10

.001

> .10

> .05

> .10

.001

.001

.001

.001

> .10

1 The significance of the values are based on a

comparison to a normal distribution table.

124

in this study. Figure 5.6 presents the research model and

summarizes the results of the individual hypotheses in

terms of significant and nonsignificant relationships.

The results provide insight into the relationships

among the constructs. Job performance and job

satisfaction are affected by the level of organizational

commitment possessed by staff accountants. life-span

career development has been shown to have an effect on job

satisfaction and is significantly related with

professional commitment among staff accountants.

Professional commitment has a significant effect on

organizational commitment and job satisfaction among staff

accountants. There are reciprocal effects between job

satisfaction and job performance for staff accountants.

These findings and their implications, additional areas of

potential research, and the limitations of this study are

discussed in the next chapter.

125

sign - significant at .001 probability level ns - not significant at .001 probability level

Figure 5.6. Research Model—Results of Hypotheses Testing

126

CHAPTER VI

SUMMARY AND CONCLUSIONS

In Chapter V, a two-step approach was used to test

the research model. First, the measurement model was

examined and refined. Next, SCDTs were performed to

assist in choosing the model that theoretically and

statistically provided the best information. Last, other

considerations for choosing the model were examined, such

as goodness-of-fit indices, root mean square residuals,

and individual paths between the constructs.

This chapter summarizes the results of the analyses.

Contributions and limitations of the study are presented,

and recommended areas of future research are given.

Summary of the Findings and Implications

The purpose of this study was to investigate

relationships among the constructs professional

commitment, life-span career development, self-monitoring,

organizational commitment, job satisfaction, and job

performance. The relationships among these constructs

were presented in three research objectives: (1)

investigate the effects of organizational commitment,

professional commitment, life-span career development, and

self-monitoring on job satisfaction and job performance

for staff accountants in public accounting firms; (2)

127

investigate the relationship between job satisfaction and

job performance of staff accountants in public accounting

firms; and (3) investigate the relationship among

organizational commitment, professional commitment, and

life-span career development for staff accountants in

public accounting firms. These research objectives will

be discussed based on the results of this study.

Research Objective One

Job performance and job satisfaction were the primary

constructs in this study. To investigate research

objective one, they were treated as endogenous constructs,

which were affected by organizational commitment,

professional commitment, self-monitoring, and life-span

career development. These relationships were posited in

the first eight hypotheses. The effects of organizational

commitment, professional commitment, self-monitoring, and

life-span career development on job satisfaction will be

discussed first, followed by their effects on job

performance.

Hypothesis seven, the effect of self-monitoring on

job satisfaction was not significant and its implications

were discussed in Chapter V. The remaining three

constructs, professional commitment, organizational

commitment, and life-span career development all had

128

significant effects on job satisfaction (see Appendix D,

Table D.4).

The significant effects of professional and

organizational commitment on job satisfaction among staff

accountants suggests that even at the entry-level

position, staff accountants' attitudes toward the

profession and organization will directly affect their

level of job satisfaction. These results support the

earlier work of Norris and Niebuhr [1983], which showed

significant correlations between professional and

organizational commitment, and job satisfaction. However,

Norris and Niebuhr [1983] used all job levels of public

accounting firms, rather than just staff accountants.

The significant positive relationship between life­

span career development and job satisfaction supports

hypothesis five. The result suggests that the further

along individuals are in their implementation stage of

career development, the more satisfied they are with their

job.

Hypotheses four, six, and eight posited a

relationship between job performance and professional

commitment, life-span career development, and self-

monitoring, respectively. These relationships were not

significant, and were discussed in Chapter V.

129

Hypothesis two, the relationship between

organizational commitment and job performance was

significant. These results are partially consistent with

the findings of Ferris [1981] who reported that junior-

level accountants were willing to exert considerable

effort on the job (job performance) on behalf of the

organization, rather than exert effort on the job because

they desired to maintain membership in the organization.

Ferris reported his significant, positive results on the

basis of correlations between the constructs. The

correlations between these two constructs in this study

were also significant and positive. However, under

structural equation modeling, this study reports a

significant but negative relationship between these two

constructs. This may be explained by the high turnover

rates reported in empirical accounting studies (for

example, Bao, Bao, and Vasarhelyi, 1986; Lampe and

Earnest, 1984; Rhode, Sorensen, and Lawler, 1976).

Ferris [1981] reported that nonperformers are

generally not retained by the firm; therefore, there is no

desire to maintain a strong sense of organizational

commitment. In addition, on average, 50% of all employees

leave the firm in the first few years, and less than 25%

attain managerial positions [Ferris, 1981]. Approximately

47% of the respondents in this study had one or more years

130

of experience with the firm. It is possible that some of

the respondents were considering leaving the firm. If

this is true, then there may not be a decrease in job

performance because of the importance of maintaining good

references and professional relationships with the firm,

for future positions outside the firm. Therefore, job

performance can increase with a simultaneous decrease in

the level of organizational commitment.

Research Objective Two

Research objective two was to investigate the

relationship between job satisfaction and job performance

of staff accountants in public accounting firms.

Hypothesis nine posited a significant, nonrecursive

relationship between these two constructs, which was

supported by the results of this study.

The relationship job satisfaction—>job performance

is positive. This suggests that as the level of job

satisfaction increases, so does the level of job

performance. Chapter V discusses the results of treating

this relationship as recursive; that is, no significant

relationship exists in the path job satisfaction—>job

performance.

The relationship job performance—>job satisfaction,

while significant, is negative. This suggests that as job

performance increases, job satisfaction decreases.

131

Job performance may be increasing based on a learning

curve, or some other factors. While staff accountants may

perceive themselves as performing better over time on the

assigned jobs, overall they may not be satisfied with

their jobs. There could be two reasons for these results.

First, it is possible that the same effects discussed

above on job performance and organizational commitment are

occurring in this relationship. While performance is

increasing over time, so is the likelihood of migration

from the firm. Another explanation is that the tasks may

not be as challenging as a staff accountant would like,

and the repetitive nature of the tasks could lead to job

dissatisfaction. Norris and Niebuhr [1983] reported that

job satisfaction decreases for accountants early in their

careers due to the work, self, and home.

Research Objective Three

The third research objective was to investigate the

relationship among organizational commitment, professional

commitment, and life-span career development for staff

accountants in public accounting firms. Hypothesis twelve

posited a positive significant relation between life-span

career development and organizational commitment. This

hypothesis was rejected and discussed in Chapter V.

"Hypothesis ten stated that a staff accountant's

organizational commitment will be directly predicted by

132

professional commitment. This hypothesis was not

rejected, supporting previous empirical studies in

accounting research (for example, Aranya, Lachman, and

Amernic, 1982; Lachman and Aranya, 1986).

Hypothesis eleven posited a significant positive

relation between professional commitment and life-span

career development. This hypothesis was not rejected as

evidenced by the statistically significant covariance

among these two exogenous concepts in matrix phi. This

covariance "parallels the correlation between predictor

variables in multiple regression" [Hayduk, 1987, p. 94].

This suggests that as staff accountants progress further

through the implementation stage of their career

development, their level of commitment to the profession

should also increase.

The use of structural equation analysis to examine

the hypothesized relationships cimong the constructs is a

positive characteristic of this study. A two-step

approach was used, which tests the measurement model prior

to testing the structural submodels. This approach is

considered superior to the traditional approach of testing

individual paths using structural equations [Anderson and

Gerbing, 1988].

The results of the study should also be considered in

light of its limitations and contributions. The next

133

section will discuss these as they apply to the research

design and results of the data analysis.

Contributions

Caplan [1989] discussed the need to examine the

behavior of individuals in public accounting firms. One

of the contributions of this study is the addition to this

body of literature by investigating the relationships of

professional commitment, life-span career development,

self-monitoring, and organizational commitment with job

satisfaction and job performance among staff accountants

in public accounting firms.

Job satisfaction and job performance are constructs

of interest in empirical studies among numerous

disciplines (see footnote 1, Chapter II). The constructs

self-monitoring and life-span career development have not

been used in published empirical studies using job

satisfaction and job performance as endogenous constructs.

Also, a research instrument was developed to measure the

implementation stage of an individual's career

development.

This study used only staff accountants as subjects,

which provides information on a large segment of

professional accountants. For public accounting firms,

this-study provides information that may be useful in

personnel decisions. The importance of maximizing job

134

satisfaction and job performance among employees within

organizations is continually sought by managers.

Maximizing job satisfaction and job performance may lead

to higher productivity, lower overall costs for the firm,

and reduced turnover of employees. While attrition may be

necessary in the public accounting profession due to the

limited number of supervisory and managerial positions,

early detection can diminish the premature or unexpected

attrition of employees whom the firm would like to retain.

In addition, better understanding of the relationship

between job satisfaction and job performance may lead to

minimizing expenditures to increase job performance, job

satisfaction, and costs associated with employee turnover.

Limitations

The subjects were all employees of one office of a

national public accounting firm. As noted earlier, all of

the subjects were staff accountants. This limits the

generalizability of the results to other firms and job

levels.

When nonresponse cannot be eliminated, there is the

possibility of nonresponse bias. Attempts were made to

maximize the response rate by including a stamped self-

addressed envelope, and a letter from a partner in the

firm encouraging participation, with each research

instrument. A follow-up letter was sent to all subjects

135

from a partner in the firm encouraging them to complete

and mail back the research instrument if they had not

already done so. Tests for nonresponse bias were

performed to see if there was a difference between early

and late respondents. While these tests did not indicate

nonresponse bias, it is always possible that nonresponse

bias is present. The overall response rate for this study

was 74%.

Lack of precision in measurement is a possibility in

survey research. With the exception of life-span career

development, the survey questions had been used in

previous empirical studies. While this does not guarantee

reliability and validity, it provides support for the use

of these questions in this study. The questions for life­

span career development were created for this study.

Pretesting was undertaken to support the use of these

questions. Attempts were made to respecify the

measurement model to increase the reliability of the

research instrument. Even with due care, the possibility

of measurement error exists when doing survey research.

Future Research

This study can lead to additional research in the

area of job satisfaction and job performance among

accountants. Comparisons can be made between staff

accountants, managers, and partners with the constructs

136

used in this study. Further stages of life-span career

development can be investigated with managers and partners

as subjects. Additional constructs could be used to

investigate job satisfaction and job performance.

The study can be expanded by surveying staff

accountants, managers, and partners employed in public

accounting firms nationwide. This would increase the

generalizability of the results.

Accountants not employed in public accounting could

be used as subjects, and comparisons can be made between

groups. This would include accountants employed in

private industry, not-for-profit organizations, and

education.

Finally, a study of this type could be used to

investigate how the constructs relate to job turnover.

This would entail a longitudinal study, in which a model

predicated on job satisfaction and job performance would

be developed to attempt to predict job turnover.

137

REFERENCES

Ajzen, I., C. Timko, and J. White. "Self-monitoring and the Attitude-Behavior Relation." Journal of Personality and Social Psychology (1982): 426-435.

Alderfer, C. P. Existence, Relatedness, and Growth (The Free Press, 1972).

Anastasi, A. Psychological Testing (Macmillan Publishers, 1976).

Anderson, J. and D. Gerbing. "Structural Equation Modeling in Practice: A Review and Recommended Two-Step Approach." Psychological Bulletin (1988): 411-423.

Aranya, N., A. Barak, and J. Amernic. "A Test of Holland's Theory in a Population of Accountants." Journal of Vocational Behavior (1981b): 15-24.

and K. Ferris. "A Reexamination of Accountants' Organizational-Professional Conflict." The Accounting Review (January 1984): 1-15.

, N., R. Lachman, and J. Amernic. "Accountants' Job Satisfaction: A Path Analysis." Accounting, Organizations and Society (1982): 201-215.

, N., J. Pollock, and J. Amernic. "An Examination of Professional Commitment in Public Accounting." Accounting, Organizations and Society (1981a): 271-280.

Armstrong, J. and T. Overton. "Estimating Nonresponse Bias in Mail Surveys." Journal of Marketing Research (August, 1977): 396-402.

Arrington, C , C. Bailey, and W. Hopwood. "An Attribution Analysis of Responsibility Assessment for Audit Performance." Journal of Accounting Research (Spring 1985): 1-20.

Bao, B., D. Bao, and M. Vasarhelyi. "A Stochastic Model of Professional Accountant Turnover." Accounting, Organizations and Society (1986): 289-296.

Bagozzi, R. "Performance and Satisfaction In An Industrial Sales Force: An Examination of Their Antecedents and Simultaneity." Journal of Marketing (1980a): 65-77.

138

Bagozzi, R. Causal Models in Marketing (John Wiley & Sons Inc., 1980b).

"A Prospectus for Theory Construction in Marketing." Journal of Marketing (1984): 11-29.

and Y. Yi. "On The Evaluation of Structural Equation Models." Journal of the Academy of Marketing Science (1988): 74-94.

Bartol, K. "Professionalism as a Predictor of Organizational Commitment, Role Stress, and Turnover: A Multidimensional Approach." Academy of Management Journal (1979): 815-821.

Bary, J. and S. Maxwell. Multivariate Analysis of Variance (Sage Publications, Inc., 1990).

Benke, R. Jr. and J. Rhode. "The Job Satisfaction of Higher Level Employees in Large Certified Public Accounting Firms." Accounting, Organizations and Society (1980): 187-201.

Berry, W. Nonrecursive Causal Models (Sage Publications, Inc., 1989).

Bline, D., D. Duchon, and W. Meixner. "The Measurement of Organizational and Professional Commitment: An Examination of the Psychometric Properties of Two Commonly Used Instruments." Behavioral Research in Accounting (1991): 1-12.

Bohrnstedt, G. "Measurement." In Rossi, P. et al. , eds. , Handbook of Survey Research (Academic Press, 1983).

Bollen, K. Structural Eguations with Latent Variables (John Wiley & Sons, Inc., 1989).

Brenner, V., C. Carmack, and M. Weinstein. "An Empirical Test of the Motivation-Hygiene Theory." Journal of Accounting Research (Autumn 1971): 359-366.

Briggs, S., J. Cheek, and A. Buss. "An Analysis of the Self-monitoring Scale." Journal of Personality and Social Psychology (1980): 679-686.

Brownell, P. "Leadership Style, Budgetary Participation and Managerial Behavior." Accounting, Organizations and Society (1983): 307-321.

139

Brownell, P. and and M. Mclnnes. "Budgetary Participation, Motivation, and Managerial Performance." The Accounting Review (October 1986): 587-600.

Bullen, M. and E. Flamholtz. "A Theoretical and Empirical Investigation of Job Satisfaction and Intended Turnover in the Large CPA Firm." Accounting, Organizations and Society (1985): 287-302.

Caplan, E. "Behavioral Accounting—A Personal View." Behavioral Research In Accounting (1989): 109-123.

Chenhall, R. and P. Brownell. "The Effect of Participative Budgeting on Job Satisfaction and Performance: Role Ambiguity as an Intervening Variable." Accounting, Organizations and Society (1988): 225-233.

Choo, F. "Job Stress, Job Performance, and Auditor Personality Characteristics." Auditing: A Journal of Practice & Theory (Spring 1986): 17-34.

Chow, C. "The Effects of Job Standard Tightness and Compensation on Performance: An Exploration of Linkages." The Accounting Review (October 1983): 667-685.

Collins, F., P. Munter, and D. Finn. "The Budgeting Games People Play." The Accounting Review (January 1987): 29-49.

Converse, J. and S. Presser. Survey Questions: Handcrafting the Standardized Questionnaire (Sage Publications, Inc., 1986).

Cook, D. "The Effect of Frequency of Feedback on Attitudes and Performance." Journal of Accounting Research (Selected Studies 1967): 213-224.

De Coster, D. and J. Rhode. "Analysis of Certified Public Accounting Subgroup and Accounting Student Personality Traits Using The California Psychological Inventory." Journal of Vocational Behavior (1972): 155-162.

Dillard, J. "Applicability of an Occupational Goal-Expectancy Model in Professional Accounting Organizations." Decision Sciences (1979): 161-176.

Dillon, W. and M. Goldstein. Multivariate Analysis: Methods and Applications (John Wiley & Sons, 1984).

140

Douglas, W. "Scripts and Self-monitoring When Does Being a High Self-monitor Really Make a Difference?" Human Communication Research (1983): 81-96.

Earnest, K. and J.C. Lampe. "An Expectancy Theory Investigation Into Causes of Employee Turnover In Public Accounting." Advances in Accounting (1987): 215-238.

Ferris, K. "Perceived Uncertainty and Job Satisfaction in the Accounting Environment." Accounting, Organizations and Society (1977a): 23-28.

"A Test of the Expectancy Theory of Motivation in an Accounting Environment." The Accounting Review (July 1977b): 605-615.

"Organizational Commitment and Performance in a Professional Accounting Firm." Accounting, Organizations and Society (1981): 317-325.

"Perceived Environmental Uncertainty, Organizational Adaptation, and Employee Performance: A Longitudinal Study in Professional Accounting Firms." Accounting, Organizations and Society (1982): 13-25.

and D. Larcker. "Explanatory Variables Of Auditor Performance In A Large Public Accounting Firm." Accounting, Organizations and Society (1983): 1-11.

Fisher, F. The Identification Problem in Econometrics (McGraw-Hill, 1966).

Freidson, E. The Professions and Their Prospects (Sage Publications, 1973).

Gaski, J. "Interrelations Among a Channel Entity's Power Sources: Impact on the Exercise of Reward and Coercion on Expert, Referent, and Legitimate Power Sources." Journal of Marketing Research (February, 1986): 62-77.

"Commentary on Howell's Observations." Journal of Marketing Research (February, 1987): 127-129.

Gruneberg, M. (Ed.). Job Satisfaction-A Reader (John Wiley and Sons, 1976).

Harrell, A., E. Chewing, and M. Taylor. "Organizational-Professional Conflict and the Job Satisfaction and Turnover Intentions of Internal Auditors." Auditing; A Journal of Practice & Theory (Spring 1986): 109-121.

141

Harrell, A. and R. Eickhoff. "Auditors' Influence-Orientation and Their Affective Responses to The "Big Eight" Work Environment." Auditing: A Journal of Practice & Theory (Spring 1988): 105-118.

and M. Stahl. "McLelland's Trichomoty of Needs Theory and The Job Satisfaction and Work Performance of CPA Firm Professionals." Accounting, Organizations and Society (1984): 241-252.

Hayduk, L. Structural Equation Modeling with LISREL (The John Hopkins University Press, 1987).

Heneman, H., III. "Comparisons of Self- and Superior Ratings of Managerial Performance." Journal of Applied Psychology (1974): 638-642

Herzberg, F., F. Mausner, and B. Snyderman. The Motivation to Work (John Wiley & Sons, Inc., 1959).

Holland, J. L. Making Vocational Choices: A Theory of Careers (Prentice-Hall, Inc., 1973).

Howell, R. "Covariance Structure Modeling and Measurement Issues: A Note on Interrelations Among a Channel Entity's Power Sources". Journal of Marketing Research (February, 1987): 119-126.

Hunt, S., L. Chonko, and V. Wood. "Organizational Commitment and Marketing." Journal of Marketing (Winter, 1985): 112-126.

, R. Sparkman Jr., and J. Wilcox. "The Pretest in Survey Research: Issues and Preliminary Findings." Journal of Marketing Research (1982): 269-273.

Huizinga, G. Maslow's Need Hierarchy in the Work Situation (Welters, 1970).

Ickes, W. and R. Barnes. "The Role of Sex and Self-monitoring in Unstructured Dyadic Interactions." Journal of Personality and Social Psychology (1977): 315-330.

James, L., S. Mulaik, and J. Brett. Causal Analysis: Assumptions, Models, and Data (Sage Publications, Inc., 1987) .

142

Jiambalvo, J. "Performance Evaluation and Directed Job Effort: Model Development and Analysis in a CPA Firm Setting." Journal of Accounting Research (Autumn 1979): 436-455.

"Measures of Accuracy and Congruence in the Performance Evaluation of CPA Personnel: Replication and Extensions." Journal of Accounting Research (Spring 1982): 152-161.

and J. Pratt. "Task Complexity and Leadership Effectiveness in CPA Firms." The Accounting Review (October 1982): 734-750.

, D. Watson, and J. Baumler. "An Examination of Performance Evaluation Decisions in CPA Firm Subunits." Accounting, Organizations and Society (1983): 13-29.

Joreskog K. and D. Sorbom. PRELIS (Scientific Software, 1988).

. LISREL VII (Scientific Software, 1989a).

LISREL VII: A Guide to the Program and Applications (SPSS Inc., 1989b).

Kerlinger, F. Foundations of Behavioral Research (Holt, Rinehart and Winston, 1986).

Kim, J. and C. Mueller. Factor Analysis: Statistical Methods and Practical Issues (Sage Publications, Inc., 1986) .

and C. Mueller. Introduction to Factor Analysis: What It Is and How To Do It (Sage Publications, Inc., 1986) .

Kulik, J. and S. Taylor. "Self-monitoring and the Use of Consensus Information." Journal of Personality (1981): 75-84.

Lachman, R. and N. Aranya. "Evaluation of Alternative Models of Commitments and Job Attitudes of Professionals." Journal of Occupational Behavior (1986): 227-243.

Lampe, J. C. and K. Earnest. "How Motivation Affects Accountants' Productivity and Turnover." Management Accounting (1984): 50-55.

143

Landy, F. and J. Farr. The Measurement of Performance (Academic Press, Inc., 1983).

, S. Zedeck, and J. Cleveland (Eds.). Performance, Measurement and Theory (Lawrence Erlbaum Publishers, 1983).

Lawler, E., III. "Ability as a Moderator of the Relationship between Job Attitudes and Job Performance." Personnel Psychology (1966): 153-164.

"A Correlational-Causal Analysis of The Relationship Between Expectancy Attitudes and Job Performance." Journal of Applied Psychology (1968): 462-468.

Pay and Organizational Effectiveness: A Psychological View (McGraw-Hill Book Co., 1971).

, W. Kuleck, Jr., J. Rhode, and J. Sorensen. "Job Choice and Post Decision Dissonance." Organizational Behavior and Human Performance (1975): 133-145.

and L. Porter. "The Effect of Performance on Job Satisfaction." Industrial Relations (1967): 20-28.

Lengermann, J. "Supposed and Actual Differences in Professional Autonomy Among CPAs as Related to Type of Work Organization and Size of Firm." The Accounting Review (October 1971): 665-675.

Locke, E. "The Nature and Cause of Job Satisfaction." In Dunnette, M. ed.. Handbook of Industrial Psychology (John Wiley and Sons, 1983).

Maslow, A. Motivation and Personality (Harper & Row, Inc., 1954).

McClelland, D. C. The Achieving Society (Van Nostrand, 1961).

Mossholder, K., A. Bedeian, J. Touliatos, and A. Barkman. "An Examination of Interoccupational Differences: Personality, Perceived Work Climate, and Outcome Preferences." Journal of Vocational Behavior (1985): 164-176.

Mowday, R. , R- Steers, and L. Porter. "The Measurement of Organizational Commitment." Journal of Vocational Behavior (1979): 224-247.

144

Norris, D. and R. Niebuhr. "Professionalism, Organizational Commitment and Job Satisfaction in an Accounting Organization." Accounting, Organizations and Society (1983): 49-59.

Nunnally, J. Psychometric Theory, Second Edition (McGraw-Hill Book Company, 1978).

Osipow, S. Theories of Career Development, Third Edition (Prentice-Hall, Inc., 1983).

Philips, J. and S. Freedman. "Situational Performance Constraints and Task Characteristics: Their Relationship to Motivation and Satisfaction." Journal of Management (1984): 321-331.

Porter, L. and E. Lawler III. Managerial Attitudes and Performance (Richard D. Irwin, Inc. and The Dorsey-Press, 1968).

, R. Steers, R. Mowday, and R. Boulian. "Organizational Commitment, Job Satisfaction and Turnover Among Psychiatric Technicians." Journal of Applied Psychology (1974): 603-609.

Price, J. and C. Mueller. Handbook of Organizational Commitment (Pitman Publishing Inc., 1986).

Prien, E. and R. Liske. "Assessments of Higher-Level Personnel Rating Criteria: A Comparative Analysis of Supervisor Ratings and Incumbent Self-Ratings of Job Performance." Personnel Psychology (1962): 187-194.

Rasch, R. and A. Harrell. "The Impact of Personal Characteristics on the Turnover Behavior of Accounting Professionals." Auditing; A Journal of Practice & Theory (1990): 90-102.

Ramanatham, K., R. Peterson, and M. Maher. "Strategic Goals and Performance Criteria in CPA Firms." The Journal of Accountancy (January 1976): 56-64.

Rasch, R. and A. Harrell. "The Impact of Personal Characteristics on the Turnover Behavior of Accounting Professionals." Auditing: A Journal of Practice & Theory (Spring, 1990): 90-102.

Reed, S. and S. Kratchman. "Job Satisfaction: Differences Throughout the Accounting Profession." The Ohio CPA Journal (Autumn 1987): 33-47.

145

Rhode, J., J. Sorensen, and E. Lawler. "An Analysis of Personal Characteristics Related to Professional Staff Turnover in Public accounting Firms." Decision Sciences (October 1976): 771-797.

Ross, T. and E. Bomeli. "A Comment on Accountants' Job Satisfaction. " Journal of Accounting Research (Autumn 1971): 383-388.

Seller, R. and R. Sapp. "Just How Satisfied Are Accountants With Their Jobs?" Management Accounting (March 1979): 18-21.

Senatra, P. "Role Conflict, Role Ambiguity, and Organizational Climate in a Public Accounting Firm." The Accounting Review (October 1980): 594-603.

Sims, H. Jr. and A. Szilagyi. "Leader Reward Behavior and Subordinate Satisfaction and Performance." Organizational Behavior and Human Performance (1975): 426-439.

Snyder, M. "Self-monitoring of Expressive Behavior." Journal of Personality and Social Psychology (1974): 526-537.

Advances in Experimental Social Psychology (Academic Press, 1979).

"The Influence of Individuals on Situations: Implications for Understanding the Links Between Personality and Social Behavior." Journal of Personality (1983): 497-516.

, E. Berscheid, and P. Click. "Focusing on the Exterior and Interior: Two Investigations of the Initiation of Personal Relationships." Journal of Personality and Social Psychology (1985): 1427-1439.

and N. Cantor. "Thinking About Ourselves and Others: Self-monitoring and Social Knowledge." Journal of Personality and Social Psychology (1980): 222-234.

and S. Gangestad. "Choosing Social Situations: Two Investigations of Self-monitoring Processes." Journal of Personality and Social Psychology (1982); 123-135.

146

Snyder, M., S. Gangestad, and J. Simpson. "Choosing Friends as Activity Partners: The Role of Self-monitoring." Journal of Personality and Social Psychology (1983): 961-972.

and T. Monson. "Persons, Situations, and the Control of Social Behavior." Journal of Personality and Social Psychology (1975): 637-644.

and J. Simpson. "Self-monitoring and Dating Relationships." Journal of Personality and Social Psychology (1984): 1281-1291.

and W. Swann Jr. "When Actions Reflect Attitudes: The Politics of Impression Management." Journal of Personality and Social Psychology (1976): 1034-1042.

and E. Tanke. "Behavior and Attitude: Some People are More Consistent Than Others." Journal of Personality (1976): 510-517.

, E. Tanke, and E. Berscheid. "Social Perception and Interpersonal Behavior: On the Self-fulfilling Nature of Social Stereotypes." Journal of Personality and Social Psychology (1977): 656-666.

Sorensen, J. "Professional and Organizational Profiles of the Migrating and Non-migrating Large Public Accounting Firm CPA." Decision Sciences (1970): 489-512.

and D. Franks. "The Relative Contribution of Ability, Self-Esteem and Evaluative Feedback to Performance: Implications for Accounting Systems." The Accounting Review (October 1972): 735-746.

and T. Sorensen. "The Conflicts of Professionals in Bureaucratic Organizations." Administrative Science Quarterly (March 1974): 98-106.

Steers, R. "Antecedents and Outcomes of Organizational Commitment." Administrative Science Quarterly (March 1977): 46-56.

Strawser, R. , J. Ivancevich, and H. Lyon. "A Note on the Job Satisfaction of Accountants in Large and Small CPA Firms." Journal of Accounting Research (Autumn 1969): 339-345.

147

Super, D. E. "Vocational Development in Adolescence and Early Adulthood: Tasks and Behavior." In Super, D. E. et al.. Career Development: Self-Concept Theory (CEEB Research Monograph No. 4, 1963).

The Psychology of Careers (Harper and Row, 1957).

and E. Knasel. "Career Development in Adulthood: Some Theoretical Problems and a Possible Solution." British Journal of Guidance and Counseling (July 1981): 194-201.

Sutton, S. G. and J. C. Lampe. "Monitoring Quality in a Knowledge-Work Environment: Integrating Structured Group Processes with Decision Support System Design." Productivity Management Frontiers (Vol. II 1988): 365-372.

Teas, K. "Supervisory Behavior, Role Stress, and the Job Satisfaction of Industrial Salespeople." Journal of Marketing Research (February 1983): 84-91.

Teel, K. "Self-Appraisal Revisited." Personnel Journal (July 1978) : 364-367.

Thornton, G., III. "Psychometric Properties of Self-Appraisals of Job Performance." Personnel Psychology (1980): 263-271.

Todd, J., P. Thompson, and G. Dalton. "Management Control of Personnel." The Journal of Accountancy (February 1974): 34-40.

Tracy, L. and T. Johnson. "What do the Role Conflict and Role Ambiguity Scales Measure?" Journal of Applied Psychology (October 1981): 464-469.

Turner, R. "Self-monitoring and Humor Production." Journal of Personality (1980): 163-172.

Vroom, V. Work and Motivation (John Wiley & Sons, Inc., 1964) .

Wahba M. and L. Bridwell. "Maslow Reconsidered: A Review of Research on the Need Hierarchy Theory." Organizational Behavior and Human Performance (April 1976): 212-237.

148

Weiss, D., R. Dawis, G. England, and L. Lofquist. Manual for the Minnesota Satisfaction Questionnaire (University of Minnesota, 1967).

Wright, A. "Performance Appraisal of Staff Auditors." The CPA Journal (November 1980): 37-43.

"An Investigation of the Engagement Evaluation Process for Staff Auditors." Journal of Accounting Research (Spring 1982): 227-239.

Zahra, S. "Antecedents and Consequences of Organizational Commitment: An Integrative Approach." Akron Business and Economic Review (Fall 1984): 26-32.

Zytowski D. and R. Hay. "Do Birds of a Feather Flock Together? A Test of the Similarities Within and the Differences Between Five Occupations." Journal of Vocational Behavior (1984): 242-248.

149

APPENDIX A

RESEARCH INSTRUMENT

150

FOR YOOa ACTIOH Memorandum

0 All Persoaael DUtribudoQ

Limiud DLstribudon: Partners Maiugers Audit • f TAX • • .MC • f Admin. 0

SuiT t •

•

0

Secrcttries 0 0 0

° TO: All Professional Personnel

FROM:

DISSERTATION QUESTIONNAIRB

Attached is a questionnaire which all office professionals are being asked to complete in order to assist a doctoral student with his dissertation. A high response rate is critical to the success of the dissertatloo. The Firm has committed Its assistance to this project and therefore, I urge each of you to take the time to complete this questionnaire. Thank you for your cooperation.

151

Peter John Poznanski, CPA 3219 Chadwick Drive Rockford, II. 61109

Certified Public Accountants

Dear

Attached is a copy of a questionnaire I will be using in my doctoral dissertation. Your assistance in answering the questions will facilitate my completing the dissertation required for my degree, and contribute to the body of knowledge in accounting.

All of your responses will be kept in the strictest of confidence, known only to myself. In no way will this questionnaire identify you, nor will any attempt be made to identify you by your responses. The first page requesting the demographic information will be detached from the questionnaire when I receive it, and the questionnaires and demographic data sheet will be destroyed as soon as the data is entered into the computer. You can return the completed questionnaire in the attached self-addressed st£unped envelope.

One of the more frustrating aspects of answering a questionnaire of this nature is not knowing the purpose behind the questions. The reason for this is that unconsciously you may bias your responses towards the answers you feel are appropriate rather than your true feelings. While I believe this point is debatable, it is a requirement imposed upon roe for this study. However, the time you take to answer the questions as best you can will greatly enhance the results of this study. This study will culminate four years of my graduate school work, so your responses are very important and very necessary to me.

The data analysis and the dissertation should be completed early next year. I will be sending

a copy of the completed study. If you are interested in receiving a copy of the results please retain my address and contact roe. I will be happy to share a copy of the results with you. Thank you for taking the time to help roe complete ray schooling and study.

Sincerely,

Peter John Poznanski, CPA

152

Demographic Data

1. Age:

2. Sex:

3

4

5

6

7

Number of years in the public accounting profession: __

Years at current Firm:

Position at current Firm (Please circle the appropriate response): ^^ ^^^ixacc

Partner or Principal

Senior Manager

Manager

Senior Consultant

Supervising Senior, Audit

Supervising Senior, Tax

Senior Accountant, Audit

Senior Specialist, Tax

Consultant

Staff Accountant

Tax Specialist

Assistant Accountant

Assistant Tax Specialist

Intern

Preprofessional

Staff Technician

Paraprofessional

Administrative

Other (please specify):

Number of years in current Position:

College Degree(s) and Major and Minor course work for each degree:

8. Are you currently enrolled in a degree program, emd if so which degree are you working on?

9. Professional Certification(s) and the number of years with the Certification (e.g., CPA, two years):

153

Please c i r c l e the nmnber-frorn one to five--which best represents your gersonal fee l ings towards the accounting

Strongly Disagree

1 Disagree

2 Neutral

3 Agree 4

Strongly Agree 5

I am willing to put in a great deal of effort beyond that normally expected m order to help make my profession successful

I talk up this profession with my friends as a great profession to be associated with

I feel very little loyalty to this profession

I would accept almost any type of job assignment in order to keep working in areas that are associated with this profession

I find that my values and the profession's values are very similar...

I am proud to tell others that I am part of this profession

I could just as well be associated with another profession as long as the type of organization in which I worked were similar

Being a member of this profession really inspires the very best in roe in the way of job performance

It would take very little change in my present circumstances to cause me to work in areas that are not associated with this profession

I am extremely glad that I chose this profession over others I was considering at the time I joined

2

2

2

2

2

4

4

4

4

5

5

5

5

5

154

Strongly Disagree

1 Disagree

2 Neutral

3 Agree

4

Strongly Agree

5

There's not too much to be gained by sticking with this profession indefinitely 1

Often, I find it difficult to agree with this profession's policies on important matters relating to its members 1

I really care about the fate of this profession 1

For me this is the best of all professions to be a member of 1

Deciding to be a member of this profession was a definite mistake on my part 1

2

2

2

3

3

3

4

4

4

5

5

5

Please c i r c l e the number--frora one to five—which best represents your personal feelings on your performance of your present job.

Needs Unsatisfactory Improvement Satisfactory Superior Outstanding

1 2 3 4 5

Maintaining quantity of work

Maintaining quality of work

Communicating orally

Communicating in writing

Accepting responsibility and initiating action

Exercising professional skills and care

Following policies and procedures

Planning and organizing work

2

2

2

2

2

2

2

5

5

5

5

5

5

5

155

Needs Unsa t i s f ac to ry Improvement Sa t i s fac tory Superior Outstanding

1 2 3 4 5

Adapting to new or d i f f e r e n t job s i t u a t i o n s 1 2 3 4 5

G e t t i n g along with o the r s in the f i rm. . 1 2 3 4 5

Dealing with c l i e n t s outs ide the f i r m . . 1 2 3 4 5

Supervis ing others 1 2 3 4 5

P l e a s e c i r c l e the number—from one t o five—which best r e p r e s e n t s your personal feel ings towards your accounting c a r e e r .

St rongly Strongly Disagree Disagree Neutral Agree Agree

1 2 3 4 5

My previous education and training have given me adequate preparation for ray current pos i t ion 1 2 3 4 5

I have completed my training for my career 1 2 3 4 5

Ky current job is relevant to my career goals 1 2 3 4 5

I am considering changing careers 1 2 3 4 5

My current career offers job security.. 1 2 3 4 5

I feel that the amount of time that I will be required to stay in ny current position before I am eligible for a promotion is reasonable 1 2 3 4 5

I am satisfied with my career choice to be an accountant 1 2 3 4 5

I am not prepared for my current position 1 2 3 4 5

156

Please circle the number--from one to five--which best represents your personal feelings towards your present job.

Very Dissatisfied

1 Dissatisfied

2 Neutral

3 Satisfied

4

Very Satisfied

Being able to keep busy all the time...

The chance to work alone on the job

The chance to do different things from time to time

The chance to have some prestige in the community because of my job

The way my supervisor handles his/her people

The competence of my supervisor in making decisions

Being able to do things that don't go against my conscience

The way the job provides for steady employment

The chance to do things for other people

The chance to supervise other people...

The chance to do things that make use of my abilities

The way organizational policies are put into practice

My pay and the amount of work I do.

The chances for advancement ,

The freedom to use my own judgment.

The chance to try ray own methods of doing the job

The general working conditions

157

Very ^ Dissa t i s f i ed Dissat i s f ied Neutral Satisfied Sat is f ied

1 2 3 4 5

The way my co-workers get along with each other X 2 3 4

The feedback I get for doing an e f f e c t i v e job 1 2 3 4 s

The feel ing of accomplishment I get from the job ; 1 2 3 4 5

Please circle the number—from one to five—which best 5??h^wh?^h r^^ E^isonal feelings towards the orga^zation with which you are presently employed. Please"nite that

Strongly Disagree

1 Disagree

2 Neutral

3 Agree 4

Strongly Agree 5

I am willing to put in a great deal of effort beyond that normally expected in order to help this organization be successful ,

I talk up this organization with my friends as a great organization to be associated with

I feel very little loyalty to this organization

I would accept almost any ty];>e of job assignment in order to keep lurking for this organization

I find that ny values and the organization's values are very similar.

I am proud to tell others that I an part of this organization

I could just as well be working for a different organization as long as the type of work was similar

4

4

4

4

4

5

5

5

5

5

158

strongly Disagree

1 Disagree

2 Neutral

3 Agree 4

Strongly Agree 5

This organization really inspires the very best in me in the way of job performance 1

It would take very little change in my present circumstances to cause me to leave this organization 1

I am extremely glad that I chose this organization to work for, over others I was considering at the time I joined. 1

There is not too much to be gained by sticking with this organization indefinitely 1

Often, I find it difficult to agree with this organization's policies on important matters relating to its employees 1

I really care about the fate of this organization 1

For me this is the best of all possible organizations for which to work 1

Deciding to work for this organization was a definite mistake on my part 1

3

3

3

3

4

4

4

4

5

5

5

5

The statements on the following pages concern your personal reactions to a number of different situations. No two statements are exactly alike, so consider each statement carefully before answering. If a statement is TRUE or MOSTLY TRUE as applied to you, circle the word True next to the question. If a statement is FALSE or NOT USUALLY TRUE as applied to you, circle the word False next to the question.

I find it hard to imitate the behavior of other people

My behavior is usually an expression of my true inner feelings, attitudes, and beliefs

True

True

False

False

159

At-parties and social gatherings, I do not attempt to do or say things that others will like True False

I can only argue for ideas which I already believe True False

I can make impromptu speeches even on topics about which I have almost no information True False

I guess I put on a show to impress or entertain people True False

When I am uncertain how to act in a social situation, I look to the behavior of others for cues True False

I would probably make a good actor True False

I rarely need the advice of my friends to choose movies, books, or music True False

I sometimes appear to others to be experiencing deeper emotions than I actually am True False

I laugh more when I watch a comedy with others thain when alone True False

In a group of people I am rarely the center of attention True False

In different situations and with different people, I often act like very different persons True False

I aro not particularly good at making other people like roe True False

Even if I aro not enjoying royself, I often pretend to be having a good time. True False

I'm not always the person I appear to be True False

I would not change my opinions (or the way I do things) in order to please someone else or win their favor True False

I have considered being an entertainer. True False

160

In order to get along and be liked, I tend to be what people expect me to be rather than anything else True False

I have never been good at games like charades or improvisational acting True False

I have trouble changing my behavior to suit different people and different situations True False

At a party I let others keep the jokes and stories going True False

I feel a bit aw)cward in company and do not show up quite so well as I should.. True False

I can look anyone in the eye and tell a lie with a straight face (if for a right end) True False

I may deceive people by being friendly when I really dislike them True False

Please circle your responses to the following two questions.

Are you planning to leave the firm at which you are currently employed?

Yes Undecided No

Axe you planning to remain in the public accounting profession?

Yes Undecided No

Thank you very much for your time and assistance.

161

APPENDIX B

RESEARCH INSTRUMENT APPROVAL

162

TEXASXECH 0«f<t ol t rwifch S«r«<ei

8oi «*,'0/U66oci. r«i4t 71*01. IP 35/(BOH N j j

July 10, 1990

Dr. reter John Pozhanski Accounting Campus

Caar OiT. Pozhanski;

The Te.' as Tech University Committee for the Protection of Human Subjec t s has approved your project, "Tne Effects of Orga.'.izat lonal Commitment, Professional Commitment, Life Span Career Oevelopraent, and Self-Monitoring on Job Sat i s fac t icn and Job Performance Among Staff Accountants." The approval v i l l extend for one year from J u l y 31, 1990. You w i l l be reminded of the pending e x p i r a t i o n one month before your approval expires so that you .-nay request an e x t e n s i o n if you wish.

The best of luc)c on your project .

Sincerely,

Jaoes Saith Chairperson, Hunan Subjects Use Coocittee

JS/ah

•4A (qu*i Cop<yc«*^/'^"'"*'<"»« '*<i»0*» 'nuMuifOM'

APPENDIX C

DEMOGRAPHIC DATA AND DESCRIPTIVE

STATISTICS

164

TABLE C.I DEMOGRAPHIC DATA

1 Category Frequency of Responses

Sex-

Age-

Years

Years

Male Female No Response

21 22 23 24 25 26 27 28 29 Over 30 No Response

Professional 1/2-1 1 - 2 2 - 3 3 - 4

Over 4

With Current 1/2-1 1 - 2 2 - 3 3 - 4

Over 4

1 All frequencies

responses.

Accountant-

Firm-

-

total 281, the number of

146 134 1

10 59 57 45 46 28 7 4 6 18 1

151 61 42 20 7

158 65 37 16 5

usable

165

Construct

Organizational Commitment

Job Satisfaction

Job Performance

Professional Commitment

Life-Span Career Development

Self-Monitoring

TABLE DESCRIPTIVE

No.

281

281

281

281

281

281

C.2 STATISTICS

Mean

3.613

3.608

3.481

3.554

3.492

1.491

Std. Dev.

.611

.459

.514

.564

.518

.500

The responses, except for self-monitoring, were made on a five point Likert-type scale, where higher responses indicate stronger levels of the construct. The values for self-monitoring were either 1 (for low self-monitoring) or 2 (for high self-monitoring).

166

APPENDIX D

LISREL RESULTS FOR INDIVIDUAL

HYPOTHESES AND PARAMETERS

167

TABLE D.l TESTING INDIVIDUAL PHI

Model Chi-Square Difference D.F

Significance Level

Null (baseline) 10693.69 106

Relationship and Hypothesis Tested:

PC<->LSCD/H11 10333.55 360.14 105 .001

Significance level based on 1 degree of freedom difference and the change in the chi-square value.

168

TABLE D.l TESTING INDIVIDUAL PHI

Model Chi-Square Difference D.F

Significance Level

Null (baseline) 10693.69 106

Relationship and Hypothesis Tested:

PC<->LSCD/H11 10333.55 360.14 105 .001

Significance level based on 1 degree of freedom difference and the change in the chi-square value.

168

TABLE D.2 TESTING INDIVIDUAL BETAS—WITH PHI

1 Chi- Significance

Model Square Difference D.F Level

Null (baseline) 10333.55 105

Relationship and Hypothesis Tested:

0C->JS/H1

0C->JP/H2

JP->JS/H9

JS->JP/H9

10109.23

10321.76

9641.96

9824.16

224.32

11.79

691.59

509.39

104

104

104

104

.001

.001

.001

.001

Significance level based on 1 degree of freedom difference and the change in the chi-square value.

169

TABLE D.3 TESTING INDIVIDUAL GAMMAS—WITH PHI

Model Chi-Square Difference D.F

Significance Level

Null (baseline) 10333.55 105

Relationship and Hypothesis Tested:

PC->OC/H10

PC->JS/H3

PC->JP/H4

LSCD->0C/H12

LSCD->JS/H5

LSCD->JP/H6

SM->JS/H7

SM->JP/H8

5899.77

10264.41

10323.41

5949.34

10294.33

10263.84

10290.87

10333.43

4433.78

69.14

10.14

4384.21

39.22

69.71

42.68

.12

104

104

104

104

104

104

104

104

.001

.001

.005

.001

.001

.001

.001

> .10

Significance level based on 1 degree of freedom difference and the change in the chi-square value.

170

TABLE D.4 LISREL VALUES FOR THE HYPOTHESES

1 Significance

Hypothesis Value Level

0C->JS/H1

0C->JP/H2

PC->JS/H3

PC->JP/H4

LSCD->JS/H5

LSCD->JP/H6

SM->JS/H7

SM->JP/H8

JP->JS/H9

JS->JP/H9

PC->OC/H10

PC<->LSCD/H11

LSCD->0C/H12

8.430

-6.253

-2.974

-.901

4.565

-1.008

1.559

-.962

-14.212

37.076

12.024

51.829

-.722

.001

.001

.001

> .10

.001

> .10

> .05

> .10

.001

.001

.001

.001

> .10

The significance of the values are based on a comparison to a normal distribution table.

171

TABLE D.5 1 LISREL VALUES FOR LAMBDA Y AND LAMBDA X

Lambda Y

OCl 0C2 0C3 JSl JS2 JS3 JPl JP2 JP3

Lambda

OC 1 127 112

X

JS

1 63 67

JP

1 76 77

PC LSCD SM PCI 1 PC2 102 PC3 108 LSCDl 1 LSCD2 43 LSCD3 37 SM 1

The value 1 indicates that the starting values was set at 1 to link the specific indicator to the concept [Hayduk, 1987].

1 All values are significant at the .001 level.

172

Lambda

OCl 0C2 0C3 JSl JS2 JS3 JPl JP2 JP3

Lambda

PCI PC2 PC3 LSCDl LSCD2 LSCD3 SM

Beta

OC JS JP

Gamma

OC JS JP

TABLE D.6 STANDARDIZED SOLUTIONS

Y OC .862 .915 .880

X PC .862 .839 .883

OC

.671 -.467

PC .884

-.662 -.151

JS JP

.813

.832

.767 •

•

•

LSCD

.761

.677

.708

JS

1.431

LSCD -.057 .906

-.141

841 847 866

SM

.894

JP

-1.007

SM

.131 -.102

173