Public Management Research Association

Work Motivation in the Senior Executive Service: Testing the High Performance Cycle
Theory
Author(s): Sally Coleman Selden and Gene A. Brewer
Source: Journal of Public Administration Research and Theory: J-PART, Vol. 10, No. 3
(Jul., 2000), pp. 531-550
Published by: Oxford University Press on behalf of the Public Management Research
Association
Stable URL: https://www.jstor.org/stable/3525627
Accessed: 28-02-2019 10:47 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Public Management Research Association, Oxford University Press are collaborating with
JSTOR to digitize, preserve and extend access to Journal of Public Administration
Research and Theory: J-PART

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service:
Testing the High Performance Cycle Theory

Sally Coleman Selden

Syracuse University
Gene A. Brewer
University of Georgia

ABSTRACT

Some scholars believe that more effort should be dev

the development of a comprehensive theory of work motiva
Several candidate theories have been advanced in recent
but they are radically different and empirical testing is alm
nonexistent. This article reports the first known effort to
of these metatheories-the high performance cycle propo
Edwin A. Locke and Gary P. Latham. Specifically, we as
empirical adequacy of the high performance cycle amon
federal executives, using attitudinal data from the U.S. O
Personnel Management's 1991-1992 Survey of Federal G
ment Employees. The overall fit of the LISREL 8 model
and there is strong empirical support for the key hypothes
relationships. The article concludes with a discussion of
implications and limitations of this research.

Work motivation is one of the most intensely studied

in the social sciences,1 and it is one of the big questions
'For reviews of the literature, see Rainey
management (Behn 1995). Scholars have devoted substant
(1994), Katzell and Thompson (1990b),
Landy and Becker (1987), Locke and and effort to developing a master theory of work motivatio
Henne (1986), and Pinder (1984).such a theory has proven to be elusive. As Katzell and T
(1990a, 65) note, empirical studies on various theories ty
Revised version of a paper presented
explain less than 20 percent of the variance in work outp
at the annual meeting of the American
Political Science Association, New see Locke and Henne 1986). The literature is peppered w
York,
September 1-4, 1994, and the twenty-fifth
flicting empirical findings that do not conclusively answer
anniversary celebration conference for
questions. Considering the vast intellectual resources spe
the doctor of public administration pro-
topic, Pinder (1984) concluded that progress has been pa
gram, University of Georgia, Athens,
May 9-10, 1997. The authors thank slow. Shamir (1990, 405) added that current reviews are
mous in their dissatisfaction with the 'state of the art.'"
Hal G. Rainey, J. Edward Kellough,
Stuart I. Bretschneider, and four anony-
mous reviewers for helpful comments on
Despite weak explanatory power and conflicting empirical
an earlier version of the manuscript.
findings, the mass of research on work motivation has improved
J-PART 10(2000):3:531-550
our understanding of the concept, and it has helped produce tools

531/Journal of Public Administration Research and Theory

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

to motivate employee performance. In practice, however, some

of these tools fail to produce the positive results that theory
predicts. For example, merit pay incentive plans are founded on
expectancy theory, but past research shows that these plans have
little motivational effect in the public or private sectors (for a
review of these studies, see Heneman 1992). Yet merit pay con-
tinues to have considerable appeal. In other words, when studies
fail to confirm our intuitive sense of what motivates employees,
we tend to stick with our intuition and conduct more studies.

Another reason for the chronic lack of progress in work

motivation research is that motivation is a difficult concept to
define and operationalize. Capturing the complexity of human
behavior in a recursive theoretical framework is difficult, and
such a framework is hard to operationalize and test in real world
settings. Besides, practical concerns about work motivation over-
shadow its theoretical significance. Managers in government and
business cope with motivational issues constantly, and they tend
to view the topic from a practical rather than a theoretical per-
spective. The sheer complexity of work motivation is a related
problem. Some recent scholars have portrayed work motivation
as a unifying concept that links together a network of subordinate
concepts (e.g., goals, individual performance, rewards, job satis-
faction, and organizational commitment). Each of these subordi-
nate concepts has generated a large research literature of its own,
and researchers must confront vexing questions about the inter-
relationships among these variables. A final problem is that past
research has depicted work motivation too simplistically. Various
theories such as the reinforcement, equity, and expectancy theo-
ries have focused primarily on the influence of one or more
values. Locke and Henne (1976) assert that some of these theo-
ries are "simply wrong," while Katzell (1982) characterizes them
as "incomplete." In reality, several of these theories may work in
concert.

2Some scholars contend that it is too early Katzell and Thompson (1990b, 151) recently concluded
to propose comprehensive theories of of motivation are individually incom-
work motivation (Campbell and Pritchard
1976; Pinder 1984; Landy and Becker plete and it would be desirable to integ
1987). Landy and Becker (1987) have hensive framework ..." for empiric
argued for less extensive middle-range 1982).2 Recently, several such metath
theories (also see Mitchell 1982). For
example, Mayes (1978) emphasized the (Evans 1986; H. Klein 1989; J. Kle
importance of individual attributes, 1990a and 1990b; Katzell and Thom
whereas Staw (1977) emphasized situa- vais, and Scholl 1995).3 The purpose o
tional factors. one promising metatheory of work motiv
3The term metatheory refers to a compre-and Gary P. Latham's (1990a) high
hensive theoretical framework that synthe- will explain the high performance
sizes existing research on a topic. the theory with attitudinal data collected

532/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms

4For a more in-depth discussion of this
model, see Locke and Latham (1990a and
result.
1990b).
5The high performance cycle does not
specify any differences in public- and
private-sector work contexts or the
motivati*on of public- and private-sector
employees. Yet there are more constraints
in the public sector, and public servants
are motivated by a stronger desire to
perform public, community, and social
service (Perry and Porter 1982; Rainey
1997; Brewer and Selden 1998; Brewer,
Selden, and Facer 1999). In a recent
study of SES members, Ingraham,
Murlis, and Peters (1999, 22) wrote:
"One of the most striking findings in
these results is a strong and idealistic
commitment to the public service." The
model partially incorporates this type of
motivation with the moderating factor of
goal commitment; however, it does not
specify important motivational foci that
lie beyond the rim of the organization
(i.e., public, community, and social
service).
This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
Work Motivation in the Senior Executive Service
executives; and third, we will discuss the implications and
limitations of this research.
THE HIGH PERFORMANCE CYCLE4
Locke and Latham's high performance cycle is a promising
model of work motivation for several reasons. First, the authors
acknowledge the strengths and limitations of past research, and
they explicitly try to integrate this research into the high
performance cycle. Second, the authors incorporate concepts into
the model that have been carefully developed in past theoretical
and empirical research, and these concepts have practical mean-
ing to managers and employees (e.g., goals, individual perform-
ance, rewards, job satisfaction, and organizational commitment).
In other words, if the model proves to be valid, it should readily
translate into practice. Third, while some research has empha-
sized differences between individuals and work settings, the
high performance cycle is a generic model of work motivation
designed to accommodate such differences.5
Locke and Latham (1990a and 1990b) integrate several theo-
ries of motivation to explain the chain of causal relationships
between goals, individual performance, rewards, job satisfaction,
and organizational commitment. These include goal setting
theory, control theory, expectancy theory, equity theory, and
reinforcement theory. For instance, the model posits that when
rewards are tied closely to performance, job satisfaction will
The relationship between performance and rewards is
derived primarily from expectancy theory, but elements of the
control, equity, and reinforcement theories are present as well.
The high performance cycle is reproduced in exhibit 1. The
model begins with specific challenges and demands placed on the
employee. Locke and Latham (1990a) justify the inclusion of this
component based on goal setting theory. Reviews of goal setting
studies have shown that people perform better in situations where
goals are specific and challenging, even if they are impossible
to accomplish (for reviews of this body of research, see Locke
1997; Locke et al. 1981; Locke and Henne 1986; Locke and
Latham 1984a; 1984b; 1990b). The model posits that challenging
goals stimulate high performance. Thus individuals confronted
with such goals are expected to perform better. The authors
acknowledge that a combination of factors contributes to creating
and maintaining challenging jobs such as the expectations of
superiors, colleagues, subordinates, and oneself.
The model specifies several variables that moderate the rela-
tionship between demands and high performance such as ability,
533/J-PART, July 2000
Exhibit 1
The High Performance Cycle

Moderators
Ability
Commitment
Self-efficacy
Self-efficac y Noncontingent Rew
Task complexity

Demands - - Performance Continge

Award

o A ,,

Mediat
Effort
Directi

Persistence Conse
Task specific strategies

Source: Locke and Latham (1990a, 4).

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

commitment, feedback, self-efficacy, and task complexity. One

assumption that underlies the positive linkage between goal
setting and performance is that individuals have the ability to
accomplish their goals. Those with adequate skills and abilities
will perform better than those with less aptitude, regardless of
challenges or demands. Therefore, it is necessary to control for
the individual's level of ability when the relationship between
goal setting and performance is tested.

Second, Locke and Latham (1990a and 1990b) contend that

demanding goals will only facilitate performance if employees are
committed to those goals. In other words, goal setting will not be
effective if employees are not committed to the established goals.
Goal commitment is higher when individuals understand their
goals, feel pressure from peers to perform well, perceive that
they can attain their goals, and believe that they will be recog-
nized for their accomplishments (Locke and Latham 1990a and
1990b). In a recent review of factors that cultivate commitment,
Latham and Seijts (1999) note that it is critical to demonstrate
the relevance and importance of goals to the individual and the
organization.

Another integral component of the high performance cycle is

job-related feedback. Feedback enables individuals to gauge their
progress toward goal attainment. Without feedback, goal setting
theory predicts that goals will have little or no impact on subse-
quent performance (Locke and Latham 1990a and 1990b). Many
studies have shown that goal setting is more effective when it is
accompanied by feedback (e.g., Chhokar and Wallin 1984; Kim
1984; Reber and Wallin 1984). Matsui, Okada, and Mizuguchi
(1983) showed that the relationship between feedback and per-
formance is complicated; feedback does not affect the perform-
ance levels of individuals who are already meeting expectations,
but it does have a significant impact on the performance levels of
underachievers. Locke and Henne (1986) concluded that feedback
does not improve performance directly; rather, it serves as an
essential condition for goal setting to work.

The role of self-efficacy differs slightly in Locke and

Latham's two accounts of the high performance cycle. In one
description of the model, the authors describe self-efficacy as a
facet of demands: such demands consist of "challenge, high goals
on meaningful, growth-facilitating tasks or series of tasks plus
high self-efficacy" (Locke and Latham 1990b, 253, fig. 11-1). In
the other characterization of the model, the authors describe self-
efficacy as a moderating influence between demands and perform-
ance (Locke and Latham 1990a). Locke and Henne (1986) argue
that the two constructs are conceptually distinct; that is, demands

535/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

convey direction and goals, whereas self-efficacy captures a self-

confidence dimension. Research suggests that self-efficacy differs
conceptually from goal setting, and studies have shown that both
have powerful effects on performance (Latham and Seijts 1999).

The final moderating factor included in the model is task

complexity. Using metanalysis to examine 125 goal setting stud-
ies, Wood, Mento, and Locke (1987) concluded that task com-
plexity is an important determinant of performance. Locke and
Latham (1990a, 11) explain: "In more complex tasks, the plans,
tactics, and strategies used by the individual play a larger role in
task performance than they do in simpler tasks where the number
of different strategies is more limited and is generally known to
all performers." Thus, less complex tasks may facilitate high
performance ceteris paribus.

In addition to the moderating factors described above, the

model specifies several mediators or mechanisms by which goals
affect performance. These mediators include direction of atten-
tion, effort, persistence, and task specific strategies. First,
employees are more likely to achieve goals that are clearly
delineated. If employees are unclear about their goals, they may
waste time and effort on less important tasks. When employees
understand their goals clearly, they are likely to perform better.

The other three mediating factors are individual attributes.

Locke and Latham (1990a and 1990b) contend that effort and
performance are positively related; the more effort an individual
exerts, the better he or she will perform. Locke and Latham
(1990a and 1990b) also allow that effort may be partly dependent
on the nature of the job or tasks assigned. Individuals with more
demanding jobs are expected to exert greater effort, and as a
result, they are expected to perform better. A closely related
concept included in the high performance cycle is persistence.
Employees who are more tenacious and willing to work longer
hours are more likely to achieve their goals and perform at
higher levels (Locke and Latham 1990a and 1990b). Faced with
challenging goals, some individuals may try to enhance their
skills and abilities. Those who develop task specific strategies to
adapt to changing goals are likely to be higher performers.

Demands, moderators, and mediators affect individual per-

formance. As exhibit 1 shows, rewards are modeled as a function
of performance. The model predicts that when performance is
commensurate with rewards, employees will be more satisfied
vis-a-vis equity theory. In turn, more satisfied employees will be
more loyal and committed to the organization.

536/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

In summary, the high performance cycle model explains

how organizations can use goals as inducements to create and
sustain high performance (Locke and Latham 1990a and 1990b).
6Hughes, Price, and Marrs (1986) explainPerformance influences job satisfaction via another concept-
two major strengths of structural equation
contingent rewards. The model does not posit a direct connection
modeling. First, technically, these models
between job satisfaction and subsequent productivity. However,
provide a method for estimating structural
relationships among unobservable con- job satisfaction is related indirectly to subsequent performance by
structs and for assessing the adequacy maintaining the employee's commitment to the organization.
with which they have been operational- Therefore, the model is recursive in nature.
ized. Second, conceptually, using struc-
tural modeling entails a model of thinking
about theory construction, measurement The high performance cycle model has not been validated
empirically.
problems, and data analysis that is helpful Hughes, Price, and Marrs (1986) suggest that
in stating theory in a more exact manner,
LISREL is an appropriate mechanism for testing such a theory.
testing theory more precisely, and provid-
ing a more thorough understanding of the
The major advantage of LISREL is that it allows researchers to
data. test both measurement and latent construct linkages.6 Harris and
Schaubroeck (1990) reviewed the use of latent variable models in
7The journals reviewed include Academy
five journals and found that studies have employed this approach
of Management Journal, Journal of
Applied Psychology, Journal of Manage-
to examine several related issues such as job satisfaction and
ment, Organizational Behavior and stress, work motivation, organizational commitment, absentee-
Human Decision Processes, and Person- ism, and turnover.7 Therefore, we judge that LISREL is the
nel Psychology.
appropriate device for testing the high performance cycle model.
8We recognize that other agencies and
segments of the federal civil service also
METHOD
participate in performance-based pay
systems. For example, the General
The Sample
Accounting Office operates an agency-
wide, performance-based pay system;
civil servants in GM 13-15 positions The data for this study come from the U.S. Office
participate in a merit-type pay system; sonnel Management's (OPM) Survey of Federal Govern
and GS 1-8 employees in the Department
ployees, November 1991-February 1992. The survey in
of Defense compete for performance-
based pay raises. However, we have sample of 56,767 federal employees. Fifty-six percent,
limited our analysis to the cadre of employees, returned completed questionnaires. Because
senior-level executives who work in all
of the Senior Executive Service (SES) are governed by
federal agencies.
pay plan and agencies typically have more flexibility in
9According to 1996 OPM figures, the financial rewards to SES members, we limited our stud
average age of SES members was 51.9 2,731 senior executives who responded to the survey.8
years and their average length of servicether limited the statistical analysis to those 2,474 indiv
was 22.8 years. About 24 percent were
answered all the questions. We recognize that the facto
eligible for regular (optional) retirement,
and 37 percent could retire under early- ing senior-level managers may differ from the factors mo
other civil servants; therefore, we will
out rules. Approximately 70 percent were refrain from gener
college graduates with advanced degrees,
to the entire civil service. Nonetheless, recent efforts to r
although 4.3 percent never graduated
neer, downsize, and eliminate layers of management in
from college. Nearly 80 percent of SES
eral
members were men and 12 percent were government may make these findings somewhat re
minorities. About half were in the admin- the larger population.9
istrative or technical fields, and the
remainder were split between science,
The data are analyzed with LISREL 8.12 using weig
engineering, and other professionals.
Most SES members (74 percent) work least squares.'1 LISREL permits the empirical adequacy
in the Washington, D.C., area model to be assessed based on the model's degree of fit
(FEDmanager 1998). sample of SES employees (Bentler and Chou 1987). The LISREL

537/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Exhibit 2
Structural Model for High Performance Cycle

00

r4
1-

to
t3
Perf .987*** Contingen?t
Performance rewa
o
\rewards~

I I represents latent variables measured with one observed variable C)

***Significant at the .001 level **Significant at the .01 level *Significant at t

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

model consists of two parts, the latent variable (or structural

equation) model and the measurement model. The latent variable
model specifies the hypothesized causal structure among the
unobserved theoretical constructs (see exhibit 2). The measure-
ment model specifies how the latent variables are measured
through the observed variables.

Variable Selection and Measures

The first step in the analysis was to translate the substanti

theory into a form that could be tested by structural equation
modeling. To do this, we related items in the survey to the
concepts in Locke and Latham's (1990a and 1990b) high per
formance cycle model. To select the items that measure eac
cept shown in exhibit 2, the following criteria were used. F
we identified measurement items based directly on the conc
descriptions provided by Locke and Latham (1990a and 199
Second, since the survey was not constructed for testing the
performance cycle, we selected approximate measures to re
'?When observed variables are measured sent concepts that were not adequately defined in the
with ordinal scales, Joreskog and Sorbom step. Third, each measurement model was individuall
(1993) recommend using a polychoric
correlation matrix rather than an ordinary deemed statistically acceptable.
product-moment correlation matrix. (See
Joreskog 1990 for a detailed discussion Exhibits 3 and 4 present indicators for each con
of this issue.) Further, the weighted leas n exhibit 2 (see the appendix for means and standard
squares method in LISREL should be nrin,
used to analyze data sets with ordinally of the observed variables).11 Due to space constr
measured concepts (Joreskog 1990; explain how two concepts were operationalized:
Joreskog and Sorbom 1993). formance and contingent rewards. Individual perfor
sists of the employee's most recent performance rating a
XMost variables, Y,, Y2, Y4 - Y1,o X, -
X3, X, - X2,, X1, - X16, are measured questions asking the employee about the accuracy
using a five-category Likert scale In the case of contingent rewards, OPM did not
anchored from strongly disagree to tion about actual performance-based pay increases.
strongly agree. X13, X19, and X20 are tap this concept with two questions that probe t
measured using a five-category Likert
scale that ranges from not at all to to a perception of fairness and equity in how perform
great extent. Performance rating, Y3, is is administered.
based on the following five categories:
unacceptable, minimally successful, fully Findings and Discussion
successful, exceeds fully successful, and
outstanding. Education, X4, has eleven
categories ranging from elementary school This analysis assesses the validity of a syste
to doctorate degree. Experience, Xs, relationships assumed to generate high performanc
includes nine categories beginning with presents the standardized coefficients and goodne
less than 1 year and ending with more
than 30 years. X7 is measured with the for the model (see exhibit 2).12 The most widely
following four categories: I rarely or fit is the X2 goodness-of-fit test; however, growing
never work overtime, 1-5 hours/week, 6- tion with this test has resulted in the developme
10 hours/week, and more than 10 hours/ alternative measures (Bollen 1989; Hoyle and P
~~~week. ~and Bentler 1995; Marsh and Balla 1994; Marsh, Ba
2The polychoric correlation matrix is McDonald 1988; Muliak et al. 1989). Except for
available from the authors. goodness-of-fit indices take on values between 0, rep

539/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

Exhibit 3
Indicators of the Endogenous Latent Variables

PERFORMANCE
Y1 Which of the following most closely describes the performance rating you received at your last appraisal?
Y2 If it were possible, I would like to be given pay increases based on the performance rating I receive from
my supervisor.
Y3 My performance rating is a fair and accurate reflection of my performance.

CONTINGENT REWARDS
Y4 I can expect to receive a pay raise or cash award if I perform exceptionally well.
Y5 Pay raises and cash rewards around here depend on how well you perform

SATISFACTION
Y6 In general, I am satisfied with my job.
Y7 I am satisfied with my pay.

CONSEQUENCES
Y8 High performers tend to stay with this organization.
Y9 I have confidence and trust in my organization.
Ylo The federal government is a great place to work.

Exhibit 4
Indicators of the Exogenous Latent Variables

DEMANDS
X1 My job is challenging.
X2 In my work unit, you know what is expected of you.
X3 People in my work group are expected to work hard.

EDUCATION
X4 What is your educational level?

EXPERIENCE
X5 How many years have you been a federal government employee?

COMMITMENT
X6 I have confidence and trust in my second-level supervisor.
X7 Employees in my work unit are encouraged to participate in decisions affecting their work.
X8 In my work unit, employees' personal development is encouraged.

FEEDBACK

Xg My supervisor discusses with me the specific reasons for my performance rating.

X,I My supervisor encourages me to help in developing work methods and job procedures.
X1n My supervisor and I agree on what good performance on my job means.

SELF-EFFICACY
X12 I am satisfied with the chances that I have to accomplish something worthwhile.

TASK COMPLEXITY
X13 To what extent do conflicting work assignments make your job difficult?

540/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

Exhibit 4 (continued)

EFFORT

X14 When I don't feel well in the morning, I still try to come to work because I know my contribution will be
missed.

DIRECTION
X15 I get enough information to do my job properly.
X16 The guidance I receive from my supervisor helps me do a better job.

PERSISTENCE
X17 On average, how much overtime do you work?

TASK SPECIFIC STRATEGIES

X18 I need training because the required skill level of my job has increased in the past two years.

NONCONTINGENT REWARDS
Xl9 To what extent are you satisfied with the amount of annual leave you receive from the federal government?
X20 To what extent are you satisfied with the amount of sick leave you receive from the federal government?

Exhibit 5
Standardized Coefficients for Model of High Performance

Endogenous Variables
Explanatory Variables Performance Contingent Rewards Satisfaction Consequences

Performance .987*** -
Contingent rewards - .819**
Satisfaction - - .957**
Demands .295* -
Education .000 - -
Experience .021 *
Commitment .411** - -
Feedback .276* - -
Self-efficacy .063*
Task complexity -.035*** - -
Effort .009 - -
Direction -.082
Persistence -.060** - - -
Task specific strategies -.017 - -
Noncontingent rewards - - -.043** -
R2 .835 .974 .863 .915

Goodness-of-Fit Indexes
Model chi-square 2,017 (df = 328, n = 2,494)
Goodness of fit index .96
Adjusted goodness of fit .95
Comparative fit index .90
Incremental fit index .90

***Significant at the .001 level **Significant at the .01 le

541/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms

3Since the chi-square is inflated when the
number of cases is large, researchers
often employ an ad hoc fit measure-
chi-square estimator divided by its degree
of freedom (Bollen 1989; Carmines and
Mclver 1981). There is no consensus on
what constitutes a good fit; recommenda-
tions range from ratios of 2 to 12 (Bollen
1989; Carmines and Mclver 1981; Brown
and Woods 1991). In this study, the ratio
is 6.15. Given the size of the data set and factors: commitment and self-efficacy. Although the OPM survey
the complexity of the model, we concludedid
that the model fits the data well. Baer
(1995, 60) makes the point with the
following example: "A well-fitting model
tested in a sample of 200 should yield
approximately the same fit index value as higher performance. (The standardized coefficient for commit-
the same model tested in a sample of
2,000 (across repeated samplings). The
chi-square statistic, on the other hand, is
expected to be 10 times as large in the
sample of 2,000."
This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
Work Motivation in the Senior Executive Service
poor fit, and 1, representing a perfect fit (Baer 1995). Although
universally agreed upon standards for assessing the goodness-of-
fit indices do not exist, some studies recommend using .90 as a
generic cutoff value for fit indices (Baer 1995; Hoyle and Panter
1995). The goodness-of-fit statistics reported in exhibit 5 show
that the estimated model yields a moderately good fit (for
excellent reviews of goodness-of-fit indices, see Hoyle and Panter
1995; Hu and Bentler 1995). The chi-square value is 2,017 with
328 degrees of freedom (ratio 6.15) and all of the fit indices
equal or exceed .90.13 Moreover, eleven of the fifteen paths
specified in the structural equation model are statistically
significant.
The results show that demands have a strong and significant
impact on individual performance, indicating that individuals with
more challenging and demanding jobs record higher performance
levels. Locke and Latham (1990b) reasoned that more demanding
goals cause employees to examine their personal goals and their
capacity for achieving those goals. The statistical analysis also
shows that feedback has a significant and positive influence on
individual performance. In a 1984 study, Reber and Wallin found
that difficult and specific goals accompanied by feedback were
more effective than either of the two alone. Goals and feedback
are critical in encouraging employees to perform because goals
provide purpose, and feedback allows the individual to monitor
progress toward achieving those goals (Locke and Latham
1990b). Locke and Latham (1990a and 1990b) contend that feed-
back is a necessary condition for high performance. Although the
cross-sectional nature of our data precludes us from examining
the conditional linkages that may exist between goal setting,
feedback, and high performance, we do find evidence that feed-
back has a strong and independent impact on SES members' per-
formances. One corollary is that a lack of feedback may inhibit
performance. In other words, the provision of supervisory feed-
back appears to moderate the linkage between goals and high
performance.
The findings provide support for two other moderating
not include questions that probe commitment to specific
goals, it did include three factors that Locke and Latham (1990a
and 1990b) believe influence goal commitment. On this index,
the analysis shows that higher levels of commitment lead to
ment is .411-by far the strongest of the moderating and mediat-
ing factors). In the case of public employees, commitment is one
form of public service motivation (PSM), and theory predicts that
PSM is positively related to performance (Romzek 1990; Perry
542/J-PART, July 2000
 Work Motivation in the Senior Executive Service

and Wise 1990; Perry 1996; Brewer, Selden, and Facer 2000).
Thus, the analysis provides evidence that PSM promotes high
performance (for additional evidence, see Brewer and Selden
1998). The analysis also shows that self-efficacy plays a sig-
nificant role in the high performance cycle. SES members who
believe they can accomplish something worthwhile consistently
outperform those with lower estimates of self-efficacy. The
analysis shows that commitment and self-efficacy are positively
linked to individual performance; thus, they are enabling
influences.

The analysis included two variables that tap individual

ability: education and experience. Although education is not a
statistically significant determinant of performance, experience is.
The findings show that SES members with more years of service
are higher performers. Perhaps these individuals have developed
more job-related competencies and better understand how to
translate goals into action. As a result, they may be more effec-
tive and efficient in achieving their goals.

The findings reveal that the nature of an individual's job

affects his or her performance. As expected, SES members who
are assigned conflicting work assignments record lower perform-
ance levels. Locke and Latham (1990a and 1990b) reasoned that
an individual who is assigned a complex task has to calculate the
most effective use of time and effort. As a result, performance is
hindered. Under conditions of conflict, tasks become more com-
plex as employees have to sort out conflicting demands and deter-
mine an appropriate course of action. Such conflict can also pit
the employee against the supervisor, which is likely to result in
lower performance ratings.

One factor-persistence-is highly significant but not in the

expected direction. In this study, persistence is measured as the
amount of overtime worked. The results show that as overtime
increases, performance decreases. While the model as a whole
supports Locke and Latham's central hypothesis that demands
drive performance, this particular finding urges moderation.
Heaping overwhelming demands on employees may actually
undercut performance and lead to burnout-an event that will
arrest the high performance cycle.

Several moderating and mediating factors do not achieve

statistical significance. These include education, effort, direction,
and task-specific strategies. Here we are reminded that the high
performance cycle is a generic model of work motivation while
SES members are a highly unique group of federal employees.
For instance, more than 95 percent of SES members are college

543/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

educated, so our education variable does not have the properties

of a normal distribution. Perhaps these variables will prove to be
more important when other populations are studied.

As exhibit 5 shows, individual performance exerts a strong,

positive effect on contingent rewards, suggesting that individuals
who perform well believe they will receive larger monetary
rewards. (The standardized coefficient of .987 is the strongest
found in the model.) When performance produces such rewards,
employees are more satisfied and thus are more motivated and
more committed to the organization.

An individual's level of job satisfaction is influenced directly

by both contingent and noncontingent rewards. The positive
relationship between contingent rewards and job satisfaction
shows that as perceptions of contingent rewards increase, an
individual's level of job satisfaction increases. Contrary to
expectations, satisfaction with noncontingent factors such as the
amount of annual and sick leave earned are inversely related to
job satisfaction. Previous research shows that individuals who are
dissatisfied with their jobs are absent and tardy more frequently
(for example, see Henne and Locke 1985; Mowday, Porter, and
Steers 1982). Annual leave and sick leave may allow these
employees to avoid work without risking disciplinary action, thus
providing short-term relief for workplace discomfort and anxiety.
However, the results clearly show that satisfaction with the
amounts of annual leave and sick leave earned does not enhance
overall job satisfaction.

Finally, job satisfaction is associated significantly with

positive organizational consequences. Individuals with higher
levels of job satisfaction are more committed to the organization,
reinforcing the linkage between job satisfaction and organiza-
tional commitment. The final link in the model suggests that
employees with high levels of organizational commitment will
remain with the organization and be receptive to future demands
placed on them, thus making the high performance cycle recur-
sive. Unfortunately, we could not test this final link because our
data set was cross sectional rather than longitudinal.

CONCLUSION

At the outset of this study, we noted that many peopl

a strong intuitive sense of what motivates employees, and
tend to look for empirical studies that confirm their views.
Fortunately, the high performance cycle model and the em
results reported in this study comport well with much of the
conventional wisdom.

544/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

The analysis indicates a good fit for the hypothesized

structural model based on Locke and Latham's high performance
cycle. Most of the individual relationships are significant and
provide insight into the mechanics of work motivation. The
analysis provides empirical support for the relationships posited
between demands and individual performance, individual per-
formance and contingent rewards, contingent rewards and job
satisfaction, and job satisfaction and organizational commitment.
In addition, the relationships between several moderating and
mediating factors and individual performance were confirmed.
These factors include experience, feedback, commitment, self-
efficacy, and task complexity.

Two of the relationships we have mentioned merit particular

attention-the relationship between individual performance and
contingent rewards, and the relationship between contingent
rewards and job satisfaction. In the high performance cycle,
rewards follow rather than precede performance. This is in
contrast to behaviorist theory, which maintains that rewards
induce performance (Locke and Latham 1990b). The model im-
plies that rewards will not necessarily encourage high per-
formance in the future unless such performance is first emitted-
an event that cannot occur without specific and challenging
demands (Locke and Latham 1990b). Once high performance is
emitted, rewards can be used as an incentive to reinforce and
perpetuate this behavior. Second, researchers have long puzzled
over the weak relationship between job satisfaction and individual
performance, so they will be interested in Locke and Latham's
respecification of this causal relationship. The model predicts that
high performance will lead to job satisfaction when rewards are
commensurate with performance, and the empirical analysis sup-
ports this respecification. Thus, the immediate value of job
satisfaction is to increase organizational commitment and reduce
turnover-factors that are necessary to create and sustain a cycle
of high performance.

These findings have important implications for policy

makers and public managers. Individual performance is driven in
part by specific and challenging demands placed on the individual
and in part by moderating and mediating factors that include
experience, feedback, commitment, self-efficacy, and task com-
plexity. Once high performance is obtained, it must be rewarded.
Such rewards may be either internal or external, and they may
include feelings of pride and accomplishment, personal recogni-
tion, and pay increases (Locke and Latham 1990b, 264). A tight
coupling between performance and rewards produces job satisfac-
tion, which in turn produces organizational commitment and
other positive net benefits.

545/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

The finding that challenging and demanding goals lead to

high performance should not be interpreted as a justification for
gutting the public service or overloading employees with impos-
sible tasks. Ingraham, Murlis, and Peters (1999, 17) note that the
erosion of the public service symbolizes a loss of recognition for
public servants. Recognition is one of the most sought after
rewards, and its loss will likely stall the high performance cycle.
Similarly, saddling employees with overwhelming demands may
undercut their performances and lead to burnout. Policy makers
who simultaneously advocate deep personnel reductions and
improved performance of public organizations need to attend
these findings carefully.

Clearly the high performance cycle will not run on auto-

matic pilot. Creating and sustaining high performance requires
increased investment in the organization and its employees.
Policy makers and public managers must create a working
environment that supports and enhances high performance. Two
prerequisites are to devote sufficient time and effort to the
management enterprise and to ensure that human and financial
resources are adequate.

Typically, one objective of administrative reform is to foster

a more productive workforce. When they formulate reforms,
change agents need to understand what it is that motivates
employees. This study provides evidence that contingent rewards
play a significant role in the high performance cycle. Flannery,
Hofrichter, and Platten (1996) warn that, in the rush to reengi-
neer and reinvent organizations, these organizations have over-
looked a critical component of change-pay. Our findings suggest
that creating an environment that facilitates and rewards high
performance leads to a more satisfied workforce, which subse-
quently leads to lower turnover (Lawler and Porter 1967;
Mowday, Porter, and Steers 1982).

One limitation of this study is that the OPM survey was not
designed to test Locke and Latham's model; consequently, some
concepts incorporated in the model may be poorly operational-
ized. In addition, the survey data consists of individual self-
reports; such reports could be biased. Another cautionary note is
that associations within a structural equation model are necessary
but not sufficient evidence of causal relations. While our empir-
ical model is consistent with Locke and Latham's set of causal
hypotheses, the data might be equally consistent with other theo-
retical frameworks (Hoyle and Panter 1995). Although the analy-
sis supports the underlying linkages in the high performance
cycle, future research should replicate this work with other data
sets and populations. Relatedly, since this study focuses on

546/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

members of the SES, the results cannot be generalized to other

public or private-sector work settings without further research.
Given the necessary linkage between performance and contingent
rewards, testing this model in work settings where employees do
not receive performance-based pay may not be appropriate.

Despite these limitations, we believe this study provides a

rich foundation for future research. Such research should investi-
gate the possibility of modifying the high performance cycle,
using both theory and empirical findings. Researchers need to
employ other data sets and populations to test the model, and
they need to test other comprehensive models of work motivation
empirically. These comprehensive models promise more insight
into the process of work motivation, and they constitute sounder
theoretical frameworks for personnel reforms. However, public
officials will be reluctant to base reforms on such metatheories
without strong and convincing empirical evidence.

Finally, the literature on research in public administration

urges more theory testing (Stallings and Ferris 1988; Perry and
Kraemer 1990). One chronic problem has been a lack of funding
to collect data and conduct research (Garson and Overman 1983;
Perry and Kraemer 1990). This article provides an exemplar for
addressing this problem. Specifically, we have tested an impor-
tant theory with data drawn from a large governmental survey
that was not designed for basic research. The limitations of this
approach are obvious, but so are the advantages. Considering the
large number of theories and data sets that are presently idle,
researchers could increase theory testing by being more oppor-
tunistic in their quest for data. Our final thought is a bit quixotic;
researchers could make large governmental data sets more mal-
leable for theory testing by participating in their design.

APPENDIX
Means and Standard Deviations of Observed Variables

Observed Variable Mean Standard Deviation

Y1 4.37 1.52
Y2 3.85 0.98
Y3 3.92 1.12
Y4 3.30 1.27
Y5 3.50 1.03
Y6 2.35 1.62
Y7 4.26 0.91
Y8 3.68 0.98
Yg 3.71 1.02
Y,0 3.37 0.99

547/J-PART, July 2

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

APPENDIX (continued)

Observed Variable Mean Standard Deviation

X1 4.47 0.76
X2 3.95 0.76
X3 4.40 0.63
X4 9.06 1.46
X5 6.98 1.69
X6 4.02 0.84
X7 3.98 0.86
X8 3.61 1.11
X9 3.45 1.19
X,o 3.73 1.06
X ll 3.83 1.06
X12 4.04 0.94
X13 3.03 3.04
X14 4.17 0.76
X15 3.93 0.89
X16 3.58 1.09
X17 3.09 0.91
X18 2.12 0.93
X19 4.33 0.78
X20 4.25 0.81

REFERENCES

Baer, D.
1995 An Introduction to Latent Vari-
able Structural Equation Models
in the Social Sciences. London,
Ontario: University of Western
Ontario.
Behn, R.D.
1995 "The Big Questions of Public
Management." Public Adminis-
tration Review 55:4:313-24.
Bender, P.M., and Chou, C.P.
1987 "Practical Issues in Structural
Modeling." Sociological Meth-
ods and Research 16:1:78-115.
Bollen, K.A.
the Public Service Ethic." Jour-
nal of Public Administration Re-
search and Theory 8:3:413-39.
Brewer, G.A.; Selden, S.C.; and
Facer, R.L. II.
2000 "Individual Conceptions of
Public Service Motivation."
Public Administration Review
60:3:254-64.
Brown, R.D., and Woods, J.A.
1991 "Toward a Model of Congres-
sional Elections." Journal of
Politics 53:2:454-73.
Campbell, J.P., and Pritchard, R.D. 1986 "Organizational Behavior: The
1976 "Motivation Theory in Industrial
Carmines, E.G., and Mclver, J.P.
1981 "Analyzing Models with Unob-
served Variables: Analysis of
Covariance Structures." In G.W.
Bohrnstedt and E.F. Borgatta,
eds. Social Measurement,
pp. 65-115. Beverly Hills,
Calif.: SAGE.
Chhokar, J.S., and Wallin, J.A.
1984 "A Field Study of the Effect of
Feedback Frequency on Per-
formance." Journal of Applied
Psychology 69:3:524-30.
Evans, M.G.
Central Role of Motivation."

1989 Structural Equations with Latent and Organizational Psychology." Journal of Management 12:2:
Variables. New York: Wiley. In M.D. Dunnette, ed. Hand- 203-22.

book of Industrial and Organiza-

Brewer, G.A., and Selden, S.C. tional Psychology. Chicago:
1998 "Whistle Blowers in the Federal Rand-McNally.
Civil Service: New Evidence of

548/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
 Work Motivation in the Senior Executive Service

FEDmanager. Ingraham, P.W.; Murlis, H.; and Landy, F.J., and Becker, W.S.
1998 "SES Profile: Do You Fit the Peters, B.G. 1987 "Motivation Theory Recon-
Mold?" Available online: 1999 "The State of the Higher Civil sidered." In L. Cummings and
< fedmanager@list.email.puh Service After Reform: Britain, B.M. Staw, eds. Research in
com> September 29. Canada and the United States." Organizational Behavior 9:1-38.
PUMA/HRM (99)1. Paris: Greenwich, Conn.: JAI.
Flannery, T.P.; Hofrichter, D.A.; and Organisation for Economic Co-
Platten, P.E. operation and Development. Latham, G.P., and Seijts, T.H.
1995 People, Performance and Pay: 1999 "The Effects of Personal and

Dynamic Compensation for Joreskog, K.G. Distal Goals on Performance on

Changing Organizations. 1990 "New Developments in LISREL: a Moderately Complex Task."

London: Hay Group.
Garson, G.D., and Overman, E.S.
1983 Public Management Research in
the United States. New York:
Analysis of Ordinal Variables
Using Polychoric Correlations
and Weighted Least Squares."
Quality and Quantity 24:387-
404.
Journal of Organizational
Behavior 20:421-29.
Lawler, E.E. III, and Porter, L.W.
1967 "The Effect of Performance on

Praeger. Job Satisfaction." Industrial

Joreskog, K.G., and Sorbom, D. Relations 7:1:20-28.

Harris, M.M., and Schaubroeck, J.
1990 "Confirmatory Modeling in
Organizational Behavior/Human
Resource Management: Issues
and Applications." Journal of
Management 16:2:337-60.
Heneman, R.L.
1992 Merit Pay: Linking Pay In-
creases to Performance Ratings.
Reading, Mass.: Addison-
Wesley.
Henne, D., and Locke, E.A.
1985 "Job Dissatisfaction: What Are
1993 LISREL 8: Structural Equation
Modeling with the Simplis
Command Language. Hillsdale,
N.J.: Lawrence Erlbaum.
Katzell, R.A.
1982 "Measuring Employee Motiva-
tion: A Systems Approach."
Paper presented at the American
Psychological Association,
Washington, D.C.
Katzell, R.A., and Thompson, D.E.
1990a "An Integrative Model of Work
Attitudes, Motivation, and
Leonard, N.H.; Beauvais, L.L.; and
Scholl, R.W.
1995 "A Self Concept-Based Model of
Work Motivation." Paper pre-
sented at the annual meeting of
the Academy of Management,
Vancouver, B.C.
Locke, E.A.
1997 "Motivation through Goal
Setting." Working paper.
Locke, E.A., and Henne, D.
1986 "Work Motivation Theories." In

the Consequences?" Interna- Performance." Human Perform- C. Cooper and I. Robertson,

tional Journal of Psychology ance 3:2:63-85. eds. Internal Review of Indus-
20:2:221-40. 1990b "Work Motivation: Theory and trial and Organizational Psy-
Practice." American Psycholo- chology, pp. 1-35. Chichester,
Hoyle, R.H., and Panter, A.T. gist 45:2:144-53. U.K.: Wiley.
1995 "Writing about Structural Equa-
tion Models." In Rick H. Hoyle,Kim, J.S. Locke, E.A., and Latham, G.P.
ed. Structural Equation Model- 1984 "Effect of Behavior Plus Out- 1984a Goal Setting. A Motivational
ing: Concepts, Issues, and come Goal Setting and Feedback Technique that Works. Engle-
Applications, pp. 158-76. New- on Employee Satisfaction and wood Cliffs, N.J.: Prentice-Hall.
bury Park, Calif.: SAGE. Performance." Academy of Man- 1984b Goal Setting for Individuals,
agement Journal 27:139-49. Groups, and Organizations.
Hu, L., and Bentler, P.M. Chicago: Science Research
1995 "Evaluating Model Fit." In Rick Klein, H.J. Associates.

H. Hoyle, ed. Structural Equa- 1989 "An Integrated Control Theory 1990a "Work Motivation: The High
tion Modeling: Concepts, Issues, Model of Work Motivation." Performance Cycle." In Uwe
and Applications, pp. 76-99. Academy of Management Review Kleinbeck et al., eds. Work
Newbury Park, Calif.: SAGE. 14:2:150-72. Motivation, pp. 3-25. Hillsdale,
N.J.: L. Erlbaum.
Hughes, M.; Price, R.L.; and Klein, J. 1990b A Theory of Goal Setting and
Marrs, D.W. 1990 "Feasibility Theory: A Task Performance. Englewood
1986 "Linking Theory Construction Resource-Munificence Model of Cliffs, N.J.: Prentice-Hall.
and Theory Testing: Models Work Motivation and Behavior."

with Multiple Indicators with Academy of Management Review

Latent Variables." Academy of 15:4:646-65.

Management Review 11:1:128-

44.

549/J-PART, July 2000

This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms

Locke, E.A.; Shaw, N.R.; Saari,
L.M.; and Latham, G.P.
1981 "Goal Setting and Task Per-
formance: 1969-1980." Psycho-
logical Bulletin 90:1:125-52.
Marsh, H.W., and Balla, J.
1994 "Goodness of Fit in Confirma-
tory Factor Analysis: The
Effects of Sample Size and
Model Parsimony." Quality and
Quantity 28:185-217.
Marsh, H.W.; Balla, J.R.; and
McDonald, R.P.
1988 "Goodness-of-Fit Indexes in
Confirmatory Factor Analysis:
The Effect of Sample Size."
Psychological Bulletin 103:
3:391-410.
Matsui, T.; Okada, A.; and
Mizuguchi, R.
1983 "Expectancy Theory Prediction
of the Goal Theory Postulate,
'the Harder the Goal, the Higher
the Performance.'" Journal of
Applied Psychology 66:1:54-58.
Mayes, B.T.
1978 "Some Boundary Conditions in
the Application of Motivation
Models." Academy of Manage-
ment Review 3:1:51-58.
Mitchell, T.R.
1982 "Motivation: New Directions for
Theory, Research, and Prac-
tice." Academy of Management Rainey, H.G.
Review 7:1:80-88.
Mowday, R.T.; Porter, L.W.; and
Steers, R.M.
1982 Employee-Organization Link-
ages. New York: Academic
Press.
This content downloaded from 103.231.241.233 on Thu, 28 Feb 2019 10:47:05 UTC
All use subject to https://about.jstor.org/terms
Work Motivation in the Senior Executive Service
Muliak, S.A. and assoc.
1989 "Evaluation of Goodness-of-fit
Indices for Structural Equation
Models." Psychological Bulletin
105:4:430-45.
Perry, J.L.
1996 "Measuring Public Service
Motivation: An Assessment of
Construct Reliability and
Validity." Journal of Public
Administration Research and
Theory 6:1:5-22.
Perry, J.L., and Kraemer, K.L.
1990 "Research Methodology in Pub-
lic Administration: Issues and
Patterns." In N.B. Lynn and A.
Wildavsky, eds. Public Admin-
istration: The State of the
Discipline, pp. 347-73.
Chatham, N.J.: Chatham House.
Perry, J.L., and Porter, L.W.
1982 "Factors Affecting the Context
for Motivation in Public
Organizations." Academy of
Management Review 7:1:89-98.
Perry, J.L., and Wise, L.R.
1990 "The Motivational Bases of
Public Service." Public Adminis-
tration Review 50:3:367-73.
Pinder, C.C.
1984 Work Motivation. Glenview, Ill.:
Scott Foresman.
1994 "Work Motivation." In Robert
T. Golembiewski, ed. Handbook
of Organizational Behavior,
pp. 19-40. New York: Marcel
Dekker.
1997 Understanding and Managing
Public Organizations, 2d ed. San
Francisco: Jossey-Bass.
550/J-PART, July 2000
Reber, R.A., and Wallin, J.A.
1984 "The Effects of Training, Goal
Setting, and Knowledge of
Results on Safe Behavior: A
Component Analysis." Academy
of Management Journal 27:
544-60.
Romzek, B.S.
1990 "Employee Investment and
Commitment: The Ties That
Bind." Public Administration
Review 50:3:374-82.
Shamir, B.
1990 "Meaning, Self, and Motivation
in Organizations." Organization
Studies 12:3:405-24.
Stallings, R.A., and Ferris, J.M.
1988 "Public Administration
Research: Work in PAR, 1940-
1984." Public Administration
Review 48:1:580-87.
Staw, B.M.
1977 "Motivation in Organizations:
Toward Synthesis." In B.M.
Staw, ed. Psychological Founda-
tions of Organizational
Behavior, pp. 77-89. Santa
Monica, Calif.: Goodyear.
Wood, R.E.; Mento, A.; and
Locke, E.A.
1986 "Task Complexity as a Modera-
tor of Goal Effects: A Meta-
Analysis." Journal of Applied
Psychology 72:3:416-25.