Content uploaded by Andrew Bennett
Author content
All content in this area was uploaded by Andrew Bennett on Nov 19, 2018
Content may be subject to copyright.
RESEARCH ARTICLE
Recovery from work‐related effort: A meta‐analysis
Andrew A. Bennett
1
|Arnold B. Bakker
2
|James G. Field
3
1
Department of Management, Strome College
of Business, Old Dominion University, Norfolk,
Virginia, U.S.A.
2
Center of Excellence for Positive
Organizational Psychology, Erasmus
University Rotterdam, Rotterdam, The
Netherlands
3
Department of Management, College of
Business and Economics, West Virginia
University, Morgantown, West Virginia, U.S.A.
Correspondence
Andrew A. Bennett, Department of
Management, Strome College of Business, Old
Dominion University, 2033 Constant Hall,
Norfolk, Virginia 23529, U.S.A.
Email: aabennet@odu.edu
Summary
This meta‐analytic study examines the antecedents and outcomes of four recovery experiences:
psychological detachment, relaxation, mastery, and control. Using 299 effect sizes from 54
independent samples (N= 26,592), we extend theory by integrating recovery experiences into
the challenge–hindrance framework, creating a more comprehensive understanding of how both
after‐work recovery and work characteristics collectively relate to well‐being. The results of
meta‐analytic path estimates indicate that challenge demands have stronger negative
relationships with psychological detachment, relaxation, and control recovery experiences than
hindrance demands, and job resources have positive relationships with relaxation, mastery, and
control recovery experiences. Psychological detachment after work has a stronger negative
relationship with fatigue than relaxation or control experiences, whereas control experiences after
work have a stronger positive relationship with vigor than detachment or relaxation experiences.
Additionally, a temporally driven model with recovery experiences as a partial mediator explains
up to 62% more variance in outcomes (ΔR
2
= .12) beyond work characteristics models, implying
that both work characteristics and after‐work recovery play an important role in determining
employee well‐being.
KEYWORDS
challenge–hindrance framework, fatigue, meta‐analysis, recovery experiences, vigor
1|INTRODUCTION
Work requires energy and effort to accomplish required tasks. Both
work conditions and task demands can deplete psychological
resources (Meijman & Mulder, 1998). After expending energy over a
period, it is necessary to recover or replenish resources that were used
up at work (Zijlstra & Sonnentag, 2006). For many employees, the
recovery process occurs each day after work. This recovery process
plays a “crucial intervening role in the relationship between stressful
work characteristics on the one hand, and health, well‐being and
performance capability on the other hand”(Sonnentag & Geurts,
2009, p. 2). This study focuses on recovery experiences because “it is
not a specific activity per se that helps [one] to recover from job stress
but its underlying attributes”(Sonnentag & Fritz, 2007, p. 204). In
other words, recovery experiences are the mechanisms through which
recovery processes occur (Sonnentag & Geurts, 2009). In this way,
after‐work recovery experiences are considered a mediator between
work characteristics and well‐being outcomes (Kinnunen, Feldt,
Siltaloppi, & Sonnentag, 2011).
The four most researched recovery experiences are psychological
detachment, not thinking about work during nonwork time; relaxation,
having a low activation level; mastery, facing a positive challenge to
learn something new; and control, having a feeling of control over
nonwork time (Sonnentag & Fritz, 2007). Although there has been
considerable research about recovery over the past two decades,
several major questions remain unanswered: How do recovery
experiences fit into recent work characteristic and employee well‐
being models? Is one recovery experience more effective for improving
individual well‐being? Does researching recovery experiences add to
our understanding of well‐being in a practically significant way beyond
work characteristics models? The current study aims to address these
unanswered questions as well as additional questions surrounding
after‐work recovery experiences.
The past three decades of employee well‐being research have
yielded important contributions to the understanding of how work
and nonwork experiences relate to individual well‐being. Lee and
Ashforth (1996) meta‐analyzed the relationship between work
characteristics and burnout, finding that work demands have the
strongest correlation with emotional exhaustion. Several years later,
the Job Demands–Resources model (JD‐R; Demerouti, Bakker,
Nachreiner, & Schaufeli, 2001) proposed that job demands and job
resources are associated with burnout in different ways. Subsequent
Received: 27 March 2015 Revised: 18 June 2017 Accepted: 10 July 2017
DOI: 10.1002/job.2217
262 Copyright © 2017 John Wiley & Sons, Ltd. J Organ Behav. 2018;39:262–275.wileyonlinelibrary.com/journal/job
research using this model suggests job demands and job resources
uniquely predict positive outcomes such as work engagement (Bakker
& Demerouti, 2007, 2014). Another research stream drew upon
Lazarus and Folkman0s (1984) seminal work to recognize that work
demands can be characterized as either a “positive”challenge or a
“negative”hindrance (Cavanaugh, Boswell, Roehling, & Boudreau,
2000). Crawford, LePine, and Rich (2010) used meta‐analytic structural
equation modeling (MASEM) to show that all demands are positively
associated with burnout. Furthermore, they reported that hindrance
demands are negatively associated with engagement whereas
challenge demands are positively associated with engagement.
As an alternative to studying work characteristics, Sonnentag
(2001) extended research on employee respite and explored how
evening leisure helped employees reduce stress and improve well‐
being. We now know that work characteristics typically explain more
variance in distal outcomes such as burnout (Podsakoff, LePine, &
LePine, 2007) and performance (LePine, Podsakoff, & LePine, 2005)
than do proximal outcomes (e.g., affect). One benefit of this newer
emphasis on after‐work recovery is that it focuses on proximal well‐
being outcomes such as fatigue and vigor (Sonnentag, 2012) that are
more volatile and fluctuate from day to day in employees (Ten
Brummelhuis & Bakker, 2012).
This study has three main objectives. First, this study integrates
the recovery literature with the work characteristics literature. This is
valuable because most studies have included work characteristic
variables but not explicitly theorized how antecedents are related to
recovery experiences. In addition, we contend that previous models
are a starting point but are incomplete. For example, Crawford et al.
(2010) provided distinctions between challenge demands and
hindrance demands with outcomes, but did not include recovery
experiences that can occur each day and are also related to well‐being.
Kinnunen et al. (2011) used a single sample to test recovery
experiences as mediators within the JD‐R model, but did not differen-
tiate work demands as challenges or hindrances, which has been
shown to have different relationships with well‐being outcomes
(Crawford et al., 2010). Similarly, Armon, Melamed, and Shirom
(2012) found that job resources were related to vigor and job demands
were not, but did not include any hindrance demands. Last, Sonnentag
and Zijlstra (2006) found that need for recovery mediated the relation-
ship between work characteristics and fatigue, but needing recovery is
not the same as experiencing recovery. Therefore, to fill these research
gaps, we use MASEM to provide a more comprehensive understanding
of the relationships between work demands and resources, after‐work
recovery experiences, and fatigue and vigor. Using such a procedure is
useful because meta‐analytic correlations provide a more comprehen-
sive understanding of the relationships between constructs than do
correlations found in a single sample, and structural models provide
more accurate conclusions of the relationships between constructs
than do zero‐order correlations.
Second, this study clarifies inconsistencies within the recovery
literature. One inconsistency concerns outcomes from recovery
experiences, as some authors have concluded that psychological
detachment is “the most powerful recovery experience”(Siltaloppi,
Kinnunen, & Feldt, 2009, p. 344), whereas others have found that
“for achieving positive activation and serenity in the morning, it does
not help to detach oneself from work during off‐job time but to engage
in mastery experiences or relaxation”(Sonnentag, Binnewies, & Mojza,
2008, p. 682). Thus, a meta‐analysis can provide a more consistent
overview of specific relationships between recovery experiences and
outcomes. Another inconsistency is the relationship between job
resources and recovery experiences, as some authors theorize that
job resources are positively related to all recovery experiences
(Kinnunen & Feldt, 2013), whereas others contend that job resources
may only be related to mastery and control (Shimazu, Sonnentag,
Kunota, & Kawakai, 2012).
Third, this study investigates the practical value of after‐work
recovery research. Primary studies vary greatly in the amount of
variance that recovery experiences explain above and beyond
work characteristics. For example, findings range from less than 1%
(e.g., Querstret & Cropley, 2012; Sonnentag, Binnewies, & Mojza,
2010) to as high as 6–10% (e.g., Fritz, Yankelevich, Zarubin, & Barger,
2010; Sonnentag, Kuttler, & Fritz, 2010) of the variance explained in
the relationship between evening psychological detachment and
fatigue. This variability could be because each primary study includes
different work characteristics and recovery experience variables rather
than all variables together. As such, we test how including recovery
experiences adds to a model of work characteristics and well‐being.
2|THEORETICAL BACKGROUND
2.1 |Work‐related antecedents of recovery
experiences
Drawing on JD‐R theory, much of the literature describes that work
demands are negatively related to recovery experiences (Demerouti,
Bakker, Geurts, & Taris, 2009). Sonnentag and Fritz (2007) outlined
how both work demands and job control are related to recovery
experiences. Consistent with the stress appraisal literature, which
suggests that work demands can be positive or negative (Lazarus &
Folkman, 1984), work characteristics are now typically categorized as
challenge demands, hindrance demands, and job resources (Cavanaugh
et al., 2000). These newer categorizations and refinements are
important because they have uncovered unique relationships with
well‐being outcomes (e.g., Crawford et al., 2010; Podsakoff et al.,
2007). The purpose of this section is to summarize previous research,
build off integrations of JD‐R with recovery (e.g., Kinnunen et al.,
2011), and extend theory to explore the potentially distinct relation-
ships of challenge demands, hindrance demands, and job resources
with recovery experiences.
Challenge demands are work demands such as time pressure and
high workload that are stressful but also produce positive feelings
and can foster psychological resources such as self‐efficacy
(Cavanaugh et al., 2000). Challenge demands are positively related to
autonomous work motivation, which is when employees choose to
work because they enjoy it and gain meaning from it (TadićVujčić,
Oerlemans, & Bakker, 2017). It has been suggested that individuals
who enjoy their work and find their work interesting may engage in
problem‐solving pondering (Cropley & Zijlstra, 2011). Problem‐solving
pondering refers to thinking about solutions for work‐related problems
BENNETT ET AL.263
during nonwork time and is negatively related to psychological
detachment (Querstret & Cropley, 2012). Greater challenge demands
are also associated with higher positive affect at the end of the day
(TadićVujčićet al., 2017). Although high positive affect can be
beneficial, it typically involves higher levels of activation. We expect
that prolonged high activation from challenge demands (Brosschot,
Pieper, & Thayer, 2005) spills over to nonwork time and will be
negatively related to relaxation experiences. Enthusiastic emotions
towards work goals have also been linked with interference into
nonwork time (Wood & Michaelides, 2016), so challenge demands
should be negatively related to control experiences. However,
challenge demands are work conditions that result in learning
(Tadić, Bakker, & Oerlemans, 2015), and this interest to explore can
spill over to after‐work time and is positively associated with mastery
experiences (Michel, Turgut, Hoppe, & Sonntag, 2016).
Hypothesis 1. Challenge demands will have a (a)
negative relationship with psychological detachment,
(b) negative relationship with relaxation, (c) positive
relationship with mastery, and (d) negative relationship
with control experiences.
Unlike challenge demands, hindrance demands are when an
employee0s work relationships or environment interferes with goal
attainment. We expect these to also prohibit after‐work recovery
experiences, although they do so in different ways. Measures for this
construct often include role conflict, role ambiguity, conflict at work,
and overload.
1
Issues such as conflict at work can create negative
emotional responses in the form of anxiety and anger (Tuckey, Searle,
Boyd, Winefield, & Winefield, 2015). In addition, anxiety at the end
of the work day is linked to lower detachment at home (Van Hooff,
2015). Martinez‐Corts, Demerouti, Bakker, and Boz (2015) showed
that daily task conflicts and interpersonal conflicts acted as
hindrance demands and spilled over to the home domain in the form
of strain‐based work‐to‐family conflicts, suggesting that those who
experienced conflicts at work could not detach psychologically from
their work. Thus, hindrance demands should be negatively related to
psychological detachment. These negative affective responses from
hindrance demands such as anger and anxiety also are highly
activated, so we expect hindrance demands to be negatively related
to relaxation experiences. Hindrance demands are also related to lower
motivation to actively cope with demands (Crawford et al., 2010), such
as engaging in mastery experiences. Last, hindrance demands such as
overload include the feeling that there is too much to do with no
control, which could spill over to experiencing low levels of control
during after‐work time.
Hypothesis 2. Hindrance demands will have a nega-
tive relationship with (a) psychological detachment,
(b) relaxation, (c) mastery, and (d) control experiences.
Job resources are work attributes that help an individual achieve
work goals or stimulate personal growth (Demerouti, Bakker, de Jonge,
Janssen, & Schaufeli, 2001). Typical job resources are job control, job
autonomy, job variety, and job growth opportunities (Sonnentag,
2015). Although resources such as job control are positively related
to well‐being, Sonnentag and Fritz (2007) also noted that increased
control means an individual will be more likely to continue thinking
about work during nonwork time and maintain a high level of
activation from work‐related issues. Indeed, individuals with higher
job control also report lower combinations of after‐work psychological
detachment and relaxation experiences (Bennett, Gabriel, Calderwood,
Dahling, & Trougakos, 2016). Nonetheless, individuals with greater job
resources are more able “to protect themselves from the strains of
further resource depletion”(Crawford et al., 2010, p. 837). In other
words, job resources can also be positively related to recovery
experiences. For example, increased competence arising from job
growth opportunities or job variety can spill over and increase the
desire for additional learning opportunities after work. Similarly,
feelings of job control may spill over as feelings of control over leisure
time. Thus, in line with previous findings (e.g., Siltaloppi et al., 2009),
we expect that job resources are positively related to mastery and
control recovery experiences.
Hypothesis 3. Job resources will have a negative
relationship with (a) psychological detachment and (b)
relaxation experiences, but a positive relationship with
(c) mastery and (d) control experiences.
2.2 |Outcomes of recovery experiences
The main basis for examining recovery experience outcomes is the
Effort‐Recovery model (ERM; Meijman & Mulder, 1998), which
extended the load–capacity model from exercise physiology to
integrate psychological effort and restoration of resources. ERM has
three main components: Individuals mobilize psychological resources
such as energy to engage in a work‐related process, this resource
mobilization leads to both task performance and resource depletion,
and recovery occurs when the work‐related processes end. If recovery
does not happen, individuals incur negative effects such as impaired
well‐being.
As described in the ERM, energy is a key resource used to engage
in a work procedure (Meijman & Mulder, 1998). Human energy is
frequently described as a subjective affective assessment of one0s
psychophysiological system (Quinn, Spreitzer, & Lam, 2012) and is a
limited resource (Hobfoll, 2011) that varies each day within
individuals (Ten Brummelhuis & Bakker, 2012). Energy fluctuations
have been linked to organizational outcomes such as job performance
(e.g., Bakker & Xanthopoulou, 2009), turnover (e.g., Wright &
Cropanzano, 1998), and citizenship behaviors (for a review, see
Sonnentag, 2015).
In the past decade, the structure of affect has been refined into a
model with 12 core dimensions around a circumplex (Yik, Russell, &
Steiger, 2011). Human energy fits best in two dimensions: pleasant
activation (e.g., vigor, vitality, energetic, and excited) and unpleasant
deactivation (e.g., fatigued, exhausted, sluggish, and tired). For
1
Although there is some discrepancy in categorizing workload (e.g., Tuckey et al.,
2015), we categorize time pressure and high workload as a challenge demand,
but overload as a hindrance demand because having an unmanageable workload
(e.g., White, 2010) prevents work goals (Tuckey et al., 2015) and provides no
growth potential (Tadićet al., 2015).
264 BENNETT ET AL.
consistency with most of the recovery literature, we use the term
vigor rather than pleasant activation and fatigue rather than unpleasant
deactivation. In accordance with this affective model, vigor and fatigue
are different affective dimensions rather than two poles on the same
dimension because individuals can simultaneously experience high
levels of vigor and fatigue (Mäkikangas et al., 2014) and vigor and
fatigue have unique antecedents and outcomes (e.g., Fritz, Lam, &
Spreitzer, 2011; Halbesleben, 2010).
In the context of the ERM, completing work processes during
formal work periods increases high activation and creates short‐term
load reactions (e.g., increases fatigue). Recovery experiences reverse
these reactions in multiple ways. Psychological detachment after work
is beneficial because it removes the individual mentally from work‐
related activation, thus providing time for resources to return to the
previous state (Meijman & Mulder, 1998). Relaxation experiences after
work contribute to employee well‐being because relaxation involves
little or no activation of the psychophysiological system, thus allowing
for previously depleted resources to be replenished (Stone, Kennedy‐
Moore, & Neale, 1995). Mastery experiences, which typically occur
when engaging in a more demanding leisure task such as learning a
new language or playing a musical instrument, both reduce activation
from work (like psychological detachment) and also increase an
individual0s personal resources such as competence. Experiencing
control can also be beneficial to employee well‐being because it
builds resources such as an individual0s feeling of self‐efficacy
(Sonnentag & Fritz, 2007).
In sum, there are two different pathways that recovery experiences
are related to fatigue and vigor. One pathway is reducing or stopping
psychological load from work tasks, typically occurring due to psycho-
logical detachment and relaxation experiences. This halts the prolonged
negative effects (e.g., fatigue) and allows this state to return to normal
levels. The second pathway is by building up additional psychological
resources, typically occurring due to mastery and control experiences.
Conservation of resources theory (Hobfoll, 2011) describes how an
increase in one personal resource (e.g., self‐efficacy) will also increase
another resource (e.g., vigor). In line with previous conceptualizations
that psychological detachment and relaxation experiences have a more
robust relationship with negative outcomes, whereas mastery and
control experiences have a more robust relationship with positive
outcomes (Siltaloppi et al., 2009), we posit that different recovery
experiences will have unique associations with outcomes.
Hypothesis 4. (a) Psychological detachment and (b)
relaxation experiences will have stronger negative
relationships with fatigue than will mastery and control
experiences.
Hypothesis 5. (a) Mastery and (b) control experiences
will have stronger positive relationships with vigor
than will psychological detachment and relaxation
experiences.
2.3 |Recovery experiences as mediators
Theoretically, recovery experiences are considered mediators between
work characteristics and well‐being (Demerouti et al., 2009). This is
especially evident when one considers how the recovery process
unfolds over time: An employee experiences demands and resources
at work and then recovery experiences after work, with both experi-
ences related to well‐being. For example, the stressor‐detachment
model (Sonnentag & Fritz, 2015) describes that work demands have
the potential to impair psychological detachment, and this lack of
detachment is related to low well‐being. The majority of findings
support this mediating role of recovery experiences (e.g., Safstrom &
Hartig, 2013; Sonnentag et al., 2010).
We build off the work of Kinnunen et al. (2011) who proposed
a partial mediation model by expanding the processes within the
challenge–hindrance framework and testing all four recovery experi-
ences at the same time. The partial mediation process is an important
distinction for two reasons. First, this continues to keep the direct
relationships between work characteristics and well‐being evident
in the JD‐R model. Second, partial mediation implies that work
characteristics are related to recovery experiences and that recovery
experiences are related to well‐being. As described in earlier
hypotheses, we contend that challenge demands, hindrance demands,
and job resources can be uniquely related to recovery experiences,
and these relationships may therefore alter the indirect associations
between work characteristics and outcomes. For example, challenge
demands have a positive direct relationship with fatigue (impaired
well‐being), but challenge demands can also have negative relationships
with after‐work detachment and relaxation experiences, so the total
relationship (both direct and indirect effects through low recovery
experiences) with fatigue may be stronger than may the direct effects
only. As a second example, job resources should have a direct negative
relationship with fatigue. However, if job resources are also negatively
related to detachment experiences, this may decrease the total effect
because of negative indirect effects through detachment. As a third
example, hindrance demands may have a lower negative relationship
with detachment, relaxation, and control experiences than challenge
demands because they create less nonwork interference (Wood &
Michaelides, 2016), so indirect effects on outcomes from hindrance
demands may be relatively minimal. Thus, we propose:
Hypothesis 6. Recovery experiences will partially
mediate the relationship between work characteristics
and outcomes.
3|METHOD
3.1 |Literature search and inclusion criteria
In June 2013, we conducted a comprehensive electronic literature
search on all forms of recovery related to work stress. Manuscripts
were identified using ABI Inform Complete, PsychInfo, and Academic
Search Complete databases with keywords of recovery,recovery
experience, and job stress. In January 2015, we conducted an electronic
search of all available Academy of Management, Society for
Industrial and Organizational Psychology, and Southern Management
Association conference proceedings, as well as dissertations through
ProQuest and requested unpublished studies through the email
listserv of the Academy of Management Organizational Behavior
BENNETT ET AL.265
division (i.e., OBlistserv). We also double checked that our search
contained all manuscripts citing Sonnentag and Fritz (2007) or were
used in previous meta‐analyses (i.e., Crawford et al., 2010; Lee &
Ashforth, 1996). This yielded 989 studies.
We then read the titles, abstracts, and/or methods sections to
discern if a study included at least one correlation between two
variables of interest. Following previous categorizations (e.g., Crawford
et al., 2010), we read items of all measures and categorized variables
into the hindrance–challenge framework. Time pressure and high
workload were categorized as challenge demands. Role conflict,
conflict at work, overload, and stress
2
were categorized as hindrance
demands. Job control, autonomy, job growth opportunities, and job
variety were categorized as job resources. Recovery experience
variables were psychological detachment, relaxation, mastery, or
control. Outcome variables of vigor, vitality, positive activated
affect, and excited were categorized as vigor. Variables of fatigue,
exhaustion, and emotional exhaustion were categorized as fatigue.
To be consistent with the temporal ordering proposed, we focused
on work‐related antecedents, after‐work recovery experiences, and
end‐of‐day outcomes. For daily diary studies, variables were assessed
in this temporal order (e.g., work characteristics measured at work,
and recovery experiences measured after work). For cross‐sectional
studies, items needed to include temporal wording, such as “during
time after work”(used in Sonnentag & Fritz, 2007) when assessing
recovery experiences. By request, three authors provided additional
correlations that were not included in published manuscripts. The
final database included 299 effect sizes from 54 independent
samples with a total sample size of 26,592.
3.2 |Coding procedures
Data pertaining to effect size, sample size, independent variable
reliability, and dependent variable reliability were independently
extracted by three coders, with one coder naïve to the hypotheses of
this study. The coders agreed on 1,001 (84%) of the 1,196 coding
assignments. In addition, consistent with previous meta‐analyses of
this type (e.g., Colquitt et al., 2013), the ICC(2) form of the intraclass
correlation was computed as a secondary assessment of interrater
reliability. More specifically, absolute agreement was calculated using
a two‐way random model in SPSS version 22. This yielded an ICC(2)
of .97 and a Pearson correlation of .94, exceeding the recommended
.70 threshold for an acceptable ICC(2) value (Bliese, 2000). All coding
discrepancies were resolved with the assistance of the first author.
For cross‐sectional data, we used the effect size provided in the
correlation table. For daily diary studies, we used the between‐person
correlation because the between‐person information in a diary study is
the mean variance across all days (not partitioning the variance for
each day), or the general amount of variance for each person, which
is what the cross‐sectional data examines. Reliability information
pertaining to each endogenous and exogenous variable was derived
from each respective correlation matrix or in the article text. If
reliability was reported on a daily level but correlations were reported
as the average over multiple days (e.g., Sonnentag & Bayer, 2005), the
mean reliability value was used because we were interested in the
between‐person correlations. If data from multiple studies were
presented, these results were coded as independent samples (e.g., a
Spanish sample and a Dutch sample, Salanova & Schaufeli, 2008).
In cases when outcomes were measured several times each day
(e.g., exhaustion at work and exhaustion at bedtime; Demerouti,
Bakker, Sonnentag, & Fullagar, 2012), we adhered to our coding rule
and used the effect size that temporally followed the recovery
experience (in this case, exhaustion at bedtime).
Given that the focus of our study is on after‐work recovery
experiences, our decision rules maintained that the effect size most
closely linked in time to the recovery experience was retained. For
example, only the measure of exhaustion at bedtime would be used
as this would capture exhaustion after evening recovery experi-
ences. We reverse coded when necessary to produce effect sizes
with consistent meaning (e.g., “inability to psychologically detach”
was reverse coded; Taris, Geurts, Schaufeli, Blonk, & Lagerveld,
2008). Last, only self‐reported data are included in this study.
Although other‐reported recovery experiences (Sonnentag et al.,
2010) and resource outcomes (Booth, 2011) were available in one
case each, other‐reported data were excluded from the present
analyses because a lack of data prohibited us from testing
other‐reported versus self‐reported moderating effects. Further-
more, we felt that including these effect sizes could potentially
distort the results.
3.3 |Meta‐analysis and path analysis procedures
This study calculated the meta‐analytic correlation between all
variables using the random effects model of meta‐analysis (Hunter &
Schmidt, 2004), which can correct for artifactual variance and, thus,
provide a more conservative effect size estimate (Kepes, McDaniel,
Brannick, & Banks, 2013). Independent effect sizes were used to
calculate the sample size‐weighted mean correlation (−
r). Corrections
for unreliability in the independent and dependent variables were
made using reliability information presented in each primary study.
As such, a corrected point estimate (ρ) and standard deviation (σ
ρ
)
of the population estimate are included, as well as the 95% confidence
interval (95% CI) around this corrected correlation. To check that
the published literature was not systematically unrepresentative of
the population, publication bias analyses were completed using trim‐
and‐fill, Egger0s regression, and funnel plot methodologies (Kepes,
Banks, McDaniel, & Whetzel, 2012). To test hypotheses and create
a comprehensive model of work characteristics, recovery experiences,
and outcomes, we used AMOS 22 software (SPSS Inc., 2013) to
perform a MASEM using the meta‐analytic correlation matrix with
corrected correlations and the harmonic mean of the sample size. In
this analysis, the harmonic mean of 2,410 is smaller than the average
mean of 4,057 and is suggestive of a more conservative approach
(Viswesvaran & Ones, 1995). We also followed recent recommenda-
tions for these analyses in the organizational sciences (e.g., Bergh
et al., 2014; Landis, 2013).
2
We based this coding decision by reading sample items (e.g., “Today I had con-
flicts with other people”; Feuerhahn, Sonnentag, & Woll, 2014), and recognizing
it was similar to conflict at work.
266 BENNETT ET AL.
4|RESULTS
4.1 |Work‐related antecedents of recovery
experiences
Table 1 presents the meta‐analytic relationships between work‐
related antecedents and recovery experiences. Hypothesis 1 pro-
posed that challenge demands have a positive relationship with
mastery experiences but a negative relationship with all other recov-
ery experiences. As expected, challenge demands have a negative
relationship with psychological detachment (ρ=−.37), relaxation
(ρ=−.30), and control (ρ=−.28) recovery experiences. For a more
rigorous test of hypotheses, we also examined the path weights and
95% CIs from the MASEM. To examine this, we organized the
corrected correlations between all variables into a correlation matrix,
as shown in Table 2, and created a structural equation model using
AMOS 22, as shown in Figure 1. Table 3 presents the meta‐analytic
regression weights of antecedents and recovery experiences. We
focus on regression weights rather than on corrected correlations
because correlations examine the singular relationships between
two variables, whereas the regression weights provide an effect size
accounting for the relationships of all other work characteristics,
recovery experiences, and outcomes. Nonoverlapping 95% CIs
suggest that the relationships are significantly different. For challenge
demands, the negative relationships are the same when examining
regression weights. However, challenge demands have no statistical
relationship with mastery recovery experiences, as the 95% CI
includes zero. Thus, Hypotheses 1a, 1b, and 1d are supported, and
1c is rejected.
Hypothesis 2 proposed a negative relationship between hindrance
demands and all recovery experiences. Meta‐analytic correlations
show that hindrance demands have a negative relationship with
psychological detachment (ρ=−.21), relaxation (ρ=−.15), and control
(ρ=−.12), but a nonsignificant relationship with mastery experiences
(ρ= .02) as the 95% CI includes zero. An examination of the regression
weights in Table 3 confirms that hindrance demands have a negative
relationship with psychological detachment (β=−.11) and relaxation
(β=−.05), supporting H2a and H2b. However, hindrance demands
have a positive relationship with mastery experiences (β= .06), and
the 95% CI for the relationship between hindrance demands and con-
trol experiences includes zero. Thus, both H2c and H2d are rejected.
Hypothesis 3 posited that job resources are negatively related to
psychological detachment and relaxation and positively related to
mastery and control. As expected, job resources are positively
correlated with mastery (ρ= .17) and control (ρ= .21) recovery
experiences. Unexpectedly, job resources are also positively related
to relaxation (ρ= .08). The relationship with psychological detachment
(ρ= .04) is not statistically different from zero (95% CI includes zero).
The regression weights in Table 3 confirm that job resources are posi-
tively related to evening relaxation, mastery, and control experiences.
In addition, job resources have a statistically stronger positive relation-
ship with after‐work control and mastery experiences than relaxation.
Thus, Hypothesis 3a and 3b are rejected, and 3c and 3d are supported.
4.2 |Outcomes of recovery experiences
Table 4 presents the meta‐analytic results of recovery experiences
with outcomes. Theoretically, higher evening recovery experiences
TABLE 1 Meta‐analytic correlations between work‐related antecedents and recovery experiences
Antecedents kN r
_ρSDρ
95% CI 80% CR %
SE χ
2
Lower Upper Lower Upper
Challenge demands
Psychological detachment 14 3,463 −.30 −.37 .13 −.37 −.24 −.45 −.16 21 66.4*
Relaxation 4 821 −.23 −.30 .19 −.40 −.06 −.44 −.02 14 28.0*
Mastery 4 821 .01 .01 .00 −.03 .05 .01 .01 100 1.5
Control 3 745 −.21 −.28 .11 −.35 −.08 −.35 −.08 25 11.9*
Hindrance demands
Psychological detachment 11 5,602 −.18 −.21 .07 −.22 −.13 −.25 −.10 38 29.3*
Relaxation 3 4,451 −.12 −.15 .00 −.13 −.11 −.12 −.12 100 0.5
Mastery 3 4,451 .02 .02 .04 −.04 .07 −.03 .06 34 8.9
Control 3 4,451 −.10 −.12 .00 −.12 −.09 −.10 −.10 100 0.7
Job resources
Psychological detachment 11 5,291 .03 .04 .03 −.01 .06 .00 .06 83 13.2
Relaxation 3 3,065 .06 .08 .00 .05 .08 .06 .06 100 0.6
Mastery 3 3,065 .14 .17 .00 .11 .17 .14 .14 100 2.0
Control 3 3,065 .17 .21 .00 .17 .17 .17 .17 100 0.1
Note.%SE is the percentage of variance accounted for by sampling error; 80% CR lower = lower limit 80% credibility interval of the corrected estimate of the
population; 80% CR upper = upper limit 80% credibility interval of the corrected estimate of the population; 95% CI lower = lower limit of 95% confidence
interval of the corrected estimate of the population; 95% CI upper = upper limit of 95% confidence interval of the corrected estimate of the population;
k= number of independent effect sizes; N= sample size; r
_= mean sample size‐weighted correlation; ρ= corrected point estimate of the population
(correcting for measurement error); SD
p
= standard deviation of corrected population point estimate; χ
2
= a chi‐square test for the variance remaining
unaccounted for.
*p< .01.
BENNETT ET AL.267
TABLE 2 Meta‐analytic correlation table with work antecedents, recovery experiences, and outcomes
Variable 1 2 3 4 5 6 7 8
1. Challenge demands (ρ)
2. Hindrance demands (ρ) .29
k(N) 4 (1,454)
3. Job resources (ρ) .03 −.19
k(N) 8 (3,118) 4 (3,907)
4. Psychological detachment (ρ)−.37 −.21 .04
k(N) 14 (3,463) 11 (5,602) 11 (5,291)
5. Relaxation (ρ)−.30 −.15 .08 .62
k(N) 4 (821) 3 (4,451) 3 (3,065) 11 (2,413)
6. Mastery (ρ) .01 .02 .17 .14 .33
k(N) 4 (821) 3 (4,451) 3 (3,065) 9 (2,270) 9 (2,270)
7. Control (ρ)−.28 −.12 .21 .41 .65 .39
k(N) 3 (745) 3 (4,451) 3 (3,065) 8 (2,104) 8 (2,104) 8 (2,104)
8. Fatigue (ρ) .35 .46 −.24 −.39 −.35 −.18 −.30
k(N) 15 (8,642) 10 (10,083) 19 (13,674) 17 (4,164) 10 (2,135) 9 (2,066) 8 (1,900)
9. Vigor (ρ) .19 −.09 .39 .14 .24 .29 .31 −.42
k(N) 6 (4,275) 10 (9,161) 15 (10,843) 11 (2,519) 6 (1,663) 5 (1,589) 4 (1,423) 14 (10,883)
Note.k= number of independent effect sizes; N= combined sample size used to determine meta‐analytic correlation between variables; ρ= corrected
correlation of the population (correcting for measurement error).
Psychological
detachment
Relaxation
Mastery
Control
Fatigue
Vig or
Challenge
stressors
Job
resources
Hindrance
stressors -.28**
.16** .34**
-.15**
.31**
-.07**
.30**
-.01*
-.11**
-.08**
-.18**
.08**
.13**
.08**
.19**
-.34**
-.11**
.03
-.29**
-.05*
.08**
-.01
.06**
.18**
-.29**
.01
.22**
.29**
-.19**
.03
FIGURE 1 Partially mediated meta‐analytic structural equation model. Note: Harmonic mean = 2,408. Values represent standardized regression
weights. ** (and solid lines) are statistically significant paths at p< .01 level, * (and dotted lines) are statistically significant paths at p< .05 level.
Dashed lines are paths at p> 0.05 level. Model includes paths between recovery experiences, but for visual clarity this figure does not include these
regression weights
268 BENNETT ET AL.
should be associated with lower fatigue. As expected, all four recovery
experiences have a negative correlation with after‐work fatigue.
However, we also sought to examine whether certain recovery
experiences are more negatively associated with fatigue. Hypothesis 4
proposed that (a) psychological detachment and (b) relaxation
experiences would have stronger negative correlations with fatigue
than would mastery and control experiences. An examination of the
corrected correlations (see Table 4) indicates that psychological
detachment (ρ=−.39) and relaxation (ρ=−.35) recovery experiences
do have stronger relationships than control (ρ=−.30) and mastery
(ρ=−.18) experiences. Examining regression weights and 95% CIs in
Table 5, we find that psychological detachment has a stronger negative
relationship (β=−.18) with fatigue than relaxation and control experi-
ences but shows no statistical difference with mastery. Relaxation
experiences do not have a stronger relationship with fatigue than the
other recovery experiences. Thus, H4a is partially supported and H4b
is rejected.
After‐work recovery experiences are also expected to have a
positive relationship with vigor. Hypothesis 5 proposed that (a)
mastery and (b) control experiences would have a stronger positive
correlation with vigor than psychological detachment or relaxation
experiences. Results show that control (ρ= .31) and mastery (ρ= .29)
experiences present higher corrected correlations with vigor in the
evening than relaxation (ρ= .24) and psychological detachment
(ρ= .14) experiences. Focusing on regression weights (Table 5), control
experiences (β= .19) have 95% CIs that do not overlap with
detachment or relaxation, but the CIs of mastery experience overlap
with all other recovery experiences. Thus, Hypothesis 5a is rejected
and 5b is supported.
TABLE 4 Meta‐analytic correlations between recovery experiences and outcomes
Outcomes kN r
_ρSDρ
95% CI 80% CR %
SE χ
2
Lower Upper Lower Upper
Fatigue
Psychological detachment 17 4,164 −.34 −.39 .20 −.43 −.25 −.26 −.12 10 177.1*
Relaxation 10 2,135 −.30 −.35 .22 −.42 −.19 −.53 −.08 11 89.9*
Mastery 9 2,066 −.15 −.18 .05 −.20 −.10 −.20 −.10 70 12.8
Control 8 1,900 −.26 −.30 .14 −.35 −.16 −.41 −.10 21 38.3*
Vigor
Psychological detachment 11 2,519 .12 .14 .09 .06 .17 .02 .23 38 29.2*
Relaxation 6 1,663 .21 .24 .08 .14 .28 .12 .29 42 14.3
Mastery 5 1,589 .25 .29 .05 .19 .31 .19 .31 57 14.3
Control 4 1,423 .26 .31 .00 .25 .28 .26 .26 100 0.5
Note.%SE is the percentage of variance accounted for by sampling error; 80% CR lower = lower limit 80% credibility interval of the corrected estimate of the
population; 80% CR upper = upper limit 80% credibility interval of the corrected estimate of the population; 95% CI lower = lower limit of 95% confidence
interval of the corrected estimate of the population; 95% CI upper = upper limit of 95% confidence interval of the corrected estimate of the population;
k= number of independent effect sizes; N= sample size; r
_= mean sample size‐weighted correlation; ρ= corrected point estimate of the population
(correcting for measurement error); SD
p
= standard deviation of corrected population point estimate; χ
2
= a chi‐square test for the variance remaining
unaccounted for.
*p< .01.
TABLE 5 Meta‐analytic regression weights predicting well‐being
outcomes
Fatigue Vigor
β
95% CI
β
95% CI
Lower Upper Lower Upper
Psychological detachment −.18 −.22 −.14 .08 .02 .14
Relaxation −.08 −.13 −.03 .08 .03 .13
Mastery −.11 −.15 −.08 .13 .09 .17
Control −.01 −.06 .04 .19 .14 .24
Challenge stressors .16 .13 .20 .31 .27 .35
Hindrance stressors .34 .31 .37 −.07 −.11 −.04
Job resources −.15 −.18 −.12 .30 .27 .34
Note. 95% CI lower = lower limit of 95% confidence interval; 95% CI
upper = upper limit of 95% confidence interval.
TABLE 3 Meta‐analytic regression weights and confidence intervals predicting recovery experiences
Psychological detachment Relaxation Mastery Control
β
95% CI
β
95% CI
β
95% CI
β
95% CI
Lower Upper Lower Upper Lower Upper Lower Upper
Challenge demands −.34 −.38 −.30 −.29 −.33 −.25 −.01 −.05 .03 −.29 −.33 −.25
Hindrance demands −.11 −.15 −.07 −.05 −.09 −.01 .06 .02 .10 .01 −.03 .05
Job resources .03 −.01 .07 .08 .04 .12 .18 .14 .22 .22 .18 .26
Note. 95% CI lower = lower limit of 95% confidence interval; 95% CI upper = upper limit of 95% confidence interval.
BENNETT ET AL.269
4.3 |Comprehensive model
Hypothesis 6 predicted that evening recovery experiences would
partially mediate the relationship between work characteristics and
after‐work outcomes. This model is shown in Figure 1. To test this
hypothesis, we used 1,000 bootstrap samples with a 95% CI and exam-
ined the statistical significance of indirect paths when the direct paths
were also modeled. As shown in Table 6, challenge demands and job
resources have statistically significant indirect and direct paths with
both fatigue and vigor. Interestingly, challenge demands have a posi-
tive direct relationship with vigor, but the total effect is lower because
challenge demands also have a negative indirect effect through the
negative relationship with most recovery experiences. Hindrance
demands have statistically significant direct and indirect paths with
fatigue, but only the direct path with vigor is statistically significant.
Overall, Hypothesis 6 is mostly supported. In addition, the indirect
effects of challenge demands were stronger than were hindrance
demands or job resources, mirroring the regression weights showing
that challenge demands have stronger associations with psychological
detachment and relaxation recovery experiences.
The last main objective of this paper was to examine the additional
impact of after‐work recovery experiences. Figure 1 shows how both
work characteristics and after‐work recovery experiences are related
to well‐being. Figure 2 provides the variance explained in outcomes
of a model with just antecedents, a model with just recovery experi-
ences, and then the partially hypothesized model. By adding recovery
experiences to the model, there is a 26% increase in variance explained
TABLE 6 Tests of mediation
Fatigue Vigor
Total effect Direct effect Indirect effect Total effect Direct effect Indirect effect
Work characteristics
Challenge demands .25 .16* .09* .20 .31* −.11*
Hindrance demands .35 .33* .02* −.08 −.07* −.01
Job resources −.18 −.15* −.03* .37 .30* .07*
Recovery experiences
Psychological detachment −.18 −.18* .08 .08*
Relaxation −.08 −.08* .08 .08*
Mastery −.11 −.11* .13 .13*
Control −.01 −.01 .19 .19*
*p< .01.
Fatigue
Vig or
Job
characteristics
Recovery
experiences
R2= .294
R2= .189
R2= .184
R2= .131
Model 1: Daily energy variance explained by perceptions of job characteristics
(challenge stressors, hindrance stressors, & job resources)
Model 2: Daily energy variance explained by recovery experiences (psychological
detachment, relaxation, mastery, & control)
Model 3: Daily energy variance explained from partial mediation model of perceptions
of job characteristics mediated by recovery experiences
Fatigue
Vig or
Job
characteristics
Recovery
experiences
Fatigue
R2= .369
R2= .307
Vig or
FIGURE 2 Variance explained from models of
work characteristics and recovery experiences
270 BENNETT ET AL.
in fatigue (ΔR
2
= .08) and a 62% increase in variance explained in vigor
(ΔR
2
= .12). Both values are practically and statistically significant.
4.4 |Publication bias analyses
Evidence for publication bias was analyzed using funnel plots, Egger0s
regression tests, and trim‐and‐fill techniques when at least 10 effect
sizes were presented for a meta‐analytic correlation (Kepes et al.,
2013). Analyses were conducted according to recommendations for
organizational research (Kepes et al., 2012). Table S7 provides a
summary of these findings, with additional figures and data available
by request from the first author. Egger0s regression tests indicate that
three relationships potentially have evidence of publication bias:
hindrance demands and psychological detachment, job resources and
fatigue, and job resources and vigor. The funnel plot and trim‐and‐fill
techniques suggest that only the relationship between hindrance
demands and psychological detachment could be smaller in magnitude,
but the adjusted effect size is not practically significant (Δr
_= .02).
When using trim‐and‐fill techniques and imputing studies on both
the left and right sides of the observed effect size, several of the
imputations showed that the relationships between variables might
be larger rather than smaller. For example, adjusted correlations for
the psychological detachment–fatigue and relaxation–fatigue relations
are larger than are the values presented in the meta‐analytic
correlation matrix (Δr
_= .07), and thus the value used to create the path
model is a more conservative estimate. Only the adjusted effect size of
job resources and fatigue has a practically significant smaller value
(Δr
_= .07) that could indicate the presence of publication bias.
However, the direction of this relationship does not change, and given
the multitude of other variables in our model, we feel confident that a
downward adjustment of this relationship would not impact our overall
conclusions regarding recovery experiences in a meaningful way.
5|DISCUSSION
This paper focused on employee after‐work recovery experiences
(i.e., psychological detachment, relaxation, mastery, and control
experiences; Sonnentag & Fritz, 2007). Whereas previous meta‐
analyses have focused on work demands and resources with well‐being
(e.g., Crawford et al., 2010; Halbesleben, 2010; Lee & Ashforth, 1996),
our objective was to provide a meta‐analytic examination of after‐work
recovery experiences, extend theory regarding JD‐R and recovery
experiences (Kinnunen et al., 2011) utilizing the challenge–hindrance
framework, and integrate our findings to see how the combination of
work characteristics and recovery experiences collectively and uniquely
influence employee well‐being. The antecedents of interest were
hindrance demands, challenge demands, and job resources. The
outcomes of interest were fatigue and vigor, the vital and proximal
outcomes of after‐work recovery experiences.
This study provides multiple contributions to scholars and
practitioners. First, we created a comprehensive model integrating
recovery experiences within models of work characteristics and well‐
being. The partial mediation model is in line with theory and item
wording focusing on recovery experiences occurring in time after the
work period. Regarding after‐work outcomes, we found that challenge
demands and hindrance demands both have a positive relationship
with fatigue, whereas hindrance demands have a negative relationship
with vigor, and challenge demands have a positive relationship with
vigor. More interesting is how these antecedents relate to recovery
experiences. Challenge demands have stronger negative relationships
with psychological detachment, relaxation, and control recovery
experiences than hindrance demands. The unexpected findings that
challenge demands have no statistical relationship with mastery
experiences and that hindrance demands have a positive relationship
with mastery experiences are worthy of future research. We are
intrigued that a perceived positive challenge in one domain has no
impact on a different domain, yet a hindrance in the work domain
potentially creates the need to experience positive challenges at home.
Consistent with the latter finding, Petrou and Bakker (2016) found
that employees who experienced high job strain during certain
workweeks were more likely to engage in leisure crafting during those
weeks (i.e., shape their leisure activities in a way that addresses their
passions and values). Although challenge demands have a positive
direct relationship with vigor, the positive direct relationship with
fatigue coupled with stronger negative path coefficients with most
recovery experiences indicates that individuals perceiving higher
challenge demands may not engage in the recovery experiences
associated with lower fatigue. Therefore, we contend that individuals
perceiving their work as positively challenging should also be aware
of the need to engage in recovery experiences. The finding that
hindrance demands have weaker negative relationships with detach-
ment and relaxation than have challenge demands, and show a positive
relationship with mastery, indicates that hindrances are associated
with nonwork interference differently and that future research should
continue to explore these mechanisms (e.g., Wood & Michaelides,
2016).
Second, we addressed inconsistent conclusions in the literature
regarding after‐work recovery experiences with outcomes. Psycholog-
ical detachment experiences have a stronger negative relationship with
fatigue than relaxation and control. Control experiences have a
stronger positive relationship with vigor than detachment and
relaxation. Although psychological detachment is the most frequently
researched recovery experience (Sonnentag & Fritz, 2015; also see
Wendsche & Lohmann‐Haislah, 2017, for a meta‐analysis), these
findings highlight that each recovery experience has a unique
relationship with outcomes. We suggest that future research continue
to explore how multiple recovery experiences may be used in
combination to provide beneficial outcomes, such as high control and
relaxation experiences working together to reduce fatigue during a
lunch break (Trougakos, Hideg, Cheng, & Beal, 2014).
Third, we assessed the practical value of studying after‐work
recovery experiences. Adding recovery experiences to a model with
work characteristics had a considerable impact on well‐being
outcomes, as these additional variables in a partial mediation model
explain 26% more variance in fatigue (ΔR
2
= .08) and 62% more
variance in vigor (ΔR
2
= .12). Clearly, both work characteristics and
after‐work recovery experiences influence employee well‐being.
Future research should continue to focus on work and after‐work
experiences collectively.
BENNETT ET AL.271
From a practical perspective, these results provide strong support
that recovery experiences during nonwork time are beneficial to
employees, and that it is necessary for employees to have different
types of recovery experiences. Organizations should consider offering
training so that employees can monitor and positively alter their
recovery experiences (e.g., Hahn, Binnewies, Sonnentag, &
Mojza, 2011). In addition, some companies have started turning off
e‐mail servers after work hours to help employees psychologically
detach from work by limiting work e‐mails during nonwork hours
(Vasagar, 2013). Further exploration of organizational policies that help
individuals engage in recovery experiences would be very interesting
from both scholarly and practical perspectives.
5.1 |Limitations and future research
There are several limitations in this study that can lead to future
research opportunities. First, all correlations coded from each study
are from self‐reported evaluations of work, recovery experiences,
and outcomes, potentially creating same‐source bias. However, an
individual evaluation of his or her work and well‐being may be
more appropriate than asking others to assess the feelings of the
target employee. In addition, we only focused on pleasant activation
(e.g., vigor) and unpleasant deactivation (e.g., fatigue), but not other
aspects of core affect (Yik et al., 2011) that have also been studied in
recovery research such as negative activated affect (e.g., anxiety and
tension). Second, the mediation model is consistent with theory, daily
time use, and item wording, such that an individual has work
experiences, then after‐work recovery experiences, which both are
associated with well‐being. Diary studies are perhaps less biased than
are regular surveys because the time between the subjective reports
and the actual experiences is limited to only a few hours. Unfortu-
nately, very few of the studies included in this analysis had a daily diary
design that measured each variable over these different periods, and
we were unable to create a true causal model (only 11 of the 36 effect
sizes in the correlation matrix contain data from at least one daily diary
study). Third, we limited our separation of work characteristics to
well‐studied categorizations. Future research may explore the
separation of work demands into threats, hindrances, and challenges
(e.g., Tuckey et al., 2015) along with differentiating resources into job
resources and personal resources (Tadićet al., 2015). Fourth, this study
modeled between‐person differences in recovery experience. Future
research should continue to explore within‐person variations in well‐
being from day to day. Last, recent evidence suggests that only some
organizational interventions have a practical impact on employee
well‐being (Maricuţoiu, Sava, & Butta, 2014). Future research should
compare both at‐work and after‐work recovery interventions.
6|CONCLUSION
This study provides a quantitative summary of after‐work recovery
experiences and refines existing models of work characteristics to
include how recovery experiences are associated with well‐being.
Results confirm previous relationships of challenge demands,
hindrance demands, and job resources with well‐being outcomes.
However, we found that these work antecedents have unique
relationships with after‐work recovery experiences. Adding recovery
experiences as partial mediators between work characteristics and
well‐being explains a significant amount of additional variance and,
thus, improves our understanding of how individuals can alter their
feelings of fatigue and vigor.
ACKNOWLEDGEMENTS
The authors would like to thank Sabine Sonnentag and three reviewers
for their constructive feedback, Melissa Virtue for her help coding
studies, as well as Mike McDaniel and Doug Pugh for their feedback
on initial versions of this manuscript.
ORCID
Andrew A. Bennett http://orcid.org/0000-0003-1991-3611
James G. Field http://orcid.org/0000-0001-8487-6648
REFERENCES
References marked with an asterisk (*) indicate studies included in the
meta‐analysis.
Armon, G., Melamed, S., & Shirom, A. (2012). The relationship of the Job
Demands–Control–Support model with vigor across time: Testing for
reciprocality. Applied Psychology: Health and Well‐Being,4, 276–298.
https://doi.org/10.1111/j.1758‐0854.2012.01074.x
Bakker, A. B., & Demerouti, E. (2007). The Job Demands–Resources model:
State of the art. Journal of Managerial Psychology,22, 309–328. https://
doi.org/10.1108/02683940710733115
Bakker, A. B., & Demerouti, E. (2014). Job Demands–Resources theory. In
P. Y. Chen, & C. L. Cooper (Eds.), Wellbeing: A complete reference guide
(Volume III (pp. 37–64)John Wiley & Sons. https://doi.org/10.1002/
9781118539415.wbwell019
*Bakker, A. B., Emmerik, H. V., & Euwema, M. C. (2006). Crossover of
burnout and engagement in work teams. Work and Occupations,33,
464–489. https://doi.org/10.1177/0730888406291310
*Bakker, A. B., Hakanen, J. J., Demerouti, E., & Xanthopoulou, D. (2007).
Job resources boost work engagement, particularly when job demands
are high. Journal of Educational Psychology,99, 274–284. https://doi.
org/10.1037/0022‐0663.99.2.274
Bakker, A. B., & Xanthopoulou, D. (2009). The crossover of daily work
engagement: Test of an actor–partner interdependence model. Journal
of Applied Psychology,94, 1562–1571. https://doi.org/10.1037/
a0017525
*Beckers, D. G. J., van der Linden, D., Smulders, P. G. W., Kompier, M. A. J.,
van Veldho ven, M. J. P. M., & van Yperen, N. W. (2004). Working over-
time hours: Relations with fatigue, work motivation, and the quality of
work. Journal of Occupational and Envinronmental Medicine,46, 1282–
1289. https://doi.org/10.1097/01.jom.0000147210.95602.50
Bennett, A. A., Gabriel, A. S., Calderwood, C., Dahling, J. J., & Trougakos, J.
P. (2016). Better together? Examining profiles of employee recovery
experiences. Journal of Applied Psychology,101, 1635–1654. https://
doi.org/10.1037/apl0000157
Bergh, D. D., Aguinis, H., Heavey, C., Ketchen, D. J., Boyd, B. K., Su, P., …
Joo, H. (2014). Using meta‐analytic structural equation modeling to
advance strategic management research: Guidelines and an empirical
illustration via the strategic leadership–performance relationship.
Strategic Management Journal.. https://doi.org/10.1002/smj.2338
Bliese, P. D. (2000). Within‐group agreement, non‐independence, and
reliability: Implications for data aggregation and analysis. In K. J. Klein,
& S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in
organizations: Foundations, extensions, and new directions (pp. 349–381).
San Francisco: Jossey‐Bass.
272 BENNETT ET AL.
*Bono, J. E., Glomb, T. M., Shen, W., Kim, E., & Koch, A. J. (2013). Building
positive resources: Effects of positive events and positive reflection on
work‐stress health. Academy of Management Journal,56, 1601–1627.
https://doi.org/10.5465/amj.2011.0272
Booth, S. M. (2011). Family supportive organization perceptions, work role
overload, and burnout: Crossover effects of burnout on recovery.
Baton Rouge: Louisiana State University.
*Boswell, W. R., Olson‐Buchanan, J. B., & LePine, M. A. (2004). Relations
between stress and work outcomes: The role of felt challenge, job
control, and psychological strain. Journal of Vocational Behavior,64,
165–181. https://doi.org/10.1016/S0001‐8791(03)00049‐6
*Bourgeois, L. R. (2011). Gambling as stress recovery? A new perspective on
the stress–gambling relationship. Nova Scotia: Saint Mary0s University,
Halifax.
Brosschot, J. F., Pieper, S., & Thayer, J. F. (2005). Expanding stress
theory: Prolonged activation and perseverative cognition.
Psychoneuroendocrinology,30, 1043–1049. https://doi.org/10.1016/j.
psyneuen.2005.04.008
Cavanaugh, M. A., Boswell, W. R., Roehling, M. V., & Boudreau, J. W.
(2000). An empirical examination of self‐reported work stress among
U.S. managers. Journal of Applied Psychology,85,65–74. https://doi.
org/10.1037//0021‐9010.85.1.65
Colquitt, J. A., Scott, B. A., Rodell, J. B., Long, D. M., Zapata, C. P., Conlon, D.
E., & Wesson, M. J. (2013). Justice at the millennium, a decade later: A
meta‐analytic test of social exchange and affect‐based perspectives.
Journal of Applied Psychology,98, 199–236. https://doi.org/10.1037/
a0031757
Crawford, E. R., LePine, J. A., & Rich, B. L. (2010). Linking job demands and
resources to employee engagement and burnout: A theoretical
extension and meta‐analytic test. Journal of Applied Psychology,95,
834–848. https://doi.org/10.1037/a0019364
Cropley, M., & Zijlstra, F. (2011). Work and rumination. In J. Langan‐Fox, &
C. L. Cooper (Eds.), Handbook of stress in the occupations (pp. 487–502).
Northampton, MA: Edward Elgar Publishing Ltd.
*Demerouti, E., Bakker, A. B., de Jonge, J., Janssen, P. P. M., & Schaufeli, W.
B. (2001). Burnout and engagement at work as functions of demands
and control. Scandinavian Journal of Work, Environment & Health,27,
279–286.
Demerouti, E., Bakker, A. B., Geurts, S. A. E., & Taris, T. W. (2009). Daily
recovery from work‐related effort during non‐work time. In P. L.
Perrewé, & D. C. Ganster (Eds.), Research in occupational stress and
well‐being (Vol. 7) (pp. 85–123)Emerald Group Publishing. https://doi.
org/10.1108/S1479‐3555(2009)0000007006
Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The
Job Demands–Resources model of burnout. Journal of Applied Psychol-
ogy,86, 499–512. https://doi.org/10.1037//0021‐9010.86.3.499
*Demerouti, E., Bakker, A. B., Sonnentag, S., & Fullagar, C. J. (2012). Work
related flow and energy at work and at home: A study on the role of
daily recovery. Journal of Organizational Behavior,33, 276–295.
https://doi.org/10.1002/job.760
*Derks, D., & Bakker, A. B. (2014). Smartphone use, work–home interfer-
ence, and burnout: A diary study on the role of recovery. Applied
Psychology: An International Review,63, 411–440. https://doi.org/
10.1111/j.1464‐0597.2012.00530.x
*Donahue, E. G., Forest, J., Vallerand, R. J., Lemyre, P., Crevier‐Braud, L., &
Bergeron, E. (2012). Passion for work and emotional exhaustion: The
mediating role of rumination and recovery. Applied Psychology: Health
and Well‐Being,4, 341–368. https://doi.org/10.1111/j.1758‐
0854.2012.01078.x
*Feldt, T., Huhtala, M., Kinnunen, U., Hyvönen, K., Mäkikangas, A., &
Sonnentag, S. (2013). Long‐term patterns of effort‐reward imbalance
and over‐commitment: Investigating occupational well‐being and
recovery experiences as outcomes. Work & Stress,27,64–87. https://
doi.org/10.1080/02678373.2013.765670
*Feuerhahn, N., Sonnentag, S., & Woll, A. (2014). Exercise after work, psy-
chological mediators, and affect: A day‐level study. European Journal of
Work and Organizational Psychology,23,62–79. https://doi.org/
10.1080/1359432X.2012.709965
Fritz, C., Lam, C. F., & Spreitzer, G. M. (2011). It0s the little things that mat-
ter: An examination of knowledge workers0energy management.
Academy of Management Perspectives,24,28–39.
*Fritz, C., Yankelevich, M., Zarubin, A., & Barger, P. (2010). Happy, healthy,
and productive: The role of detachment from work during nonwork
time. Journal of Applied Psychology,95, 977–983. https://doi.org/
10.1037/a0019462
*Hahn, V. C., Binnewies, C., Sonnentag, S., & Mojza, E. J. (2011). Learning
how to recover from job stress: Effects of a recovery training program
on recovery, recovery‐related self‐efficacy, and well‐being. Journal of
Occupational Health Psychology,16, 202–216. https://doi.org/
10.1037/a0022169
*Hakanen, J., Rodríguez‐Sánchez, A. M., & Perhoniemi, R. (2012). Too good
to be true? Similarities and differences between engagement and work-
aholism among Finnish judges. Ciencia & Trabajo,14,72–80.
*Hakanen, J. J., Bakker, A. B., & Schaufeli, W. B. (2006). Burnout and work
engagement among teachers. Journal of School Psychology,43, 495–
513. https://doi.org/10.1016/j.jsp.2005.11.001
Halbesleben, J. R. B. (2010). A meta‐analysis of work engagement: Rela-
tionships with burnout, demands, resources, and consequences. In A.
B. Bakker, & M. P. Leiter (Eds.), Work engagement: A handbook of essen-
tial theory and research (pp. 102–117). New York, NY: Psychology Press.
*Hallberg, U. E., Johansson, G., & Schaufeli, W. B. (2007). Type A behavior
and work situation: Associations with burnout and work engagement.
Scandinavian Journal of Psychology,48, 135–142. https://doi.org/
10.1111/j.1467‐9450.2007.00584.x
Hobfoll, S. E. (2011). Conservation of resource caravans and engaged set-
tings. Journal of Occupational and Organizational Psychology,84, 116–
122. https://doi.org/10.1111/j.2044‐8325.2010.02016.x
Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta‐analysis: Correcting
error and bias in research findings (2nd ed.). Thousand Oaks, CA: Sage.
*Jackson, L. T. B., Rothmann, S., & van de Vijver, F. J. R. (2006). A model of
work‐related well‐being for educators in South Africa. Stress and Health,
22, 263–274. https://doi.org/10.1002/smi.1098
Kepes, S., Banks, G. C., McDaniel, M., & Whetzel, D. L. (2012). Publication
bias in the organizational sciences. Organizational Research Methods,
15, 624–662. https://doi.org/10.1177/1094428112452760
Kepes, S., McDaniel, M. A., Brannick, M. T., & Banks, G. C. (2013). Meta‐
analytic reviews in the organizational sciences: Two meta‐analytic
schools on the way to MARS (the Meta‐Analytic Reporting Standards).
Journal of Business and Psychology,28, 123–143. https://doi.org/
10.1007/s10869‐013‐9300‐2
*Kinnunen, U., & Feldt, T. (2013). Job characteristics, recovery experiences
and occupational well‐being: Testing cross‐lagged relationships across
one year. Stress & Health,29, 369–382. https://doi.org/10.1002/
smi.2483
Kinnunen, U., Feldt, T., Siltaloppi, M., & Sonnentag, S. (2011). Job
Demands–Resources model in the context of recovery: Testing
recovery experiences as mediators. European Journal of Work and
Organizational Psychology,20, 805–832.https://doi.org/10.1080/1359
432X.2010.524411
*Kinnunen, U., Mauno, S., & Siltaloppi, M. (2010). Job insecurity, recovery
and well‐being at work: Recovery experiences as moderators. Economic
and Industrial Democracy,31, 179–194. https://doi.org/10.1177/
0143831X09358366
Landis, R. S. (2013). Successfully combining meta‐analysis and structural
equation modeling: Recommendations and strategies. Journal of Busi-
ness and Psychology,28(3), 251–261. https://doi.org/10.1007/
s10869‐013‐9285‐x
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York:
Springer.
BENNETT ET AL.273
Lee, R. T., & Ashforth, B. E. (1996). A meta‐analytic examination of the cor-
relates of the three dimensions of job burnout. Journal of Applied
Psychology,81, 123–133.
LePine, J. A., Podsakoff, N. P., & LePine, M. A. (2005). A meta‐analytic test
of the challenge stressor–hindrance stressor framework: An explana-
tion for inconsistent relationships among stressors and performance.
Academy of Management Journal,48, 764–775. https://doi.org/10.
5465/AMJ.2005.18803921
*Llorens, S., Bakker, A. B., Schaufeli, W., & Salanova, M. (2006). Testing the
robustness of the Job Demands–Resources model. International Journal
of Stress Management,13, 378–391. https://doi.org/10.1037/1072‐
5245.13.3.378
Mäkikangas, A., Kinnunen, S., Rantanen, J., Mauno, S., Tolvanen, A., &
Bakker, A. B. (2014). Association between vigor and exhaustion during
the workweek: A person‐centered approach to daily assessments. Anx-
iety, Stress, & Coping,27, 555–575. https://doi.org/10.1080/
10615806.2013.860968
Maricuţoiu, L. P., Sava, F. a., & Butta, O. (2014). The effectiveness of
controlled interventions on employees0burnout: A meta‐analysis.
Journal of Occupational and Organizational Psychology. https://doi.org/
10.1111/joop.12099
Martinez‐Corts, I., Demerouti, E., Bakker, A. B., & Boz, M. (2015). Spillover
of interpersonal conflicts from work into nonwork: A daily diary study.
Journal of Occupational Health Psychology,20, 326–337. https://doi.
org/10.1037/a0038661
*Mauno, S., Kinnunen, U., & Ruokolainen, M. (2007). Job demands and
resources as antecedents of work engagement: A longitudinal study.
Journal of Vocational Behavior,70, 149–171. https://doi.org/10.1016/
j.jvb.2006.09.002
Meijman, T. F., & Mulder, G. (1998). Psychological aspects of workload. In P.
J. D. Drenth, & C. J. de Wolff (Eds.), Handbook of work and organizational
psychology: Volume 2: Work psychology (pp. 5–33). Hove, England:
Psychology Press.
Michel, A., Turgut, S., Hoppe, A., & Sonntag, K. (2016). Challenge and threat
emotions as antecedents of recovery experiences: Findings from a diary
study with blue‐collar workers. European Journal of Work and Organiza-
tional Psychology.. https://doi.org/10.1080/1359432X.2015.1128414
*Montgomery, A. J., Peeters, M. C. W., Schaufeli, W. B., & Den Ouden, M.
(2003). Work–home interference among newspaper managers: Its
relationship with burnout and engagement. Anxiety, Stress & Coping,
16, 195–211. https://doi.org/10.1080/1061580021000030535
*Moreno‐Jiménez, B., Rodríguez‐Muñoz, A., San‐Vergel, A. I., & Garrosa, E.
(2012). Elucidating the role of recovery experiences in the Job
Demands–Resources model. Spanish Journal of Psychology,15,659–669.
*Nohe, C., Michel, A., & Sonntag, K. (2013). Family–work conflict and job
performance: A diary study of boundary conditions and mechanisms.
Journal of Organizational Behavior,35, 339–357. https://doi.org/
10.1002/job.1878
Petrou, P., & Bakker, A. B. (2016). Crafting one0s leisure time in response to
high job strain. Human Relations,69, 507–529. https://doi.org/
10.1177/0018726715590453
Podsakoff, N. P., LePine, J. A., & LePine, M. A. (2007). Differential challenge
stressor–hindrance stressor relationships with job attitudes, turnover
intentions, turnover, and withdrawal behavior: A meta‐analysis. Journal
of Applied Psychology,92, 438–454. https://doi.org/10.1037/0021‐
9010.92.2.438
*Potok, Y., & Littman‐Ovadia, H. (2014). Does personality regulate the
work stressor–psychological detachment relationship? Journal of Career
Assessment,24,43–58. https://doi.org/10.1177/1069072713487853
*Querstret, D., & Cropley, M. (2012). Exploring the relationship between
work‐related rumination, sleep quality, and work‐related fatigue. Jour-
nal of Occupational Health Psychology,17, 341–353. https://doi.org/
10.1037/a0028552
Quinn, R. W., Spreitzer, G. M., & Lam, C. F. (2012). Building a sustainable
model of human energy in organizations: Exploring the critical role of
resources. Academy of Management Annals,6, 337–396. https://doi.
org/10.1080/19416520.2012.676762
*Richardson, K. M., & Thompson, C. A. (2012). High tech tethers and work–
family conflict: A conservation of resources approach. Engineering Man-
agement Research,1,29–43. https://doi.org/10.5539/emr.v1n1p29
*Safstrom, M., & Hartig, T. (2013). Psychological detachment in the rela-
tionship between job stressors and strain. Behavioral Sciences,3, 418–
433. https://doi.org/10.3390/bs3030418
*Salanova, M., & Schaufeli, W. B. (2008). A cross‐national study of work
engagement as a mediator between job resources and proactive behav-
iour. The International Journal of Human Resource Management,19, 116–
131. https://doi.org/10.1080/09585190701763982
*Sanz‐Vergel, A. I., Demerouti, E., Bakker, A. B., & Moreno‐Jimenez, B.
(2011). Daily detachment from work and home: The moderating effect
of role salience. Human Relations,64, 775–799. https://doi.org/
10.1177/0018726710393368
*Sanz‐Vergel, A. I., Sebastián, J., Rodríguez‐Muñoz, A., Garrosa, E., Moreno‐
Jiménez, B., & Sonnentag, S. (2010). Adaptation of the recovery experi-
ence questionnaire in a Spanish sample. Psicothema,22, 990–996.
*Schaufeli, W. B., & Bakker, A. B. (2004). Job demands, job resources, and
their relationship with burnout and engagement: A multi‐sample study.
Journal of Organizational Behavior,25, 293–315. https://doi.org/
10.1002/job.248
*Schaufeli, W. B., Taris, T. W., & van Rhenen, W. (2008). Workaholism, burn-
out, and work engagement: Three of a kind or three different kinds of
employee well‐being? Applied Psychology: An International Review,57(2),
173–203. https://doi.org/10.1111/j.1464‐0597.2007.00285.x
*Shimazu, A., Sonnentag, S., Kunota, K., & Kawakai, N. (2012). Validation of
the Japanese version of the recovery experience questionnaire. Journal
of Occupational Health,54(3), 196–205.
Siltaloppi, M., Kinnunen, U., & Feldt, T. (2009). Recovery experiences as
moderators between psychosocial work characteristics and occupa-
tional well‐being. Work & Stress,23, 330–348. https://doi.org/
10.1080/02678370903415572
Sonnentag, S. (2001). Work, recovery activities, and individual well‐being: A
diary study. Journal of Occupational Health Psychology,6, 196–210.
https://doi.org/10.10377/1076‐8998.6.3.1
Sonnentag, S. (2012). Psychological detachment from work during leisure
time: The benefits of mentally disengaging from work. Current Direc-
tions in Psychological Science,21, 114–118. https://doi.org/10.1177/
0963721411434979
Sonnentag, S. (2015). Dynamics of well‐being. Annual Review of Organiza-
tional Psychology and Organizational Behavior,2, 261–293. https://doi.
org/10.1146/annurev‐orgpsych‐032414‐111347
*Sonnentag, S., & Bayer, U. (2005). Switching off mentally: Predictors and
consequences of psychological detachment from work during off‐job
time. Journal of Occupational Health Psychology,10, 393–414. https://
doi.org/10.1037/1076‐8998.10.4.393
*Sonnentag, S., Binnewies, C., & Mojza, E. J. (2008). “Did you have a nice
evening?”A day‐level study on recovery experiences, sleep, and affect.
Journal of Applied Psychology,93, 674–684. https://doi.org/10.1037/
0021‐9010.93.3.674
*Sonnentag, S., Binnewies, C., & Mojza, E. J. (2010). Staying well and
engaged when demands are high: The role of psychological detach-
ment. Journal of Applied Psychology,95, 965–976. https://doi.org/
10.1037/a0020032
*Sonnentag, S., & Fritz, C. (2007). The recovery experience questionnaire:
Development and validation of a measure for assessing recuperation
and unwinding from work. Journal of Occupational Health Psychology,
12, 204–221. https://doi.org/10.1037/1076‐8998.12.3.204
Sonnentag, S., & Fritz, C. (2015). Recovery from job stress: The stressor‐
detachment model as an integrative framework. Journal of Organiza-
tional Behavior,36, S72–S103. https://doi.org/10.1002/job.1924
Sonnentag, S., & Geurts, S. A. E. (2009). Methodological issues in recovery
research. In P. L. Perrewé, & D. C. Ganster (Eds.), Research in
274 BENNETT ET AL.
occupational stress and well‐being (Vol. 7) (pp. 1–36)Emerald Group Pub-
lishing. https://doi.org/10.1108/S1479‐3555(2009)0000007004
*Sonnentag, S., & Kruel, U. (2006). Psychological detachment from work
during off‐job time: The role of job stressors, job involvement, and
recovery‐related self‐efficacy. European Journal of Work and Organiza-
tional Psychology,15, 197–217. https://doi.org/10.1080/
13594320500513939
*Sonnentag, S., Kuttler, I., & Fritz, C. (2010). Job stressors, emotional
exhaustion, and need for recovery: A multi‐source study on the benefits
of psychological detachment. Journal of Vocational Behavior,76, 355–
365. https://doi.org/10.1016/j.jvb.2009.06.005
*Sonnentag, S., Mojza, E., Binnewies, C., & Scholl, A. (2008). Being engaged
at work and detached at home: A week‐level study on work engage-
ment, psychological detachment, and affect. Work & Stress,22, 257–
276. https://doi.org/10.1080/02678370802379440
*Sonnentag, S., Unger, D., & Nägel, I. J. (2013). Workplace conflict and
employee well‐being: The moderating role of detachment from work
during off‐job time. International Journal of Conflict Management,24,
166–183. https://doi.org/10.1108/10444061311316780
*Sonnentag, S., & Zijlstra, F. R. H. (2006). Job characteristics and off‐job
activities as predictors of need for recovery, well‐being, and fatigue.
Journal of Applied Psychology,91, 330–350. https://doi.org/10.1037/
0021‐9010.91.2.330
SPSS Inc (2013). AMOS 22.0. Chicago: SPSS Inc.
*Stevens, S. (2010). Understanding how employees unwind after work:
Expanding the construct of “Recovery”. Halifax, Nova Scotia: Saint Mary0s
University.
Stone, A. A., Kennedy‐Moore, E., & Neale, J. M. (1995). Association
between daily coping and end‐of‐day mood. Health Psychology,14,
341–349. https://doi.org/10.1037/0278‐6133.14.4.341
Tadić, M., Bakker, A. B., & Oerlemans, W. G. M. (2015). Challenge versus
hindrance job demands and well‐being: A diary study on the moderat-
ing role of job resources. Journal of Occupational and Organizational
Psychology,88, 702–725. https://doi.org/10.1111/joop.12094
TadićVujčić, M., Oerlemans, W. G., & Bakker, A. B. (2017). How challenging
was your work today? The role of autonomous work motivation.
European Journal of Work and Organizational Psychology.26,81–93.
https://doi.org/10.1080/1359432X.2016.1208653
*Taris, T. W., Geurts, S. A. E., Schaufeli, W. B., Blonk, R. W. B., & Lagerveld,
S. E. (2008). All day and all of the night: The relative contribution of
two dimensions of workaholism to well‐being in self‐employed
workers. Work & Stress,22(2), 153–165. https://doi.org/10.1080/
02678370701758074
*Ten Brummelhuis, L. L., & Bakker, A. B. (2012). Staying engaged during the
week: The effect of off‐job activities on next day work engagement.
Journal of Occupational Health Psychology,17, 445–455. https://doi.
org/10.1037/a0029213
Trougakos, J. P., Hideg, I., Cheng, B. H., & Beal, D. J. (2014). Lunch breaks
unpacked: The role of autonomy as a moderator of recovery during
lunch. Academy of Management Journal,57, 405–421. https://doi.org/
10.5465/amj.2011.1072
Tuckey, M. R., Searle, B. J., Boyd, C. M., Winefield, A. H., & Winefield, H. R.
(2015). Hindrances are not threats: Advancing the multidimensionality
of work stress. Journal of Occupational Health Psychology,20, 131–
147. https://doi.org/10.1037/a0038280
*Van Hooff, M. L. M. (2015). The daily commute from work to home: Exam-
ining employees0experiences in relation to their recovery status. Stress
& Health,31, 124–131. https://doi.org/10.1002/smi.2534
Vasagar, J. (2013). Out of hours working banned by German labour minis-
try. The Telegraph. Retrieved from http://www.telegraph.co.uk/news/
worldnews/europe/germany/10276815/Out‐of‐hours‐working‐
banned‐by‐German‐labour‐ministry.html
Viswesvaran, C., & Ones, D. S. (1995). Theory testing: Combining psycho-
metric meta‐analysis and structural equations modeling. Personnel
Psychology,48, 865–885. https://doi.org/10.1111/j.1744-6570.1995.
tb01784.x
*Waite, E. (2012). Running to work: Marathon training, replenishment, and
worker well‐being. Houston, Texas: University of Houston.
Wendsche, J., & Lohmann‐Haislah, A. (2017). A meta‐analysis on anteced-
ents and outcomes of detachment from work. Frontiers in Psychology,
7. https://doi.org/10.3389/fpsyg.2016.02072
*White, E. (2010). Helping to promote psychological well‐being at work:
The role of work engagement, work stress and psychological detach-
ment using the Job Demands–Resources model. The Plymouth Student
Scientist,4, 155–180.
Wood, S. J., & Michaelides, G. (2016). Challenge and hindrance stressors
and wellbeing‐based work‐nonwork interference: A diary study of port-
folio workers. Human Relations,69, 111–138. https://doi.org/10.1177/
0018726715580866
Wright, T. A., & Cropanzano, R. (1998). Emotional exhaustion as a predictor
of job performance and voluntary turnover. Journal of Applied
Psychology,83, 486–493. https://doi.org/10.1037/0021-9010.83.3.
486
*Xanthopoulou, D., Bakker, A. B., Demerouti, E., & Schaufeli, W. B. (2007).
The role of personal resources in the Job Demands–Resources model.
International Journal of Stress Management,14, 121–141. https://doi.
org/10.1037/1072‐5245.14.2.121
Yik, M., Russell, J. A., & Steiger, J. H. (2011). A 12‐point circumplex structure
of core affect. Emotion,11,705–731. https://doi.org/10.1037/a0023980
Zijlstra, F. R. H., & Sonnentag, S. (2006). After work is done: Psychological
perspectives on recovery from work. European Journal of Work and
Organizational Psychology,15, 129–138. https://doi.org/10.1080/
13594320500513855
Andrew Bennett is an Assistant Professor of Management at Old
Dominion University. He completed a post‐doc at the University
of Alabama after earning his PhD from Virginia Commonwealth
University, MA from Gonzaga University, and BS from Clemson
University. His research focuses on employee well‐being, research
methods, and management education.
Arnold Bakker is a full professor of Work and Organizational
Psychology at Erasmus University Rotterdam, the Netherlands.
Dr Bakker0s research interests include positive organizational
behavior (e.g., job crafting and playful work design), Job
Demands–Resources theory, and the work–family interface.
James Field is an Assistant Professor of Organizational Behavior
and Human Resource Management at West Virginia University.
He earned his PhD from Virginia Commonwealth University,
MBA from Marshall University, and BS from Glenville State Col-
lege. He is also an active contributor to the metaBUS project. Dr
Field0s research interests include organizational research methods,
organizational staffing, and employee well‐being.
SUPPORTING INFORMATION
Additional Supporting Information may be found online in the
supporting information tab for this article.
How to cite this article: Bennett AA, Bakker AB, Field JG.
Recovery from work‐related effort: A meta‐analysis. J Organ
Behav. 2018;39:262–275. https://doi.org/10.1002/job.2217
BENNETT ET AL.275