Copyright 2001 by the Educational Publishing Foundation

1076-8998/01/S5.00 DOI: 10.1037//1076-8998.6.2.91

Journal of Occupational Health Psychology

2001, Vol. 6, No. 2, 91-100

Career Involvement and Family Involvement as Moderators

of Relationships Between Work-Family Conflict
and Withdrawal From a Profession
Karen M. Collins

Jeffrey H. Greenhaus and

Saroj Parasuraman

Lehigh University

Drexel University
This study extended prior analyses by J. H. Greenhaus, K. M. Collins, R. Singh, and S.
Parasuraman (1997) by examining relationships between 2 directions of work-family conflict
(work-to-family conflict and family-to-work conflict) and withdrawal from public accounting.
The sample consisted of 199 members of the American Institute of Certified Public Accountants
(135 men and 64 women) who were married or in a long-term relationship and who had 1 or more
children. It was found that work-to-family conflict (but not family-to-work conflict) was positively related to withdrawal intentions. In addition, relationships of work-to-family conflict with
withdrawal intentions and withdrawal behavior were stronger for individuals who were relatively
uninvolved in their careers than for those who were highly involved in their careers. The
implications of the findings for future research are discussed.

the field of public accounting itself. These developments emphasize the need to gain a better understanding of the dynamics underlying withdrawal
from a professional career field.
Accounting firms have assumed that the high rate
of withdrawal from the profession is due, in part, to
the stress generated by extensive work-family conflict (American Institute of Certified Public Accountants, 1992). This is not an unreasonable assumption.
Work-family conflict tends to produce low levels of
job satisfaction and career satisfaction (Adams, King,
& King, 1996; Higgins, Duxbury, & Irving, 1992;
Netemeyer, Boles, & McMurrian, 1996; Parasuraman, Purohit, Godshalk, & Beutell, 1996), both of
which may stimulate withdrawal tendencies (Blau &
Lunz, 1998; Neapolitan, 1980; Rhodes & Doering,
1983, 1993). Extensive work-family conflict has also
been associated with a diminished quality of life
(Higgins et al., 1992; Rice, Prone, & McFarlin, 1992)
and a variety of dysfunctional behaviors (Frone, Russell, & Barnes, 1996; Frone, Russell, & Cooper,
1993, 1997; MacEwen & Barling, 1994; Stewart &
Barling, 1996).
Although work-family conflict has been mentioned as a potential determinant of the decision to
withdraw from a profession or career field, this topic
has received little attention in the research literature.
Recently, however, Greenhaus, Collins, Singh, and
Parasuraman (1997) examined the impact of work
and family factors on withdrawal from public accounting. They observed significant relationships between work-family conflict and withdrawal. However, the relationships were smaller than what might

Organizational downsizing and restructuring in response to global competitive pressures have been
accompanied by marked changes in the nature of
careers. The uncertainty of continuous employment
within a single organization and extensive work pressures in the new "lean and mean" organizations have
led to the growing prevalence of "boundaryless" careers (Arthur & Rousseau, 1996). In light of recent
changes in the psychological contract between employers and employees as well as a favorable labor
market, individuals pursuing boundaryless careers
are likely to move across organizational and occupational boundaries with increasing frequency.
Although high rates of voluntary turnover have
been observed among diverse groups of occupations
(e.g., nursing, information technology), the public
accounting profession has recently witnessed turnover rates in excess of 20% per year (Connor, Hooks,
& McGuire 1997). A disturbing trend observed by
the profession is that public accountants are not only
leaving their employer but are also withdrawing from

Jeffrey H. Greenhaus and Saroj Parasuraman, Department of Management, LeBow College of Business, Drexel
University; Karen M. Collins, College of Business and
Economics, Lehigh University.
An earlier version of this article was presented in August
1999 at the annual meeting of the American Accounting
Association, San Diego, California. We thank Romila Singh
for her assistance with the research.
Correspondence concerning this article should be addressed to Jeffrey H. Greenhaus, Department of Management, Drexel University, Philadelphia, Pennsylvania 19104.
Electronic mail may be sent to greenhaus@drexel.edu.

91

92

GREENHAUS, PARASURAMAN, AND COLLINS

have been expected given prior beliefs regarding the

role of work-family stress in withdrawal from a
career field (American Institute of Certified Public
Accountants, 1992; Hooks, Thomas, & Stout, 1995;
Rhodes & Doering, 1983). The present study was
motivated by our desire to gain a deeper understanding of the relationships between work-family conflict
and withdrawal intentions and withdrawal behavior.
The study, which is based on Greenhaus et al.'s
(1997) data set, goes beyond prior analyses in two
respects.
First, Greenhaus et al. (1997) studied global
work-family conflict using a composite measure
that confounded the direction of the conflict. Recognizing the bidirectionality of work-family conflict, recent studies have distinguished work-tofamily conflict, in which work interferes with
family life, from family-to-work conflict, in which
family responsibilities interfere with work. Because the two directions of conflict can have different consequences (Frone, Russell, & Cooper,
1992), it is possible that withdrawal from a profession is not affected in the same way by workto-family conflict and family-to-work conflict.
Therefore, combining both directions of conflict
into one composite measure may have masked a
stronger relationship between withdrawal and one
or the other direction of conflict. This research
extends Greenhaus et al.'s analyses by examining
the separate effects of work-to-family conflict and
family-to-work conflict on withdrawal intentions
and behavior.
Second, it is possible that the strength of the
relationship between work-family conflict and
withdrawal from a profession is contingent on personal characteristics that were not examined by
Greenhaus et al. (1997). Just as organizational
turnover is produced by interactions between situational and personal variables (Steel & Ovalle,
1984), so too may withdrawal from a profession.
Using social identity theory as a theoretical framework, the present research extends the prior analyses by investigating the conditions under which
work-family conflict is related to withdrawal intentions and behavior.
In sum, the goals of the present study were to
examine the effects of work-to-family conflict and
family-to-work conflict on withdrawal from public
accounting and to investigate the conditions under
which conflict is associated with withdrawal from the
profession. The rationale for the hypotheses is developed in the following section.

Work-Family Conflict, Role Involvement, and

Withdrawal From a Profession
The present study examined both directions of
work-family conflict to determine whether one direction of conflict has a stronger impact on withdrawal
from a profession than the other direction of conflict.
There is good reason to believe that work-to-family
conflict is more likely to have an impact on the
withdrawal process than family-to-work conflict. Because the most substantial determinants of work-tofamily conflict reside within the work domain (Frone
et al., 1992; Frone, Yardley, & Markel, 1997), individuals who experience extensive work-to-family
conflict may attempt to reduce the conflict by leaving
the work environment that has produced the stress.
However, individuals who experience extensive family-to-work conflict may not necessarily withdraw
from their profession because the roots of the conflict
lie not in the work environment but rather in the
family domain. Although Frone et al. (1992) suggested that work-to-family conflict, rather than family-to-work conflict, may bear a direct relationship to
voluntary turnover, we are not aware of any research
that has tested this proposition. Therefore, Hypothesis 1 predicts that there is a positive relationship
between work-to-family conflict and withdrawal intentions and behavior. For the reason discussed
above, we do not expect to observe a relationship
between family-to-work conflict and withdrawal.
It is also possible that the modest relationships
between work-family conflict and withdrawal observed by Greenhaus et al. (1997) reflect the presence
of other factors that moderate the strength or direction of the relationship between conflict and withdrawal. Consistent with social identity theory, we
propose that the impact of work-to-family conflict on
withdrawal intentions is dependent on an employee's
level of involvement in family and career.
According to social identity theory, social roles
(e.g., professional, spouse, or parent) form the basis
of a person's self-identity (Burke, 1991; Frone, Russell, & Cooper, 1995), and the effectiveness of role
performance has consequences for one's self-evaluation (Thoits, 1991). Because individuals participate
in a variety of social roles, they have multiple identities (Lobel, 1991) that are organized in a hierarchy
of centrality (Thoits, 1991). Therefore, there are individual differences in the salience of a particular
role as a determinant of self-identification (Frone et
al., 1995).
Thoits (1991, pp. 105-106) suggested that "the
more salient the role identity, the more meaning,

93

WITHDRAWAL FROM A PROFESSION

purpose, and behavioral guidance the individual

should derive from its enactment, and thus the more
that identity should influence psychological well-being." Thus, some role requirements have the potential
to reinforce one's self-identity, and others may diminish an individual's sense of identity. The latter are
thought to be more psychologically damaging when
they threaten one's identity in highly salient roles
(identity-threatening stressors) than when they
threaten one's identity in less central roles (Thoits,
1991).
Therefore, Hypothesis 2 predicts that the relationships between work-to-family conflict and withdrawal intentions and behavior are stronger for employees who are highly involved in their families
than for those who are relatively uninvolved in their
families. Extensive work-to-family conflict is likely
to be an identity-threatening stressor for individuals
who are highly involved in family life because work
interferes with an aspect of their livesthe family
that is a central part of their self-identity. Familyinvolved professionals may respond to this identity
threat by seeking to leave the profession that is the
source of the conflict. Employees who are relatively
uninvolved in their families, however, may tolerate
high levels of work-to-family conflict and see no
need to leave their profession as a result of the
interference. For this latter group, work-to-family
conflict is an identity-irrelevant experience (Thoits,
1991) that is unlikely to trigger withdrawal. We made
no prediction regarding family involvement as a
moderator of the relationships between family-towork conflict and withdrawal.
On the basis of similar reasoning, Hypothesis 3
predicts that the relationships between work-to-family conflict and withdrawal intentions and behavior
are weaker for employees who are highly involved in
their careers than for those who are relatively uninvolved in their careers. Because effective performance in highly salient roles is important to one's
self-identity (Thoits, 1991), career-involved employees may tolerate extensive work-to-family interference to achieve career success. For this reason, the
withdrawal of career-involved professionals may not
be highly dependent on the level of work-to-family
conflict they experience. However, individuals relatively uninvolved in their careers who experience
extensive work-to-family conflict may be inclined to
leave the profession that has produced the conflict
because their career is not sufficiently salient to their
self-identity to justify its continual intrusion into their
family life. We made no prediction regarding career

involvement as a moderator of the relationships between family-to-work conflict and withdrawal.

Method
Sample
The present research was part of a larger study on the
quality of life in public accounting (Greenhaus et al., 1997).
Participants were members of the American Institute of
Certified Public Accountants (AICPA) who were employed
in public accounting below the partnership level, excluding
sole practitioners and those who had joined the institute
within the previous 12 months. Surveys were mailed to
1,000 randomly selected members of the AICPA who met
the above-mentioned criteria. To relate survey responses to
subsequent withdrawal from public accounting (discussed
below), we assigned a confidential research number that
identified the respondent to each outgoing survey. Respondents were informed about the presence of the research
number, were assured that their responses would be kept
confidential, and were instructed to return their completed
surveys directly to the researchers.
Completed surveys were returned by 428 accountants,
which represents a response rate of 43%. We were able to
compare our respondents with the AICPA membership on
four background characteristics: sex, age, number of years
in public accounting, and size of their firm. Compared with
the public practice membership of the AICPA, the present
sample included a larger percentage of women (37% in the
present sample vs. 24% in the AICPA), was younger (median age of 33.0 years vs. 40.9 years), had spent fewer years
in public accounting (median time of 8.0 years vs. 13.6
years), and was more likely to be employed by a national or
international public accounting firm (33% vs. 24%). It is
likely that our strategy of excluding public accountants at
the partner level and sole practitioners yielded a younger,
less-experienced sample with a greater percentage of
women than the national membership as a whole.
The present analyses were conducted on the 199 respondents who were married or in a long-term relationship and
who had one or more children. This sample included 135
men and 64 women, 98% of whom were married. The
number of children ranged from 1 to 7, with 82% of the
sample having one or two children. The average age of the
respondents' children was 7.52 years (SD = 8.19 years).
Thirty-five percent of the respondents were senior accountants and 65% were managers. The vast majority of the
sample worked in auditing (27%), tax (36%), or multiple
areas of specialization (31%). Respondents averaged 4.74
years in their position, 6.31 years in their firm, and 9.64
years in public accounting.

Measures
Work-to-family conflict was measured by nine items
(e.g., "Because my work is demanding, at times I am
irritable at home" and "My work schedule often conflicts
with my family life") that assessed time-based conflict and
strain-based conflict (Greenhaus & Beutell, 1985). Five of
the items were used previously by Greenhaus et al. (1997)
and assessed time-based and strain-based interference of

94

GREENHAUS, PARASURAMAN, AND COLLINS

work with home life. We also included four additional items

that assessed time-based and strain-based interference of
work with family life. Responses to the nine items, which
were made on a 5-point scale from 1 (strongly disagree) to
5 (strongly agree), were averaged to form a work-to-family
conflict scale (a = .90). Family-to-work conflict was measured by four items that assessed time-based and strainbased interference of home and family with work activities
(e.g., "The demands of my family life make it difficult to
concentrate at work"). Responses to the four items, which
were made on a 5-point scale from 1 (strongly disagree) to
5 (strongly agree), were averaged to form a family-to-work
conflict scale (a = .57).
Career involvement was assessed with three items from
Lodahl and Kejner's (1965) job involvement scale, with the
word career substituted for job (e.g., "Most of the important
things that happen to me involve my career"). Responses to
the three items were averaged to form a career involvement
score (a = .79). Family involvement was assessed with a
parallel set of three items (with the word family substituted
for career) that has been used previously in the literature
(Parasuraman, Greenhaus, & Granrose, 1992). Responses
were averaged to produce a family involvement score
(a = .84).
Intentions to leave public accounting were measured by
three items (e.g., "I will probably look for a new job outside
of public accounting in the next year"). Two of the items
were adapted from Cammann, Fichman, Jenkins, and Klesh
(1979; cited in Seashore, Lawler, Mirvis, & Camman,
1982), and the third item was written for this research. Each
item was answered on a 5-point scale from 1 (strongly
disagree) to 5 (strongly agree), and responses to the three
items were averaged to form a total intention score (a =
.92). The measure of withdrawal from public accounting
was derived from information contained on a computer tape
provided by the AICPA. The tape indicated the employment
status of each member of the sample 22 months after the
original survey data were collected. A dichotomous variable
was created with individuals who had withdrawn from
public accounting coded 1 and individuals who had remained in public accounting coded 0.

Data Analyses
Preliminary analyses revealed that two demographic variables were significantly related to intentions to leave and
withdrawal from public accounting: size of firm (national or
international vs. regional or local) and number of years in
public accounting. Because they were also significantly
related to work-to-family conflict, firm size and tenure in
public accounting were controlled in all subsequent analyses.
Hypothesis 1 predicted a positive relationship of workto-family conflict with withdrawal intentions and withdrawal behavior, and Hypotheses 2 and 3 predicted interactions of work-to-family conflict with family involvement
and career involvement, respectively. The predictions of
withdrawal intentions were tested with hierarchical multiple
regression analysis, and the predictions of withdrawal behavior were tested with hierarchical logistic regression analysis. All variables were standardized prior to conducting the
regression analyses (Jaccard, Turrisi, & Wan, 1990). Withdrawal intentions and withdrawal behavior were separately
regressed on the two demographic controls in Step 1. The
independent variables (work-to-family conflict, family-towork conflict, career involvement, and family involvement)
were entered in Step 2. In Step 3, the interactions of workto-family conflict and family-to-work conflict with career
involvement and family involvement were entered. All significant interactions were plotted by calculating regression
lines from the full equation at high (+1 SD) and low (-1
SD) levels of the moderator variable (Cohen & Cohen,
1983).

Results
Table 1 presents the means, standard deviations,
and intercorrelations for the study variables. During
the 22 months between the administration of the
survey and the collection of the withdrawal data, 18%
of the sample had withdrawn from public accounting.

Table 1
Intercorrelations Among Study Variables
Variable
1.
2.
3.
4.
5.
6.
7.
8.

Size of firm"
Years in public accounting
Family-to-work conflict
Work-to-family conflict
Career involvement
Family involvement
Intention to leave public accounting
Withdrawal from public accounting11
M
SD

1
-.10
-.02
.35**

-.04
-.17*
.11
-.06
-.23**
-.19**

.09
.04
.31**
.31**
0.30
0.46

Note. Pairwise Ns range from 191 to 199.

a
1 = national/international; 0 = regional/local.
accounting.
* p < .05, two-tailed. ** p s .01, two-tailed.

9.64

5.64
1

.34**
.01
-.07
-.17*
-.06
.02
.40**
.02
.19**
2.22
3.21
0.61
0.81

-.26**
-.31**
-.09
3.14
0.76

.01
.11
4.35
0.64

.38**
2.68
1.16

0.18

0.38

1 = withdrew from public accounting; 0 = remained in public

95

WITHDRAWAL FROM A PROFESSION

Work-to-family conflict was positively related to
withdrawal intentions and subsequent withdrawal
from public accounting, whereas family-to-work
conflict was unrelated to both measures of withdrawal. In addition, career involvement was negatively associated with withdrawal intentions.
Tables 2 and 3 show the results of the multiple
regression analysis and the logistic regression
analysis, respectively. Hypothesis 1 was partially
supported. Work-to-family conflict was positively
related to withdrawal intentions, family-to-work
conflict was unrelated to intentions, and a comparison of regression coefficients (Cohen & Cohen,
1983) revealed that work-to-family conflict was
more strongly related to intentions than was
family-to-work conflict, f(l) = 3.03, p < .01.
Inconsistent with Hypothesis 1, however, work-tofamily conflict was not related to subsequent withdrawal behavior.
Hypothesis 2 predicted that the relationships of
work-to-family conflict with withdrawal intentions
and behavior would be stronger for employees who
are highly involved in their families than for those
who are relatively uninvolved in their families. The
findings provided no support for the hypothesis.
There was no interaction between work-to-family
conflict and family involvement predicting either indicator of withdrawal from public accounting. In
addition, there was no evidence that family involvement moderated the relationships between familyto-work conflict and withdrawal intentions and
behavior.

Hypothesis 3 predicted that the relationships between work-to-family conflict and withdrawal intentions and behavior would be weaker for employees who are highly involved in their careers
than for those who are relatively uninvolved in
their careers. The regression analyses revealed significant interactions between work-to-family conflict and career involvement predicting withdrawal
intentions and withdrawal behavior. Figure 1 indicates that the relationship between work-to-family
conflict and withdrawal intentions was not as substantial for employees who are highly involved in
their careers as for those who are relatively uninvolved in their careers. The interaction predicting
withdrawal behavior, shown in Figure 2, reflects a
somewhat similar pattern. There was a strong positive relationship between work-to-family conflict
and withdrawal for employees uninvolved in their
careers, and a weak negative relationship between
conflict and withdrawal for those highly involved
in their careers.
Although it was not predicted that family-towork conflict would interact with career involvement, there was a significant interaction between
these variables predicting withdrawal intentions.
As shown in Figure 3, there was a weak positive
relationship between family-to-work conflict and
withdrawal intentions for employees who are relatively uninvolved in their careers and a weak
negative relationship for employees who are highly
involved in their careers.

Table 2
Hierarchical Multiple Regression Analysis Predicting the Intention
to Leave Public Accounting
Step & predictors

A/?2

F change

dfa

1. Controls
Size of firm"
Years in public accounting
2. Independent variables
Work-to-family conflict (WFC)
Family-to-work conflict (FWC)
Career involvement (CI)
Family involvement (FI)
3. Interactions
WFC x CI
WFC X FI
FWC x CI
FWC X FI

.14

16.24**

2, 193
.30**
-.20**

.19

13.18**

4,189
.28**
-.06
-.33**
-.09

.04

2.80*

4, 185

Note. N = 196.
a

1 = national/international; 0 = regional/local.

t p .06, two-tailed. * p < .05, two-tailed. ** p < .01, two-tailed.

-.12t
-.07
-.14*
-.01

96

GREENHAUS, PARASURAMAN, AND COLLINS

Table 3
Hierarchical Logistic Regression Analysis Predicting Withdrawal From
Public Accounting
Step & predictors

X2 improvement

1. Controls
Size of firm6
Years in public accounting
2. Independent variables
Work-to-family conflict (WFC)
Family-to-work conflict (FWC)
Career involvement (CI)
Family involvement (FI)
3. Interactions
WFC X CI
WFC X FI
FWC X CI
FWC X FI

30.69**

Note. N = 111.
' Unstandardized coefficient.

0.80**
-1.30**
4.47

0.20
0.14
-0.20
0.31
5.72

-0.50t
-0.60
0.29
0.44

1 = national/international; 0 = regional/heal.

t p = .06, two-tailed. **p .01, two-tailed.

Discussion and Implications for

Future Research
This study was conducted to address two related
questions. Do work-to-family conflict and family-towork conflict differ in their impact on the decision to
withdraw from a profession? Do the levels of family
involvement and career involvement moderate the
relationships between conflict and withdrawal? Our
findings indicate that work-to-family conflict has a
more substantial impact on withdrawal than does
family-to-work conflict and that career involvement
moderates the relationship between conflict and
withdrawal.
Our research supports these conclusions in several
ways. First, work-to-family conflict had a stronger
effect on withdrawal intentions than did family-towork conflict, which was unrelated to intentions.
Apparently, when considering whether to withdraw
from a professional career field, individuals take into
account the source of the stress and the likelihood
that departure from the profession will relieve the
stress. Because work-to-family conflict is aroused
primarily from stressors within the work environment
(Prone et al., 1992, 1997), leaving the profession that
interferes with family life is likely to reduce an
individual's level of stress. Individuals who experience extensive family-to-work conflict are not likely
to withdraw from their profession because the source
of the interference is coming from the family domain,
not the workplace. Thus, a major contribution of this

study derives from our finding that an individual's

intention to leave a professional career field is sensitive to the direction of the interference between work
and family roles.
Although work-to-family conflict was unrelated to
subsequent withdrawal for the total sample, it did
predict withdrawal behavior for individuals who are
not highly involved in their careers. Career involvement interacted with work-to-family conflict to predict withdrawal intentions and withdrawal behavior.
The interactions had two common elements. First, the
relationships of work-to-family conflict with withdrawal intentions and behavior were strongly positive
only for those individuals low in career involvement.
Second, withdrawal intentions and the likelihood of
withdrawal from public accounting were strongest
for employees who experienced extensive work-tofamily conflict and who were relatively uninvolved in
their careers.
Therefore, work-to-family conflict, in conjunction
with career involvement, does influence withdrawal
from a profession. Apparently, there is less of a
reason to put up with extensive work-to-family conflict if work is not a highly salient part of one's life.
However, career-involved professionals are not
overly disturbed by the interference of work with
family life and are willing to tolerate the interference
for the sake of their careers. Perhaps they accept the
work demands and the interference with family life
as the price to be paid for the prestige and economic
rewards that come with a career in public accounting.

WITHDRAWAL FROM A PROFESSION

97

3.75

CO

3.5

B 3.25
I
.E
3

15
| 2.75 -\
~ 2.5
2.25 ^

2
Low

High

Work-to-Family Conflict
- Low career involvement

- Ugh career involvement

Figure 1. Interaction between work-to-family conflict and career involvement predicting

withdrawal intentions. The dependent variable is expressed in raw scores.

Thus, an intense involvement in a professional career

not only reduces the intention to withdraw from the
field (Blau & Lunz, 1998) but also weakens the
impact of work-to-family conflict on withdrawal.
Although we have concluded that family-to-work
conflict plays little or no role in the withdrawal
process, we should note that family-to-work conflict
did interact with career involvement to predict withdrawal intentions (see Figure 3). Despite the interaction, the relationship between family-to-work conflict
and withdrawal intentions was not strong either for
individuals who are relatively uninvolved in their
careers or for those who are highly involved in work.
Moreover, although it is possible that the relationship between family-to-work conflict and withdrawal
intentions was attenuated by the relatively low reliability of the family-to-work conflict scale, applying
a correction for attenuation due to unreliability (Ferguson, 1966) did not substantially increase the correlation. A more likely explanation, as noted earlier,
is that the direction of the interference between work
and family roles determines the impact of conflict on
withdrawal (Frone et al., 1992).

The findings of the present study did not support

the hypothesized role of family involvement in the
withdrawal process. Although we had expected that
family-oriented individuals would withdraw from
their profession when their work lives extensively
interfered with their family, there was no interaction
between family involvement and work-to-family
conflict. It is possible that employees who are highly
involved with then- families not only want to protect
the time they have available for their families but are
also motivated to maintain their families' financial
security. In this event, high levels of work-to-family
conflict, although upsetting, may also be seen as a
necessary byproduct of the level of effort that is
required to assure the material well-being of the
family. Additional research is required to understand
the role, if any, that family involvement plays in the
decision to withdraw from a professional career field.
Additional research in several other areas is also
warranted. It would be useful to determine whether
family-to-work conflict influences withdrawal decisions in ways that could not be examined in the
present study. For example, when individuals who

98

GREENHAUS, PARASURAMAN, AND COLLINS

0.8 -,

i
2

0.7

T3
^~
JL.

0.6 0.5 0.4

-a
o 0.3 o
"o 0.2

in

0.1 -

Low

High

Work-to-Family Conflict
- Low career involvement

- High career involvement

Figure 2. Interaction between work-to-family conflict and career involvement predicting the
likelihood of withdrawal from public accounting. The dependent variable, logit (withdrawal),
is converted to the likelihood of withdrawal through exponentiation and division as described
by Menard (1995, p. 13).

experience extensive family-to-work conflict fail in

their attempt to relieve the stressors within the family
role, do they withdraw from interactions with other
members of their family, as suggested by the findings
of MacEwen and Barling (1994, Figure 2)1 Or do
they attempt to leave their professional role for a less
demanding or more flexible work environment that is
less likely to be the object of family-induced interference? Answers to these questions require a research design that is capable of detecting the sequence of responses designed to cope with workfamily conflict.
It would also be helpful to examine the linkage
between withdrawal intentions and subsequent withdrawal behavior. The organizational turnover literature has revealed that the relationship between turnover intentions and turnover behavior is moderated
by a variety of variables (Maertz & Campion, 1998;
Steel & Ovalle, 1984). It is equally plausible that
situational and personal factors can either strengthen
or weaken the impact of intentions to withdraw from
a profession on subsequent withdrawal.
The present findings should be used cautiously to

predict withdrawal from public accounting for several reasons. First, our sample had a higher percentage of women and was somewhat younger and less
experienced than the national membership of the
AICPA. Second, the interactions we observed between conflict and career involvement explained
modest percentages of the variance in withdrawal
tendencies. We view our findings in general, and the
interactions in particular, as providing theoretical insight into the role of the work-family interface in
professional withdrawal and stimulating future research to enhance theoretical development further.
Moreover, future research should examine the impact of work and family life on withdrawal from
professional fields other than public accounting.
Many professional career fieldssuch as nursing,
law, and consultingsuffer high withdrawal rates,
and the new emphasis on boundaryless careers that
move across functional, organizational, and occupational boundaries will undoubtedly increase the frequency of major career shifts.
Finally, it is important to acknowledge that the
present research was not designed to test a compre-

WITHDRAWAL

FROM

PROFESSION

Low

High

Famify-to-Work Conflict

-Low career involve;

Figure 3.

Interaction between family-to-work conflict

h e n s i v e model of flillaiiav. al from a profession. Such

and c

Burke, P. J.

(1991). Identity

processes

and social

itjcis

a model would necessarily include a variety of workCohen, J., & Cohen, P. (1983). Applied multiple regression/
related variablesstressors, job

characteristics, and
correlation

attitudesthat were beyond the scope of the present

study. Having established in this study that the workfamily interface can play a role in withdrawal from a
profusion,

we

tncouraiji. 1

researchers

lor

the

bcliadnral

sciences

(2nd

Connor, M., Hooks, K., & McGuire, T. (1997). Gaining

legitimacy

for

flexible

work

arrangements

and

career

to examine a
fessional

broad array of work and family

analysis

ed.X HilKdale, NJ: Hribaum.

influences

in

future

services

Greenhaus

firms.

In

S.

(Eds.), Integrating

Parasurarnan

work

&

and family:

J.

H.

Chal-

studies of the withdrawal process.

Wssip;)rt
cryjsa:]

References

(i

CT
A

Qu;n. H o o k ,
il^i

and education, V-.i

.^;;;^;<.;( imtil\sii in p u t ht,lt,?,\

Y o i k : NkGraw-Hia

Frone, M. R., Russell, M., & Barnes, G. M. (1996)- WorkAdams, G. A., King, L. A., & King, D. W. (1996). Relationships of job

ami

fa:i:ilj

support, and work-famjiy

isfaction. Journal

i/j

LIU olvemtiH, family social

conflict with job and life

sat-

-Ir.i.W fv.; iitilagy, 81, 411-420.

Journa! ofOffupntirwii!

flfiliii

1'^i.li/i/ogy,

1, 57-69.

Frone, M. R., Russell, M., & Cooper, M. L. (1992). Prev-

boundaries a s y m n i e m j a l l - . p c r r i e a h l e " 1 Jnurnat

mobility

of

IDiocliuu-l

women and family

C.Yu

York

issues in the

workplace

Authar

nizational Behavior,

ofOrga

13, 723-729.

Frone, M. R.. Russell. M.. & Cooper, M. L. (1993). Rela-

Arthur. M. ., & Rousseau. D. M. (1996). A career lexicon

10(4), 28-39.
Blau, G., &Lunz, M. (1998). Testing the incremental cliK;

fi.-i'j,'

flil^-./.jr,

14. 545-558.

Frone, M. R., Russell. M,, & Cooper, M. L. (1995). Job

stressors, job involvement and employee health: A test of

profession beyond the effects of external, personal, and

identity theory. Journal of Occupational and

work-related variables. Journal of Vocational Behavior,

tional Psychology,

52, 260-269.

l-r;oe. M

Organiza-

68, 1-11.

K . Ru-,u-:i. V , & Coopci, M. L. 11997). Relation

100

GREENHAUS, PARASURAMAN, AND COLLINS

of work-family conflict to health outcomes: A four-year

longitudinal study of employed parents. Journal of Occupational and Organizational Psychology, 70, 325-335.
Frone, M. R., Yardley, J. K., & Market, K. S. (1997).
Developing and testing an integrative model of the workfamily interface. Journal of Vocational Behavior, 50,
145-167.
Greenhaus, J. H., & Beutell, N. J. (1985). Sources of conflict between work and family roles. Academy of Management Review, 10, 76-88
Greenhaus, J. H., Collins, K. M., Singh, R., & Parasuraman,
S. (1997). Work and family influences on departure from
public accounting. Journal of Vocational Behavior, 50,
249-270.
Higgins, C. A., Duxbury, L. E., & Irving, R. H. (1992).
Work-family conflict in the dual-career family. Organizational Behavior and Human Decision Processes, 51,
51-75.
Hooks, K. L., Thomas, P. B., & Stout, W. D. (1995).
Work-personal life balance variables and turnover intent
in big six public accounting firms. Unpublished manuscript.
Jaccard, J., Turrisi, R., & Wan, C. K. (1990). Interaction
effects in multiple regression. Thousand Oaks, CA: Sage.
Lobel, S. A. (1991). Allocation of investment in work and
family roles: Alternative theories and implications for
research. Academy of Management Review, 16, 507521.
Lodahl, T. M., & Kejner, M. (1965). Definition and measurement of job involvement. Journal of Applied Psychology, 49, 24-33.
MacEwen, K. E., & Barling, J. (1994). Daily consequences
of work interference with family and family interference
with work. Work & Stress, 8, 244-254.
Maertz, C. P., & Campion, M. A. (1998). 25 years of
voluntary turnover research: A review and critique. International Review of Industrial and Organizational Psychology, 13, 49-81.
Menard, S. (1995). Applied logistic regression analysis.
Thousand Oaks, CA: Sage.
Neapolitan, J. (1980). Occupational change in mid-career:

An exploratory investigation. Journal of Vocational Behavior, 16, 212-225.

Netemeyer, R. G., Boles, J. S., & McMurrian, R. (1996).
Development and validation of work-family conflict and
family-work conflict scales. Journal of Applied Psychology, 81, 400-410.
Parasuraman, S., Greenhaus, J. H., & Granrose, C. S.
(1992). Role stressors, social support, and well-being
among two-career couples. Journal of Organizational
Behavior, 13, 339-356.
Parasuraman, S., Purohit, Y. S., Godshalk, V. M., & Beutell, N. J. (1996). Work and family variables, entrepreneurial success, and psychological well-being. Journal of
Vocational Behavior, 48, 275-300.
Rhodes, S. R., & Doering, M. (1983). An integrated model
of career change. Academy of Management Review, 8,
631-639.
Rhodes, S. R., & Doering, M. (1993). Intention to change
careers: Determinants and process. Career Development
Quarterly, 42, 76-92.
Rice, R. W., Frone, M. R., & McFarlin, D. B. (1992).
Worknonwork conflict and the perceived quality of life.
Journal of Organizational Behavior, 13, 155-168.
Seashore, S. E., Lawler, E. E., Mirvis, P. H., & Cammann,
C. (Eds.). (1982). Observing and measuring organizational change: A guide to field practice. New York:
Wiley.
Steel, R. P., & Ovalle, N. (1984). A review and metaanalysis of research on the relationship between behavioral intentions and employee turnover. Journal of Applied Psychology, 69, 673-686.
Stewart, W., & Barling, J. (1996). Fathers' work experiences effect children's behaviors via job-related affect
and parenting behaviors. Journal of Organizational Behavior, 17, 221-232.
Thoits, P. A. (1991). On merging identity theory and stress
research. Social Psychology Quarterly, 54, 101112.

Received September 2, 1999

Revision received June 21, 2000
Accepted August 2, 2000