Do We Recover from Vacation? : Meta-analysis of Vacation Effects on Health and Well-being

Abstract

Objectives:
The aim of this meta-analysis is to investigate to what extent vacation has positive effects on health and well-being, how long such effects endure after work resumption, and how specific vacation activities and experiences affect these relationships.

Methods:
Based on a systematic literature search (PsycInfo, Medline) and methodological exclusion criteria, in a stepwise approach, 7 studies were selected and reviewed. Effect sizes (Cohen's d) were calculated i) for every outcome variable within every study, ii) for every study by averaging the effect sizes per study, and iii) for homogeneous categories of outcome variables (exhaustion, health complaints, life satisfaction).

Results:
The results suggest that vacation has positive effects on health and well-being (small effect, d=+0.43), but that these effects soon fade out after work resumption (small effect, d=-0.38). Our research further demonstrated that vacation activities and experiences have hardly been studied. Therefore, their contribution to vacation effect and fade out remains unclear.

Discussion:
Progress in future vacation research will depend on strong research designs that incorporate repeated measurements pre-, inter- and post-vacation.

Full text

J Occup Health 2009; 51: 13–25
Received Jun 27, 2008; Accepted Nov 7, 2008
Published online in J-STAGE Dec 19, 2008
Correspondence to: J. de Bloom, Department of Work and
Organizational Psychology, Behavioural Science Institute,
Radboud University Nijmegen, P.O. Box 9104, NL-6500 HE
Nijmegen, The Netherlands (e-mail: j.debloom@psych.ru.nl)
Journal of
Occupational Health
Do We Recover from Vacation? Meta-analysis of Vacation Effects
on Health and Well-being
Jessica de BLOOM1, Michiel KOMPIER1, Sabine GEURTS1, Carolina de WEERTH2, Toon TARIS1
and Sabine SONNENTAG1, 3
1Department of Work & Organizational Psychology, 2Department of Developmental Psychology, Behavioural
Science Institute, Radboud University Nijmegen, The Netherlands and 3Department of Work & Organizational
Psychology, University of Konstanz, Germany
Abstract: Do We Recover from Vacation? Meta-
analysis of Vacation Effects on Health and Well-
being: Jessica de BLOOM,
et al
. Department of Work
& Organizational Psychology, Behavioural Science
Institute, Radboud University Nijmegen, The
Netherlands—Objectives: The aim of this meta-
analysis is to investigate to what extent vacation has
positive effects on health and well-being, how long such
effects endure after work resumption, and how specific
vacation activities and experiences affect these
relationships. Methods: Based on a systematic
literature search (PsycInfo, Medline) and
methodological exclusion criteria, in a stepwise
approach, 7 studies were selected and reviewed. Effect
sizes (Cohen’s
d
) were calculated i) for every outcome
variable within every study, ii) for every study by
averaging the effect sizes per study, and iii) for
homogeneous categories of outcome variables
(exhaustion, health complaints, life satisfaction).
Results: The results suggest that vacation has positive
effects on health and well-being (small effect,
d
=+0.43),
but that these effects soon fade out after work
resumption (small effect,
d
=–0.38). Our research
further demonstrated that vacation activities and
experiences have hardly been studied. Therefore, their
contribution to vacation effect and fade out remains
unclear. Discussion: Progress in future vacation
research will depend on strong research designs that
incorporate repeated measurements pre-, inter- and
post-vacation.
(J Occup Health 2009; 51: 13–25)
Key words: Allostatic load, Holiday, Recovery, Review,
Satisfaction, Stress
Time off is crucial for workers to recover from load
effects built up at work. A core assumption of Effort-
Recovery theory1, 2) and Allostatic Load Theory3–5) is that
initial normal load reactions (e.g., accelerated heart rate
and fatigue) can develop into more chronic load reactions
(e.g., prolonged fatigue, sleep complaints, high blood
pressure) in cases of continued exposure to workload and
incomplete recovery during time off2). The essence of
recovery is a process of psycho-physiological unwinding
after working, opposite to the activation of the
sympathetic-adrenal-medullary system and the
hypothalamic-pituitary-adrenal system during effort
expenditure (work), particularly under demanding or
stressful conditions2). Earlier research addressing rest
breaks6), long work hours7–9), and shift work10) has
acknowledged the role of recovery from work in
preserving individual well-being, health and performance
capabilities. Furthermore, over the years labour unions
have emphasized the importance of sufficient recovery
time in their endeavours for a shorter working week, rest
breaks and vacation rights, and both national and
international working time legislations have been enacted
to enable recovery possibilities for employees.
Recent studies have revealed that workers often recover
insufficiently during time off work due to, for instance,
working overtime, and that day-to-day incomplete
recovery may have serious adverse health consequences
in the long run (2, for an overview). Sluiter et al.11)
distinguished 4 different types of recovery based on
duration and time span after work: microrecovery (first
minutes after task performance), mesorecovery (10 min
to 1 h after task performance), metarecovery (1 h to 2
days after work) and macrorecovery (more than 2 days
after work).
Vacation as a form of macrorecovery is a prime
Review

14 J Occup Health, Vol. 51, 2009
candidate for helping workers to recover more completely
from work. Vacation is likely to be a more powerful
recovery opportunity than regular free evenings and
weekends because of two mechanisms underlying the
recovery process. The first ‘passive’ mechanism reflects
a direct release from daily job demands: vacation is ideally
a relatively long period of rest that is mostly spent in a
different and more relaxing environment that may help
workers to detach psychologically from work and from
other daily demands and routines. The second ‘active’
mechanism reflects the active engagement in potentially
recovering activities: vacation is a pre-eminent
opportunity to spend time on valued non-work activities
of one’s own choice, such as hobbies and family activities.
This article reviews the empirical literature with regard
to the recovering impact of this prototypical recovery
possibility, i.e. a vacation from work. The term ‘vacation’
stems from the Latin ‘vacatio’: ‘being free from work,
being at leisure, having time for’. We hypothesize that
vacation, as a relatively long and uninterrupted period of
respite from work may be a major contributor to the
recovery process, and therefore may be beneficial for
health and well-being.
Following a vacation, employees return to work, and
we are also interested in how long potential vacation
effects last, assuming that due to this renewed exposure
to work demands vacation effects will be temporary and
thus ‘fade out’.
Thirdly, from a work psychological point of view it is
important not to treat a vacation as a black box, but rather
to find out whether vacation activities (e.g., sports or
exercise) and vacation experiences (e.g., vacation
satisfaction) play a role in the relationship between
vacation and well-being.
In sum, this meta-analysis aims to answer 3 related
research questions:
1) What empirical evidence exists for an improvement
of health and well-being due to a vacation from work
(vacation effect)?
2) In the case of a positive effect of vacation, how long
does this last (fade out)?
3) a. Do vacation activities play a role in these potential
relationships?
b.Do vacation experiences play a role in these
potential relationships?
Methods
A systematic literature search was carried out in 2
bibliographical databases: PsycInfo and Medline. No
publication year limits were set and the final search date
was June 15th, 2008.
We used the following search terms within the fields
‘title’ or ‘keywords’:
1) vacation OR holiday (1,702 hits), and
2) well-being OR health OR quality of life OR
satisfaction OR stress OR burnout OR recovery
OR sleep OR mood OR affect (829,536 hits)
The combination of these 2 searches resulted in 125
hits (see Fig. 1). In a first selection round, the following
exclusion criteria were used:
—Language: non-English papers (minus 22)
—Publication type: dissertations, short
communications, letters, non-empirical and/or non-
peer-reviewed papers (minus 38)
After application of these criteria 65 hits remained.
All 65 abstracts were retrieved and read by the first 3
authors. Exclusion criteria in this second round were:
—Sample: papers not dealing with healthy, working
sample (e.g. school children, psychiatric patients):
minus 14
—Research purpose: studies irrelevant for the research
questions, i.e. studies not investigating vacation
effects and/or fade out on health and/or well-being
(e.g. weight gain during vacation, holiday shopping
intentions, sleepiness in drivers during summer
vacation): minus 35
—Design: studies not using an interrupted time series
design with at least a pre-test, i.e. before vacation
and a post-test, i.e. after vacation per subject as such
studies do not permit the evaluation of a vacation
(e.g. only post-vacation measure during annual
doctor visit): minus 5
Based on these criteria, 54 articles were excluded, and
11 papers remained. Studies that were referred to in the
11 selected papers were also examined but no additional,
relevant papers were detected. The first 3 authors of the
present article studied the remaining 11 papers and
excluded 4 more papers. In 2 cases, papers were written
by the same authors12–15), based on the same sample with
the second paper not offering extra information for our
research purposes. Therefore the second paper was
excluded in both cases13, 15). A third paper was excluded16)
because it investigated cross-over and thus compared
outcome variables on pre- and post-vacation in spouses
instead of vacationers themselves. Finally, a fourth paper
had to be excluded17) as it did not fit our research purposes:
the ‘vacation’ in this study was a compulsory off work
period, ‘a brief interlude during an acutely stressful
computer crisis’18). This resulted in a final selection of 7
studies (see Table 1).
To mathematically quantify the empirical evidence for
vacation effects in the 7 different studies we calculated
the effect size d for paired observations as described in
Cohen19). First we calculated, within every study, effect
sizes for all outcome variables in that study. Secondly,
we calculated average effect sizes for all studies by
averaging all effect sizes within each study.
Thirdly, in order to obtain a more detailed picture for
specific homogeneous outcome categories, we computed
a mean d for those outcome variables that were used in 3

15Jessica de BLOOM, et al.: Meta-analysis of Vacation Effects on Health and Well-being
or more different studies. This fine-grained analysis was
performed for the following outcome categories:
exhaustion (4 studies), health complaints (3 studies), and
life satisfaction (3 studies).
Following Cohen19) we distinguished among small (0
to 0.5), medium (0.5 to 0.8) and large effect sizes (>0.8).
Positive effect sizes indicate a beneficial effect of vacation
(improvement of health and well-being), whereas
negative effects denote the opposite (decrease in health
after vacation as compared to pre-vacation levels).
Results
Table 1 provides an overview of the 7 studies, by
characterizing sample and design characteristics, pre-
vacation measurement(s), measurements during vacation,
post-vacation measurement 1, and post vacation
measurement 2.
Sample and design characteristics
Number of participants: Sample sizes of the reviewed
studies were mostly small. Attrition from the pre-vacation
to the first post-vacation measurement varied between
5% 20) and 59%12). The loss of participants from the pre-
vacation to the post-vacation 2 measure varied between
5%20) and 86%12).
Sex, age and occupation: The distributions of sex, age
and occupation were diverse in the reviewed studies.
Control group: 5 studies12, 14, 18, 22, 23) did not include a
control group. Gilbert and Abdullah21) incorporated a
non-vacationing control group of 249 respondents
(opposed to 355 holiday-takers) that reported a lower
well-being than the holiday-takers on pre-vacation.
Etzion20) used a matched-pairs technique to create a
comparable control group of 55 respondents (age, marital
status and job function). This control group’s pre-vacation
scores on exhaustion resembled the vacation group’s
scores.
Duration: 3 of the 7 studies20–22) did not report the
duration of the vacation of their respondents. The average
duration of the vacation in the other studies was 914), 1023),
and 14 days12, 18).
Timing: In 2 studies, the timing of the vacation was not
Fig. 1. Systematic literature search on vacation and health or well-being from 125 to final 7 studies.

16 J Occup Health, Vol. 51, 2009
Table 1. Design characteristics of the 7 reviewed studies
Author(s), year Sample and design Pre-vacation During vacation Post-vacation Post-vacation
of publication* characteristics measurement(s) measurement measurement 1 measurement 2
Lounsbury & N pre-vac: 168 1–14 days before vacation, –1st wk after vacation, –
Hoopes, 1986
14)
N post-vac 1: 128 median 7 days median 3 days
N post-vac 2: –
Country: USA Outcome variables Outcome variables
: 28% Life satisfaction Same as pre-vacation
Age: 39 yr Job involvement/
Occ: variety central life interest Activities:
Contr.group: No Job involvement/valued self Activities during vacation
Vacation features Organizational commitment Experiences:
Duration: 9 days on average Turnover intention Overall vacation satisfaction
Timing: summer vacation Job satisfaction Satisfaction with aspects of
Location: NR vacation
Westman & N pre-vac1: 88 6 wk before vacation 2nd wk of vacation 3 days after vacation 3 wk after vacation
Eden, 1997
18)
N pre-vac2: 76 3 days before vacation (18 days after post-vacation 1)
N post-vac 1: 76
N post-vac 2: 76 Outcome variables Outcome variables Outcome variables
Country: Israel Exhaustion (physical, Same as pre-vacation Same as pre-vacation
: 59% emotional, mental) Outcome variables
Age: NR Activities: Activities: Same as pre-vacation
Occ: Administrative clerks ––
Contr. group: No Experiences: Experiences:
Vacation features – Vacation satisfaction
Duration: 14 days
Timing: summer vacation
Location: 87% at home,
13% away from home
Strauss-Blasche, N pre-vac: 130 10 days before vacation –3 days after vacation 5 wk after vacation
Ekmekcioglu & N post-vac 1: 53 (32 days after post-vacation 1)
Marktl, 2000
12)
N post-vac 2: 18 Outcome variables Outcome variables
Country: Austria Life satisfaction Same as pre-vacation
: 70% Physical complaints Outcome variables
Age: 34 yr Quality of sleep Activities: Same as pre-vacation
Occ: 57% manual workers, Positive mood Activities during vacation
43% white collar workers Negative mood Experiences:
Contr.group: No Recuperation
Vacation features
Duration: 14 days
Timing: summer vacation
Location: 76 % at home,
24 % at holiday resort

17Jessica de BLOOM, et al.: Meta-analysis of Vacation Effects on Health and Well-being
Table 1. Continued
Author(s), year Sample and design Pre-vacation During vacation Post-vacation Post-vacation
of publication* characteristics measurement(s) measurement measurement 1 measurement 2
Westman N pre-vac: 126 10 days before vacation –3 days after vacation 4 wk after vacation
& Etzion, N post-vac 1: 87 (25 days after post-vacation 1)
2001
23)
N post-vac 2: 87 Outcome variables Outcome variables
Country: Israel Exhaustion (physical, Same as pre-vacation
: 61% emotional, mental) Outcome variables
Age: 41 yr Absenteeism for health reasons Activities: Same as pre-vacation
Occ: blue collar industrial Absenteeism for other reasons –
workers (Company records, on
Contr. group: No aggregate level, not Experiences:
Vacation features individually) –
Duration: 10 days
Timing: Passover vacation (spring)
Location: NR
Etzion, 2003
20)
N pre-vac: 58 NR – NR 3 wk after vacation
N post-vac 1: 55 (“(...)before the individual (...) (“immediately after he/she
N post-vac 2: 55 went on vacation”) returned to work”)
N control: 55
Country: Israel Outcome variables Outcome variables Outcome variables
: 49% Exhaustion (physical, Same as pre-vacation Same as pre-vacation
control: 49% emotional, mental)
Age: 45 yr
Age control: 43 yr Activities:
Occ: employees at industrial enterprise –
Contr. group: Yes Experiences:
Vacation features Vacation satisfaction
Duration: NR, at least 1 wk Detachment from workplace
Timing: summer vacation
Location: NR

18 J Occup Health, Vol. 51, 2009
Table 1. Continued
Author(s), year Sample and design Pre-vacation During vacation Post-vacation Post-vacation
of publication* characteristics measurement(s) measurement measurement 1 measurement 2
Gilbert & N pre-vac: NR NR – 2–6 mo after first questionnaire –
Abdullah, 2004
21)
N post-vac 1: 355 (“...3,541 questionnaires were (“(...) within a period of 2–6 mo
N post-vac 2: – distributed at 2 points in time, after completion of the first questionnaire”)
N control: 249 during a 12-mo period”)
Country: United Kingdom
: 50% Outcome variables
control: 50% Life satisfaction Outcome variables
Age: Satisfaction with... Same as pre-vacation
16–24:14%; 25–34:22%; Interpersonal relationships
35–44:17%, 45–54:21%; Self Activities:
55–64:14%; 65–x:13% Services and facilities –
Age control: Nation
16–24:16%; Economic situation Experiences:
25–34:18%; 35–44:16%; Leisure –
45–54:15%; 55–64:11%; Friends
65–x: 23% Family
Occ: variety Home
Occ control: variety Neighborhood
Contr.group: Yes Health
Vacation features Positive affect
Duration: NR, at least 4 nights Negative affect
Timing: whole year round Current affect -
Location: NR Job satisfaction
Fritz & Sonnentag, N pre-vac: 414 7 days before vacation NR 1–2 days after vacation 2 wk after vacation
2006
22)
N post-vac 1: 221 (“The survey booklet had to (12–13 days after
N post-vac 2: 221 Outcome variables be filled in (...) during vacation”) Outcome variables post-vacation 1)
Country: Germany Exhaustion Same as pre-vacation
: 15% Disengagement
Age: 46 yr Health complaints Activities: Outcome variables
Occ: Non-academic Work effort – Same as pre-vacation
University employees Task performance Experiences:
Contr.group: No Relaxation experience
Vacation features Mastery experience
Duration: NR, at least 1 wk Negative work reflection
Timing: NR Positive work reflection
Location: NR Non-work hassles
(e.g. conflicts, financial problems)
Studies presented in order of publication date, *= Number in brackets refers to number in reference list, NR= Not reported,– = Not measured, Occ= Occupation of participants, Contr.
group: Control group.

19Jessica de BLOOM, et al.: Meta-analysis of Vacation Effects on Health and Well-being
reported21, 22). In 1 case, vacation time of the respondents
was in spring23). In the remaining 4 studies12, 14, 18, 20),
participants went on vacation in summertime.
Location: In 5 of the 7 studies14, 20–23), vacation location
of the respondents was not reported14, 20–23). In 2 studies,
more than 75% of the participants stayed at home during
their vacation (76%12) and 87%18)).
Pre-vacation measure
Timing of measurement: Two studies20, 21) did not report
when the pre-vacation measure took place. Of the
remaining 5 studies, Westman and Eden18) was the only
study that collected measures at 2 pre-vacation time points
(6 wk, and 3 days before vacation; they found no evidence
for major differences between these 2 time points). In
the study of Lounsbury and Hoopes14) data were collected
1 to 14 days (7 days on average) prior to vacation, and
the 3 remaining studies fell into the same time range:
1012, 23), and 7 days22).
Outcome variables: All studies measured the same
health and well-being parameters pre- and post-vacation
but the type of variables used varied: 4 studies measured
exhaustion18, 20, 22, 23), 3 measured health complaints12, 21, 22)
and 3 measured general life satisfaction12, 14, 21). Job
satisfaction was measured in 2 different studies14, 21), and
several parameters were measured in only one study: e.g.,
negative mood12), turnover intention14), and self-reported
work effort22).
During vacation measure
Timing of measurement: Only 2 papers18, 22)
incorporated a during vacation measurement. Westman
and Eden18) scheduled their during vacation measure in
the second week of the vacation. Fritz and Sonnentag22)
did not report when exactly their vacation measurement
took place.
Activities and experiences during vacation: Only Fritz
and Sonnentag22) asked their respondents about their
vacation experiences when actually on vacation. They
investigated experiences during vacation in a detailed
way, by gathering information on relaxation and mastery
experiences, positive and negative work reflection and
non-work hassles.
Post-vacation measure 1
Timing of measurement: Etzion20) did not report when
the first post-vacation measure took place. Gilbert and
Abdullah21) took their only post-vacation measure 2 to 6
months after the pre-vacation measure. The remaining 5
studies12, 14, 18, 22, 23) scheduled their first post-vacation
measure within the first week of returning to work (3
days on average).
Activities during vacation: At post-vacation 1, i.e.
retrospectively, 2 studies collected information about
vacation activities12, 14).
Experiences during vacation: 3 papers21–23) gathered
no information about vacation experiences post hoc.
Three of the remaining studies14, 18, 20) asked respondents
about their vacation satisfaction in retrospect. Strauss-
Blasche et al.12) and Etzion20) included questions about
recuperation during vacation and detachment from the
workplace respectively.
Post-vacation measure 2
Timing of measurement: Five studies12, 18, 20, 22, 23) adopted
a second post-vacation measure. Post-vacation 2
measures were collected 2 wk after vacation (12–13 days
after post vac 1) in Fritz and Sonnentag22), 3 wk (18 days
after post-vacation 1) in Westman and Eden18) and
Etzion20), and 4 wk (25 days after post-vacation 1) in
Westman & Etzion23). Strauss-Blasche et al.12) had the
longest time interval: 5 wk after vacation (32 days after
post-vacation 1).
Research question 1: Vacation effect?
We calculated the pre-vacation–post-vacation 1
difference in health and well being indicators
(‘vacation effect’) in all 7 studies. The time span
between these 2 time points was unknown in 3 studies:
there was no data available on vacation duration22) or
pre-vacation time point, vacation duration and post
vacation 1 time point20, 21). The time span between the
pre- and post-vacation 1 time points in the other 4
studies ranged between 1914), 2018), 2323) and 27 days12).
First we calculated, within every study, effect sizes for
every outcome variable in that study. Then, we calculated
general effect sizes for every study, i.e. averaged the
number of effect sizes in each study (Table 2).
The minimum number of outcome variables per study
was 118, 20, 23) and the maximum number was 1721). Within
the 7 papers 36 outcome variables were studied, hence
36 effect sizes were calculated. Thirty of these of these
were positive (improvement in well-being) and 6 negative
(decrease in well-being). The 6 negative effect sizes were
small (mean d =–0.18) and of the positive effect sizes,
21 were small, 6 were medium, and 3 were large.
Large effect sizes were found for health complaints
(d =+1.0119) and +0.8212)) and exhaustion (d =+0.9218)).
The average effect sizes per study varied from –0.0514) to
+0.9218). The overall mean d across 7 studies was +0.43,
indicating a small positive vacation effect.
In the 2 control group studies, Etzion20) found a small
“pre–post vacation 1” increase in exhaustion in the control
group (d =–0.12). The “post-vacation 1” difference in
exhaustion between vacationers and non-vacationers was
small (d=+0.35), with non-vacationers reporting more
exhaustion. Gilbert and Abdullah21) found negative
changes for the control group on all outcome variables,
indicative of deterioration in well-being (mean d=–0.27).
The difference between non-holiday and holiday takers

20 J Occup Health, Vol. 51, 2009
Table 2. Means, standard deviations and effect sizes for vacation effect on all outcome variables for each study
Study * Outcome variable Mean SD Mean SD Cohen dMean d
pre-vac pre-vac post-vac 1 post-vac 1
Lounsbury & Life satisfaction 24.87 5.68 23.83 6.36 + 0.40 – 0.05
Hoopes, 198614) Job involvement/interest 21.17 4.88 22.10 4.61 – 0.48
Job involvement/valued self 11.43 3.18 11.68 3.21 + 0.16
Organizational commitment 10.51 2.56 10.65 2.69 + 0.10
Turnover intention 3.80 0.98 3.67 0.99 – 0.26
Job satisfaction 22.08 5.71 21.51 5.69 – 0.23
Westman & Exhaustion 3.30 0.60 3.03 0.62 + 0.92 + 0.92
Eden, 199718)
Strauss–Blasche, Life satisfaction NR NR NR NR + 0.04 + 0.53
Ekmekcioglu & Physical complaints NR NR NR NR + 0.82
Marktl, 200012) Quality of sleep NR NR NR NR + 0.45
Positive mood NR NR NR NR + 0.66
Negative mood NR NR NR NR + 0.67
Westman & Exhaustion 2.89 0.65 2.70 0.99 + 0.35 + 0.35
Etzion, 2001 23)
Etzion, 2003 20) Exhaustion 2.59 0.54 2.44 0.59 + 0.46 + 0.46
Gilbert & Life satisfaction (1 item) 6.99 1.23 7.11 1.20 + 0.23 + 0.33
Abdullah, Life satisfaction (scale) 30.78 7.12 31.78 7.59 + 0.31
2004 21) Positive affect 60.47 12.24 63.58 11.79 + 0.45
Negative affect 31.22 14.28 30.21 14.44 + 0.12
Current affect 29.30 22.77 33.29 23.13 + 0.30
Satisfaction friends 7.25 1.15 7.24 1.07 –0.01
Satisfaction family 7.22 1.50 7.20 1.42 –0.03
Satisfaction home 6.85 1.31 6.94 1.26 +0.16
Satisfaction relationships 6.80 1.06 7.02 1.02 +0.49
Satisfaction econ. situation 6.75 1.43 6.97 1.22 +0.37
Satisfaction leisure 6.34 1.45 6.53 1.22 +0.33
Satisfaction neighborhood 6.29 1.36 6.49 1.30 +0.34
Satisfaction self 6.22 1.22 6.55 1.20 + 0.62
Satisfaction services 6.12 1.23 6.39 1.11 +0.54
Satisfaction health 5.97 1.42 6.22 1.44 +0.21
Satisfaction nation 4.75 1.19 5.15 1.32 +0.73
Job satisfaction 6.42 1.29 6.67 1.19 +0.47
Fritz & Health complaints 1.94 0.47 1.59 0.35 +1.01 +0.46
Sonnentag, Exhaustion 2.18 0.55 2.05 0.55 +0.45
2006 22) Disengagement 2.10 0.53 2.06 0.53 +0.15
Task performance 4.51 0.49 4.49 0.54 –0.05
Work effort 2.90 1.14 2.26 1.15 + 0.74
Total +0.43
*= Number in brackets refers to number in reference list, + = positive effect, improvement in health and/or well-being, – = negative
effect, decrease in health and/or well-being, Mean pre-vac= mean at pre-vacation, SD pre-vac= standard deviation at pre-vacation,
Mean post-vac 1= mean at post vacation 1, SD post-vac 1= standard deviation at post-vacation1, NR= Not reported in study.

21Jessica de BLOOM, et al.: Meta-analysis of Vacation Effects on Health and Well-being
at “post-vacation 1” was small (mean d = +0.50), the
former reported a lower well-being.
Next, a fine-grained analysis for the homogenous
outcome categories exhaustion, life satisfaction, and
health complaints was conducted (Table 3). Effect sizes
for the category exhaustion (4 studies) varied from
+0.3523) to +0.9218). The average d was +0.55, indicating
a medium vacation effect.
Concerning health complaints, effect sizes were
+1.0122), +0.2121) and +0.8212). The average effect size
was +0.68, indicating a medium effect.
Finally, a small average effect size (d =+0.24) was
found for the category life satisfaction. Cohen’s d ranged
between +0.0412), +0.2721) and +0.4014).
Research question 2: Fade out?
The concept of ‘fade out’ supposes the a priori existence
of an effect. Vacation effects can only disappear when
they were present in the first place, i.e. at post-vacation
1. Our analysis was thus based upon those 4 studies that
employed 2 post-vacation measures, and found a positive
vacation-effect18, 20, 22, 23). Note that Strauss-Blasche et
al.12) included a post-vacation 2 measure too, but they
neither compared their outcome variables at this time
point with those at post-vacation 1, nor reported means
and standard deviations at the different measurement
occasions.
In 4 studies that compared post-vacation 1 and 218, 20, 22, 23)
effect sizes could be calculated for exhaustion. In addition,
in the study of Fritz and Sonnentag22) also 3 other effect
sizes could be calculated. Single outcome effect sizes per
study were –0.0220), –0.1222), –0.2823) and –1.0818). In the
study of Fritz and Sonnentag22) effect sizes ranged from +0.08
to –0.49.
From the total of 7 different outcome variables, 1 had
a positive sign, 1 was 0, and 5 had a negative sign meaning
that in most cases well-being decreased between post-
vacation 1 and 2. The positive effect size was negligibly
small (d=+0.08). Within the 5 negative effect sizes, 4
were small and 1 large. This large effect size was found
for exhaustion (d=–1.0818)).
The overall mean d across 4 studies was –0.38,
indicating a small fade out effect. Table 4 further shows
the time span between the 2 post-vacation measures that
varied between approximately 2 to 4 wk. As there were
only 2 post-vacation measures in all 4 studies and the
minimum fade out interval was 12–13 days after
vacation22), it was impossible to study the specific course
of fade out and to determine when fade out began and
when pre-vacation base levels were reached again.
Only Etzion20) compared scores on 2 measures in a
non-vacation group taken at the same time as post-
vacation 1 and 2 in the vacation group. She found a
small positive effect (d =+0.22) meaning that exhaustion
decreased in the control group in the time between the
second and the third measurement occasion. The
difference between vacationers and non-vacationers was
+0.19 on “post-vacation 2”, meaning that non-vacationers
Table 3. Effect sizes for vacation effect in homogeneous outcome variables used in 3 or more different studies
Outcome variables Study* Cohen dMean Cohen dMean Cohen d
corrected for
more than 1 indicator
per study
Exhaustion (4 studies) + 0.55
Exhaustion Westman & Eden, 199718) + 0.92 + 0.92
Exhaustion Westman & Etzion, 200123) + 0.35 + 0.35
Exhaustion Etzion, 200320) + 0.46 + 0.46
Exhaustion Fritz & Sonnentag, 200622) + 0.45 + 0.45
Health complaints (3 studies) + 0.68
Physical complaints Strauss-Blasche et al., 200012) + 0.82 + 0.82
Satisfaction with health Gilbert & Abdullah, 200421) + 0.21 + 0.21
Health complaints Fritz & Sonnentag, 200622) + 1.01 + 1.01
Life satisfaction (3 studies) + 0.24
Life satisfaction Lounsbury & Hoopes, 198614) + 0.40 + 0.40
Life satisfaction Strauss-Blasche et al., 200012) + 0.04 + 0.04
Life satisfaction (item) Gilbert & Abdullah, 200421) + 0.23
Life satisfaction (scale) Gilbert & Abdullah, 200421) + 0.31 + 0.27
*= Number in brackets refers to number in reference list, + = Positive effect, improvement in health and/or well-being, –= Negative
effect, decrease in health and/or well-being.

22 J Occup Health, Vol. 51, 2009
were slightly more exhausted than their vacation taking
fellows.
Again, we performed a fine-grained analysis of
homogeneous outcome variables, measured in 3 or more
different studies. Only exhaustion met this criterion (4
studies). The average effect size was small (d= (–0.02)
+ (–1.08) + (–0.28) + 0.08)/4= –0.33).
Research question 3a: Activities on vacation?
Only 2 of 7 studies collected data during vacation.
However, neither study18, 22) collected information about
what vacationers actually did during their holiday. Two
other studies12, 14) collected information on vacation
activities in retrospect, i.e. at post vacation 1. These
studies reported percentages that were spent on certain
activities (e.g., traveling, reading, sight seeing) but did
not relate these percentages to the outcome variables.
This means that research question 3a could not be
addressed.
Research question 3b: Experiences on vacation?
One study22) collected information on vacation
experiences during the vacation itself. Four other
studies12, 14, 18, 20) collected information on vacation
experiences at post vacation 1 when respondents had
already resumed working.
Vacation satisfaction was measured in 3 studies14, 18, 20)
and appeared to be positively related to job satisfaction
and life satisfaction14) and negatively to exhaustion18),
whereas Etzion20) found no such relationship with
exhaustion. Etzion20) also retrospectively collected
information on detachment from work during the vacation
and did not find a relationship with post-vacation
exhaustion, whereas Strauss-Blasche et al.12) found that
well-being at post-vacation was higher among those
respondents who reported sufficient recuperation during
vacation as compared to those who indicated that
recuperation during vacation was insufficient.
In the only ‘during vacation study’ Fritz and
Sonnentag22) tested the effect of vacation experiences on
health indicators after vacation. Positive (e.g. relaxation)
as well as negative experiences (e.g. negative work
reflexion) were related to almost all outcome variables.
Within these experiences, negative work reflexion seemed
to play a major role: respondents engaging in negative
work reflexion during vacation reported also lower well-
being on post-vacation 1.
In sum, only 1 study22) measured vacation experiences
when actually on vacation. This study found evidence in
support of a temporal relation between vacation
experiences and outcome variables: positive experiences
were related to improved well-being after vacation
whereas negative experiences had the opposite effect. Of
the 4 studies that collected information on vacation
experiences after returning to work (mostly vacation
satisfaction), 2 studies reported positive cross-sectional
associations between vacation satisfaction and outcome
variables14, 18), whereas 1 study20) did not.
Table 4. Means, standard deviations and effect sizes for fade out on all outcome variables for each study
Study* Outcome variable Time span Mean SD Mean SD Cohen dMean d
post 1-post 2 post-vac 1 post-vac 1 post-vac 2 post-vac 2
Westman & Exhaustion 18 days 3.03 0.62 3.35 0.62 –1.08 –1.08
Eden, 199718)
Westman & Exhaustion 25 days 2.70 0.99 2.92 0.94 –0.28 –0.28
Etzion, 200123)
Etzion, 200320) Exhaustion 21 days 2.44 0.59 2.45 0.66 –0.02 –0.02
(post–vacation 1
immediatly after
returning to work)
Fritz & Health complaints 12–13 days 1.59 0.35 1.71 0.42 – 0.49 – 0.12
Sonnentag, Exhaustion 2.05 0.55 2.03 0.56 +0.08
200622) Disengagement 2.06 0.53 2.06 0.54 0.00
Work effort 2.26 1.15 2.31 1.15 –0.07
Total –0.38
* =Numbers in brackets refer to number in reference list, time span post 1–post 2= time span between post-vacation 1 and post-
vacation 2, + =Positive effect, improvement in health and/or well-being, – =Negative effect, decrease in health and/or well-being,
Mean post–vac 1=mean at post-vacation 1, SD post-vac 1= standard deviation at post-vacation 1, Mean post-vac 2=mean at post-
vacation 2, SD post-vac 2=standard deviation at post-vacation 2.

23Jessica de BLOOM, et al.: Meta-analysis of Vacation Effects on Health and Well-being
Discussion
The aim of this meta-analysis was to find out if vacation
had a positive impact on health and well-being, how long
such beneficial effects would last, and whether vacation
activities and experiences were related to these outcomes.
In a stepwise approach 7 studies were identified that could
shed light on these questions.
Vacation effect
There was evidence for a small effect of vacation on
health and well-being. The average d was + 0.43,
indicating that well-being improved slightly following a
vacation. In accordance with effort-recovery theory1, 2),
the vacation effect was more prominent among outcome
variables that were closer to the core of the concept ‘health
and well-being’, than among more distal variables. Thus,
health complaints and exhaustion as proximal health
indicators improved more than life satisfaction as a more
distal indicator.
As only 4 studies reported the duration of the vacation,
the relation between the magnitude of effects and vacation
length could not be determined. Future research should
address this relation, eventually pointing to an “optimum
point of recovery”. Subsequently, such knowledge could
be applied to develop guidelines for the scheduling and
duration of vacations.
Fade out
There was also evidence for the post-vacation
disappearance of vacation effects 2 to 4 wk post-vacation.
The average d was –0.38. Regrettably the available
information was too limited to evaluate the precise course
of fade out and hence the duration of vacation effects. It
seems that (entire or partial) fade out took place within 2
to 4 wk post-vacation but since the second post-vacation
measure was scheduled at least 2 wk after vacation in all
4 studies, we were not able to determine when beneficial
effects on different variables exactly started to diminish
and were erased. Simple and frequent measures from
the day of return until 8 wk after vacation would
contribute to a better understanding of the course of fade
out.
Another interesting question is which factors might
prolong vacation effects and delay fade out24). Methods
borrowed from cognitive therapy (e.g., brief daily writing
about positive vacation experiences) could be useful for
this purpose.
Vacation activities and experiences
The role of vacation activities and experiences on
vacation effects remains unclear hitherto. Vacation
activities as moderators of vacation effects have not been
studied yet, while they may be important behavioural
determinants of positive and negative vacation outcomes.
The few results regarding vacation experiences suggest
that vacation satisfaction as well as negative work
reflection do play a major role in influencing vacation
outcomes in a positive or negative way respectively. But
until now most reports on vacation experiences were
potentially biased because data were collected after
returning home. To overcome this problem, researchers
need to include measurement occasions during vacation
and ask respondents about vacation expectations,
activities (e.g. active versus passive, voluntary versus
involuntary activities), uplifts, hassles and
(dis)satisfaction.
Surprisingly there was even very limited information
on basic vacation features like timing and location
available. Even the (average) vacation duration was not
reported in 3 cases. Most studies dealt with summer
vacations. Furthermore, it remains unclear in 5 studies
if participants stayed at home during their vacation or
‘left their house and went away’14, 20–23). As spending time
at a holiday resort may well differ from spending time in
one’s regular surroundings, future vacation research
should report vacation timing and location, to interpret
findings in this light and to compare different vacation
features.
Methodological considerations
An intriguing issue in vacation research is the question
of causality, i.e. were differences in outcome variables
before and after vacation indeed due to vacation? In many
cases there were plausible rival hypotheses, e.g., that pre-
post vacation changes in work demands may account for
pre-post differences in health outcomes. Eden24) called
this tendency of attributing changes in outcomes to
vacation the “post hoc ergo propter hoc inference fallacy”.
Only an intensified repeated measure strategy can
overcome this problem of limited internal validity in the
future.
Another frequent problem of earlier studies is the small
number of respondents and the accompanying attrition,
possibly due to difficult recruitment and low compliance.
This might be counteracted by close collaborations with
travel agencies, attractive rewards for participants and
devoted respondent care. The use of different kinds of
attractive new media (e.g. palm pilots, online surveys,
mobile phones) could also support participant compliance
and prevent attrition.
The absence of a control group in most of the studies
is also problematic. This deficiency may partly be due
to the fact that randomization into experimental and
control groups is difficult, if not impossible in vacation
research. Accordingly holiday and non-holiday takers
will differ anyway because non-vacationers may have
many reasons for not going on vacation like illness, lack
of funds or abundance of work. The use of an internal
referencing strategy instead of a control group might be

24 J Occup Health, Vol. 51, 2009
a better way to strengthen internal validity25). In this
approach, additional variables are included that are similar
to the outcome variables but that are theoretically not
expected to change because of a treatment (i.e. vacation
in our example). If these control variables do not change,
whilst ‘real’ outcome variables do, this is interpreted as
empirical support for a true vacation effect. An example
for such a variable is teamwork competency.
A final shortcoming is the use of only self-reports in
vacation research. With most reviewed authors we agree
that the use of other ‘objective’ measures like performance
ratings and physiological measures would be desirable.
Suggestions for future vacation research
Vacation research will profit from better designs, which
boils down to the principle of repeated measurements.
Vacation research necessarily requires research on
vacation: the assessment of vacation activities and
experiences during vacation itself. A suitable framework
for structuring diverse measurement occasions around a
vacation period was developed by Westman and Eden18)
and consists of 2 pre-, 1 inter- and 2 post-vacation
measurements. Its application may well contribute to
the comparability of future vacation research findings.
As discussed above, resolutions for earlier
methodological problems, the detailed investigation of
the fade out process by means of brief daily measures,
studies on optimal vacation duration, frequency and
timing, and the design and evaluation of interventions to
prolong positive vacation effects, deserve a place on the
vacation research agenda.
Although in general neuroendocrine and cardiovascular
measures are quite difficult and costly to apply in field
settings, applications in vacation research may even be
more difficult as participants are out of sight of the
researcher for a relatively long period and daytime activity
cannot be controlled for. However, as chronic incomplete
recovery may manifest itself in a disturbed balance of
sympathetic and parasympathetic activity, also during
sleepe.g. 26–30), a possibility for collecting physiological
measures during a vacation period would be, for instance,
during night time. During sleep, parasympathetic
activation with its main restorative function should be
dominant, but high blood pressure levels, high heart rate,
low heart rate variability and high levels of catecholamine
in morning urine would be strong markers of high
sympathetic and low parasympathetic activation, and thus,
indicative of disturbed restorative functions and
incomplete recovery.
Typically, moderators of vacation effects have hardly
been studied. Still, vacation research will benefit from
the inclusion of moderators in the work context (e.g., job
stressors, job type), the non-work context (e.g., culture,
relational problems, economic hardship) and person
characteristics (e.g., self-efficacy, workaholism).
Moreover, different vacation features (duration, timing
and location) should be investigated and reported
accurately to compare the effect of different vacation
types on outcome variables.
In conclusion, much has been learned from previous
vacation studies. The general picture that emerges from
these pioneering studies is that vacation positively, though
weakly, impacts well-being but that those positive effects
do not last long. Future vacation research may benefit
from multiple measurements: pre-vacation and post-
vacation but especially during vacation.
References
1) Meijman TF, Mulder G. Psychological aspects of
workload. In: Drenth PJD, Thierry H, De Wolff CJ,
editors. Handbook of Work and Organizational
Psychology. East Sussex (UK): Psychology Press,
Hove: 1998. p.5–33.
2) Geurts SAE, Sonnentag S. Recovery as an explanatory
mechanism in the relation between acute stress
reactions and chronic health impairment. Scand J Work
Env Hea 32, 482–492 (2006)
3) Clow A. The physiology of stress. In: Jones F, Bright
J, editors. Stress, Myth, Theory, and Research. Harlow
(UK): Prentice Hall; 2001. p.47–61.
4) McEwen BS. Stress, adaptation, and disease: allostasis
and allostatic load. Ann NY Acad Sci 1998; 840: 33–
44.
5) Sterling P, Eyer J. Allostasis: a new paradigm to explain
arousal pathology. In: Fisher S, Reason J, editors.
Handbook on Life Stress, Cognition, and Health.
Chichester (UK): Wiley; 1990. p.629–49.
6) Tucker P. The impact of rest breaks upon accident risk,
fatigue and performance: a review. Work Stress 2003;
17: 123–137.
7) Härmä MD. Workhours in relation to work stress,
recovery and health. Scand J Work Env Hea 2006; 32:
502–514.
8) Van der Hulst M. Long work hours and health. Scand
J Work Env Hea 2008; 29: 171–88.
9) Beckers D, Van der Linden D, Smulders PGW,
Kompier M, Van Veldhoven MJPM, Van Yperen NW.
Working overtime hours: relations with fatigue, work
motivation, and the quality of work. J Occup Environ
Med 2004; 46: 1282–9.
10) Totterdell P, Spelten E, Smith L, Barton J, Folkard S.
Recovery from work shifts: how long does it take? J
Appl Psychol 1995; 80: 43–57.
11) Sluiter JK, Frings-Dresen MH, Meijman TF, Van der
Beek AJ. Reactivity and recovery from different types
of work measured by catecholamines and cortisol: a
systematic literature review. Occup Environ Med 2000;
57: 289–315.
12) Strauss-Blasche G, Ekmekcioglu C, Marktl W. Does
vacation enable recuperation? Changes in well-being
associated with time away from work. Occup Med
2000; 50: 167–72.
13) Strauss-Blasche G, Ekmekcioglu C, Marktl W.
Moderating effects of vacation on reactions to work

25Jessica de BLOOM, et al.: Meta-analysis of Vacation Effects on Health and Well-being
and domestic stress. Leisure Sci 2002; 24: 237–49.
14) Lounsbury JW, Hoopes LL. A vacation from work:
changes in work and nonwork outcomes. J Appl
Psychol 1986; 71: 392–401.
15) Hoopes LL, Lounsbury JW. An investigation of life
satisfaction following a vacation: a domain specific
approach. J Community Psychol 1989; 17: 129–40.
16) Etzion D, Westman M. Job stress, vacation, and the
crossover of strain between spouses- stopping the
vicious cycle. Man and Work 2001; 11: 106–18.
17) Eden D. Acute and chronic job stress, strain, and
vacation relief. Organ Behav Hum Dec 1990; 45: 175–
93.
18) Westman M, Eden D. Effects of a respite from work
on burnout: vacation relief and fade-out. J Appl Psychol
1997; 82: 516–27.
19) Cohen J. Statistical power analysis for the behavioral
sciences. Hillsdale (New Jersey): Lawrence Erlbaum
Associates; 1988.
20) Etzion D. Annual vacation: duration of relief from job
stressors and burnout. Anxiety Stress Copin 2003; 16:
213–26.
21) Gilbert D, Abdullah J. Holidaytaking and the sense of
well-being. Ann Tourism Res 2004; 31: 103–21.
22) Fritz C, Sonnentag S. Recovery, well-being, and
performance-related outcomes: the role of workload
and vacation experiences. J Appl Psychol 2006; 91:
936–45.
23) Westman M, Etzion D. The impact of vacation and job
stress on burnout and absenteeism. Psychol Health
2001; 16: 95–106.
24) Eden D. Vacations and other respites: studying stress
on and off the job. In: Cooper C, Robertson IT, editors.
Well-being in organizations. West Sussex (UK): John
Wiley & Sons, Ltd; 2001. p.305–30.
25) Haccoun RR, Hamtiaux T. Optimizing knowledge tests
for inferring learning acquisition levels in single group
training evaluation designs: the internal referencing
strategy. Pers Psychol 1994; 47: 593–604.
26) Akerstedt T. Psychosocial stress and impaired sleep.
Scand J Work Environ Health 2006; 32: 493–501.
27) Dahlgren A, Kecklund G, Akerstedt T. Overtime work
and its effects on sleep, sleepiness, cortisol and blood
pressure in an experimental field study. Scand J Work
Environ Health 2006; 32: 318–27.
28) Hall M, Vasko R, Buysse D, et al. Acute stress affects
heart rate variability during sleep. Psychosom Med
2004; 66: 56–62.
29) Brosschot JF, Van Dijk E, Thayer JF. Daily worry is
related to low heart rate variability during waking and
the subsequent nocturnal sleep period. Int J
Psychophysiol 2007; 63: 39–47.
30) Rau R, Triemer A. Overtime in relation to blood
pressure and mood during work, leisure, and night time.
Soc Indic Res 2004; 67: 51–73.