Content uploaded by Russell A. Matthews
Author content
All content in this area was uploaded by Russell A. Matthews on Aug 24, 2016
Content may be subject to copyright.
Work and Personal Life Boundary Management: Boundary Strength,
Work/Personal Life Balance, and the
Segmentation–Integration Continuum
Carrie A. Bulger
Quinnipiac University Russell A. Matthews
University of Connecticut
Mark E. Hoffman
Quinnipiac University
While researchers are increasingly interested in understanding the boundaries surrounding the
work and personal life domains, few have tested the propositions set forth by theory. Boundary
theory proposes that individuals manage the boundaries between work and personal life through
processes of segmenting and/or integrating the domains. The authors investigated boundary
management profiles of 332 workers in an investigation of the segmentation–integration contin-
uum. Cluster analysis indicated consistent clusters of boundary management practices related to
varying segmentation and integration of the work and personal life domains. But, the authors
suggest that the segmentation–integration continuum may be more complicated. Results also
indicated relationships between boundary management practices and work–personal life interfer-
ence and work–personal life enhancement. Less flexible and more permeable boundaries were
related to more interference, while more flexible and more permeable boundaries were related to
more enhancement.
Keywords: boundary management, boundary theory, work family interface, work-life balance,
cluster analysis
Although the idea that the boundaries between
work and personal life are becoming blurred has been
discussed by scholars for more than a decade (e.g.,
Hall & Richter, 1988; Nippert-Eng, 1996), only re-
cently have researchers begun to carefully investigate
those boundaries. Hall and Richter (1988) discussed
the idea that workers navigate the boundaries be-
tween work and personal life in various ways, such as
by fostering boundaries that are permeable in an
attempt to balance work and home lives. Boundary
theory as applied to the work/personal life interface
was further developed by Nippert-Eng (1996) and the
similar concept, work–family border theory, was fur-
ther developed theoretically by Clark (2000). This
theoretical work has since been expanded into test-
able propositions by other researchers (e.g., Ashforth,
Kreiner, & Fugate, 2000). These researchers propose
that working adults develop boundaries around both
their work and personal life domains that vary in
strength. The strength of the boundaries then influ-
ences outcomes of the interaction between work and
personal life (e.g., work–family conflict).
Much is still to be learned about boundaries,
boundary management, and the impact of various
boundary management strategies on worker out-
comes. Because it is proposed that various boundary
management practices may influence the experience
of work–family conflict (Ashforth et al., 2000), one
of the leading sources of occupational stress in the
United States (Sauter, Murphy, & Hurrell, 1990),
learning more about boundary management and its
outcomes is important for researchers and practitio-
ners interested in occupational health and stress. In
this study, we present an investigation of boundary
management profiles, as well as the relationship be-
Carrie A. Bulger, Department of Psychology, Quinnipiac
University; Russell A. Matthews, Department of Psychol-
ogy, University of Connecticut; Mark E. Hoffman, Depart-
ment of Computer Science and Interactive Digital Design,
Quinnipiac University.
Parts of this research were supported by a Quinnipiac
University Faculty Grant-in-Aid. We thank Kathy Alagno
for her help with data collection. This study was presented
at the 21st Annual Conference of the Society for Industrial
and Organizational Psychology, Dallas, TX, May 2006.
Russell A. Matthews is now located in the Department of
Psychology, Louisiana State University.
Correspondence concerning this article should be addressed
to Carrie A. Bulger, Department of Psychology, Quinnipiac
University, 275 Mount Carmel Avenue, CLAC-1, Hamden,
CT 06518. E-mail: carrie.bulger@quinnipiac.edu
Journal of Occupational Health Psychology
2007, Vol. 12, No. 4, 365–375
Copyright 2007 by the American Psychological Association
1076-8998/07/$12.00 DOI: 10.1037/1076-8998.12.4.365
365
tween boundary management strategies and work/
personal life balance.
Managing Multiple Roles and
Boundary Theory
The idea that people hold multiple important roles,
and that the management of those roles has important
outcomes, is not new (Barnett & Baruch, 1985;
Marks, 1977; Sieber, 1974). Researchers interested in
the work/family, or work/personal life, interface have
devoted decades to investigating these issues. Much
of that research has focused on the negative impact of
holding multiple roles. For instance, the conflict be-
tween work and family roles has received a great
deal of attention from researchers (Parasuraman &
Greenhaus, 2002) and for good reason. Work–family
conflict has been shown to negatively impact job
satisfaction (e.g., Allen, Herst, Bruck, & Sutton,
2000), life satisfaction (e.g., Kossek & Ozeki, 1998),
work withdrawal (Hammer, Bauer, & Grandey,
2003), and other stress outcomes such as depression
(e.g., Frone, Russell, & Cooper, 1992). Recently, a
few researchers have begun to heed calls to investi-
gate the positive impact of holding multiple roles
(e.g., Grzywacz & Marks, 2000).
While much of the research on the positive out-
comes of juggling multiple roles in the work and
personal life domains is recent, the possibility has
long been discussed in the literature. Sieber (1974)
acknowledged that role accumulation might result in
role strain, but suggested that it also affords individ-
uals opportunities for rewards such as self-enrich-
ment. Marks and MacDermid (1996) cite studies in
which some individuals reported strain as a result of
balancing multiple roles, but they point out that oth-
ers in those same studies reported little or no strain.
Given that, they suggested that role theories need to
acknowledge that individuals actively organize them-
selves within their roles and that the construction of
roles might explain the positive and negative out-
comes of attempts to balance multiple roles.
This idea of active organization is echoed by
Nippert-Eng (1996) in her extensive case study of
employees of a large organization. She suggested that
individuals create boundaries around their work and
their personal life in a deliberate fashion. Nippert-
Eng suggests that some individuals construct the
boundaries around their work and their personal life
to ensure that these domains remain separate, or
segmented, from one another while others construct
boundaries that are not as strong so that the domains
can be interconnected, or integrated. She further sug-
gests that the construction of a boundary that allows
for segmentation or integration depends upon various
factors including occupation type, coworkers, family
members, or individual preference.
These ideas were developed further by Clark
(2000) in her work–family border theory and sepa-
rately by Ashforth et al. (2000) in their examination
of boundary theory and role transitions. In work–
family border theory, Clark developed propositions
pertaining to multiple aspects of the work–family
interface, including the nature and strength of the
border between work and family, as well as other
characteristics of the work and family domains that
influence the ability to balance work and family.
Boundary theory, as originally developed by Nippert-
Eng (1996), was also intended to offer a means of
understanding the ways individuals negotiate differ-
ent domains. However, she emphasized not just work
and family, but other personal life domains as well.
Ashforth et al. (2000) focused on transitions between
roles held within and between domains as a way of
understanding boundary management. Despite these
few differences, there are key similarities between
work–family border theory and boundary theory.
Most recent research has used the term “boundary”
(e.g., Chesley, 2005; Desrochers, Hilton, & Larwood,
2005; Rothbard, Phillips, & Dumas, 2005); for the
sake of clarity, we will do the same.
In terms of similarities between border and bound-
ary theories, a defining idea behind work–family
border theory is the notion that individuals are more
enactive than reactive in defining their work and
family lives (Clark, 2000). She suggested that people
actively construct a boundary around each domain
that varies in strength. This is echoed by Ashforth et
al. (2000), who suggested that individuals have a
preference for the relative strength of their bound-
aries. According to these researchers, strong bound-
aries are constructed in order to maintain work and
family as separate domains, whereas weak bound-
aries are constructed to facilitate ease of interaction
between domains (Ashforth et al., 2000; Clark,
2000).
Both theories suggest that the strength of the
boundary can be characterized by permeability and
flexibility. A boundary is permeable if elements from
one domain are readily found in the other domain
(Ashforth et al., 2000; Clark, 2000; Hall & Richter,
1988). Permeability might be thought of as actual
interruptions or intrusions from one domain into the
other, over which the employee may have little con-
trol. Thought of another way, boundary permeability
366 BULGER, MATTHEWS, AND HOFFMAN
relates to being physically located in one domain, but
actually behaviorally responding to the other domain
(Clark, 2000, 2002). For example, a work boundary
is permeable if the employee is contacted by family
while at work.
The second element of boundary strength is flex-
ibility. A boundary is flexible if hypothetically it
could be relaxed to meet the demands of the other
domain (Ashforth et al., 2000; Clark, 2000; Hall &
Richter, 1988). For example, a work boundary is
flexible if the employee perceives that he or she could
leave work to attend to a family matter. Matthews
and Barnes-Farrell (2004) further refined the defini-
tion of boundary flexibility. They suggested that be-
cause flexibility entails the hypothetical, or per-
ceived, capacity to strengthen or weaken a boundary
(Hall & Richter, 1988; Clark, 2002), it should ac-
count both for an individual’s willingness and an
individual’s ability to leave one domain to attend to
the demands of the other domain. They demonstrated
that domain boundary strength is a function of flexi-
bility–ability, flexibility–willingness, and permeabil-
ity.
Because both border theory and boundary theory
suggest that boundary strength is related to work/
personal life balance, we will test the relationships
between boundary strength and the positive and neg-
ative aspects of work/personal life balance. Ashforth
et al. (2000) propose that the primary benefit of
segmenting is maintaining clarity between domains.
However, when domains are segmented the transi-
tions between these domains are hindered, which
could function as a cost to the individual (Ashforth et
al., 2000). They also suggest that integration facili-
tates smoother transitions between work and personal
life domains, but the domains also become more
blurred (Ashforth et al., 2000). Previous theoretical
work also suggested that attempts to balance multiple
roles would have both positive and negative conse-
quences (Marks & MacDermid, 1996; Sieber, 1974).
Therefore, we hypothesize that lower flexibility–
ability, lower flexibility–willingness, and lower per-
meability will be related to higher domain interfer-
ence. We further hypothesize that higher domain
flexibility–ability, higher flexibility–willingness, and
higher permeability will be related to higher domain
enhancement.
Hypothesis 1: Work flexibility–ability, work
flexibility–willingness, and work permeability
are negatively correlated with work interference
with personal life.
Hypothesis 2: Personal life flexibility–ability,
personal life flexibility–willingness, and per-
sonal life permeability are negatively correlated
with personal life interference with work.
Hypothesis 3: Work flexibility–ability, work
flexibility–willingness, and work permeability
are positively correlated with work enhance-
ment of personal life.
Hypothesis 4: Personal life flexibility–ability,
personal life flexibility–willingness, and per-
sonal life permeability are positively correlated
with personal life enhancement of work.
We make these predictions specifically for with-
in-domain relationships. That is, we expect to
see relationships between work domain bound-
ary strength and work interference with personal
life, and between personal life domain boundary
strength and personal life interference with
work. While it is possible that work domain
boundary strength is related to personal life in-
terfering with work, or vice versa, previous re-
search on the bidirectional nature of the work/
family interface suggests that relationships tend
to be stronger within domain (Frone, 2003).
Boundary Strength and the
Segmentation–Integration Continuum
As stated above, the idea that individuals actively
segment or integrate their work and personal life
domains is a seminal concept within boundary theory
(Nippert-Eng, 1996). Specifically, the strategies indi-
viduals use to manage their work and personal life
domain boundaries (i.e., flexibility and permeability)
create what Nippert-Eng (1996) first termed a seg-
mentation–integration continuum.Segmentation ex-
ists when there is low flexibility to leave one domain
to attend to the other and low permeability of the
domain boundaries. An individual at this end of the
continuum would maintain work and personal life as
separate domains. For example, such an individual
would not interrupt work to attend to personal mat-
ters.
In the case of integration, there is high flexibility
and high permeability of domain boundaries. At this
end of the continuum, an individual would maintain
work and personal life as freely interacting domains.
Such an individual would be likely to bring work
home or leave a family event to attend to work
matters. It should be noted here that Ashforth et al.
367WORK AND PERSONAL LIFE BOUNDARY MANAGEMENT
(2000) proposed that role identity, specifically role
contrast, is associated with flexibility and permeabil-
ity. They proposed that role contrast influences flex-
ibility and permeability and may therefore influence
how individuals array along the segmentation-
integration continuum. Because role contrast may be
seen as separate from flexibility and permeability,
and because our focus is on the ways in which
flexibility–ability, flexibility–willingness, and perme-
ability fall into patterns of boundary management
practices, we did not include role contrast in this
study.
Although there has been research on segmentation
and integration as boundary management strategies,
there is still little empirical evidence for the segmen-
tation-integration continuum. In her study on organi-
zation responses to the work/nonwork boundary,
Kirchmeyer (1995) found that perceptions that the
organization encouraged segmentation were nega-
tively related to organizational commitment, whereas
perceived organizational respect for multiple roles
was positively related to organizational commitment.
Rothbard et al. (2005) found that the fit between an
individual’s desire for segmentation or integration
and available organizational policies impacted job
attitudes. For example, individuals with a desire for
segmentation who had access to integrating organi-
zational policies, such as onsite childcare, had lower
job satisfaction and organizational commitment than
individuals with a desire for integration. Another
study showed that the number of hours working both
at home and work, distractions, and work–family
conflict were related to higher integration of work
and family (Desrochers et al., 2005). These studies
contribute to our understanding of boundary manage-
ment, but none was designed to directly investigate
the segmentation–integration continuum.
In this study, we explore the segmentation–
integration continuum using a cluster analysis ap-
proach. That is, one way of investigating the segmen-
tation–integration continuum would be to classify
individuals in terms of their boundary management
practices. For instance, individuals might construct a
flexible and permeable boundary around work, but an
inflexible and impermeable boundary around their
personal lives, or they might construct flexible and
permeable boundaries around both domains. The
combination of boundary management practices dic-
tates where they are placed on the segmentation–
integration continuum. We propose that, while
boundary management practices are enacted by indi-
viduals and are therefore individualistic in nature,
there are prototypical boundary management profiles
into which working adults might “cluster.” Given the
exploratory nature of this analysis, we do not offer
specific hypotheses about the types of clusters, or
profiles, which might emerge.
Method
Participants
Participants were employees of 24 different orga-
nizations in a small city in the Northeast United
States. Surveys were distributed to organizations by
investigators. In total, 593 surveys were distributed to
employees of these organizations; 399 surveys were
returned completed. The initial response rate was
67%. Five participants failed to complete 80% of the
total survey and were removed from the data set.
Another 48 were removed from the data for either
failing to report work hours or, given the theoretical
focus of the study, for not working at their current
organization for at least three months. An additional
19 participants were removed due to missing data for
the core constructs. The final sample consisted of 332
participants.
More than half of the respondents were women
(N⫽59%). Respondents were predominantly Cau-
casian (84%) with small numbers of Asian (2%),
Black (3%), and Hispanic (2%) respondents. The
remaining respondents did not indicate their ethnic/
racial background. Participants ranged in age from 20
to 84 years, with a mean age of 44.27 years (SD ⫽
11.5). More than half of the respondents held a bach-
elor’s degree or higher (61.1%). Most respondents
were married or currently partnered (70.9%) and
slightly more than half (51.4%) reported having at
least one child living at home; more than three quar-
ters (80.1%) of those respondents with at least one
child living at home reported living with a child
under 18 years old. More than half of the respondents
(54%) indicated being involved in a nonwork orga-
nization (e.g., church).
Mean job tenure for participants was 6.7 years
(SD ⫽6.8), and mean organization tenure was 9.0
years (SD ⫽8.4). Participants reported jobs in sales,
programming, marketing, engineering, analysis, and
reporting, among others.
Procedure
The investigators partnered with the Chamber of
Commerce of the city in which the study was con-
ducted to recruit organizations for a larger survey
368 BULGER, MATTHEWS, AND HOFFMAN
study of the impact of the use of Internet communi-
cations on work. We contacted representatives of
member organizations from a list furnished by the
Chamber to arrange permission to conduct the re-
search and to agree upon a day and time to drop
surveys off. Representatives were asked to distribute
surveys to any employee who used a computer while
at work and to avoid focusing on any particular type
of job. Accompanying each survey was a scratch-off
lottery ticket as an incentive to participate in the
survey. All participants were ensured that neither
they nor their employing organization would be iden-
tified and only the researchers would have access to
the data. Participants were asked to seal completed
surveys in a provided envelope, to be collected by the
investigators.
Study Measures
Domain boundary strength. We used the 28-item
instrument developed by Matthews and Barnes-
Farrell (2004) to assess the strength of the work
domain boundary (14 items) and the personal life
domain boundary (14 items). Responses are made on
a seven-point Likert scale (1 ⫽strongly disagree;
7⫽strongly agree). Four items measure flexibility–
willingness of the work domain (e.g., “I am not
willing to take time off from work to deal with my
family and personal life responsibilities.”), five items
measure flexibility–ability of the work domain (e.g.,
“If the need arose, I could leave work early to attend
to family related issues.”), and five items, originally
reported by Clark (2002), measure permeability of
the work domain (e.g., “My family contacts me while
I am at work.”). Items related to the personal life
domain boundary follow the same pattern. Previous
research demonstrated the reliability and construct
validity of this measure (Matthews & Barnes-Farrell,
2004). For this study, all subscales demonstrate ad-
equate reliability. Internal consistency reliability co-
efficients are shown in the diagonal of Table 1.
Work/personal life balance. To measure the out-
comes associated with the management of the work–
personal life boundaries, we used the 18-item scale
developed by Fisher (2001). This scale was devel-
oped to capture the bidirectional interaction between
work and personal life in terms of both the interfer-
ence between those domains and the enhancement
that can occur as a result of active participation in
multiple domains. Subscales measure work interfer-
ence with personal life (e.g., “I come home from
work too tired to do things I would like to do”),
personal life interference with work (e.g., “My per-
sonal life drains me of the energy I need to do my
job”), work enhancement of personal life (e.g., “Be-
cause of my job, I am in a better mood at home”), and
personal life enhancement of work (e.g., “My per-
sonal life helps me relax and feel ready for the next
day’s work”). Responses are made on a five-point
Likert scale (1 ⫽not at all; 5 ⫽almost all the time).
Internal consistency reliability coefficients for this
study are reported in the diagonal of Table 1; all
subscales demonstrate adequate reliability.
Results
Table 1 reports means, standard deviations, reli-
abilities, and intercorrelations of the domain bound-
ary measures and the work/personal life balance mea-
sures. Because the measures we used are relatively
new to the literature, we also report skewness and
kurtosis values along with the associated standard
errors for each of the variables. With the exception of
personal life flexibility-willingness, each of the vari-
ables demonstrated significant skewness or kurtosis,
or both.
1
Because Pearson correlation is inappropri-
ate in skewed distributions, we report Spearman cor-
relation coefficients. Intercorrelations between the
domain boundary measures indicates that boundary
measures pertaining to the same domain (i.e., work or
personal life) demonstrate modest to strong intercor-
relations, but not so strong as to indicate multicol-
linearity.
Boundary Strength and Work/Personal
Life Balance
The correlations in Table 1 suggest partial support
for each of our hypotheses. For instance, in partial
support of Hypothesis 1, work flexibility–ability and
work flexibility–willingness are negatively correlated
with work interference with personal life (⫽
⫺0.23, p⬍.01; ⫽⫺0.14, p⬍.05, respectively),
but work permeability was not. Similar partial sup-
port is demonstrated for the other hypotheses
To better explore the relationships between bound-
ary strength and work/personal life balance within
and across domains, we conducted a series of hier-
archical regressions. In each regression equation one
1
Significance of skewness and kurtosis is often assessed
by comparing the skewness or kurtosis value with its stan-
dard error. If the ratio exceeds 2, then significant deviation
from normal is indicated.
369WORK AND PERSONAL LIFE BOUNDARY MANAGEMENT
of the four work/personal life balance measures were
regressed onto all six domain boundary measures.
Boundary measures were entered in two blocks: one
block of work domain measures and one block of
personal life domain measures. We used hierarchical
regression to allow for examination of the individual
effects of each of the domain boundary strength
measures on the outcome of interest while controlling
for the other five measures. In addition, by entering
the constructs as separate blocks, we facilitate theory
development on the relative contribution of work and
personal life boundaries to work/personal life bal-
ance.
Regression results are reported in Table 2. It
should be noted that when study variables do not
meet the assumption of normality for a given analy-
sis, data transformations should be considered
(Tabachnik & Fidell, 1989). However, where inter-
pretation of the results might be made difficult, trans-
formation is not always recommended (Tabachnik &
Fidell, 1989). Because some of our variables exhib-
ited significant skew, we considered data transforma-
tion and ran our regression analyses separately with
transformed and nontransformed variables. The pat-
tern of results was identical: significant R
2
estimates
were observed for all four outcome variables and the
same variables were significant predictors. Given
those findings, the exploratory nature of our investi-
gation, as well as the caveat about interpretation
noted above, we report regression results using non-
transformed variables.
2
Significant regression coefficients were observed
for each outcome variable. Total variance explained
for each of the outcome variables ranged between 5%
and 12% as observed by R
2
estimates. Across the
outcome variables, the work domain block explained
either 4% or 5% of the variance; the personal life
domain block explained between 1% and 9% of the
variance. This suggests that neither work nor per-
sonal life boundary strength is particularly dominant
in explaining these aspects of work/personal life bal-
ance.
Results suggest that different aspects of the work
and personal life domain boundaries are related to
different outcomes. Significant predictors for work
interference with personal life were work flexibility-
ability (⫽⫺0.22, p⬍.01) and personal life per-
meability (⫽0.28, p⬍.01). Higher work interfer-
ence with personal life was associated with higher
personal life domain permeability and with lower
2
A regression table containing results using transformed
variables can be obtained by contacting Carrie A. Bulger.
Table 1
Descriptive Statistics, Spearman Correlations, and Reliability Coefficients for Study Variables
12345678910
1. Work flexibility–ability (0.86)
2. Work flexibility–willingness 0.52
*
(0.73)
3. Work permeability 0.34
*
0.43
**
(0.81)
4. Personal life flexibility–ability 0.28 0.18
**
0.07 (0.82)
5. Personal life
flexibility–willingness 0.17
**
0.02 0.07 0.35
**
(0.76)
6. Personal life permeability 0.17
**
0.07 0.32
**
0.24
**
0.35
**
(0.94)
7. Work interference with
personal life ⫺0.23
**
⫺0.14
**
0.02 ⫺0.13
*
0.00 0.22
**
(0.86)
8. Personal life interference with
work ⫺0.08 ⫺0.09 0.12
*
⫺0.28
**
⫺0.09 0.03 0.42
**
(0.85)
9. Work enhancement of personal
life 0.24
**
0.04 0.07 0.09 0.21
**
0.11
*
⫺0.28
**
⫺0.03 (0.76)
10. Personal life enhancement of
work 0.13
*
0.21
**
0.17
**
0.12
*
0.05 ⫺0.01 ⫺0.21
**
⫺0.19
**
0.43
**
(0.91)
Mean
a
5.62 5.65 4.78 5.25 4.29 4.35 2.75 1.72 2.76 3.52
SD 1.31 1.19 1.41 1.39 1.40 1.93 0.88 0.59 0.87 0.92
Skewness
b
⫺1.49 ⫺1.30 ⫺0.46 ⫺0.81 ⫺0.10 ⫺0.29 0.57 0.94 ⫺1.60 ⫺0.48
Kurtosis
c
2.40 2.21 ⫺0.27 0.33 ⫺0.31 1.05 0.08 0.98 0.17 0.26
Note. Internal consistency coefficients are reported along the diagonal.
a
Means reported are uncentered.
b
SE ⫽0.13.
c
SE ⫽0.27.
*
p⬍.05.
**
p⬍.01.
370 BULGER, MATTHEWS, AND HOFFMAN
work flexibility–ability. Significant predictors of per-
sonal life interference with work included work flexi-
bility–willingness (⫽⫺0.16, p⬍.05), work per-
meability (⫽0.18, p⬍.01), and personal life
flexibility–ability (⫽⫺0.26, p⬍.01). Higher
personal life interference with work was associated
with lower willingness to flex the work boundary,
lower ability to flex the personal life boundary, and
higher work permeability.
Significant predictors of work enhancement of per-
sonal life were work flexibility–ability (⫽0.21,
p⬍.01) and personal life flexibility–willingness
(⫽0.21, p⬍.01). Higher work enhancement of
personal life was associated with higher ability to flex
the work boundary and higher willingness to flex the
personal life boundary. Work flexibility–willingness
was the only significant predictor of personal life
enhancement of work (⫽0.16, p⬍.01). High
personal life enhancement of work was related to
higher willingness to flex the work boundary.
Investigating the
Segmentation–Integration Continuum
Because the number and the nature of possible
boundary management profiles is unknown, tradi-
tional clustering algorithms (i.e., k-means clustering)
where the user must specify the number of clusters
(k) were not appropriate here. For this reason we used
TwoStep clustering analysis within SPSS 12. As im-
plied by its name, TwoStep clustering involves two
steps. In step 1, cases are grouped into subclusters
based on a distance criterion (e.g., Schwarz’s Bayes-
ian Criterion). In step 2, these subclusters are auto-
matically clustered. This second step does not require
specification of the number of clusters when there is
no a priori rationale for the optimal number of clus-
ters. Interested readers may find more information on
TwoStep clustering in Zhang, Ramakrishnon, and
Livny (1996) and Chiu, Fang, Chen, Wang, and Jeris
(2001). To facilitate interpretation of results, the six
boundary measures were centered on the scale mid-
point (i.e., participant’s score ⫺4) prior to entry into
the cluster analysis. Positive scores reflect greater
flexibility and permeability; negative scores indicate
lower flexibility and permeability.
The TwoStep clustering analysis resulted in a four-
cluster solution. The centered variable means for the
four clusters are reported in Table 3 along with the
results of a series of one-way analyses of variance
(ANOVAs). While cluster analysis generates clusters
that are maximally distant from one another,
ANOVA indicates the specific dimensions on which
the clusters differ.
Individuals in cluster 1 (n⫽107) tend to exhibit
higher mean ability and willingness to flex both do-
main boundaries, as well as higher permeability of
both domain boundaries. These individuals would
Table 2
Hierarchical Regression Analyses for Work-Personal Life Interface Outcomes
Dependent variables
Work
interference
with
personal life
Personal
life
interference
with work
Work
enhancement
of personal
life
Personal life
enhancement
of work
Work domain block
Work flexibility–ability ⫺0.22
**
0.02 0.21
**
⫺0.02
Work flexibility–willingness 0.00 ⫺0.16
*
⫺0.08 0.16
*
Work permeability 0.03 0.18
**
⫺0.01 0.08
Personal life domain block
Personal life flexibility–ability ⫺0.08 ⫺0.26
**
⫺0.01 0.08
Personal life flexibility–willingness ⫺0.05 ⫺0.03 0.21
**
0.04
Personal life permeability 0.28
**
0.05 0.03 ⫺0.06
Work domain block R
2
0.05 0.05 0.04 0.04
Personal life domain block R
2
0.07 0.09 0.06 0.01
Total R
2
0.11 0.12 0.09 0.05
Adjusted R
2
0.10 0.10 0.07 0.03
Total F6.85
**
7.03
**
5.16
**
2.95
**
df 325 325 325 325
Note. Values in table are standardized beta weights for each predictor, unless otherwise indicated.
*
p⬍.05.
**
p⬍.01.
371WORK AND PERSONAL LIFE BOUNDARY MANAGEMENT
fall on the high integration end of the segmentation-
integration continuum. Individuals in cluster 2 (n⫽
128) tend toward higher mean scores on ability and
willingness to flex their work domain boundary, but
lower scores on work permeability. Individuals in
cluster 2 also exhibit higher ability to flex the per-
sonal life domain boundary, but much lower willing-
ness and permeability. These individuals would be
difficult to place on the continuum: they somewhat
integrate the work domain and somewhat segment
the personal life domain.
Cluster 3 (n⫽32) consists of individuals who
score near the midpoint on most of the boundary
strength measures. This cluster might be best placed
on the midpoint of the segmentation-integration con-
tinuum but for the work flexibility-ability dimension,
where these participants exhibit lower scores. The
individuals in cluster 4 (n⫽65) report both ability
and willingness to flex the work domain boundary,
but not the personal life domain boundary. Although
these individuals should not be placed on the extreme
segmentation end of the segmentation–integration
continuum, they fall closer to segmentation than any
of the other three clusters.
Describing Cluster Members
To better understand each cluster, we investigated
cluster membership in terms of demographic vari-
ables, as shown in Table 4. Cluster 1 (integrators) is
slightly less than half female while cluster 4 (work
integrators/personal life segmenters) is nearly three-
quarters female. About half of the cluster 1 members
indicated having children under 18 at home, whereas
about two thirds of the cluster 4 members indicated
the same. Cluster 1 members average about seven
Table 4
Cluster Member Demographics
Cluster 1
(N⫽107)
Cluster 2
(N⫽128)
Cluster 3
(N⫽32)
Cluster 4
(N⫽65)
Percentage of women 48.6 60.2 62.5 73.8
Percentage married/partnered 84.1 67.5 53.1 75.4
Percentage BA/BS or higher 72.6 56.9 60.0 64.1
Percentage with children under 18 49.5 28.1 53.1 63.1
Percentage involved nonwork
organizations 62% 49.6 54.8 50.8
Mean age 45.74 (10.76) 45.54 (12.4) 38.67 (9.67) 41.22 (10.84)
Average work hours 42.78 (7.3) 41.28 (5.8) 41.84 (6.71) 39.38 (10.05)
Average job tenure 7.26 (5.96) 4.0 (7.65) 5.99 (5.76) 4.85 (6.23)
Average organizational tenure 11.14 (8.86) 9.05 (8.65) 8.48 (6.56) 7.59 (7.56)
Note. Where means are reported, they are uncentered. Standard deviations are reported in parentheses where appropriate.
Table 3
Cluster Membership Means, Standard Deviation, and Analysis of Variance Results
Cluster 1 Cluster 2 Cluster 3 Cluster 4
F
a
Mean SD Mean SD Mean SD Mean SD
Work flex–ability 2.35
b
0.63 1.69
a
0.82 ⫺1.39
c
1.13 1.74
a
0.87 176.66
**
Work flex–willing 2.17
a
0.74 1.40
b
0.97 ⫺0.22
c
1.72 2.19
a
0.67 61.44
**
Work permeability 1.82
a
0.82 ⫺0.18
b
1.16 ⫺0.48
b
1.21 1.44
a
1.07 90.51
**
Personal life flex–ability 2.09
b
0.71 1.62
c
0.99 0.21
a
1.64 ⫺0.36
a
1.16 93.64
**
Personal life flex–willing 1.09
b
1.19 0.21
a
1.06 0.32
a
1.37 ⫺0.88
c
1.44 35.83
**
Personal life permeability 1.86
b
1.24 ⫺0.57
a
1.79 ⫺0.26
a
1.63 ⫺0.04
a
1.79 47.52
**
Note. All means reported are centered based on the scale midpoint. Across rows, means that share subscripts do not differ
significantly from one another; means with different subscripts are significantly different. Results for mean differences
indicated by Scheffe test of mean differences.
a
df are 3, 328 for all analyses.
**
p⬍.01.
372 BULGER, MATTHEWS, AND HOFFMAN
years at their jobs, whereas cluster 4 members aver-
age slightly less than five years at their jobs. Cluster
1 members have the highest job and organizational
tenure of all four clusters. More cluster 1 members
have a college degree or higher, are involved in
nonwork organizations, and are married or partnered
than any of the other clusters. The most complex
group, cluster 2, has the lowest percentage of chil-
dren under 18 at home. Cluster 3, the relatively
neutral group, is the youngest group. This group also
has a smaller percentage of married/partnered mem-
bers.
Discussion
To our knowledge, this study was the first direct
investigation of the segmentation–integration contin-
uum proposed by various boundary theorists
(Ashforth et al., 2000; Clark, 2000; Nippert-Eng,
1996). Thus, one contribution of our study is the
identification of possible boundary management pro-
files. This study also contributes to our understanding
of the ways in which boundary management practices
are related to work/personal life balance.
Relationships Between Boundary Strength and
Work/Personal Life Balance
Our data suggested interesting relationships be-
tween boundary strength and work/personal life bal-
ance. Attending first to the two measures of work/
personal life interference, we showed that lower
ability to flex the work domain and more permeation
of personal life into work are predictive of experi-
encing work interference with personal life. In addi-
tion, lower ability and willingness to flex the personal
life boundary and more permeation of work into
personal life are related to more personal life inter-
ference with work. Although these regressions were
exploratory, it is interesting to note that for all four
outcomes, there were significant predictors both
within and across domains. For example, work inter-
ference with personal life was significantly predicted
by one work domain boundary measure and by one
personal life domain boundary measure. Further, it is
interesting to note that, speaking broadly, lower flex-
ibility and higher permeability seem to be related to
higher interference between work and personal life.
Flexibility-ability is likely to be related to constraints
placed on the individual by norms, rules, or others in
that domain. Clark (2000) refers to these others as
“border keepers.” It seems likely that preferences or
rules set by people in one domain would influence the
permeability-related intrusions that ultimately occur.
This suggests a need to further investigate the inter-
relationships between flexibility–ability, flexibility–
willingness, and permeability. Although we conceive
of these as three measures of boundary strength, it is
possible that the two flexibility measures predict per-
meability. Future research should further investigate
these relationships.
In terms of the two measures of work/personal life
enhancement, the results are somewhat different than
for interference. Individuals who are more able to
leave work to attend to their personal lives and who
are more willing to flex their personal life boundary,
report higher work enhancement of personal life. The
only significant predictor of personal life enhance-
ment of work was work flexibility–willingness. Thus,
it appears that willingness to flex the work domain
boundary is related to experiencing more personal
life enhancement of work. What is interesting about
the regression results for domain enhancement is that
a major predictor in each domain was flexibility–
willingness of that domain. Flexibility–willingness
suggests that the individual would leave one domain
to attend to the demands of the other and is therefore
likely to be related to individual preference for seg-
menting or integrating. Our data support a recent
assertion that flexibility in one work/family role is
one contributor to successful performance in the
other role and that role flexibility is therefore one
means by which work and nonwork enrich one an-
other (Greenhaus & Powell, 2006).
In the main, the results of our examination of the
relationships between boundary strength and work/
personal life balance suggest that the ways in which
individuals manage their boundaries have implica-
tions for their experiences of work/personal life bal-
ance. Consistent with past theoretical work, we show
that there are both positive and negative conse-
quences of managing multiple roles or domains (e.g.,
Marks & MacDermid, 1996). It would be interesting
to examine other outcomes. For instance, do individ-
uals who experience enhancement as a result of flex-
ible, permeable boundaries have additional positive
outcomes, such as lower job stress or family stress?
This question has implications for researchers in occu-
pational health psychology who seek ways to reduce
occupational stress and enhance worker well-being.
Boundary Management Profiles
The cluster analysis suggested four boundary man-
agement profiles. An examination of boundary man-
373WORK AND PERSONAL LIFE BOUNDARY MANAGEMENT
agement within each cluster provides some support
for a segmentation–integration continuum. However,
our data raise questions as to whether the continuum
may not be as straightforward as initially proposed
(Ashforth et al., 2000; Clark, 2000; Nippert-Eng,
1996). Cluster 1 was made up of individuals who
would be placed at the integration end of the contin-
uum because they scored relatively high on all
boundary strength measures. Cluster 2 was made up
of individuals who could and would leave work to
attend to personal life demands, but for whom per-
sonal life issues only sometimes permeate the work
domain. These cluster members also could leave their
personal life domain to attend to work demands, but
are less willing to do so. And, as with the work
domain, work issues only sometimes permeate the
personal life domain. One interpretation of this clus-
ter is that they could integrate rather than segment
domains, but do not have a strong preference or
reason to do so. This group might be placed toward
the integration end of the segmentation–integration
continuum, but they would be closer to neutral than
to the endpoint. Of note, this is the largest group (n⫽
128).
Individuals in cluster 3 showed relatively neutral
scores on all measures of boundary strength except
for negative work flexibility–ability. Given their low
work flexibility–ability, it might be said that they
tend toward segmentation for that particular dimen-
sion, but looking across all scores, it is most appro-
priate to place this cluster at or near the midpoint of
the segmentation-integration continuum. Of note, this
was the smallest group (n⫽32). Cluster 4 members
showed high work flexibility and permeability, but
low personal life flexibility and permeability. These
are people who are neither fully segmenters nor fully
integrators; instead, they flex their work domain but
shield their personal lives.
Because we do not have a clear group that would
be at the segmentation end of the continuum, our
results raise questions. It may be that boundary man-
agement is not necessarily a linear function of seg-
menting or integrating both domains simultaneously.
Perhaps it is the case that a segmentation–integration
continuum exists for each domain separately and that
boundary management is best construed in a two-
dimensional, rather than one-dimensional, space. It
would also be interesting to know the degree to
which each element of boundary strength contributes
to placement on the continuum. For instance, it is
possible that flexibility–ability is the gateway to in-
tegrating or segmenting boundaries for some, while
for others flexibility–willingness carries more
weight. Future research should expand this analysis
to find ways to further investigate the nature of seg-
mentation and integration.
The demographic information about our cluster
members raises more questions about boundary man-
agement than it provides answers. For instance, the
highest percentage of women was in the most seg-
menting cluster (cluster 4) and the lowest percentage
of women was in the most integrating cluster (cluster
1). Perhaps this implies that men and women make
use of different boundary management techniques.
But, the integrating group was also the group with the
highest average job and organization tenure. Perhaps
integration becomes more a possibility or more a
preference with more tenure. Cluster 2 members had
the fewest children under age 18 at home and also
had a complex pattern of boundary management
practices, which may suggest that having young chil-
dren at home impacts boundary management choices.
Future research should investigate individual differ-
ence predictors of boundary management styles.
Limitations
As in any study, ours has a few limitations. While
our sample size was more than adequate for the
nature of our exploration of boundary management
profiles and boundary strength, we were not able to
use a holdout sample to confirm the structure of the
profiles and the relationships our results suggested.
This may be a factor in the small cell sizes within
clusters. We intend, in future studies, to obtain a
larger, more nationally representative sample to bet-
ter investigate the segmentation–integration contin-
uum.
In addition, our respondents were recruited as part
of a larger study. Though the respondents hold many
different types of jobs at different types of organiza-
tions, they could all generally be called knowledge
workers. This might raise the question as to whether
occupation type influences boundary management
profiles. While we see this as a possibility, we also
suggest that occupation type might be an organiza-
tional constraint on ability to flex boundaries, and not
willingness to flex boundaries, which is beyond the
scope of this study.
Implications
Despite these limitations, the results of the current
study indicate that this is a promising avenue of
research. We are just beginning to understand the
practices of boundary management and how such
374 BULGER, MATTHEWS, AND HOFFMAN
practices impact the interface between work and per-
sonal life. Because work and personal life balance
continues to be an important issue for employees,
who are attempting to balance, and for employers and
researchers, who are attempting to help, it is imper-
ative that we deepen our understanding of the bound-
aries workers set up for their competing domains.
The implications for future research are clear: more
work is needed testing the propositions of boundary
theory, boundary strength, and boundary manage-
ment.
References
Allen, T. D., Herst, D. E. L., Bruck, C. S., & Sutton, M.
(2000). Consequences associated with work-to-family
conflict: A review and agenda for future research. Jour-
nal of Occupational Health Psychology, 5, 278–308.
Ashforth, B. E., Kreiner, G. E., & Fugate, M. (2000). All in
a day’s work: Boundaries and micro role transitions.
Academy of Management Review, 25, 472–491.
Barnett, R. C., & Baruch, G. K. (1985). Women’s involve-
ment in multiple roles and psychological distress. Jour-
nal of Personality and Social Psychology, 49, 135–145.
Chesley, N. (2005). Blurring boundaries? Linking technol-
ogy use, spillover, individual distress, and family satis-
faction. Journal of Marriage and Family, 67, 1237–1248.
Chiu, T., Fang, D., Chen, J., Wang, Y., & Jeris, C. (August,
2001). A robust and scalable clustering algorithm for
mixed type attributes in large database environment. Pro-
ceedings of the Seventh Association for Computing
Machinery Special Interest Group Knowledge Discovery
and Data Mining International Conference on Knowl-
edge Discovery and Data Mining, 263–268.
Clark, S. C. (2000). Work/family border theory: A new
theory of work/family balance. Human Relations, 53,
747–770.
Clark, S. C. (2002, August). Border between work and
home, and work/family conflict. Paper presented at the
annual meeting of the Academy of Management, Denver,
CO.
Desrochers, S., Hilton, J. M., & Larwood, L. (2005). Pre-
liminary validation of the Work-Family Integration-
Blurring Scale. Journal of Family Issues, 26, 442–466.
Fisher, G. G. (2001). Work-life balance: A construct devel-
opment study. Unpublished doctoral dissertation. Bowl-
ing Green State University, Bowling, OH.
Frone, M. R. (2003). Work-family balance. In J. C. Quick &
L. E. Tetrick (Eds.), Handbook of occupational health
psychology (pp. 143–162). Washington, DC: American
Psychological Association.
Frone, M. R., Russell, M., & Cooper, M. L. (1992). Ante-
cedents and outcomes of work-family conflict: Testing a
model of the work-family interface. Journal of Applied
Psychology, 77, 65–78.
Greenhaus, J. G., & Powell, G. N. (2006). When work and
family are allies: A theory of work-family enrichment.
Academy of Management Review, 31, 72–92.
Grzywacz, J. G., & Marks, N. F. (2000). Reconceptualizing
the work-family interface: An ecological perspective on
the correlates of positive and negative spillover between
work and family. Journal of Occupational Health Psy-
chology, 5, 111–126.
Hall, D. T., & Richter, J. (1988). Balancing work life and
home life: What can organizations do to help? Academy
of Management Executive, 3, 213–223.
Hammer, L. B., Bauer, T. N., & Grandey, A. A. (2003).
Work-family conflict and work-related withdrawal be-
haviors. Journal of Business and Psychology, 17, 419–
436.
Kirchmeyer, C. (1995). Managing the work-nonwork
boundary: An assessment of organizational responses.
Human Relations, 48, 515–536.
Kossek, E. E., & Ozeki, C. (1998). Work-family conflict,
policies, and the job-life satisfaction relationship: A re-
view and directions for organizational behavior-human
resources research. Journal of Applied Psychology, 83,
139–149.
Marks, S. R. (1977). Multiple roles and role strain: Some
notes on human energy, time, and commitment. Ameri-
can Sociological Review, 42, 921–936.
Marks, S. R., & MacDermid, S. M. (1996). Multiple roles
and the self: A theory of role balance. Journal of Mar-
riage and the Family, 58, 417–432.
Matthews, R. A., & Barnes-Farrell, J. L. (2004, August).
Development of a comprehensive measure of boundary
strength for work and family domains. Paper presented at
the annual meeting of the Academy of Management,
New Orleans, LA.
Nippert-Eng, C. E. (1996). Home and work. Chicago, IL:
The University of Chicago Press.
Parasuraman, S., & Greenhaus, J. H. (2002). Toward reduc-
ing some critical gaps in work-family research. Human
Resource Management Review, 12, 299–312.
Rothbard, N. P., Phillips, K. W., & Dumas, T. L. (2005).
Managing multiple roles: Work-family policies and in-
dividuals’ desire for segmentation. Organization Science,
16, 243–258.
Sauter, S. H., Murphy, L. R., & Hurrell, J. J. (1990).
Proposed national strategy for the prevention of psycho-
logical disorders in the workplace. American Psycholo-
gist, 45, 1146–1158.
Sieber, S. D. (1974). Toward a theory of role accumulation.
American Sociological Review, 39, 567–578.
Tabachnik, B. G., & Fidell, L. S. (1989). Using multivariate
statistics (2nd ed.). New York: HarperCollins Publishers.
Zhang, T., Ramakrishnon, R., & Livny, M. (June, 1996).
BIRCH: An efficient data clustering method for very
large databases. Proceedings of the Association for Com-
puting Machinery Special Interest Group: Management
of Data Conference on Management of Data, Montreal,
Quebec, 103–114.
Received June 5, 2006
Revision received January 18, 2007
Accepted February 9, 2007 y
375WORK AND PERSONAL LIFE BOUNDARY MANAGEMENT
