Is telework effective for
A meta-analysis of empirical research on
perceptions of telework and organizational
Brittany Harker Martin and Rhiannon MacDonnell
Haskayne School of Business, University of Calgary, Calgary, Canada
Purpose – Telework is an alternative work relationship with demonstrated positive beneﬁts for
individuals and society, yet it has not been implemented with enthusiasm by most organizations. This
could be due to the lacking, consolidated evidence for management regarding whether or not telework
is a good thing for the ﬁrm. The purpose of this paper is to integrate multidisciplinary literature that
reports effects of telework on organizational outcomes with the aim of providing a clearer answer to
the question: is telework effective for organizations?
Design/methodology/approach – Meta-analytical methods were used, beginning with an
interdisciplinary search for effect sizes in eight databases. Limited to scholarly journals and
dissertations, results included 991 articles scanned for inclusion criteria. The independent variable is
telework, measured as a dichotomous variable. Dependent variables are outcomes of interest to
organizations: productivity, retention, turnover intention, commitment, and performance. In total,
22 studies were double coded and meta-analyzed using Hunter and Schmidt’s approach, followed by ﬁve
exploratory moderator analyses: level of analysis, level of the employee, response rate, proportion of
females, and country of the study. Signiﬁcant results are discussed.
Findings – Review and metaanalysis of 32 correlations from empirical studies ﬁnd that there is a small
but positiverelationship betweentelework and organizational outcomes.Telework is perceivedto increase
productivity, secure retention, strengthen organizational commitment, and to improve performance within
the organization. In other words, it is indeed beneﬁcial for organizations. All ﬁve hypotheses are
supported. H1 (productivity), rc¼0.23 (k ¼5, n¼620), (95% CI¼0.13 20.33). H2 (retention), r¼0.10
(k¼6, n¼1652), (95% CI¼0.04 20.16). H3(commitment),r¼0.11 (k¼8, n¼3144), (95% CI¼0.03 20.18);
moderator analysis shows sampleage is signiﬁcant (F(1,4)¼4.715, p,0.05, R2¼0.80). H4 (performance),
r¼0.16 (k¼10, n¼2522). H5 (organizational outcomes), r¼0.17 (k¼19, n¼5502), (95% CI¼0.1 20.20).
Originality/value – To the authors’ knowledge, this is the ﬁrst meta-analysis of telework research at
the organizational level, providing a unique contribution to the ﬁeld in ﬁlling the gap between research
on effects to the individual and society. Additional contributions resulted from the moderator analyses:
ﬁrst, in ﬁnding that the relationship between telework and performance is moderated by whether or
not the sample was one individual per ﬁrm, or many individuals from one; and second, in ﬁnding that
the relationship between telework and organizational commitment is moderated by age. Thus, the
paper provides unique contributions with both scholarly and practical implications.
Keywords Teleworking, Organizational structures, Telework beneﬁts, Organizational outcomes,
Paper type Research paper
It used to be a scene from science ﬁction: humans working at interactive machines that
calculated, documented, and sent information in instant transmissions with the press of
a button. However, since the advent of the personal computer and the emergence
The current issue and full text archive of this journal is available at
Management Research Review
Vol. 35 No. 7, 2012
qEmerald Group Publishing Limited
of the internet, this scene has not only become one of the present, but is itself fading into
memory as technology advances at a frenzied pace. From hand-held devices that receive
e-mail, to smart phones that link photos to social media web sites, information and
communication technologies (ICT) have integrated into mainstream society. Indeed, one
might expect that with these powerful tools, much of today’s information-based work
would take place outside a central ofﬁce, saving millions from the daily commute. Add to
that, the power of ICT to enable business continuity in times of pandemics, acts of
terrorism, or natural disasters, and one might also expect that organizations would be
adopting ICT-based off-site programs with rapid enthusiasm. This, however, is not
Despite several decades of research on the beneﬁts of telework (Hill et al., 1996; Nilles,
1994; Olson, 1988; Verbeke et al., 2008), it has not been adopted on the scale anticipated
by scholars (Berger, 1996; Salopek, 1998; Tofﬂer, 1981). In fact, data from academic
research, census reports and think-tank reports alike all show the average frequency of
telework by organizational employees to be somewhere between 2.2 and 12 percent
(Akyeampong and Nadwodny, 2001; Akyeampong, 2007; Duxbury and Higgins, 2002;
Jones, 2005; US Census Bureau, 2004; Wallace, 2003; Welz and Wolf, 2010). Central to a
possible explanation of why this might be is the fact that much of the empirical research
on telework has been conducted at the individual level of employee outcomes, whereas
the decision to implement a telework program ultimately lies at the top of the ﬁrm, and is
driven by organizational outcomes. Although beneﬁts of telework for employees
aggregate into beneﬁts for organizations (Verbeke et al., 2008), organizations may
require more empirical evidence that shows whether or not teleworking is more effective
than traditional work arrangements. Such empirical studies exist, but are scattered
across disciplines such as computer science, economics, education, information
technology, and sociology. Hence, the purpose of this meta-analysis is to integrate
multidisciplinary literature that reports effects of telework on organizational outcomes
with the aim of providing a clearer answer to the question: is telework effective for
Telework deﬁnition and research
For the purposes of this paper, the terms telework and telecommuting will be used
interchangeably as is accepted practice in this area of research. Drawing from a recent
literature review (Bailey and Kurland, 2002), a recent meta-analysis (Gajendran and
Harrison, 2007), and other literature from the ﬁeld, telework is deﬁned here as the
substitution of communication technology for work-related travel, and can include paid
work from home, a satellite ofﬁce, a telework centre or any other work station outside of
the main ofﬁce for at least one day per work week (Verbeke et al., 2008).
Jack Nilles was the ﬁrst scholar to coin the term “telecommuting” in 1975. Nearly
30 years later, it is deﬁned by Bailey and Kurland (2002, p. 384) as, “working outside the
conventional workplace and communicating with it by way of telecommunications or
computer-based technology,”. In their 2002 review of 80 telework studies, they looked
beyond the advantages and disadvantages of telework to seek the answers to three
questions: who participates in telework, why, and what happens when they do? Limited
by a claim that much of the literature focuses on the individual level, they state that
work-related factors are the most predictive of an individual’s choice whether
or not to telework. In answer to the question of why individuals telework, they turned to
studies of managers, and found that concerns about cost and control, as well as little
perceived need, led to managerial reluctance to create telework programs. These authors
further suggest that research move beyond studying teleworkers as a class of employees
and study telework as a practice with a broad range of outcomes. In their review of
managerial perceptions, they found that managers see no need for the change to
telework. A strong enough business case has not been provided, and ambiguity as to
whether or not the organizational beneﬁts outweigh the costs leaves managers in a
At another unit of analysis, Gajendran and Harrison’s (2007) meta-analyzed
46 studies on psychological mediators and individual consequences of telework to ﬁnd
that telework is beneﬁcial for individuals. In line with the deﬁnition used here, they
deﬁne telecommuting as:
[...] an alternative work arrangement in which employees perform tasks elsewhere that are
normally done in a primary or central workplace, for at least some portion of their work
schedule, using electronic media to interact with others inside and outside the organization
(2007, p. 1525).
There is a broad scope of research interests in the telework literature. It ranges from the
inﬂuence of gender on telework outcomes (Hill et al., 1998; Hyland et al., 2005) to the
economic impact of telework on society (Mokhtarian et al., 1998; Salomon et al., 1991).
Union issues are studied (Di Martino and Wirth, 1990), as well as the attitudes that are
either in favor of, or against, telework (Yap and Tng, 1990; Duxbury and Haines, 1991).
Work/life balance is another major branch of the research (Duxbury et al., 1992; Madsen,
2003), as is a managerial perspective regarding supervision and control (Illegems and
Verbeke, 2004; Valsecchi, 2006). Finally, there is a stream of research dedicated to
telework’s potential to resolve issues of equity and employability for disabled or
home-bound individuals due to physical or care-giving reasons (Matthes, 1992;
Verbeke et al., 2008). In summary, there is a wealth of information within each level of
analysis that can potentially be applied at the others. Indeed, there is often overlap
between outcomes at each level, such as turnover intentions of employees and retention
within the organization.
In other research, telework adoption has been predicted to increase at the
organizational level if beneﬁcial organizational outcomes are perceived to outweigh the
costs and consequences (Illegems and Verbeke, 2003; Verbeke et al., 2008). In this view,
managerial perception is key to whether or not telework will be adopted, and doing so
requires evidence that telework is more effective for the organization than the alternative.
Thus, this meta-analysis meets a critical need in organization studies, by providing
statistical rationale for decision makers that includes the managerial perceptions of
telework and organizational outcomes.
Among the top outcomes of interest to organizations are productivity, retention,
organizational commitment, and performance (Bailey and Kurland, 2002; Byrd, 2005;
Mokhtarian and Sato, 1994; Verbeke et al., 2008). As such, these are the organizational
outcomes examined in this meta-analysis, as summarized in the theoretical framework
of Figure 1 (Gajendran and Harrison, 2007). In this ﬁgure, the arrows indicate
the relationship measured in the included studies, between telework and each
organizational outcome of interest. In addition, these outcomes are enclosed within a box
to indicate an overall predicted relationship between telework and organizational
outcomes in general.
Organizational outcomes of the framework
Productivity is regularly reported as a perceived beneﬁt of telework (Callentine, 1995;
Pitt-Catsouphes and Marchetta, 1991; Hill et al., 1998). Reasons cited include working at
peak efﬁciency hours, reducing distractions and interruptions, being in an environment
conducive to increased concentration, and reducing incidental absence (Belanger, 1999;
Baruch, 2000). Productivity is often measured in terms of respondents’ perceived actual
or potential increase or decrease in work output associated with telework adoption, often
in comparison to a non-adoption state. Naturally, increased productivity is of high
organizational interest. Hence, the ﬁrst hypothesis is derived as follows:
H1. Telework will be positively associated with perceptions of increased
Retention is looked at in the telework literature by measuring perceptions of retention
and turnover intentions, from both the employee and managerial perspectives. Both
constructs are commonly reported as beneﬁts of telework, where retention is predicted to
increase in teleworkers, and turnover intentions to decrease. For the purposes of this
meta-analysis, these constructs are collapsed into the retention variable through reverse
coding of turnover intentions with the rationale that, if employees do not intend to leave
the organization they will be better retained by the organization. Thus, the second
hypothesis of the framework is derived:
H2. Telework will be positively associated with perceptions of employee
Organizational commitment in the telework literature is a multidimensional construct. In
some studies, it is broken down into speciﬁc types of commitment, such as the
three-dimensional model of Meyer and Allen (1991) that measures affective commitment,
normative commitment, and continuance commitment (Desrosiers, 2001; Piper, 2004).
In other studies, only one of these types of commitment is measured, or a general
Theoretical framework of
telework relationship with
commitment variable is deﬁned that is something of a hybrid from the ﬁeld (Belanger,
1999; Lee, 2004). Despite historical predictions that teleworkers may be less committed to
the organization, recent studies have reported organizational commitment as one of the
beneﬁts of telework whereby commitment is exchanged in reciprocity for a more ﬂexible
work arrangement than a typical ofﬁce job. As such, the third hypothesis of the framework
H3. Telework will be positively associated with perceptions of organizational
The fourth organizational outcome, performance, is also of high importance to
organizations. Related to productivity, performance is the assessment of the work being
done within the ﬁrm. As such, it is not output oriented like productivity, and is measured
differently. “Typically it canbe measured by quality of outputs, job knowledge, leadership,
judgment, innovation, goal setting or teamwork, among others” (Belanger, 1999, p. 143).
Thus, performance can be viewed as the perception of how wellemployees and the ﬁrm are
doing. In the telework literature, it is commonly measured in comparison between
teleworkers and non-teleworkers, with the predominant stance being that performance is
higher amongst the former. Since increased performance of the employees translates into
increased performance of the ﬁrm, the fourth hypothesis is as follows:
H4. Telework will be positively associated with perceptions of performance.
Thus, far, all four organizational outcomes of the fr amework have been predicted to have
a positive association with telework. Indeed, they are all regularly reported as beneﬁts of
a telework program. If this is true, then it would seem that a telework program would be
beneﬁcial, not only at the level of the individual, but also at the level of the organization.
As such, the ﬁfth hypothesis captures the purpose of this meta-analysis, and is stated
H5. Overall, there will be a positive relationship between telework and the
organizational outcomes of this framework.
Collecting effect sizes
As mentioned, telework research on organizational outcomes has taken place across a
number of disciplines. As such, the literature search was interdisciplinary conducted in:
ABI/Inform, Applied Science Index, Business Source Complete EBSCO,
GoogleScholar.com, JSTOR, ProQuest, PsycINFO, and Wilson Web. Keywords for the
search included: telework, telecommuting, mobile-work, virtual workplace and
organization. Results were limited to scholarly journals and dissertations. Despite the
large number of results, it was discovered that there is actually little empirical research
on telework to be meta-analyzed (Hill et al., 1996; Workman, 2000). As such, reference lists
of empirical studies on the related variablesand level of analysis were searched through a
backward approach. Through the entire search process, 991 articles were located.
Inclusion, exclusion and selection
Papers retained for this meta-analysis aligned with the following inclusion criteria. First
and foremost, included papers needed to include an effect size for telework or report
enough data to compute one. Next, papers needed to study one of the hypothesized
relationships of the theoretical framework for this analysis. Third, papers needed to
measure the variables of the theoretical framework based on teleworker or managerial
perceptions. Finally, papers needed to be from 1991 or later, based on the drastic change
in telework through the introduction of the internet and the world wide web.
Article selection and inclusion involved a three-part process:
(1) scanning abstracts of papers that resulted from the literature search for
(2) scanning the selected papers that showed potential to determine if inclusion
criteria was met; then
(3) reading and coding the included studies.
In total 68 studies were pulled from the database through reading abstracts for inclusion
criteria. From that, three studies were eliminated because they were experimental, and
65 studies were read for coding. Studies were further eliminated from the meta-analysis
for the following reasons: although the variables of interest appeared to be part of the
study, telework was a dependent, rather than independent, variable (e.g. adoption of
telework was being predicted (9); although the study appeared to include the right
variables, the wrong relationship was being measured (10); although the study met all
inclusion criteria, no empirical data was reported (8); although the study measured the
right variables, the study was not based on perceptions (4); although the study measured
the relationships of interest, not enough statistical information was provided to calculate
effect size (12).
Data from 22 studies met all of the inclusion criteria. These studies included
15 published scholarly journal articles and seven doctoral dissertations. Three studies
were discovered to be based on data sets already coded within the analysis, and were
therefore eliminated, leading to a total of 19 studies (highlighted by an asterisk in the
reference list) with 32 effect sizes related to the hypotheses of this meta-analysis. Each of
the studies used survey (questionnaire or interview) techniques to measure the perceived
outcomes of telework. Where reported, the average response rate was 49.7 percent, the
mean age of the sample was 38.94, and the average proportion of women was
It should be noted that, although some studies indicated their unit of analysis to be
the ﬁrm, all studies were based on surveys that captured managerial or employee
perceptions as opposed to objective ﬁrm-level outcomes. In some cases the managers
served as proxies for the ﬁrm and the number of employees was equal to the number of
ﬁrms (Martinez-Sanchez et al., 2007a, b, 2008); whereas in other cases managerial
perceptions still represented the organization, but the number of employees outnumbered
the number of ﬁrms (Illegems and Verbeke, 2004). Managerial or employee perceptions
were either reported separately or aggregated together, depending on study-speciﬁc
hypotheses (Hyland et al., 2005). For the purposes of this meta-analysis, the average
sample size is reported in number of employees, 290.
Studies were double coded for source type (e.g. peer-reviewed journal, dissertation,
unpublished), sample size, proportion of women, average age of sample, rater type
(e.g. employee, manager), method (e.g. self-report vs interview), year of publication,
and authors. Both authors coded all studies independently, and any discrepancies were
reconciled through discussion. Caution was observed to ensure the correct direction of
the reported correlation coefﬁcient, such as where turnover intentions were reversed to
be included under the retention construct.
In each of the studies, telework was the independent variable, measured as a
dichotomous variable (yes/no). Studies also speciﬁed a level of intensity ranging from
once a week to full-time. Although there is variability in the intensity captured in the
studies, there is agreement between all studies that teleworking requires working from
an alternate location from the central ofﬁce at least one day a week. In line with
Gajendran and Harrison (2007), widely accepted deﬁnitions for most of the dependent
constructs and their construct-label synonyms were observed. Organizational outcomes
included productivity, retention, turnover intention, commitment, and performance.
Using Hunter and Schmidt’s (1990) approach, correlations for telework and each
proposed outcome variable were meta-analyzed. First, each reported statistic was
identiﬁed and, if necessary, transformed into a correlation. Correlations were weighted
according to sample size and corrected for reliability and sampling error in the measures
at the aggregate level. Where reliability was not reported, an average reliability was
calculated by imputing reliabilities reported from the average of other studies using the
same constructs (Gajendran and Harrison, 2007; Lipsey and Wilson, 2001). Next,
calculations were performed to attain the estimated population effect size, its variance,
conﬁdence intervals, and credibility intervals.
Correlations were deemed signiﬁcant if the conﬁdence interval did not include zero.
Where the credibility interval is large or includes zero, the mean corrected effect size may
represent a number of subpopulations and a moderating effect(s) may be present. When
the credibility interval is small or does not include zero, it is more likely that the mean
corrected effect size represents the population of interest (see Whitener, 1990 for a
review). For meta-analytic computations, the program MetaExcel (Steel, 2009) was used.
For the purposes of this paper, meta-analytic estimates were calculated where there was
a minimum of three independent effect sizes.
Five exploratory moderator analyses were conducted, where a minimum of three data
points were required to evaluate a moderator for a given telework-outcome relationship.
The ﬁrst analysis looks at whether performance was evaluated at the level of the ﬁrm or
subordinate to the ﬁrm (e.g. work group, business unit, etc.); this moderator applies only
to the performance variable. Additional possible moderators include the source of the
report (e.g. managers or general employees), the response rate of the sample, the
proportion of females in the sample, and the country in which the study was conducted
(coded as USA and non-USA).
The signiﬁcance of the ﬁve moderators identiﬁed above was computed using weighted
least squared (WLS) regression. WLS enables a weighting variable to be created from the
inverse of the sampling error for each moderator test, and is considered superior to
ordinary least squared (OLS) models when there is heteroscedasticity among observations
(Steel and Kammeyer-Mueller, 2002). Where a moderator analysis was signiﬁcant, the
results are discussed.
The meta-analytic relationships between telework and the outcome variables are
presented in Table I. There is an international, though predominantly Western,
representation of countries where the studies were conducted. These do not necessarily
reﬂect the location of the institution conducting the study, and include: Australia,
Brussels, Ireland, Spain, and the USA. Four studies did not indicate the country of the
study sample. Nearly two-thirds of the data for this analysis comes from peer-reviewed
academic journals; the other from doctoral dissertations. In order to address the issue of
publication bias (the “ﬁle drawer” problem), Failsafe-N values were calculated for each
of the variables. The Failsafe-N estimates the number of unpublished studies with an
average effect of zero that would be required to reduce a given meta-analytic coefﬁcient
to ^0.10 (i.e. a small correlation with lower practical signiﬁcance, as per Cohen, 1969).
These results are depicted in Table II, which suggest that the current ﬁndings are not
likely to be substantively affected by publication bias.
Five correlations were reported between telework and productivity. The number of
individual perceptions of this sample was 620. H1 proposed that telework will be
positively associated with perceptions of increased productivity. Meta-analytic ﬁndings
support this proposition therein the average correlation for unreliability was r
(k¼5,n¼620), and the conﬁdence interval for the uncorrected correlation did not
include zero (95 percent CI ¼0.13 20.33). The credibility interval did not include zero
(95 percent CrI ¼0.09 20.37), and none of the moderator tests were signiﬁcant.
Variable kn r
LU L U
Productivity 5 620 0.213 0.230 0.13 0.33 0.09 0.37
Retention 6 1,652 0.098 0.103 0.16 0.04 0.19 0.02
Commitment 8 3,144 0.098 0.106 0.03 0.18 20.07 0.28
Performance 10 2,522 0.153 0.163 0.09 0.23 20.02 0.34
Overall 19 5,502 0.163 0.173 0.15 0.20 0.04 0.31
Notes: k– number of samples; n– total number of data points; r
– uncorrected weighted mean
– weighted mean correlation corrected for unreliability; 95 percent conﬁdence –
conﬁdence interval; 95 percent credibility – credibility interval
telework and perceptions
Variable r-original Failsafe-N
Productivity 0.21 6.74
Retention 0.1 0.19
Commitment 0.1 0.53
Performance 0.15 6.4
Overall 0.17 14.5
Notes: r-original – meta-analytic correlation generated in current study; Failsafe-N – number of
unpublished papers with an average correlation of zero required to equal r-criterion; for all variables
above, r-criterion – ^0.10 (proposed “true score” correlation due to publication bias); and r-
Failsafe ¼0.00 (estimated average value for unpublished studies)
Failsafe-N estimates for
H2 proposed that telework will be positively associated with perceptions of employee
retention. To examine this hypothesis, six correlations were used to analyze the
relationship between telework and retention. The number of employee perceptions of
this sample was 1,652. The average correlation corrected for unreliability was r¼0.10
(k¼6, n¼1,652), and the conﬁdence interval for the uncorrected correlation did not
include zero (95 percent CI ¼0.04 20.16). None of the exploratory moderator analyses
yielded signiﬁcant results, and the credibility interval did not include zero (95 percent
CrI ¼0.02 20.19). Thus, as expected there was a small, but positive, effect size.
Eight correlations were reported between telework and organizational commitment,
representing 3,144 perceptions of this sample. H3 proposed that telework will would be
positively associated with perceptions of organizational commitment. The
meta-analysis yielded weak support for this hypothesis. The average correlation
corrected for unreliability was r¼0.11 (k¼8, n¼3,144) with a conﬁdence interval that
did not include zero (95 percent CI ¼0.03 20.18). In this analysis, the credibility
interval did include zero (95 percent CrI ¼20.07 20.28), indicating the presence of a
moderator. Thus, a moderator analysis was conducted to ﬁnd that the average age of the
sample was signiﬁcant (F(1,4) ¼4.715, p,0.05, R
¼0.80), with a negative direction
of the standardized
indicating that the higher the average age of the sample, the lower
the correlation between telework and organizational commitment (
Last, ten correlations were reported between telework and performance. The number
of individual perceptions of this sample was 2,522 employees. H4 proposed that
telework will be positively associated with perceptions of increased performance.
Meta-analytic ﬁndings support this proposition wherein the average correlation
corrected for unreliability was r¼0.16 (k¼10,n¼2,522). The conﬁdence interval for
the uncorrected correlation did not include zero (95 percent CI ¼0.09 20.23). Of note,
because the credibility interval was found to include zero (95 percent
CrI ¼20.02 20.34), a moderator analysis was conducted and the level of focus was
found to be near signiﬁcance (F(1,7) ¼4.715, p¼,0.0710, R
¼0.40). Intriguingly, the
exploratory moderator analysis yielded different strengths of correlations, depending on
whether the sample was comprised of one individual from many ﬁrms
(Martinez-Sanchez et al., 2007a, b, 2008), or multiple individuals from one ﬁrm
(Hyland et al., 2005). Correlations were stronger when the sample consisted of the same
number of raters as ﬁrms (rc ¼0.19, k¼3, n¼791) than when the sample consisted of
multiple raters from the same ﬁrm (rc ¼0.13, k¼4, n¼1,049).
H5 proposed that telework will be positively associated with the organizational
outcomes of the framework. The total number of effect sizes (aggregated by study) for
telework and organizational outcomes that were positive and signiﬁcant was 19, taking
reverse coding of turnover intentions into account. None of the effect sizes gathered
was negative and non-signiﬁcant. The average correlation corrected for unreliability
was r¼0.17 (k¼19, n¼5,502) with a conﬁdence interval that did not include zero
(95 percent CI ¼0.15 20.20). Thus, based on the effect sizes in our analysis,
there is support for H5, that, overall, there is a signiﬁcant and positive relationship
between telework and organizational outcomes.
This meta-analysis sought to determine whether or not telework is beneﬁcial for
organizations. Based on these meta-analytical ﬁndings, there appears to be support that,
overall, thereis a positive relationship between telework and perceptions of organizational
outcomes. In answer to the question, is telework effective for organizations, the
meta-analytical answer seems to be yes.There is a small but positive relationship between
telework and all four organizational outcomes of the framework: productivity, retention,
organizational commitment, and performance.
The small but positive effects of this study are in line with the ﬁndings of Gajendran
and Harrison’s (2007, p. 1535) meta-analysis, who found that, for individuals,
“telecommuting is mainly a good thing”. Hence, this meta-a nalysis adds value to the ﬁeld
in conﬁrming that, for organizations, telework is also “a good thing.” This is important
information for decision makers who are trying to determine whether or not to
implement a telework program, as it shows there are clearly no negative relationships
between telework and the organizational variables analyzed here.
Additionally, contributions of this meta-analysis lie in ﬁndings of the moderator
analyses. Regarding correlations between telework and organizational commitment
(H3), the average age of the sample was found to be a signiﬁcant moderator
(F(1,4) ¼4.715, p,0.05, R
¼0.80) where the higher the average age of the sample, the
lower the correlation. This suggests that commitment is more positive for younger
samples when telework is an option, and could implicate telework as a tool for attracting
young talent. Additionally, a moderator was also identiﬁed for the relationship between
telework and performance (H4) where correlations were moderated by whether or not
the sample was comprised of one individual from many ﬁrms, or many individuals from
one ﬁrm. This ﬁnding may suggest that outcomes are perceived to be more substantive
for the ﬁrm than for the subordinate-level groups, and should be considered in future
studies of telework perceptions.
There is a caution about the generalizability of these ﬁndings as they may be
contingent on individual variables such as job satisfaction, role stress, or perceived career
prospects (Gajendran and Harrison, 2007). On the positive side, organizations promoting
telework as an option may attract talent among younger employees while establishing
positive relations with external stakeholders, particularly amidst societal pressures to
“be more green.” Although an environmental responsibility outcome is not measured as
part of this study, it may be viewed as an added bonus for ﬁrms that adopt telework based
on organizational outcomes. Additionally, business continuity amidst natural disasters,
terrorist attacks and pandemics has pushed considerations of telework to the forefront.
Thus, this study provides organizational support for the decision to implement programs
that enable employees to continue their work from alternate locations in times when
going to the head ofﬁce is dangerous or impossible.
Limitations and future research
This meta-analysis is limited by several factors. First, there is what is referred to in
meta-analysis as the “apples and oranges” issue, where confounds may arise from
each study using slightly different deﬁnitions, constructs and operalizations.
Second, although the deﬁnition of telework as presented here is in line with much of
the ﬁeld, there are still conﬂicting deﬁnitions within the telework literature that limit
uniﬁcation of the work. Add to that, a hodge-podge of theoretical frameworks, or in some
cases the absence of one, and the complexity of different assumptions muddies the
opportunity for comparison from one study to the next. Third, there is a time stamp on
this research as it analyzes a snapshot of research within a speciﬁc time frame.
Just as the telework research that came before the inclusion criteria was considered
irrelevant, contemporary and future research also has the potential to be very different
from studies included here. This is a limitation of all meta-analysis, but particularly in a
ﬁeld where change is the status quo.
Next, discrepancies over what actually constitutes a telework arrangement is another
limitation of this analysis (Bailey and Kurland, 2002). In some studies, multiple types of
telework are included such as home-based workers (Nguyen, 2004), whereas in others,
home-based workers are excluded and/or satellite ofﬁces are included (Gajendran and
Harrison, 2007; Illegems and Verbeke, 2003). Thus, this analysis is limited in its ability to
report effects for speciﬁc telework arrangements, and maintains the goal of reporting
effects for telework in general as they pertain to organizational outcomes.
In addition, the fact that some studies lumped managerial and employee perspectives
together, prohibits us from seeing the true picture of the organizational perspective.
Ideally, the data would be solely managerial, and perhaps qualitative to investigate the
possibility that there are variables impacting the decision to implement that are, as yet,
unidentiﬁed. Stimulated by Bailey and Kurland’s (2002) observation that managers are
reluctant to implement telework programs, future research would do well to dig deeper
into the source of this reluctance by looking at social and cognitive barriers to adoption.
Future research would also do well to look at the metrics of the organizationally
measured outcomes studied here, for a comparison to the perception research.
Productivity in terms of physical output, retention in terms of employees who remain per
annum, and performance in terms of company ratings would give a clearer picture of the
relationship between these variables and telework. In terms of sustainability, future
research might focus on telework as an important part of disaster recovery. Investigation
into the resilience and efﬁcacy of teleworking ﬁrms amidst disaster, perhaps in
comparison to non-teleworking ﬁrms, would provide meaningful data regarding
telework’s potential as an essential practice for contemporary business continuity.
Finally, the literature search of this meta-analysis has revealed that, although there is
a plethora of telework research at the individual and societal levels, somehow the
organizational level has been largely missed. Even the work that claims to be about the
organization is based on individual perceptions. As such, there is a need for future
research in telework to focus on the organization as the unit of analysis, in order to truly
capture telework’s impacts at that level.
Despite individual and societal research showing positive beneﬁts of telework, it has not
been embraced by organizational decision makers as an effective, mainstream alternative
to current work arrangements. This meta-analysis provides evidence that it may be
exactly that, in ﬁnding that there is indeed a small but positive relationship between
telework and organizational outcomes; with the caveat that certain variables may lead to
greater beneﬁts, where others may diminish them. As such, organizational decision
makers may want to take heed, for if the use of ICT outside the central ofﬁce can indeed
help organizations be more productive, strengthen organizational commitment, secure
employee retention, and improve performance, then the future discussed at the beginning
of the paper, is now. If so, then the decision of whether or not to implement a telework
program could be the factor that determines whether or not an organization becomes part
of that future, or a thing of the past.
Akyeampong, E.B. (2007), Working at Home: An Update, Perspectives, Statistics Canada –
Catalogue No. 75-001-XIE, June, available at: www.statcan.gc.ca (accessed 28 April 2011).
Akyeampong, E.B. and Nadwodny, R. (2001), “Evolution of the Canadian workplace: work from
home”, Perspectives on Labour and Income, Vol. 2 No. 9, Statistics Canada Catalogue
no. 75-001-XIE, available at: www.statcangc.ca/english/freepub/75-001-XIE/0090175-001-
XIE.html (accessed 22 April 2011).
Bailey, D.E. and Kurland, N.B. (2002), “A review of telework research: ﬁndings, new directions,
and lessons for the study of modern work”, Journal of Organizational Behavior, Vol. 23,
Baruch, Y. (2000), “Teleworking: beneﬁts and pitfalls as perceived by professionals and
managers”, New Technology, Work, and Employment, Vol. 15 No. 1, pp. 34-49.
*Belanger, F. (1999), “Workers’ propensity to telecommute: an empirical study”,
Information & Management, Vol. 35, pp. 139-53.
Berger, M. (1996), “Making the virtual ofﬁce a reality”, Sales & Marketing Management, Vol. 21,
Byrd, S.C. (2005), “The use of telecommuting as an alternative workplace strategy for Missouri
businesses”, unpublished doctoral dissertation, Southern Illinois University at Carbondale,
available at: ProQuest Dissertations, ID 920927651.
*Desrosiers, E.I. (2001), “Telework and work attitudes: the relationship between telecommuting
and employee job satisfaction, organizational commitment, perceived organizational
support, and perceived co-worker support”, unpublished doctoral dissertation, Purdue
University, available at: ProQuest Dissertations, ID 765166111.
Di Martino, V. and Wirth, L. (1990), “Telework: a new way of working and living”, International
Labour Review, Vol. 129 No. 5, pp. 529-54.
Duxbury, L. and Haines, G. Jr (1991), “Predicting alternative work arrangements from salient
attitudes: a study of decision makers in the public sector”, Journal of Business Research,
Vol. 23, pp. 83-97.
Duxbury, L. and Higgins, C. (2002), “Telework: a primer for the millennium introduction”,
in Cooper, C.L. and Burke, R.J. (Eds), The New World of Work: Challenges and
Opportunities, Blackwell, Oxford, p. 157.
Duxbury, L., Higgins, C. and Mills, S. (1992), “After-hours telecommuting and work-family
conﬂict: a comparative analysis”, Information Systems Research, Vol. 3 No. 2, pp. 173-90.
Gajendran, R.S. and Harrison, D.A. (2007), “The good, the bad, and the unknown about
telecommuting: meta-analysis of the psychological mediators and individual
consequences”, Journal of Applied Psychology, Vol. 92 No. 6, pp. 1524-41.
Hill, E.J., Hawkins, A.J. and Miller, B.C. (1996), “Work and family in the virtual ofﬁce: perceived
inﬂuences of mobile telework”, Family Relations, Vol. 45 No. 3, pp. 293-301.
*Hill, E.J., Miller, B.C., Weiner, S.P. and Colihan, J. (1998), “Inﬂuences of the virtual ofﬁce on
aspects of work and work/life balance”, Personnel Psychology, Vol. 51 No. 3, pp. 667-82.
Hunter, J.E. and Schmidt, F.L. (1990), Methods of Meta-analysis: Correcting Error and Bias in
Research Findings, Sage, Newbury Park, CA.
*Hyland, M.H., Rowsome, C. and Rowsome, E. (2005), “The integrative effects of ﬂexible work
arrangements and preferences for segmenting or integrating work and home roles”,
Journal of Behavioral and Applied Management, Vol. 6 No. 2, pp. 141-60.
Illegems, V. and Verbeke, A. (2003), Moving towards the Virtual Workplace, Edward Elgar,
*Illegems, V. and Verbeke, A. (2004), “Telework: what does it mean for management?”,
Long Range Planning, Vol. 37 No. 4, pp. 319-34.
Jones, C. (2005), Telework: The Quiet Revolution, The Gartner Group, available at: www.gartner.
com/DisplayDocument?doc_cd¼122284 (accessed 21 April 2011).
*Lee, S. (2004), “Organizational commitment and job satisfaction: the effect of alternative ofﬁcing
strategies on teleworkers’ organizational behavior”, unpublished doctoral dissertation,
Michigan State University, available at: ProQuest Dissertations, ID 765931721.
Lipsey, M.W. and Wilson, D.B. (2001), Practical Meta-analysis, Sage, Thousand Oaks, CA.
Madsen, S.R. (2003), “The effects of home-based teleworking on work and family conﬂict”,
unpublished doctoral dissertation, University of Minnesota, available at: ProQuest
Dissertations, ID 728982381.
Martinez-Sanchez, A., Perez-Perez, M., de-Luis-Carnicer, P. and Jimenez, M.J.V. (2007a),
“Telework, human resource ﬂexibility and ﬁrm performance”, New Technology, Work and
Employment, Vol. 22 No. 3, pp. 208-23.
*Martinez-Sanchez, A., Perez-Perez, M., de Luis Carnicer, P. and Jimenez, M.J.V. (2007b),
“Teleworking and workplace ﬂexibility: a study of impact on ﬁrm performance”,
Personnel Review, Vol. 36 No. 1, pp. 42-64.
*Martinez-Sanchez, A., Perez-Perez, M., Vela-Jimenez, M.J. and de-Luis-Carnicer, P. (2008),
“Telework adoption, change management, and ﬁrm performance”, Journal of
Organizational Change Management, Vol. 21 No. 1, pp. 7-31.
Meyer, J.P. and Allen, N.J. (1991), “A three-component conceptualization of organizational
commitment”, Human Resource Management Review, Vol. 1 No. 1, pp. 61-89.
Mokhtarian, P.L. and Sato, K. (1994), “A comparison of the policy, social, and cultural contexts
for telecommuting in Japan and the United States”, Social Science Computer Review, Vol. 12
No. 4, pp. 641-58.
Mokhtarian, P.L., Bagley, M., Hulse, L. and Salomon, I. (1998), “The inﬂuence of gender and
occupation on individual perceptions of telecommuting”, The Journal of the American
Society for Information Science, Vol. 49 No. 12, pp. 691-711.
*Nguyen, N.T. (2004), “The consequences of spatial distance and electronic communication for
teleworkers: a multi-level investigation”, unpublished doctoral dissertation,
Temple University, Philadelphia, PA.
Nilles, J.M. (1994), Making Telecommuting Happen: A Guide for Telemanagers and
Telecommuters, Van Nostrand Reinhold, New York, NY.
*Piper, H.H.M. (2004), “Telework and organizational commitment: a test of the Meyer and Allen
three-dimensional model of commitment”, unpublished doctoral dissertation,
Nova Southeastern University, available at: ProQuest Dissertations, ID 885613531.
Pitt-Catsouphes, M. and Marchetta, A. (1991), A Coming of Age: Telework, Boston University,
Center on Work and Family, Boston, MA.
Salomon, I., Schneider, H. and Schofer, J. (1991), “Is telecommuting cheaper than travel?
An examination of interaction costs in a business setting”, Transportation, Vol. 18 No. 4,
Salopek, P. (1998), Census Brief: Increase in At-home Workers Reverses Earlier Trend,
US Department of Commerce: BUREAU OF THE CENSUS, available at: www.pio.census.
gov (accessed 28 April 2011).
Steel, P. (2009), MetaExcel (Version 13.0) (Computer software), available at: www.ucalgary.ca/
Steel, P. and Kammeyer-Mueller, J. (2002), “Comparing meta-analytic moderator search techniques
under realistic conditions”, Journal of Applied Psychology, Vol. 87 No. 1, pp. 96-111.
Tofﬂer, A. (1981), The Third Wave, Bantam Books, London.
US Census Bureau (2004), Survey of Income and Program Participation, available at: www.
(accessed 1 August 2011).
Valsecchi, R. (2006), “Visible moves and invisible bodies: the case of teleworking in an Italian call
center”, New Technology, Work and Employment, Vol. 21 No. 2, pp. 123-38.
Verbeke, A., Schultz, R., Greidanus, N. and Hambley, L. (2008), Growing the Virtual Workplace,
Edward Elgar, Northampton.
Wallace, D. (2003), Part-time Work and Family-friendly Practices in Canadian Workplaces,
The Evolving Workplace Series, Statistics Canada, Human Resources Development Canada,
Catalogue No. 71-584-MIE, available at: www.statcan.cg.ca (accessed 22 April 2011).
Welz, C. and Wolf, F. (2010), Telework in the European Union, available at: www.eurofound.
europa.eu (accessed 28 June 2011).
Whitener, E.M. (1990), “Confusion of conﬁdence intervals and credibility intervals in
meta-analysis”, Journal of Applied Psychology, Vol. 75 No. 3, pp. 315-21.
Workman, M. (2000), “The effects of cognitive style and communications media on commitment
to telework and virtual team innovations among information systems teleworkers”,
unpublished doctoral dissertation, Georgia State University, Atlanta, GA.
Yap, C.S. and Tng, H. (1990), “Factors associated with attitudes towards telecommuting”,
Information & Management, Vol. 19 No. 4, pp. 227-35.
*Baker, E., Avery, G.C. and Crawford, J. (2006), “Home alone: the role of technology in
telecommuting”, Information Resource Management Journal, Vol. 19 No. 4, pp. 1-22.
*Brenner, E.S. (2006), “Retaining teleworkers: telework as a moderator of typical relationships
between turnover antecedents and turnover intentions”, unpublished doctoral
dissertation, George Washington University, available at: ProQuest Dissertations,
*Golden, T.D. (2006), “Avoiding depletion in virtual work: telework and the intervening impact
of work exhaustion on commitment and turnover intentions”, Journal of Vocational
Behavior, Vol. 69 No. 1, pp. 176-87.
*Golden, T.D., Viega, J.F. and Dino, R.N. (2008), “The impact of professional isolation on
teleworker job performance and turnover intentions: does time spent teleworking,
interacting face-to-face, or having access to communication-enhancing technology
matter?”, Journal of Applied Psychology, Vol. 93 No. 6, pp. 1412-21.
Hartman, R.I., Stoner, C.R. and Arora, R. (1991), “An investigation of selected variables affecting
telecommuting productivity and satisfaction”, Journal of Business and Psychology, Vol. 6
No. 2, pp. 207-25.
*Hill, E.J., Ferris, M. and Martinson, V. (2003), “Does it matter where you work? A comparison of
how three work venues (traditional ofﬁce, virtual ofﬁce, and home ofﬁce) inﬂuence
aspects of work and personal/family life”, Journal of Vocational Behavior, Vol. 63 No. 2,
Hunton, J. (2005), “Behavioral self-regulation of telework locations: interrupting interruptions!”,
Journal of Information Systems, Vol. 19 No. 2, pp. 111-40.
*Kemerling, K.R. (2002), “The effects of telecommuting on employee productivity: a perspective
from managers, ofﬁce co-workers and telecommuters”, unpublished doctoral dissertation,
Colorado Technical University, available at: ProQuest Dissertations, ID 764850061.
*Kossek, E.E., Lautsch, B. and Eaton, S.C. (2006), “Telecommuting, control, and boundary
management: correlates of policy use and practice, job control, and work-family
effectiveness”, Journal of Vocational Behavior, Vol. 68 No. 2, pp. 347-67.
*O’Neill, T.A., Hambley, L.A., Greidanus, N.S., MacDonnell, R. and Kline, T.J.B. (2009),
“Predicting teleworker success: an exploration of personality, motivational, situational,
and job characteristics”, New Technology, Work and Employment, Vol. 24 No. 2, pp. 144-62.
*Poissonnet, S.P. (2002), “Proﬁles of ﬁt for successful telework outcomes”, unpublished doctoral
dissertation, University of California, Los Angeles, CA, available at: ProQuest
Dissertations, ID 764926411.
Yehuda, B. and Yuen, Y. (2000), “Inclination to opt for teleworking: a comparative analysis of
United Kingdom versus Hong Kong employees”, International Journal of Manpower,
Vol. 21 No. 7, pp. 521-39.
Brittany Harker Martin can be contacted at: email@example.com
To purchase reprints of this article please e-mail: firstname.lastname@example.org
Or visit our web site for further details: www.emeraldinsight.com/reprints
Reproducedwith permission of thecopyright owner. Further reproductionprohibited without permission.