Journal of Vocational Behavior 59, 406–424 (2001)

doi:10.1006/jvbe.2001.1800, available online at http://www.idealibrary.com on

An Examination of the Roles of Career Uncertainty,

Flexibility, and Control in Predicting Emotional Exhaustion

Jack K. Ito and Céleste M. Brotheridge

University of Regina, Regina, Saskatchewan, Canada

The present study employs an uncertainty framework to develop a career stress model
in which career control mediates the effects of career stressors on emotional exhaustion.
This model is tested through the use of structural equation modeling with data obtained
from 204 government workers. Results indicate that a model which includes career con-
trol as a mediator of the effects of career flexibility and dependence on emotional ex-
haustion demonstrates a better fit with the data than a nested model in which job con-
text factors, the traditional predictors of emotional exhaustion, solely predict emotional
exhaustion. °C 2001 Academic Press
Key Words: career stress; uncertainty; emotional exhaustion.

Stress research has addressed many facets of organizational life using diverse,
but related, conceptualizations of the stress process and stress (e.g., Kahn &
Byosiere, 1992). McGrath (1976, p. 1352) conceptualized stress as being experi-
enced “when an environmental situation is perceived as presenting a demand which
threatens to exceed the person’s capabilities and resources for meeting it.” In devel-
oping a conceptual framework to classify stress research, Kahn and Byosiere used
a general model in which stressors (such as role ambiguity) cause stress which, in
turn, is manifest in strain (physiological and psychological outcomes, including
emotional exhaustion).
The stress associated with career uncertainties, including career transitions and
job insecurity, has received increasing attention (e.g., Brockner, 1992; Greenhalgh
& Rosenblatt, 1984; Latack, 1984, 1989; Latack, Kinicki, & Prussia, 1995). This
topic is becoming more salient given widespread organizational change, the de-
cline of employment security, and the emergence of more calculative relationships
between employees and employers (e.g., Brockner, 1992). The body of knowledge
of how individuals cope with career uncertainties is primarily conceptual in na-
ture, largely based on extending work stress research to career issues (e.g., Latack,
1989; Latack et al., 1995). Latack (1984) developed and tested a model in which
coping strategies mediated the relationship between job-related stress and the role
ambiguity and role overload caused by career transitions (mainly promotions).
This mediational role was not supported, nor was any relationship found between
Address correspondence and reprint requests to Jack Ito, Faculty of Administration, University of
Regina, Regina Saskatchewan S4S OA2, Canada. E-mail: jack.ito@uregina.ca.

406
0001-8791/01 $35.00
Copyright ° C 2001 by Academic Press

All rights of reproduction in any form reserved.

 CAREER UNCERTAINTY AND EMOTIONAL EXHAUSTION 407

stress and the magnitude of career transitions or types of coping strategies. Latack
proposed that competencies may moderate the relationships between transitions
and role overload and, by extension, between transitions and stress. Thus, de-
veloping generalizable knowledge and skills may be an effective coping strategy
in anticipation of career transitions, including demotions and potential job loss
(Latack, 1989). Latack et al. (1995) developed a model relating job loss to coping
strategies, proposing that situationally specific coping strategies would be more
appropriate that generic ones. However, our literature review found little empirical
research concerned with how career strategies may enable people to cope with the
prospect of transitions, provide a sense of control over their careers, and, therefore,
indirectly influence stress and strain at work.
The present study addresses this apparent gap through both theoretical and
empirical contributions. This article provides a cohesive and parsimonious means
of understanding the important contribution of career factors in the experience of
stress in the workplace. By incorporating job context stressors the study addresses
the question of whether career variables explain variation in emotional exhaustion
over and above the variance associated with traditional predictors of stress and
strain. By providing evidence of the relative importance of selected career factors,
this research directs attention to a set of heretofore neglected variables which
explain workplace stress levels. Understanding these predictors of workplace stress
is particularly important given its pernicious effects on workers’ psychological and
physiological health (Burke & Greenglass, 1995; Cherniss, 1992; Lee & Ashforth,
1993; Leiter, Clark, & Durup, 1994; Maslach & Leiter, 1997) and organizational
outcomes such as job performance, turnover, and work attitudes (Lee & Ashforth,
1996; Wright & Cropanzano, 1998).
The present study draws upon existing research to develop a model which links
career stressors, career coping, career control, and strain. Strain is measured by
emotional exhaustion, the core component of burnout and a chronic form of stress
(Maslach, 1982; Lee & Ashforth, 1996). Emotional exhaustion reflects “feelings
of being emotionally overextended and exhausted by one’s work. It is manifested
by both physical fatigue and a sense of feeling psychologically and emotionally
‘drained’” (Wright & Cropanzano, 1998, p. 486). Emotional exhaustion serves as
the antecedent to the remaining two dimensions of burnout—depersonalization
of others and reduced personal accomplishment (Cordes & Dougherty, 1993;
Maslach, 1982). Although originally validated in samples of workers in the helping
professions, the Maslach Burnout Inventory (Maslach & Jackson, 1986) has been
used to measure strain in diverse occupations, including computer company em-
ployees (Evans & Fischer, 1993), salespeople (Klein & Verbeke, 1999), Canadian
government employees (Fitzgerald, 1991), and police personnel (Stearns, 1992).
The use of emotional exhaustion as an indicator of work strain is appropriate given
that the career variables and contextual factors incorporated in this study also
reflect long-term processes.
This model is evaluated using survey responses obtained from a sample of
204 employees of a Canadian government department. Following Anderson and
408 ITO AND BROTHERIDGE

Gerbing (1988), structural equation modeling is employed to examine the relative

fit of several alternative models to the data. The results of these tests are presented
and then discussed in terms of their contributions to our understanding of the stress
process.

UNCERTAINTY AS A CAUSE OF STRESS

Uncertainty regarding many aspects of work life has been found to be a cause of
stress and its consequences including burnout (Cordes & Dougherty, 1993; Lee &
Ashforth, 1996; Sutton & Kahn, 1987). Uncertainty has also served as a mediating
variable to explain how facets of organizational life are associated with stress. For
example, Ferris et al. (1996) found that uncertainty in decision-making processes
and outcomes mediated the relationship between organizational politics and stress.
The notion of uncertainty is central in understanding how career issues may be
related to stress (e.g., Latack, 1989). Latack argued that stressor situations such
as career transitions are integrally linked with cognitions regarding uncertainty
and important outcomes, coping strategies, and stress reactions. Career transitions
raise the possibility of failure and its negative outcomes such as demotion or
job loss and the need to seek new employment. For example, the shift from a
professional to a managerial position requires the development of administrative
process skills which may not be well established, difficult to evaluate, and subject to
social definition. Meanwhile, the previous base of expertise, technical skills, may
be becoming outdated (Pazy, 1990). Moreover, today’s “turbulent environments,”
by accelerating organizational change, may upset normal patterns of personnel
movement and strain the capacity of individuals and organizations to cope with
career transitions.
Job insecurity is also associated with stress (DeFrank & Pliner, 1987). It is
a function of the perceived probability of job loss and its severity (a function
of the value of the job and the dependence of the individual upon the organi-
zation) (Greenhalgh & Rosenblatt, 1984; Klandermans, van Vuuren, & Jacobson,
1991). Job insecurity is associated with organizational restructuring or downsizing
(Greenhalgh & Rosenblatt, 1984), changes in staffing requirements (Klandermans
et al., 1991), and the increasing use of the external labor market versus internal
redeployment to adapt to change (Brockner, 1992). Stress may be exacerbated as
people jockey for a declining number of positions (Ferris et al., 1996).

COPING WITH UNCERTAINTY

Workers may try to cope with stress in their work environment by attempting
to understand, control, and ultimately reduce role stressors or to address the stress
outcomes themselves (e.g., through physical exercise) (Ferris et al., 1996; Ganster,
1988; Lazarus & Folkman, 1984). The literature suggests that control is important
in reducing stress and strain and in moderating the relationship between stressors
and their outcomes. The demands–control model of job stress posits that control or
decisional latitude (skill discretion and decisional authority) will act as a moderator
of the demands–strain relationship (Karasek, 1979; Karasek & Theorell, 1990).
 CAREER UNCERTAINTY AND EMOTIONAL EXHAUSTION 409

FIG. 1. Hypothesized career stress model with standardized parameter estimates.

As argued by Wall, Jackson, Mullarkey, and Parker (1996, p. 155), “increased

control reduces the effects of stressors by allowing individuals to face demands
when they are best able to do so and in ways they find most acceptable.” Although
the findings regarding the moderator effects have been mixed, the salutary effect of
decisional latitude as well as decisional authority and other more specific measures
of control on stress, have been supported (Wall et al., 1996; Ferris et al., 1996).
Latack (1989, p. 258) argued that coping which “reflects efforts to restore a sense of
personal control” would reduce stress arising from uncertainty in career transitions,
including adapting to new role requirements and job insecurity. The foregoing
discussion suggests that, as a result of organizational transitions, an increasing
number of employees may be exposed to career uncertainty and, consequently, job
stress and strain. The review also suggests that perceived control is an important
mechanism for coping with stress. The present study develops a model which
incorporates these perspectives (see Fig. 1).

A CAREER STRESS MODEL

The three dimensions at the core of the career stress model are career flexibil-
ity, dependence of employees upon the organization, and career control. Career
flexibility represents the perceived ability of employees to occupy other positions
410 ITO AND BROTHERIDGE

within and outside the organization. It enables individuals to cope with employment
uncertainties by providing alternative ways of meeting career goals. In contrast,
individuals who believe that their career options are severely limited are more
dependent on their existing jobs and more highly exposed to uncertainty given to-
day’s environment. Career control reflects the extent to which individuals believe
they can predict and influence the direction of their careers. It is the central variable
in the model and is expected to reduce the extent to which employees experience
stress, that is, report less emotional exhaustion.
The model proposes that career flexibility predicts career dependence and that
they both predict career control. Two job variables, task uncertainty and inter-
actions with other work groups, are expected to predict career flexibility. As a
form of uncertainty, task uncertainty is also expected to be directly associated with
emotional exhaustion. Employees’ length of service is related to investments in
organizationally specific knowledge and, therefore, to dependence on the organi-
zation for career progress. Finally, the model proposes that factors in the work
environment predict career flexibility, career control, and emotional exhaustion.
These factors include social support and work characteristics, which are tradition-
ally negatively associated with job strain.

Career Flexibility
Career flexibility may increase one’s sense of career control because it enables
people to cope with the uncertainty of career transitions. The development of
skills and knowledge which facilitate mobility and adaptability is central to career
resilience (London, 1993). Mobility in the “boundaryless career pattern” is facil-
itated by occupational expertise, reputation, and networks (Arthur, Claman, &
DeFillippi, 1995; DeFillippi & Arthur, 1994). Adaptability may be enhanced
through greater self-awareness (London), for example, by identifying unique per-
sonal competencies and developing transferable abilities such as the synthesis of
complex information (Weick & Berlinger, 1989). This emphasis on personal de-
velopment is an integral part of the “new paradigm” of individual responsibility
for career management. Therefore, as career flexibility may assist people in coping
with career transitions, it is expected to contribute to perceptions of career control.
Interaction with people in other work units may help employees develop process
knowledge and skills such as interpersonal competence, group decision making,
and conflict resolution. These skills may help build their confidence in the ability
to cope with the interpersonal and organizational dynamics involved in career
transitions. The development of a network of contacts and exchange relationships
may increase employees’ awareness of career options and access to other positions
(e.g., Kanter, 1977). Therefore, interunit interactions are expected to predict career
flexibility.
In the technology-structure literature, task uncertainty consists of variability
and analyzability. Variability refers to unpredictable work flows and exceptions
to the rule; analyzability addresses the degree of task difficulty as evidenced in
the degree to which employees rely upon standard routines, use trial and error in
 CAREER UNCERTAINTY AND EMOTIONAL EXHAUSTION 411

problem solving, and experience difficulty in predicting the outcomes of alterna-

tive courses of action (Perrow, 1967; Van de Ven & Delbecq, 1974; Van de Ven,
Delbecq, & Koenig, 1976). Although conceptually distinct, these dimensions are
correlated and are frequently treated as a single factor (e.g., Hage & Aiken, 1967).
Task uncertainty, conceptualized as role ambiguity, unanticipated fluctuations in
workload, and the inability to predict the effort–outcome relationship, has been
identified as a stressor (e.g., Kahn & Byosiere, 1992).
High levels of task uncertainty compel employees to deal with novel situations
in which successful performance requires developing, analyzing, and selecting al-
ternatives. This experience may help employees develop the general insights and
skills in problem solving which Weick and Berlinger (1989) suggest can be trans-
ferred across organizational boundaries. The use of challenging assignments is part
of many development programs. For example, McCauley, Ruderman, Ohlott, and
Morrow (1994) reported on the development and validation of the Development
Challenge Profile instrument. This instrument identifies transitions, task charac-
teristics, and obstacles providing challenging experiences which contribute to the
growth and development of transferable skills. Thus, task uncertainty may facil-
itate career flexibility. The foregoing suggests that task uncertainty evokes strain
directly, but has a negative effect through the task uncertainty–career flexibility–
career control–emotional exhaustion linkage. Therefore, the following hypotheses
are proposed:
Hypothesis 1: Career flexibility is positively associated with career control.
Hypothesis 2: Interaction with people in other work units is positively associated with career
flexibility.
Hypothesis 3: Task uncertainty is positively associated with career flexibility.
Hypothesis 4: Task uncertainty is positively associated with emotional exhaustion.

Career Dependence
Organizational commitment reflects the degree to which workers feel “attached”
to their organizations. It has evolved from being viewed and measured as an uni-
dimensional construct to a multifaceted one which also distinguishes between
continuance and affective commitment (Meyer & Allen, 1984; Meyer, Allen, &
Smith, 1993). McGee and Ford (1987) conceived of continuance commitment
as the availability of alternative employment and “side bets”—life disruptions
caused by taking other employment. Similarly, Carson, Carson, and Bedian (1995)
developed the concept of career entrenchment consisting of career investments,
emotional costs, and limitedness of career alternatives. The concept of career de-
pendence employed in the present article refers to the availability of career options
outside the organization—a component of both continuance commitment and ca-
reer entrenchment. Since career flexibility is associated with the development of
skills and abilities providing employment options, it is likely to reduce individual
feelings of dependence on the employer. However, these are not opposite sides
of the same coin, since people may be reluctant to leave an organization which
affords internal mobility.
412 ITO AND BROTHERIDGE

In the internal labor market literature, length of service has been associated with
the development of organizationally specific knowledge (e.g., Barney & Lawrence,
1989; Doeringer, 1967). Examples of such knowledge include (a) knowledge
of codified processes and procedures unique to the organization (Perry, 1988),
(b) fine-grained understandings of client needs (Malos & Campion, 1995; Who-
ley, 1985), and (c) tacit knowledge such as internalized beliefs about cause and
effect (know why) and internal networks (know whom) (DeFillippi & Arthur, 1994;
Kanter, 1977). These investments are likely to be perceived as being optimally uti-
lized in work performance and in career progress within rather than outside the
organization (Barney & Lawrence, 1989). Although process skills may facilitate
flexibility within an organization, the specific skills and knowledge gained over
time may limit interorganizational mobility, that is, cause workers to feel dependent
on an organization for employment as well as for career progress.
Dependence on the organization for career progress exposes workers to uncer-
tainties as the organization adapts to environmental changes through downsizing,
subcontracting, or mergers and acquisitions. In the case of the public sector, bud-
get decreases caused by shifts in government policy may result in changes in a
department’s domain. Such changes inevitably pose uncertainties with respect to
career transitions and job security (Greenhalgh & Rosenblatt, 1984; Klandermans
et al., 1991). Moreover, the responses (e.g., retraining versus subcontracting) may
lie outside the influence of most employees. Therefore, the following hypotheses
are offered:
Hypothesis 5: Career flexibility is negatively associated with career dependence.
Hypothesis 6: Length of service is positively associated with career dependence.
Hypothesis 7: Career dependence is negatively associated with career control.

Career Control
This section presents career control as the key link between career flexibility
and dependence and strain. Although several studies have addressed the issue of
control over events, career control is defined as an overall perception of being able
to predict the unfolding of one’s career and the ability to influence one’s career
trajectory. It is, therefore, a summary measure of the extent to which individuals
feel exposed to career uncertainty and their ability to cope with this uncertainty.
Similarly, Greenhalgh and Rosenblatt (1984) identified subjective threat as a key
variable mediating the relationship between objective threat and job insecurity.
Subjective threat is composed of employees’ perceptions of an event’s severity
(e.g., being temporarily versus permanently laid off) and their sense of power-
lessness to counteract the threat. A previous section discussed how perceptions of
control over situations, including one’s career, may be negatively associated with
strain. Therefore, the following hypothesis is proposed:
Hypothesis 8: Career control is negatively associated with emotional exhaustion.

Hypotheses 1 and 7 predicted that career flexibility and a lack of dependence

would be associated with career control. Therefore, it is expected that career control
 CAREER UNCERTAINTY AND EMOTIONAL EXHAUSTION 413

would fully mediate the effects of both career flexibility and dependence on levels
of strain. In other words, the variation in strain explained by career flexibility and
career dependence is expected to be due to their relationship with career control.
Therefore, the following hypotheses are proposed:
Hypothesis 9a: Career control fully mediates the relationship between career dependence and
emotional exhaustion.
Hypothesis 9b: Career control fully mediates the relationship between career flexibility and
emotional exhaustion.

Job Context
This section proposes that job context factors traditionally associated with stress
and strain are also associated with career variables. The job context factors included
in this study are peer cohesion, supervisory support, autonomy, and clarity of work
expectations. They were chosen because of their conceptual similarity to constructs
found to be associated with emotional exhaustion (Kahn & Byosiere, 1992; Lee
& Ashforth, 1996) and because they may also be linked to career flexibility and
career control. The inclusion of these job context factors allows an examination of
the “added value” provided by the inclusion of career variables in the prediction
of emotional exhaustion. Moreover, their inclusion meets the criterion for self-
contained models to incorporate variables which (a) are related to other causes
explicitly included in the model and (b) make a unique and important contribution
to the prediction of endogenous variables in the model (James, Mulaik, & Brett,
1982).
Whereas peer cohesion refers to supportive coworker behaviors, supervisor sup-
port includes direct (e.g., providing feedback on performance) and indirect support
(e.g., encouraging workers to be supportive of one another). Social support in the
workplace is known to reduce stress and may moderate the stressor–strain linkage
(Kahn & Byosiere, 1992). Both sources of support are expected to be related to
career flexibility since supportive behavior may encourage individuals to learn new
tasks, for example, by reducing the risks and penalties of failure (Numerof, 1987).
Autonomy is conceptually similar to decision-making authority which measures
the degree of individual control in task performance (Karasek, 1979). Autonomy
may provide opportunities for employees to select challenging task assignments
which develop their skills and abilities, including those which enhance transferable
decision-making abilities. Thus, autonomy may be linked with both career flexi-
bility and career control. Clarity refers to the extent to which role expectations and
policies and procedures have been clearly communicated. Clear human resource
management policies and procedures, such as training requirements for different
jobs, may facilitate career planning and therefore contribute to perceptions of ca-
reer control. Work environments which are interpersonally supportive and which
provide workers with autonomy in work performance and clear work roles (i.e.,
“supportive environments”) have been found to be associated with lower levels of
strain (Lee & Ashforth, 1996). The foregoing discussion suggests the following
hypotheses:
414 ITO AND BROTHERIDGE

Hypothesis 10a: More supportive work environments are associated with greater career flex-
ibility.
Hypothesis 10b: More supportive work environments are associated with greater career
control.
Hypothesis 11: More supportive work environments are negatively associated with
emotional exhaustion.

METHOD
Sample and Procedure
This study was undertaken in a branch of a Canadian government agency (also
referred to as the Department) as part of a consulting project in which the com-
pensation for the researchers’ access to the subjects was feedback of results. The
subjects were informed that this was a research project on organizational climate
and that the summary results would be made available to them. Participation was
voluntary. The questionnaires were administered by the second author at the work
site during working hours. Approximately 71% (or 204 individuals) of the eligible
set of employees completed the questionnaires. Summary reports provided by the
Human Resource Branch indicated that the demographic characteristics of this
sample were representative of those of the population of employees in this agency.
The sample included senior and middle management (5.9%); clerical workers
(15.7%); line professionals (70.6%); and professionals offering personnel, finan-
cial, or computer services (7.8%). Approximately 56% of the respondents were
female. Also, 22.1% of the respondents had 5 or fewer years of service within the
agency, 53% had 6 to 10 years of service, and 25% had 16 or more years of service.
Meetings with management, union, and employee groups revealed that numerous
changes were occurring within the organization that would cause employees to
feel uncertain about their careers within the department and within the federal
civil service.

Measures
Career dimensions and their antecedents. All the items for these measures used
a 5-point Likert type scale ranging from 1 (strongly agree) to 5 (strongly disagree).
The responses for these measures were transformed, where necessary, so that high
scores indicate that items strongly represent a work situation. For example, a score
of 4 for career control suggests that the respondent experiences a high degree
of control over aspects of his/her career. The measures of career control (three
items, α = .74), flexibility (four items, α = .75), and dependence (three items,
α = .72) were developed specifically for this study from concepts elaborated in
London (1983, 1993), Latack (1989), and DeFillippi and Arthur (1994). Sample
items are as follows: (a) career control: “I feel that I am in control of my career”;
(b) flexibility: “My work assignments have helped to broaden my skills and knowl-
edge which provide me with flexibility in career choice”; and (c) dependence:
“Most of my career options are with the Department.” Task uncertainty (four
items, α = .69) and interunit interactions (three items, α = .68) were assessed by
items developed by Van de Ven and Delbecq (1974) and Ito and Peterson (1986).
 CAREER UNCERTAINTY AND EMOTIONAL EXHAUSTION 415

Sample items are (a) task uncertainty: “How much variety in cases, claims, clients,
or things do you generally encounter in your working day?” (variability) and “In
some jobs, initial approaches to handling a case or client or task may not lead
to satisfactory results, and other alternatives must be applied until a satisfactory
result is achieved” (very frequently to almost never) (difficulty); and (b) interunit
interactions: “I frequently work with members of other work groups in order to
perform my duties.” Length of service was assessed by a single item “How long
have you been working with the Department?”
Work environment variables. Four work environment dimensions (peer cohe-
sion, supervisory support, clarity, and autonomy—a total of 36 items, α = .82)
were measured using the relevant scales of the Work Environment Scale (WES—
Form R, Moos, 1981). Cronbach’s α for the subscales in the study ranged from
.69 to .80 and are very similar to the .69 to .79 range reported by Moos (1981).
Each dimension is scaled from 1 to 9, and the higher the score, the higher the level
of that dimension. Given their strong zero-order correlations (ranging from .45 to
.62, with an average of .53) and a single factor solution, the four WES dimensions
were combined to form a block variable. These results are consistent with those
of Hershberger, Lichtenstein, and Knox (1994), who found a two factor structure
for all nine WES subscales, one of which included the four subscales used in the
present study. The higher the WES score, the more positive the job context.
Emotional exhaustion. The nine-item Likert type emotional exhaustion subscale
of the Maslach Burnout Inventory was used as the indicator of strain (Maslach &
Jackson, 1986). Consistent with previous research (Fitzgerald, 1991), a 5-point
scale (1 = almost never to 5 = almost always) was employed rather than the
original 7-point scale, with higher scores reflecting higher levels of emotional
exhaustion. The level of internal reliability in the present study (α = .85) was
consistent with the level previously reported (.77, on average but ranging from .52
to .91; Byrne, 1993).

Analysis
The single-indicator modeling approach employed in this study has been used
to test relatively large models (e.g., Frone, Russell, & Cooper, 1992). Given the
large number of paths to be estimated relative to the sample size, single summary
indicators of the constructs were employed rather than the individual items them-
selves. This approach is supported by a study which found the substantive results of
a latent variable model and a single indicator model to be equivalent (Netemeyer,
Johnston, & Burton, 1990). Adjustments were made for measurement error by
setting the error variance of each indicator to 1 − α multiplied by its variance
(Bollen, 1989). The few missing data were replaced with the mean score for the
item obtained from the rest of the sample, an approach consistent with Cohen and
Cohen (1983) and Frone, Yardley, and Markel (1997).
The unidimensionality of the constructs was established through principal com-
ponents extractions with varimax rotation on the items from (a) the career control,
dependence, flexibility, task uncertainty, and interunit interaction scales; (b) the
416 ITO AND BROTHERIDGE

WES scales; and (c) the emotional exhaustion scales of the Maslach Burnout In-
ventory. The items loaded on their respective factors (most loadings ranged from
.61 to .85; with some between .45 and .59) with small cross factor loadings, thus
confirming the unidimensionality of the scales.
Version 3.6 of Arbuckle’s (1997) AMOS (Analysis of Moment Structures) pro-
gram was used to evaluate the fit of the model (Fig. 1) with the empirical data.
Following Anderson and Gerbing (1988), a series of five nested models was as-
sessed: a saturated model in which all parameters were estimated (Model A), a
null model in which all parameters were set at zero (Model B), the model specified
in Fig. 1 (Model C), a model in which the path between career control and stress
was constrained to equal zero (Model D), and a model in which the path between
WES and stress was constrained to equal zero (Model E).
The last two models represent potential alternatives in the prediction of emo-
tional exhaustion. Model D emphasizes the traditional relationship between WES
and emotional exhaustion to the exclusion of career variables; model E suggests
that only career variables account for significant variance in emotional exhaus-
tion. A chi-square difference test was used to compare the fit of these models, but,
given its sensitivity to sample size, additional goodness of fit indices were also
employed including the goodness-of-fit index (GFI; Tanaka & Huba, 1985), the
Tucker–Lewis Index (TLI; Tucker & Lewis, 1973), and the Comparative Fit Index
(CFI; Bentler, 1995). For all three indices, values of .90 or greater indicate a good
fit between a model and the data.

RESULTS
Table 1 presents the means, standard deviations, and zero-order correlations
for the variables employed in this study. The results of the chi-square difference
tests and the goodness of fit indices for the five models are provided in Table 2.
Results indicate that the hypothesized career stress model (Model C) represents a
good fit with the data, especially in comparison with the alternative models under
consideration. In particular, the chi-square value of Model C was significantly
lower than that of Model D, which postulated that career variables would not predict
emotional exhaustion, and Model E, which postulated that variance in emotional
exhaustion could be solely explained by career variables. Thus, the hypothesized
career stress model is clearly a superior model of the sample covariance matrix
than the remaining models.
The parameter estimates for all the paths in the hypothesized career stress model
were significant at .05 or better (Fig. 1). The predictions that interaction with peo-
ple in other work units (Hypothesis 2), task uncertainty (Hypothesis 3), and more
supportive work environments (Hypothesis 10a) would be positively linked to
career flexibility were upheld. Also supported were Hypotheses 5 and 6, which
predicted that career flexibility would be negatively related to career dependence
and that length of service would be positively linked to career dependence. The
results also supported the proposed positive relationship of career flexibility and
work environments with career control (Hypotheses 1 and 10b, respectively) and
 CAREER UNCERTAINTY AND EMOTIONAL EXHAUSTION 417

TABLE 1
Means, Standard Deviations, and Correlations (n = 204)

Mean SD 1 2 3 4 5 6 7

1. Task uncertainty 14.14 2.79

2. Inter-unit interaction 9.09 2.62 .26∗∗
3. Length of service 2.58 1.25 .03 .00
4. Career flexibility 11.28 3.07 .23∗∗ .32∗∗ .04
5. Career dependence 10.15 2.74 −.13∗ −.12∗ .25∗∗ −.32∗∗
6. Career control 7.25 2.68 .06 .11 .05 .44∗∗ −.33∗∗
7. Work environment 15.58 7.41 .04 .22∗∗ .07 .52∗∗ −.07 .33∗∗
8. Emotional exhaustion 26.07 6.10 .17∗∗ −.01 −.04 −.29∗∗ .20∗∗ −.40∗∗ −.47∗∗
Note. Task uncertainty, interunit interaction, career flexibility, career dependence, and career control
were measured on 5-point scales. Most of the scales were described as strongly agree to strongly
disagree; some were described as very frequently to almost never. Two task uncertainty scales had
specific descriptors. Scales were reversed, where appropriate, so that the higher the value, the more of
the characteristic. Length of service was measured by a 5-point scale as follows: (1) up to and including
5 years, (2) 6–10 years, (3) 11–15 years, (4) 16–20 years, and (5) 21 or more years. The WES score is
the sum of responses to four dimensions. Each dimension was composed of nine items, each of which
was measured on a scale from 0 to 1. The higher the score, the more supportive the environment. The
maximum score is 36. Emotional exhaustion was measured by a 5-point scale (almost always to almost
never). Higher scores represent higher levels of emotional exhaustion.
∗ p < .05.
∗∗ p. < .01.

the negative relationship between career dependence and career control (Hypoth-
esis 7). Similarly, the predictions that task uncertainty would be positively related
to emotional exhaustion (Hypothesis 4) and that career control and more positive
work environments would be negatively related to emotional exhaustion (Hypothe-
ses 8 and 11, respectively) were all upheld. Finally, Hypotheses 9a and 9b, that
career control would mediate the relationship of career dependence and career
flexibility with emotional exhaustion, were supported. There were no independent
paths from these variables to emotional exhaustion.
As shown in Fig. 1, work environment and career control each make an in-
dependent and equivalent contribution to the prediction of emotional exhaustion.

TABLE 2
Model Fit Indices

Model χ df p GFI CFI TLI χ 2 differencea

A 0.00 0 — — 1.00 1.00 —

B 299.29 28 0.00 .61 .00 .00 276.96∗∗∗
C 22.33 16 0.13 .97 .98 .96 —
D 43.15 17 0.00 .95 .90 .84 20.82∗∗∗
E 66.13 18 0.99 .94 .82 .72 43.80∗∗∗
a This column reflects the χ 2 difference test between Model C and the noted model.
∗∗∗ p < .001.
418 ITO AND BROTHERIDGE

Approximately 33% of the variance in emotional exhaustion was explained by

the model. Also, work environment, task uncertainty, and interunit interaction ac-
counted for approximately 31% of the variation in career flexibility. Finally, 18
and 26% of variance in career dependency and career control respectively were
explained.

DISCUSSION
The present study sought to determine the extent to which career dimensions
influence perceptions of job strain as measured by emotional exhaustion. The ex-
isting theoretical and empirical literature was employed to develop a career stress
model. The results support the view that career dimensions are associated with
strain and are consistent with the general theoretical models posed by researchers
such as Sutton and Kahn (1987) and Latack (1989). The finding that career control
explains variance in emotional exhaustion over and above that explained by job
context variables provides evidence for the importance of career issues in under-
standing the stress process within organizations. Although a broader set of job
context and individual variables would likely reduce the incremental impact of
the career dimensions on strain, the present results provide evidence of the im-
portance of career variables in understanding the stress process. Given the many
career uncertainties in today’s work environment, these results suggest that the
career–stress–strain field may be a practically significant area for future research.
The role of career control as a predictor of strain is consistent with research
on the relationship between strain and control (Lee & Ashforth, 1996). In this
study, career control also served as a significant mediator of the effects of career
dependence and career flexibility on emotional exhaustion. This finding suggests
that future research should examine the role of perceived control as a mediating
variable in models which examine how coping mechanisms have an effect on strain
and on the stressor–strain relationship.
The findings regarding the role of career flexibility are also germane to stress
research. Career flexibility is important as a coping mechanism because it is directly
associated with career control and because it reduces perceptions of dependence,
a variable negatively associated with career control. The findings support the well-
established linkage between a positive work environment and strain. Moreover, the
support for the work environment–career flexibility path has identified a heretofore
unrecognized mechanism through which work environment indirectly influences
strain in organizations.
The results indicate that higher levels of task uncertainty and frequency of
interunit interactions are associated with higher career flexibility. This finding
supports the widely held belief that developing skills and abilities in complex tasks
and building personal networks are important to developing mobility within and
outside the organization (e.g., McCauley et al., 1994). More generally, this study
suggests that career issues are intimately bound within the general work context and
are influenced by the nature of the task in addition to human resource management
policies. The career variables also may moderate established relationships. For
 CAREER UNCERTAINTY AND EMOTIONAL EXHAUSTION 419

example, job insecurity may have more of an effect on stress where career flexibility
is low, and few career options exist, than where it is high.
The seniority–dependence–control linkage also has implications for career pro-
cesses. It has been noted that organizations have decreased their employment
commitments yet expect that individuals will remain loyal to the organization
(e.g., Capelli, 1999). Such loyalty, as expressed in seniority, may result in accu-
mulating organizationally specific knowledge which contributes to competitive
advantage (Barney, 1991). Yet, this study has found that seniority is positively
linked to feelings of dependence (since career options may become more limited
over time) and these feelings are associated with a lack of perceived career control.
Thus, loyalty appears to have a price: employees may experience lowered influ-
ence over career directions and experience strain as a result. Given the importance
of the job security–stress–strain relationship, research incorporating career flex-
ibility and its correlate, career dependence, may prove fruitful in furthering our
understanding of the employee–employer relationship. For example, equity and
inducement–contributions perspectives suggest that employees may wish to pur-
sue training and development opportunities (which increase career flexibility) to
offset their increased dependence arising from loyalty. This argument is consistent
with Waterman, Waterman, and Collard (1994).
The results support the direct task uncertainty–strain relationship and an indi-
rect negative effect through the career flexibility–control–strain linkage. This find-
ing may help extend research into the demands-control model discussed above,
in which the proposed moderator effects have enjoyed only mixed support. The
model assumes that both components of decisional latitude—decisional author-
ity and skill discretion—facilitate coping with demands. The present study found
that task uncertainty, which is conceptually similar to skill discretion and is as-
sociated with decisional authority/autonomy (e.g., Hage & Aiken, 1967; Ito &
Peterson, 1986), has a positive and direct effect on strain. Thus, task uncertainty
may compound the effects of job demands on strain. Whereas job demands may be
handled through established procedures in routine jobs, they may increase strain
in nonroutine jobs, for example, by reducing the time available for analysis of
alternatives. Therefore, the two effects may be working in opposite directions.
The job demands–strain relationship may be stronger in the low than in the high
decisional authority situation, but may be weaker in the low, than in the high skill
discretion context. The result is that the differences in the relationship between
low and high decisional latitude are less than those proposed in the model and few
moderator effects are found. Moreover, a taxonomy may be useful in extending
demands–control theory. The combinations where decisional authority and skill
discretion reinforce each other are more likely to show moderator effects. This
supports the recent focus on the decisional authority component and other mea-
sures of direct control where moderating effects have been found, thus suggesting
a reformulation of the model (Wall et al., 1996).
Future investigations which extend the career stress model may contribute to
our understanding of stress processes. One promising extension concerns the role
420 ITO AND BROTHERIDGE

of personality variables, in particular, internal–external locus of control (Rotter,

1966), which measures the degree to which workers seek to actively manage
their environment. In this context, de Rijk, Le Blanc, Schaufeli, and de Jonge
(1998) found that decisional latitude moderated the demands–strain relationship
for nurses high in active coping. Individuals who seek control over their environ-
ment are likely to strive to achieve a measure of career flexibility in planning their
careers.
The investigation of two feedback loops may enrich theories relating career
issues with strain and provide valuable insights into human resource management
practices. First, emotional exhaustion may result in avoidance or withdrawl rather
than an active coping response (Lee & Ashforth, 1996; Leiter, 1991, 1993; Wright
& Cropanzano, 1998) and may focus attention on short-term issues and routinized
responses (e.g., Sutton, 1990). This suggests that emotional exhaustion may limit
employees’ attempts to build career flexibility through longer term efforts in skill
development. Thus, empirical support for a negative feedback loop may suggest
that career development programs should use stress reduction as part of the learning
process. Second, there are two feasible feedback loops from career control to career
flexibility. The first is positive, wherein feelings of having control motivate further
career development, in the same way as goal attainment may raise aspiration levels.
The second is negative, wherein the recognition of low career control may result
in actions which influence career directions. Variables such as locus of control
and level of strain may act as moderating variables in how people respond. For
example, people with an internal locus of control who are experiencing low levels
of control may respond actively, whereas those with an external locus might choose
a more passive approach.
There are a number of methodological limitations to this study which should
be addressed in future research. First, a cross-validation study is required to de-
termine if the model remains valid across samples. The study was undertaken in
a government bureaucracy where longevity of service is the norm (though it re-
cently experienced downsizing and interviews suggested that jobs were no longer
as secure as is commonly assumed for civil service jobs). The employer–employee
relationship in other industries where job insecurity is common and long standing
may play a significant role in the relationships in the model. Thus, a wider study in
settings where relative job security is not the norm may yield somewhat different
results. Second, the cross-sectional nature of the current study limits the extent to
which conclusions of a causal nature may be drawn. A longitudinal study incor-
porating feedback loops would address this limitation. Third, a single self-report
questionnaire approach was used, which may contribute to common methods vari-
ance. However, since the career dimensions, task variables, work environment,
and emotional exhaustion tap different aspects of work life, methods variance may
not have played a strong role in the relationships among the variables. Moreover,
the questionnaire included measures of several variables which are not included in
this study, and very few of the items for the independent variables were listed con-
tiguously. Finally, strain was measured by emotional exhaustion, and other mea-
sures of stress and strain may produce different results. More objective measures,
 CAREER UNCERTAINTY AND EMOTIONAL EXHAUSTION 421

including physiological indicators, should be used to supplement self-report psy-

chological measures of strain.
The results have several practical implications, many of which have been dis-
cussed above. Although challenging jobs may serve as an integral component of
development programs and contribute to career flexibility (e.g., McCauley et al.,
1994; Ohlott Ruderman, & McCauley, 1994), they also may induce stress and
strain. The results of this study suggest that such experiences may need support
to be effective. For example, educational programs and mentoring may provide
knowledge and skills to deal effectively with uncertain and stressful situations
such as closing plants as part of a retrenchment strategy. In addition, our findings
suggest that improving the working environment (e.g., coworker and supervisory
support, role clarity, and autonomy) may enhance workers’ experience of career
flexibility and control and reduce their levels of emotional exhaustion. The op-
portunities for development on several key dimensions associated with a high
level of responsibility have been found to be different between males and females
(Ohlott et al., 1994). Combined with a lack of support and resources, females’
experience of fewer positive challenges may lead to lower levels of career flex-
ibility and control and, hence, strain. An application of a model of career stress
with more specific measurements of uncertainty, career flexibility and control,
with comparative analyses for males and females, may provide insights into how
organizations can balance the challenge–stress and –strain relationships. Finally,
programs seeking to enhance knowledge and skills tend to be evaluated in terms
of their effects on performance. However, a broader set of intermediate criteria
may be more appropriate. Such programs also may contribute to personal feelings
of career flexibility and control and, indirectly, to a reduction in stress and its
outcomes, such as emotional exhaustion, job dissatisfaction, and turnover.

REFERENCES
Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and
recommended two-step approach. Psychological Bulletin, 103(3), 411–423.
Arbuckle, J. L. (1997). AMOS users’ guide version 3.6. Chicago, IL: SmallWaters Corporation.
Arthur, M. B., Claman, P. H., & DeFillippi, R. J. (1995). Intelligent enterprise, intelligent careers.
Academy of Management Executive, 9(4), 7–20.
Barney, J. (1991). Firm resources and sustainable competitive advantage. Journal of Management, 17,
99–120.
Barney, J., & Lawrence, B. S. (1989). Pin stripes, power ties, and personal relationships: The economics
of career strategy. In M. B. Arthur, D. T. Hall, & B. S. Lawrence (Eds.), Handbook of career theory
(pp. 417–436). Cambridge, Australia: Cambridge Univ. Press.
Bentler, P. M. (1995). Comparative fit indices in structural equations. Psychological Bulletin, 107,
238–246.
Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley.
Brockner, J. (1992). Managing the effects of layoffs on survivors. California Management Review, 34,
9–28.
Burke, R. J., & Greenglass, E. (1995). A longitudinal study of psychological burnout in teachers.
Human Relations, 48(2), 187–202.
Byrne, B. M. (1993). The Maslach burnout inventory: Testing for factorial validity and invariance across
elementary, intermediate and secondary teachers. Journal of Occupational and Organizational
Psychology, 66, 197–212.
422 ITO AND BROTHERIDGE

Capelli, P. (1999) The new deal at work. Boston: Harvard Univ. Press.
Carson, K. D., Carson, P. P., & Bedian, A. G. (1995). Development and construct validation of a career
entrenchment measure. Journal of Occupational and Organizational Psychology, 68, 301–320.
Cherniss, C. (1992). Long-term consequences of burnout: An exploratory study. Journal of Organiza-
tional Behavior, 13(1), 1–11.
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral
sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Cordes, C. L., & Dougherty, T. W. (1993). A review and an integration of research on job burnout.
Academy of Management Review, 18, 621–656.
DeFrank, R. S., & Pliner, J. E. (1987). Job security, job loss, and outplacement: Implications for stress
and stress management. In J. C. Quick, R. S. Bhagat, J. E. Dalton, & J. D. Quick (Eds.), Work
stress: Health care systems in the workplace (pp. 195–209). New York: Praeger.
DeFillippi, R. J., & Arthur, M. B. (1994). The boundaryless career: A competency-based perspective.
Journal of Organizational Behavior, 15, 307–324.
de Rijk, A. E., Le Blanc, P. M., Schaufeli, W. B., & de Jonge, J. (1998) Active coping and need for control
as moderators of the job demand–control model: Effects on burnout. Journal of Occupational and
Organizational Psychology, 71, 1–18.
Doeringer, P. B. (1967). Determinants of the structure of industrial type internal labor markets. Industrial
and Labor Relations Review, 20, 206–220.
Evans, B. K., & Fischer, D. G. (1993). The nature of burnout: A study of the three-factor model of
burnout in human service and non-human service samples. Journal of Occupational and Organi-
zational Psychology, 66, 29–38.
Ferris, G. R., Frink, D. D., Galand, M. C., Zhou, J., Kacmar, M., & Howard, J. L. (1996). Perceptions of
organizational politics: Prediction, stress-related implications, and outcomes. Human Relations,
49, 233–260.
Fitzgerald, S. (1991). The stress–illness relationship among executives. Unpublished doctoral disser-
tation, University of Regina, Regina, Canada.
Frone, M. R., Russell, M., & Cooper, M. L. (1992). Antecedents and outcomes of work–family conflict:
Testing a model of the work-family interface. Journal of Applied Psychology, 77, 65–78.
Frone, M. R., Yardley, J. K., & Markel, K. S. (1997). Developing and testing an integrative model of
the work–family interface. Journal of Vocational Behavior, 50, 145–167.
Ganster, D. C. (1988). Improving measures of worker control in occupational stress research. In J. J.
Hurrell, L. R., Jr., Murphy, S. L. Sauter, & C. L. Cooper (Eds.), Occupational stress: Issues and
developments in research (pp. 88–99). New York: Taylor & Francis.
Greenhalgh, L., & Rosenblatt, Z. (1984). Job insecurity: Toward conceptual clarity. Academy of Man-
agement Review, 9, 438–448.
Hage, J., & Aiken, M. (1967). Relationship of centralization to other structural properties. Administra-
tive Science Quarterly, 12, 72–91.
Hershberger, S. L., Lichtenstein, P., & Knox, S. S. (1994). Genetic and environmental influences on
perceptions of organizational climate. Journal of Applied Psychology, 79(1), 24–33.
Ito, J. K., & Peterson, R. B. (1986). Effects of task difficulty and interdependence on information
processing systems. Academy of Management Journal, 29, 139–149.
James, L. R., Mulaik, S. A., & Brett, J. M. (1982). Causal analysis: Assumptions, models, and data.
Beverly Hills, CA: Sage.
Kahn, R. L., & Byosiere, P. (1992). Stress in organizations. In M. D. Dunnette & L. M. Hough (Eds.),
Handbook of industrial and organizational psychology (2nd ed., Vol. 3, pp. 571–650). Palo Alto,
CA: Consulting Psychologists Press.
Kanter, R. M. (1977). Men and women of the corporation. New York: Basic Books.
Karasek, R. A. (1979). Job demands, job decision latitude, and mental stress: Implications for job
redesign. Administrative Science Quarterly, 24, 285–307.
Karasek, R., & Theorell, T. (1990). Healthy work: Stress, productivity and the reconstruction of working
life. New York: Basic Books.
 CAREER UNCERTAINTY AND EMOTIONAL EXHAUSTION 423

Klandermans, B., van Vuuren, T., & Jacobson, D. (1991). Employees and job insecurity. In J. Hartley,
D. Jacobson, B. Klandermans, & T. van Vuuren with L. Greenhalgh & R. Sutton (Eds.), Job
insecurity: Coping with jobs at risk (pp. 40–64). London: Sage.
Klein, D. J., & Verbeke, W. (1999). Autonomic feedback in stressful environemnts: How do individ-
ual differences in autonomic feedback relate to burnout, job performance, and job attitudes in
salespeople? Journal of Applied Psychology, 84, 911–924.
Latack, J. C. (1984). Career transitions within organizations: An exploratory study of work, non-
work, and coping strategies. Organizational Behavior and Human Performance, 34, 296–
322.
Latack, J. C. (1989). Work, stress, and careers: A preventive approach to maintaining organiza-
tional health. In M. B. Arthur, D. T. Hall, & B. S. Lawrence (Eds.), Handbook of career theory
(pp. 252–274). New York: Cambridge Univ. Press.
Latack, J. C., Kinicki, A. J., & Prussia, G. E. (1995). An integrative process model of coping with job
loss. Academy of Management Review, 20, 311–342.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal and coping. New York: Springer-Verlag.
Lee, R. T., & Ashforth, B. E. (1993). A longitudinal study of burnout among supervisors and managers:
Comparisons between the Leiter and Maslach (1988) and Golembiewski et al. (1986) models.
Organizational Behavior and Human Decision Processes, 54, 369–398.
Lee, R. T., & Ashforth, B. E. (1996). A meta-analytic examination of the correlates of the three
dimensions of job burnout. Journal of Applied Psychology, 81, 123–133.
Leiter, M. P. (1991). Coping patterns as predictors of burnout: The function of control and escapist
coping patterns. Journal of Organizational Behavior, 12, 123–144.
Leiter, M. P. (1993). Burnout as a developmental process: Consideration of models. In W. B. Schaufeli,
C. Maslach, & T. Marek (Eds.), Professional burnout: Recent developments in theory and research
(pp. 237–250). Washington, DC: Taylor & Francis.
Leiter, M. P., Clark, D., & Durup, J. (1994). Distinct models of burnout and commitment among men
and women in the military. Journal of Applied Behavioral Science, 30(1), 63–82.
London, M. (1983). Toward a theory of career motivation. Academy of Management Review, 8, 620–630.
London, M. (1993). Relationships between career motivation, empowerment and support for career
development. Journal of Occupational and Organizational Psychology, 66, 55–69.
Malos, S. B., & Campion, M. A. (1995). An options-based model of career mobility in professional
service firms. Academy of Management Review, 20, 611–644.
Maslach, C. (1982). Understanding burnout: Definitional issues in analyzing a complex phenomenon.
In W. S. Paine (Ed.), Job stress and burnout: Research, theory, and intervention perspectives
(pp. 29–40). Beverly Hills, CA: Sage.
Maslach, C., & Jackson, S. E. (1986). Maslach burnout inventory: Manual (2nd ed.). Palo Alto, CA:
Consulting Psychologists Press.
Maslach, C., & Leiter, M. P. (1997). The truth about burnout: How organizations cause personal stress
and what to do about it. San Francisco: Jossey-Bass.
McCauley, C. D., Ruderman, M. N., Ohlott, P. J., & Morrow, J. E. (1994). Assessing the developmental
components of managerial jobs. Journal of Applied Psychology, 79, 544–561.
McGee, G. W., & Ford, R. C. (1987). Two (or more?) dimensions of organizational commitment: Re-
examination of the affective and continuance commitment scales. Journal of Applied Psychology,
72, 638–642.
McGrath, J. E. (1976). Stress and behavior in organizations. In M. D. Dunnette (Ed.), Handbook of
industrial and organizational psychology (pp. 1351–1395). Chicago: Rand McNally.
Meyer, J. P., & Allen, N. J. (1984). Testing the ‘side-bet theory’ of organizational commitment: Some
methodological considerations. Journal of Applied Psychology, 69, 372–378.
Meyer, J. P., Allen, N. J., & Smith, C. A. (1993). Commitment to organizations and occupations:
Extension and test of a three-component conceptualization. Journal of Applied Psychology, 78,
538–551.
Moos, R. H. (1981). Work environment scale manual. Palo Alto, CA: Consulting Psychologists Press.
424 ITO AND BROTHERIDGE

Netemeyer, R. G., Johnston, M. W., & Burton, S. (1990). Analysis of role conflict and role
ambiguity in a structural equations framework. Journal of Applied Psychology, 75, 148–
157.
Numerof, R. E. (1987). Team building interventions: An organizational stress moderator. In J. C.
Quick, R. S. Bhagat, & J. E. Dalton (Eds.), Work stress: Health care systems in the workplace
(pp. 171–194). New York: Praeger.
Ohlott, P. J., Ruderman, M. N., & McCauley, C. D. (1994). Gender differences in managers’ develop-
mental job experiences. Academy of Management Journal, 37, 46–67.
Pazy, A. (1990). The threat of professional obsolescence: How do professionals at different career
stages experience it and cope with it? Human Resource Management, 29, 251–269.
Perrow, C. (1967). A framework for comparative organizational analysis. American Sociological
Review, 32, 194–208.
Perry, L. T. (1988). Least-cost alternatives to layoffs in declining industries. Organizational Dynamics,
14(4), 48–61.
Rotter, J.B. (1966). Generalized expectancies for internal versus external control of reinforcement.
Psychological Monographs, 80, 1–28.
Sonnenfeld, J. A., & Peiperl, M. A. (1988). Staffing policy as a strategic response: A typology of career
systems. Academy of Management Review, 13, 588–600.
Stearns, G. M. (1992). A multivariate approach to the investigation of stress in the Royal Canadian
Mounted Police. Unpublished doctoral dissertation. University of Regina, Regina, Canada.
Sutton, R. I. (1990). Organizational decline processes: A social psychological perspective, Research
in Organizational Behavior, 12, 205–253.
Sutton, R. I., & Kahn, R. L. (1987). Prediction, understanding, and control as antidotes to organizational
stress. In J. W. Lorsch (Ed.), Handbook of organizational behavior (pp. 272–285). Englewood
Cliffs, NJ: Prentice-Hall.
Tanaka, J. S., & Huba, G. J. (1985). A fit index for covariance structure models under arbitrary GLS
estimation. British Journal of Mathematical and Statistical Psychology, 28, 197–201.
Tucker, L. R., & Lewis, C. (1973). A reliability coefficient for maximum likelihood factor analysis.
Psychometrica, 38, 1–10.
Van de Ven, A. H., & Delbecq, A. (1974). A task contingent model of work—Unit structure. Admin-
istrative Science Quarterly, 19, 183–197.
Van de Ven, A. H., Delbecq, A., & Koenig, R. (1976). Determination of coordination modes within
organizations. American Sociological Review, 41, 322–338.
Wall, T. D., Jackson, P. R., Mullarkey, S., & Parker, S. K. (1996). The demands–control model of
job strain: A more specific test. Journal of Occupational and Organizational Psychology, 69,
153–166.
Waterman, R. H., Waterman, J. A., & Collard, B. A. (1994). Toward a career-resilient workforce.
Harvard Business Review, 72, 87–96.
Weick, K. E., & Berlinger, L. R. (1989). Career improvisation in self-designing organizations. In M. B.
Arthur, D. T. Hall, & B. S. Lawrence (Eds.), Handbook of career theory (pp. 313–328). Cambridge,
Australia: Cambridge Univ. Press.
Wholey, D. R. (1985). Determinants of firm internal labor markets in large law firms. Administrative
Science Quarterly, 30, 318–355.
Wright, T. A., & Cropanzano, R. (1998). Emotional exhaustion as a predictor of job performance and
voluntary turnover. Journal of Applied Psychology, 83, 486–493.

Received April 12, 2000; published online April 26, 2001