Content uploaded by Maike E. Debus
Author content
All content in this area was uploaded by Maike E. Debus on Feb 03, 2018
Content may be subject to copyright.
RESEARCH REPORT
Making Flow Happen: The Effects of Being Recovered on Work-Related
Flow Between and Within Days
Maike E. Debus
University of Zurich Sabine Sonnentag
University of Mannheim
Werner Deutsch
Technical University of Braunschweig Fridtjof W. Nussbeck
Bielefeld University
This article examines variations of work-related flow both between and within days. On the basis of the
effort-recovery model (Meijman & Mulder, 1998), we hypothesized that a person’s relative day-specific
state of being recovered (i.e., feeling refreshed) in the morning is positively related to subsequent
day-level flow experiences during work. Taking into account research on circadian rhythms of human
functioning, we further hypothesized that flow experiences follow a U-shaped pattern within the working
day and that feeling recovered will affect this pattern. One hundred and twenty-one software profes-
sionals provided data on recovery at the start of the working day and on flow at 3 occasions during the
day, for a period of 5 consecutive working days (resulting in 493 day-level and 1,340 occasion-level data
points). Three-level multilevel models showed that relative day-level state of being recovered predicted
day-level flow experiences in the hypothesized direction. The data did not support a general curvilinear,
U-shaped main effect of flow experiences within the day. However, people in a relatively high state of
being recovered in the morning experienced the predicted U-shaped pattern, whereas poorly recovered
people experienced a gradual decrease in flow experiences over the course of the working day. This study
emphasizes the importance of recovery during nonwork time for flow experiences within the entire
working day, thereby extending research on task characteristics with personal resources when examining
predictors of flow.
Keywords: recovery, flow experiences, curvilinear effects, experience sampling study
Flow is an engrossing and enjoyable state of mind that occurs
when people feel optimally challenged and are fully absorbed in
their current activity (e.g., Csikszentmihalyi, 1999). Work-related
flow has been shown to be related to positive, organizationally
relevant outcomes (e.g., self-efficacy; Salanova, Bakker, & Llo-
rens, 2006; job performance under certain conditions, Demerouti,
2006;Eisenberger, Jones, Stinglhamber, Shanock, & Randall,
2005;Kuo & Ho, 2010), and even to cross over between people
(Bakker, 2005). However, as we know from other areas in our
daily lives, good things usually do not last forever. The same in
fact applies here: Flow is considered to be a fragile state and a
short-term peak experience (Bakker, 2005,2008;Csikszentmi-
halyi, 1996). In support of this notion, research has shown that a
large amount of variance in flow accrues from daily and momen-
tary characteristics (Ceja & Navarro, 2011;Fullagar & Kelloway,
2009;Nielsen & Cleal, 2010). Moreover, the literature on circa-
dian rhythms in humans suggests that the state of flow does not
vary completely at random, but may exhibit a curvilinear,
U-shaped pattern within the working day (e.g., Van Dongen &
Dinges, 2000).
To gain a deeper understanding of this within-person variability
in flow, it is imperative to examine predictors that vary within
similar time frames. In doing so, previous studies have focused on
fluctuations in daily or momentary task characteristics and activ-
ities to predict fluctuations in daily or momentary work-related
flow experiences (Fullagar & Kelloway, 2009;Nielsen & Cleal,
2010). However, by definition flow is a demanding and challeng-
ing state (e.g., Csikszentmihalyi, Abuhamdeh, & Nakamura,
2005), suggesting that people need personal, energetic resources in
This article was published Online First February 10, 2014.
Maike E. Debus, Department of Psychology, University of Zurich,
Zurich, Switzerland; Sabine Sonnentag, Department of Psychology, Uni-
versity of Mannheim, Mannheim, Germany; Werner Deutsch, Institute of
Psychology, Technical University of Braunschweig, Braunschweig, Ger-
many; Fridtjof W. Nussbeck, Department of Psychology, Bielefeld Uni-
versity, Bielefeld, Germany.
Werner Deutsch, who contributed to developmental psychology through
his research and teaching, passed away on October 12, 2010. He was
involved in the conceptualization of this study.
Correspondence concerning this article should be addressed to Maike E.
Debus, Arbeits- & Organisationspsychologie, Universität Zürich, Binz-
mühlestrasse 14/12, 8050 Zürich, Switzerland. E-mail: m.debus@
psychologie.uzh.ch
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Journal of Applied Psychology © 2014 American Psychological Association
2014, Vol. 99, No. 4, 713–722 0021-9010/14/$12.00 DOI: 10.1037/a0035881
713
order to experience this pleasurable state of mind. For this reason,
the current study expands previous knowledge by investigating a
person’s state of being recovered (i.e., feeling refreshed; Binnew-
ies, Sonnentag, & Mojza, 2009) as a personal resource, and its
relation with daily flow experiences, above and beyond the impact
of task type.
More precisely, we connect the literatures on recovery and flow
to examine whether a person’s relative state of being recovered at
the beginning of a working day relative to other days (i.e.,
between-day variation) will (a) be related to that person’s average
degree of flow experiences that day and (b) affect the time course
of flow experiences within that same day. Hence, this study
focuses on deviations of a person’s day-specific state of being
recovered relative to his or her average; by doing so, we are able
to model within-person fluctuations by simultaneously excluding
all effects that would be due to differences between persons
(Bolger, Davis, & Rafaeli, 2003;Ohly, Sonnentag, Niessen, &
Zapf, 2010).
As such, this study will lead to a better understanding of daily
preconditions enabling flow at work, thereby contributing to the
literature in several ways. First, previous research has analyzed
how a workday as a whole may benefit from being well recovered
(i.e., between-day variations; Binnewies et al., 2009;Sonnentag,
Mojza, Demerouti, & Bakker, 2012)—a principle that the current
research applies to the study of flow. More important, however,
this study extends previous research in that we add a dynamic
perspective by examining potential benefits of recovery that unfold
during the day. By connecting the literature on flow with the
literatures on recovery and biological, circadian processes, we are
able to explain fluctuations in flow within the course of the
working day. Moreover, we examine the state of being recovered
as a day-specific personal resource. In doing so, our study provides
insight for individuals and managers about to what extent flow
experiences can be fostered through personal resources, beyond
focusing on strategies that target work and task design issues.
The State of Being Recovered as a Predictor of
Variations in Flow Between Days
In general, flow has been described as consisting of nine ele-
ments (e.g., Csikszentmihalyi, 1975/2000,1990): (a) a balance
between high skills and high challenges (i.e., skill–challenge bal-
ance); (b) clear goals; (c) clear and immediate feedback; (d)
concentration; (e) a merging of action and awareness, meaning that
the activity becomes almost automatic; (f) a sense of control over
the action; (g) a feeling that the activity is intrinsically rewarding;
(h) the loss of self-consciousness; and (i) the transformation of
time, that is, hours seem to pass by like minutes.
Besides the fact that flow is a state of deep enjoyment and
total immersion in a task, the literature suggests that it is also a
state that requires initial energy or resource expenditure (Csik-
szentmihalyi, 1996;Csikszentmihalyi et al., 2005;Nakamura &
Csikszentmihalyi, 2002). On the basis of interviews with sur-
geons, chess players, and rock climbers, Csikszentmihalyi
(1996, p. 110) concluded that flow did not occur during unde-
manding activities but when being engaged in “difficult activ-
ities that stretched the person’s capacity and involved an ele-
ment of novelty and discovery.” During flow, people are
successfully applying above-average skills to meet the demands
of above-average tasks. In other words, the extent to which a
person experiences flow on a certain day depends not only on
characteristics of the task or the activity he or she is engaging
in (e.g., Fullagar & Kelloway, 2009;Nielsen & Cleal, 2010),
but also on the extent of personal, energetic resources that the
person is able to invest on that day.
In the current study, we use the effort-recovery model (ERM;
Meijman & Mulder, 1998), a resource-oriented approach, as an
overall framework when examining work-related flow. The ERM
theorizes that people need to invest resources in order to fulfill
their work demands. Investing these resources, however, leads to
load reactions and resource depletion (see also Hobfoll, 1989).
During the subsequent recovery process (as should happen during
leisure time), demands are reduced and invested resources are
replenished and restored. Related to this, Binnewies et al. (2009)
argued that a successful recovery process results in the subjective
state of being recovered, that is, a state of feeling refreshed and full
of energy. This state of being recovered in the morning of a
specific working day is a consequence of the previous rest period,
indicating that energetic and affective resources have been replen-
ished. At the same time, it is the initial state that makes people well
equipped for the following working period (Sonnentag & Zijlstra,
2006).
Empirical research has already supported the assumption that
a person’s relative day-specific state of being recovered in the
morning predicts performance-related behaviors on the same
day, such as task performance, personal initiative, and organi-
zational citizenship behavior (Binnewies et al., 2009). Further-
more, a recent study (Sonnentag et al., 2012) has also demon-
strated a link with daily work engagement as a state that
captures a person’s subjective perception of his or her work. As
such, these findings make clear that the extent of being well
recovered not only contributes to behavior-related indicators,
but is also reflected in the way a person experiences the
workday; that is, how he or she is “living through” it (Weiss &
Rupp, 2011, p. 87). Hence, in an attempt to extend previous
findings in the context of subjective experiences at work, we
propose that when a person comes to work in a more well-
recovered state than on average, he or she will be more mentally
and physically refreshed and will experience a relatively high
level of energy. This state of recovery and high energy will
enable the person to invest his or her resources into demanding
tasks of above-average difficulty and to become fully immersed
and concentrated on these tasks. On these days, as a conse-
quence, the individual will be able to experience the highly
pleasant state of flow.
Generally, we argue that the postulated effect of being re-
covered will emerge above and beyond the effects of activity
type. As mentioned previously, type of activity has been shown
to be a crucial predictor of flow experiences (e.g., Nielsen &
Cleal, 2010). However, we argue that task type is not the only
route to flow, but that personal resources, especially being
recovered, matter as well. In sum, our first hypothesis states:
Hypothesis 1: A person’s relative day-level state of being
recovered is positively related to his or her day-level flow
experiences (i.e., the person’s average flow levels on that
same day), over and above the effects of activity type.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
714 DEBUS, SONNENTAG, DEUTSCH, AND NUSSBECK
The State of Being Recovered as a Predictor of
Within-Day Fluctuations of Flow
As mentioned in the introduction, flow has also been shown to
exhibit substantial variation within days (Ceja & Navarro, 2011).
Research on circadian rhythms in humans (e.g., Van Dongen &
Dinges, 2000) suggests that over different time points (i.e., occa-
sions) within the working day, this variation does not occur com-
pletely at random, but may exhibit a systematic, U-shaped pattern.
Specifically, it has been shown that several measures of alertness
and cognitive functioning temporarily decline around noontime
and then rise again (e.g., Hildebrandt, Rohmert, & Rutenfranz,
1974;Monk, 2005;Rodríguez-Sánchez, Schaufeli, Salanova, Ci-
fre, & Sonnenschein, 2011), which is mirrored by body tempera-
ture, hormone levels, and other physiological processes (e.g.,
Monk, 2005;Van Dongen & Dinges, 2000). This particular
U-shaped pattern in cognitive and physiological functioning is
commonly known as the “postlunch dip” (e.g., Blake, 1967;
Colquhoun, 1971;Monk, 2005) and is mostly an effect of the
circadian clock, a tiny cluster of about 10,000 nerve cells located
in the suprachiasmatic nucleus in the hypothalamus (e.g., Klein,
Moore, & Reppert, 1991). As flow is a state that goes along with
alertness and optimal cognitive functioning (e.g., Csikszentmi-
halyi, 1990,1996), it is reasonable to assume that flow may follow
a curvilinear pattern within the working day. Again, we propose
that this pattern will emerge above and beyond the impact of task
type. Hence, our second hypothesis states:
Hypothesis 2: A person’s flow levels exhibit a curvilinear (i.e.,
U-shaped) pattern over different occasions within the working
day, over and above the effects of activity type.
Nonetheless, research has shown that this curvilinear pattern is
not as robust and universal as one might think (e.g., Lack &
Lushington, 1996;Van Dongen & Dinges, 2000). In fact, it ap-
pears that the postlunch dip does not necessarily occur in all
individuals or in all cognitive and performance measures likewise
(e.g., Carrier & Monk, 2000;Lack & Lushington, 1996;Smith &
Maben, 1993;Van Dongen & Dinges, 2000). Moreover, studies
have shown that certain factors can change the shape of the curve
(e.g., Smith & Maben, 1993;Van Dongen & Dinges, 2000). As an
example, Hildebrandt et al. (1974) showed that the extent of train
conductors’ early afternoon peak in error frequency was dependent
on their accumulated tiredness.
In line with these findings, researchers have suggested that in
addition to the aforementioned circadian process, which affects
human functioning by driving wakefulness and alertness, there is
a parallel homeostatic process. This parallel process impairs hu-
man functioning by driving a person’s need for sleep and recovery
(e.g., Borbély, 1982;Edgar, Dement, & Fuller, 1993;Van Dongen
& Dinges, 2000). Beginning with a person’s awakening, the ho-
meostatic process gradually increases over the day. Hence, this
reasoning suggests that stimuli and personal states relating to
resource availability and fatigue may heighten one process relative
to the other (cf. Van Dongen & Dinges, 2000).
On the basis of the above arguments, a person’s recovery level
appears to be an important boundary condition for the pattern of
cognitive functioning and subjective states of mind within the
working day. Hence, besides the effects on day-level flow pro-
posed above, being better recovered in the morning should predict
the specific time course of flow experiences within the working
day. In particular, we propose that being poorly recovered consti-
tutes an additional burden for the organism during the entire
working day (e.g., Van Dongen & Dinges, 2000). Hence, when
people are poorly recovered, the homeostatic process should more
strongly interfere with the curvilinear pattern of flow experiences
(i.e., the circadian process). As a consequence, the curvilinear
pattern should be weaker, that is, emerge less markedly among
those people. In extreme cases, a person’s functioning might
become dominated by the homeostatic process. Again, we propose
that the hypothesized interaction effect emerges beyond the impact
of task type. Thus, our third hypothesis states:
Hypothesis 3: A person’s relative day-level state of being
recovered predicts the pattern of flow experiences over the
course of the working day: The U-shaped pattern of flow
experiences should be stronger on days with relatively higher
recovery levels compared to days with relatively lower recov-
ery levels, over and above the effects of task type.
Method
Sample and Procedure
The study was conducted with programmers, software engi-
neers, and web designers. We first approached software companies
by phone and informed them about the study. In total, 219 people
expressed interest in participating, but only 121 completed the
study with a sufficiently high response rate (specific criteria for
data inclusion are reported below).
Data were collected by one general survey before, and by four
surveys on each day during the study week (i.e., Monday to
Friday). We administered all surveys on the Internet by sending
participants personalized e-mail links to the questionnaires. Par-
ticipants received the link to the general survey immediately after
their registration; we sent links to the four surveys for each day
(i.e., one prework questionnaire and three questionnaires during
work) during a 5-day interval (Monday to Friday). This resulted in
20 daily surveys. We used experience sampling methodology
(ESM) to assess flow with the three questionnaires during work.
ESM tries to sample randomly from a person’s everyday experi-
ences (e.g., Csikszentmihalyi & Rathunde, 1993;Hormuth, 1986)
and is often used in studies assessing flow experiences (e.g.,
Csikszentmihalyi & Figurski, 1982). To be able to map measure-
ment occasions with participants’ individual circadian rhythms, we
asked participants when they usually start and finish work (in the
general survey). Accordingly, we sent ESM signals with links to
the daily questionnaires at three distinct time points during a
typical workday, approximating individual circadian rhythms. De-
pending on the time participants started work, the signal for
Occasion 1 was sent during the morning hours (between 9:06 and
11:44 a.m.), for Occasion 2 around noontime (between 11:18 a.m.
and 2:35 p.m.), and for Occasion 3 in the afternoon (between 1:15
and 5:53 p.m.). Generally, we sent signals to the ESM question-
naires at least 55 min after the previous one (but not more than 5
hours later).
Before analyzing the data, we deleted all questionnaires that
were answered less than 30 min after the previous one and ques-
tionnaires that had been answered before the signal (indicating that
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
715
RECOVERY AND FLOW AT WORK
participants were catching up a previously missed questionnaire).
Furthermore, we only included participants in the final analysis
who provided complete data sets from at least 3 days and had only
missed one of the three questionnaires during work. Among the
121 persons included in the final analyses, 18.2% were female.
Participants’ age ranged from 21 years to 63 years (M⫽35.5,
SD ⫽9.7); mean professional experience was 10.1 years (SD ⫽
8.9). Furthermore, 61.2% of participants held a university degree,
10.7% had completed 2- to 3-year professional training and ob-
tained an additional professional degree, 18.2% had completed 2-
to 3-year professional training, and 9.9% did not hold any profes-
sional degree. On average, participants worked 5.3 days (SD ⫽
0.6) a week and completed 8.7 hr (SD ⫽10.1) of overtime work
per week. Based on the timestamps that were collected in tandem
with the surveys, we calculated average answering delays to the
three questionnaires during work that measured flow, which were
9 min 53 s (SD ⫽8 min 11 s) for the first, 10 min 29 s (SD ⫽7
min 45 s) for the second, and 10 min 15 s (SD ⫽8 min 5 s) for the
third questionnaire.
Measures
For each participant, we collected data at three levels of analy-
sis, namely the level of the person (Level 3; i.e., variables that vary
between persons), the day (Level 2, i.e., variables that vary be-
tween days), and the occasion (Level 1, i.e., variables that vary
within days). We assessed occasion-level variables (i.e., flow
experiences and type of activity) with the three questionnaires
during work each day. Day-level state of being recovered was
assessed with the daily prework questionnaire, and all person-level
variables (i.e., general level of flow experiences and all person-
level control variables) were assessed with the general survey.
Occasion-level measures (i.e., Level 1). Occasion-level flow
was assessed with nine items from the Flow Short Scale by
Rheinberg, Vollmeyer, and Engeser (2003; for the English version,
see Rheinberg, 2008; for studies published in English using this
measure, see Engeser & Rheinberg, 2008;Schüler & Brunner,
2009) to be answered on a 7-point Likert scale, ranging from 1 ⫽
not true at all to7⫽very true. Participants were instructed to
report their flow experience at the present moment. Sample items
were “At the moment, I feel just the right amount of challenge”
and “At the moment, I am totally absorbed in what I am doing.”
Cronbach’s alpha ranged from .85 to .95 over all measurements.
Occasion number was coded as 1, 2, or 3 for every day and
every participant, depending on when he or she received the
respective signal.
Day-level measures (i.e., Level 2). State of being recovered
in the morning was assessed with the four-item scale by Sonnentag
and Kruel (2006). On a 5-point Likert scale from 1 ⫽not true at
all to5⫽very true, the scale measures the degree to which a
person feels recovered in the morning. A sample item was “This
morning, I feel well rested.” Cronbach’s alpha ranged from .91 to
.94 over the 5 days.
Measures in the general survey (i.e., Level 3). General level
of flow was assessed with the same nine items used for the
day-level surveys. Participants were instructed to indicate how
much the items reflect their experience of flow at work in general.
Cronbach’s alpha was .89.
Control Variables
To rule out alternative interpretations, we assessed a number of
control variables. To control for type of activity at the occasion
level, we asked participants to write down their current activity
when receiving the ESM signals. We then coded three activity
categories, namely programming activities (programming, fixing
bugs, etc.), administrative activities (bookkeeping, writing in-
voices, etc.), and personal activities (taking a break, chatting with
a colleague, etc.). Activities were then dummy coded, taking the
category “programming activities” as the reference category (both
dummies, d1 and d2, were coded as 0). “Administrative activities”
were coded d1 ⫽1 and d2 ⫽0, and “personal activities” were
coded d1 ⫽0 and d2 ⫽1.
At the person level, we controlled for person-level flow
experiences because they may impact the respective day-level
flow score (cf. Petrou, Demerouti, Peeters, Schaufeli, & Het-
land, 2012). On the basis of previous research concerning
facilitators and inhibitors of flow (e.g., Demerouti, 2006;Tri-
emer, 2001), we also controlled for general time pressure (Sem-
mer, 1984) and job control (Zapf, 1993), yielding Cronbach’s
alphas of .86 and .80, respectively. Finally, we also controlled
for age, days of work per week, start of work, end of work, and
overtime work (each measured with one item) at the person
level and for the respective day of the week at the day level
(e.g., Daniels & Harris, 2005).
Analytical Strategy
Due to the nested data structure (Level 1: occasions; Level 2:
days; Level 3: persons), we applied multilevel modeling using
the HLM 6.0 software (Raudenbush, Bryk, Cheong, Congdon,
& du Toit, 2009). In line with methodological recommendations
regarding diary studies, we centered person-level variables at
the grand mean and day-level variables at the respective person
mean (in order to remove all variance that is due to between-
person differences; Ohly et al., 2010;Petrou et al. 2012). We
recoded the occasion numbers prior to entering them into the
regression equation (Occasion 1 ⫽0, Occasion 2 ⫽1, Occasion
3⫽2, and 0, 1, and 4 for the quadratic trend), suggesting that
the intercept denotes the starting level in flow at the first
measurement occasion (e.g., Hox, 2010). To test the proposed
main effect and interaction hypotheses, we followed the rec-
ommendations by Aiken and West (1991) and compared a set of
seven nested models by means of a likelihood ratio statistic.
Unless otherwise noted, we entered all variables as fixed effects
into the regression equation. We chose this approach in order to
(a) keep the models as parsimonious as possible (e.g., Hox,
2010) and (b) avoid estimation problems specifically at Level 2
because there are only three measurement occasions at Level 1
for every day (for an analogous case in dyadic data analysis, see
Kenny, Kashy, & Cook, 2006). The literature suggests time
point variables to be entered along with the control variables in
order to establish a baseline for calculating any further effect
sizes (Bliese & Ployhart, 2002;Hox, 2010). Thus, our analysis
first provides a test for Hypothesis 2, and then for Hypothesis
1, followed by the test for Hypothesis 3.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
716 DEBUS, SONNENTAG, DEUTSCH, AND NUSSBECK
Results
Descriptive Statistics
Means, standard deviations, and zero-order correlations are dis-
played in Tables 1 and 2. We started by examining the variance
parts on each level of analysis by inspecting the null model,
including only the intercept (Bliese, 2000;LeBreton & Senter,
2008). For occasion-level flow experiences, person-level variance
(Level 3) accounted for 44.5%, day-level variance (Level 2) ac-
counted for 8.3%, and occasion-level variance (Level 1) accounted
for 47.2% of the total variance.
Hypotheses Testing
Table 3 presents the results of the multilevel regression analysis
used to test our hypotheses. The table also includes McFadden’s R
2
(McFadden, 1974), a conventional effect size measure for maxi-
mum likelihood procedures. In Models 1 and 2, we entered the
respective control variables, with general level of flow experiences
being a significant predictor (t⫽9.98, p⬍.001) in Model 1, and
both task-related dummy variables emerging as significant predic-
tors of flow experiences in Model 2 (t
d1
⫽⫺4.12, t
d2
⫽⫺3.41,
ps⬍.001). The finding suggests that flow occurs more often in
programming activities than in administrative and personal activ-
ities. Model 3 included occasion number (to control for a potential
linear trend; Aiken & West, 1991), which emerged as a significant
predictor (t⫽⫺2.76, p⬍.01). Due to the later test of the
cross-level interaction with day-level state of being recovered, we
entered linear occasion number with a random slope. The chi-
square test of residuals performed by the HLM 6.0 software
indicated that the respective random slope varied significantly
between days,
2
(492) ⫽592.34, p⬍.01. Contrary to Hypothesis
2, quadratic occasion number did not yield a significant effect (t⫽
1.46, ns) in Model 4. Hence, flow experiences did not exhibit a
general curvilinear pattern within the working day. Again, due to
the later cross-level interaction with day-level state of being re-
covered, we entered quadratic occasion with a random slope. This
random slope was not significant,
2
(408) ⫽447.88, p⬍.09,
based on the traditional 5% significance level.
1
In line with Hy-
pothesis 1, day-level state of being recovered significantly pre-
dicted day-level flow experiences in Model 5 (t⫽3.70, p⬍.001).
In Model 6, the Linear Occasion ⫻Day-Level State of Being
Recovered interaction (which served as a further control term;
Aiken & West, 1991) was not significant (t⫽0.89, ns).
Because interaction effects can occur without a significant main
effect of the respective predictors (Cohen, Cohen, West, & Aiken,
2003; see also Goltz & Smith, 2010), we proceeded with testing
the Quadratic Occasion ⫻Day-Level State of Being Recovered
interaction in Model 7 (for the same approach, see Baer & Old-
ham, 2006). Moreover, cross-level interactions may occur also in
cases where no random slopes can be found (see Snijders &
Bosker, 2012; see also Bliese & Jex, 2002;Ilies, Scott, & Judge,
2006;La Huis & Ferguson, 2009). Thus, we tested the interaction
between quadratic occasion and day-level state of being recovered,
which proved to be statistically significant (t⫽2.26, p⬍.05).
2
An inspection of the interaction plot (see Figure 1) revealed that
the shape of the curve concerning high and low relative state of
being recovered looked somewhat different from what we had
hypothesized, however. In the case of a high relative state of being
recovered, participants experienced the hypothesized U-shaped
pattern with a dip in flow experiences around noontime and a
subsequent uplift. However, persons with a low relative state of
being recovered experienced no Ushape at all, but reported a
relatively steady level of flow until the second measurement oc-
casion (though not reaching the flow levels of the relatively
well-recovered participants) and then a decline. Thus, Hypothesis
3 was not supported.
3
Discussion
In line with the ERM (Meijman & Mulder, 1998), this
experience-sampling study showed that the more a person felt
recovered in the morning of a specific day (relative to his or her
average of feeling recovered), the more flow he or she experienced
on average during work that day. Although the data did not support
a curvilinear main effect of flow experiences within the working
day (above the effect of task type), we could show that a person’s
within-day course of flow experiences was affected by his or her
relative state of being recovered. More specifically, when being
well recovered, people experienced the predicted U-shaped pattern
of flow experiences across the working day, while we found a
different shape for relatively poorly recovered people (i.e., they
started low, stayed low around noontime, and experienced a de-
crease in flow in the afternoon).
As mentioned before, the shape of the flow patterns looked
somewhat different from our initial predictions, as the U-shaped
pattern only emerged among well-recovered persons. In fact, this
finding might be interpreted on the basis of the two-process model
(e.g., Van Dongen & Dinges, 2000) that we introduced earlier. Put
differently, when people are well recovered, their flow experiences
follow a curvilinear pattern, reflecting the previously mentioned
circadian process. However, when being poorly recovered, it ap-
1
Readers might be interested in whether the linear occasion term and the
quadratic occasion term also vary between persons. For this reason, we
added random slopes to the respective predictors in Model 3 (for the linear
occasion term) and in Model 4 (for the quadratic occasion term). Both
slopes did not vary significantly between persons:
2
(120) ⫽133.99, ns,
for the linear occasion term and
2
(120) ⫽128.85, ns, for the quadratic
occasion term. For this reason and to keep the models as parsimonious as
possible, we again removed these random slopes from Models 3 and 4 and
all further models of our analysis.
2
Due to the restricted number of Level 1 units (three measures per day),
only a restricted number of Level 2 random effects can be estimated
(Kenny et al., 2006). Thus, we excluded the Level 2 random intercept from
Model 4 onward. However, to control for spurious effects, we also ran all
models with a random Level 2 intercept, but without a random slope of
occasion number. The analyses yielded the same results; that is, the
Quadratic Occasion ⫻Day-Level State of Being Recovered term was a
significant predictor of occasion-level flow experiences (t⫽2.46, p⬍
.05).
3
Following the suggestion of one reviewer, we tested an alternative set
of models in which we entered linear occasion and quadratic occasion (in
Models 2 and 3, respectively) before entering the dummy-coded activity
types (in Model 4). In these models, both linear occasion (t⫽⫺2.90, p⬍
.01) and quadratic occasion (t⫽1.96, p⬍.05) were significant. In line
with Hypothesis 2, this finding suggests a curvilinear pattern of flow over
the course of the working day. After entering the two dummy variables,
however, quadratic occasion did not yield a significant effect any longer.
Thus, the specific U-shaped pattern of flow seems to be accounted for by
activity type.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
717
RECOVERY AND FLOW AT WORK
pears that the parallel homeostatic process dominates over the
circadian process, thereby leading to a more or less gradual de-
crease of flow experiences during the day. Hence, one may argue
that a certain level of recovery is essential for the “natural”
curvilinear pattern to emerge at all (for a similar argument on
workload and fatigue, see Grech, Neal, Yeo, Humphreys, & Smith,
2009).
In sum, our study is one of the first (cf. Demerouti, Bakker,
Sonnentag, & Fullagar, 2012) to connect the literature on flow at
work with the literatures on recovery and circadian processes. As
such, our study extends previous findings in several ways. First, as
mentioned above, we investigated a personal resource at the day-
level instead of focusing on situational aspects, as has been done
previously (Fullagar & Kelloway, 2009;Nielsen & Cleal, 2010).
Moreover, we applied the perspective of how a workday as a
whole may benefit from recovery (cf. Binnewies et al., 2009)tothe
context of flow. Even more important, we clearly extended this
literature by demonstrating how being recovered can affect the
time course of flow as a subjective, work-related perception over
the course of the day. Although the respective effects sizes were
relatively small in size (which might in part be due to the fact that
we applied a rather conservative testing strategy including several
control variables), theoretically, our findings suggest that the cir-
cadian and homeostatic processes, and the effect of recovery on
them, not only are relevant for flow, but might generalize to a more
broader range of on-the-job experiences—and ultimately also per-
formance outcomes.
There are two findings relating to the linear and curvilinear
effect of occasion that merit further consideration. First, although
the linear effect of occasion merely served as a control term in our
analysis (Aiken & West, 1991), the data revealed that its slope
significantly varied between days, suggesting that the strength of
this decline is affected by higher level variables. Although day-
level recovery failed to reach significance as a moderating vari-
able, future research might examine other variables (e.g., sleep
indicators) that particularly impact this decline. Second, our data
revealed that the slopes of both the linear and quadratic occasion
numbers did not vary between persons. Put differently, it appears
that the overall pattern of flow is relatively uniform across indi-
viduals.
Limitations, Directions for Future Research, and
Practical Implications
Our study has some limitations. First, as we only used self-
report measures, we cannot rule out that common method variance
(Podsakoff, MacKenzie, Lee, & Podsakoff, 2003) has inflated the
relationships. However, self-report measures are most likely the
best way to catch subjective states of feelings that were investi-
gated in this study. Furthermore, we obtained our focal constructs
at different times (i.e., in the morning and at three occasions during
work) and in a “right now” frame of reference, which is more
robust to avoiding common method bias than concurrent, retro-
spective measurements (see Bolger et al., 2003).
Second, our sample of software professionals was highly spe-
cific, which might limit the generalizability of our findings. We
chose this particular sample because the experience of flow is very
well documented among programmers and software engineers
Table 1
Means, Standard Deviations, and Zero-Order Correlations of All Person- and Day-Level Variables
Variable MSD 12345678910
1. Day-level state of being recovered 3.67 0.93 — .40
ⴱⴱ
2. Day-level flow experiences 5.14 0.91 .50
ⴱⴱ
—
3. General level of flow experiences 4.97 0.96 .40
ⴱⴱ
.69
ⴱⴱ
—
4. Time pressure 2.85 0.93 ⫺.03 .15 .16 —
5. Job control 4.16 0.62 .06 .26
ⴱⴱ
.34
ⴱⴱ
.23
ⴱ
—
6. Age 35.45 9.73 ⫺.05 .05 .19
ⴱ
.21
ⴱ
.36
ⴱⴱ
—
7. Start of work
a
1.36 0.48 ⫺.06 ⫺.11 ⫺.09 ⫺.15 .12 .04 —
8. End of work
b
1.60 0.49 ⫺.11 .03 .18
ⴱ
.32
ⴱⴱ
.37
ⴱⴱ
.21
ⴱ
.24
ⴱⴱ
—
9. Days of work per week 5.31 0.64 ⫺.04 .01 .12 .29
ⴱⴱ
.41
ⴱⴱ
.21
ⴱ
.11 .34
ⴱⴱ
—
10. Overtime work 8.67 10.09 .06 .09 .12 .49
ⴱⴱ
.11 .12 ⫺.04 .33
ⴱⴱ
.32
ⴱⴱ
—
Note. Correlations below the diagonal are person-level correlations (N⫽121), with day-level measures being aggregated to the person level. Correlations
above the diagonal are day-level correlations (N⫽493), with occasion-level flow (N⫽1,340) being aggregated to the day level.
a
1⫽before 9 a.m.; 2 ⫽between 9 a.m. and 12 noon.
b
1⫽between 3 and 6 p.m.; 2 ⫽after 6 p.m.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01 (two-tailed).
Table 2
Means, Standard Deviations, and Zero-Order Correlations of All Occasion-Level Variables
Variable MSD 123 45
1. Occasion number —
2. Quadratic occasion number (i.e., occasion
2
).99
ⴱⴱ
—
3. Occasion-level flow 5.14 1.08 ⫺.06
ⴱ
⫺.06
ⴱ
—
4. Occasion-level activities Dummy 1 0.64 0.48 .04 .04 ⫺.11
ⴱⴱ
—
5. Occasion-level activities Dummy 2 0.09 0.28 ⫺.01 ⫺.03 ⫺.05 ⫺.41
ⴱⴱ
—
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01 (two-tailed).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
718 DEBUS, SONNENTAG, DEUTSCH, AND NUSSBECK
(e.g., Roque, 2011; see also Chen, 2006). Related to this, however,
it might have been helpful to use a more fine-grained measure of
task type in the current study (instead of only differentiating
between three categories). Hence, it might be the case that results
differ for different samples and more fine-grained operationaliza-
tions of task type.
Third, although we demonstrated systematic patterns of flow
experiences over the course of the working day, our occasion-
related study design is a first but rough attempt to model the
proposed within-day processes. For this reason, future research
might benefit from sampling more occasions within the day and/or
sampling richer person-specific information. Apart from a person’s
start of work (which we took into account), circadian processes
may also be affected by several additional variables, such as hours
since waking up (e.g., Van Dongen & Dinges, 2000), food intake
(e.g., Karnani et al., 2011), and activities that were pursued before
coming to work. Hence, future research might model circadian
(work-related) processes more accurately by taking these variables
into account.
For future research dealing with recovery, flow, and other subjec-
tive and behavioral processes at work, it appears promising to take a
greater “magnifying lens” as has been done previously. Generally,
models that focus on the day level (i.e., diary studies; e.g., Mehl &
Conner, 2012) are currently experiencing a strong upsurge. However,
it might be helpful to examine whether certain states also fluctuate
within even shorter time frames. Besides the fact that the lowest level
of analysis also includes error variance (e.g., Bryk & Raudenbush,
1992), our data showed that the majority of variance in flow actually
accrued from the occasion level. Thus, ignoring such lower levels
might imply that researchers are only tapping the tip of the iceberg for
some psychological processes.
Table 3
Multilevel Models Predicting Occasion-Level Flow
Variable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7
Level 3 predictors (t)
Age ⫺1.38 ⫺1.19 ⫺1.16 ⫺1.10 ⫺1.09 ⫺1.08 ⫺1.07
Time pressure 1.08 1.07 1.12 1.18 1.18 1.18 1.17
Job control 1.62 1.54 1.53 1.43 1.41 1.41 1.40
Days of work per week ⫺1.30 ⫺1.38 ⫺1.33 ⫺1.21 ⫺1.22 ⫺1.21 ⫺1.22
Overtime work 0.61 0.44 0.43 0.50 0.52 0.52 0.52
Start of work 0.09 ⫺0.13 ⫺0.08 ⫺0.07 ⫺0.06 ⫺0.07 ⫺0.07
End of work ⫺1.68 ⫺1.49 ⫺1.52 ⫺1.49 ⫺1.46 ⫺1.47 ⫺1.45
General level of flow experiences 9.98
ⴱⴱⴱ
9.82
ⴱⴱⴱ
9.84
ⴱⴱⴱ
9.86
ⴱⴱⴱ
9.89
ⴱⴱⴱ
9.89
ⴱⴱⴱ
9.91
ⴱⴱⴱ
Level 2 predictors (t)
Day of the week 1.00 0.99 0.69 0.55 0.53 0.55 0.56
Day-level state of being recovered 3.70
ⴱⴱⴱ
1.98
ⴱ
2.72
ⴱⴱ
Level 1 predictors (t)
Dummy 1 (activity type) ⫺4.12
ⴱⴱⴱ
⫺3.99
ⴱⴱⴱ
⫺3.85
ⴱⴱⴱ
⫺3.75
ⴱⴱⴱ
⫺3.74
ⴱⴱⴱ
⫺3.74
ⴱⴱⴱ
Dummy 2 (activity type) ⫺3.41
ⴱⴱⴱ
⫺3.41
ⴱⴱⴱ
⫺2.88
ⴱⴱ
⫺2.80
ⴱⴱ
⫺2.82
ⴱⴱ
⫺2.84
ⴱⴱ
Occasion ⫺2.76
ⴱⴱ
⫺2.18
ⴱ
⫺2.24
ⴱ
⫺2.24
ⴱ
⫺2.23
ⴱ
Occasion
2
1.46 1.52 1.53 1.53
Cross-level interactions (t)
Occasion ⫻Day-Level State of
Being Recovered 0.89 ⫺1.90
Occasion
2
⫻Day-Level State of
Being Recovered 2.26
ⴱ
⫺2ⴱlog-likelihood 3,379.20 3,359.53 3,347.04 3,354.72 3,341.22 3,340.44 3,335.35
Differential ⫺2ⴱlog 89.29
ⴱⴱⴱ
19.67
ⴱⴱⴱ
12.50
ⴱⴱ
7.68
ⴱⴱ
13.50
ⴱⴱⴱ
0.78 5.09
ⴱ
⌬df 9231111
McFadden’s pseudo-R
2
.025 .031 .035 .033 .037 .037 .038
McFadden’s ⌬pseudo-R
2
.006 .004 .000 .004 .000 .001
Note. Data denote tvalues. Occasion ⫽linear occasion; Occasion
2
⫽quadratic occasion.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
4
5
6
7
noisaccO hgiHnoisaccO woL
Flow
Low sta te of b eing recovered
High state of being recovered
Occasion 1 Occasi on 2 Occasion 3
Flow experiences
Figure 1. Interaction of quadratic occasion number by day-level state of
being recovered on occasion-level flow experiences.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
719
RECOVERY AND FLOW AT WORK
Finally, future research might also examine in more detail the
exact mechanism and relevant boundary conditions of the
recovery–flow link. Although we have shown that recovery is
related to flow, the translation of available resources into flow
experiences might happen via intermediary variables, such as
effort or resource allocation (e.g., Pierro, Kruglanski, & Higgins,
2006), and may be enhanced or buffered by factors such as
intrinsic motivation originating from task characteristics. Hence,
future studies might delve more deeply into the mechanisms un-
derlying the relationship between recovery and flow.
4
Practically speaking, our study highlights that to experience
flow it is important to recover well during nonwork time. Research
suggests that leisure time experiences such as mental detachment
and relaxation help people to recover (e.g., Sonnentag, Binnewies,
& Mojza, 2008). Hence, managers should encourage their employ-
ees to disengage from work when at home and to find some way
of relaxing and mentally detaching from work. Moreover, on
low-recovery days, people might be well advised to engage in rest
periods to “reload” their resources (Trougakos, Hideg, Cheng, &
Beal, 2013). By deliberately engaging in breaks, employees might
thus be able to prevent a gradual decrease in flow experiences.
In sum, we showed that day-level resources in the form of
feeling recovered affect not only a person’s average level of flow
that day, but also the particular course of flow experiences within
that same day. In doing so, our study highlights the necessity of
examining variation in flow within the working day and the
importance of moving from task characteristics to personal re-
sources when studying flow at work.
4
We thank an anonymous reviewer for pointing out these ideas.
References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and
interpreting interactions. Thousand Oaks, CA: Sage.
Baer, M., & Oldham, G. R. (2006). The curvilinear relation between
experienced creative time pressure and creativity: Moderating effects of
openness to experience and support for creativity. Journal of Applied
Psychology, 91, 963–970. doi:10.1037/0021-9010.91.4.963
Bakker, A. B. (2005). Flow among music teachers and their students: The
crossover of peak experiences. Journal of Vocational Behavior, 66,
26–44. doi:10.1016/j.jvb.2003.11.001
Bakker, A. B. (2008). The work-related flow inventory: Construction and
initial validation of the WOLF. Journal of Vocational Behavior, 72,
400–414. doi:10.1016/j.jvb.2007.11.007
Binnewies, C., Sonnentag, S., & Mojza, E. J. (2009). Daily performance at
work: Feeling recovered in the morning as a predictor of day-level job
performance. Journal of Organizational Behavior, 30, 67–93. doi:
10.1002/job.541
Blake, M. J. F. (1967). Time of day effects on performance in a range of
tasks. Psychonomic Science, 9, 349–350. doi:10.3758/BF03327842
Bliese, P. D. (2000). Within-group agreement, non-independence, and
reliability: Implications for data aggregation and analysis. In K. J. Klein
& S. W. Kozlowski (Eds.), Multilevel theory, research, and methods in
organizations (pp. 349–381). San Francisco, CA: Jossey-Bass.
Bliese, P. D., & Jex, S. M. (2002). Incorporating a multilevel perspective
into occupational stress research: Theoretical, methodological, and prac-
tical implications. Journal of Occupational Health Psychology, 7, 265–
276. doi:10.1037/1076-8998.7.3.265
Bliese, P. D., & Ployhart, R. E. (2002). Growth modeling using random
coefficient models: Model building, testing, and illustrations. Organiza-
tional Research Methods, 5, 362–387. doi:10.1177/109442802237116
Bolger, N., Davis, A., & Rafaeli, E. (2003). Diary methods: Capturing life
as it is lived. Annual Review of Psychology, 54, 579–616. doi:10.1146/
annurev.psych.54.101601.145030
Borbély, A. A. (1982). A two process model of sleep regulation. Human
Neurobiology, 1, 195–204.
Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models:
Application and data analysis methods. Newbury Park, CA: Sage.
Carrier, J., & Monk, T. H. (2000). Circadian rhythms of performance: New
trends. Chronobiology International, 17, 719–732. doi:10.1081/CBI-
100102108
Ceja, L., & Navarro, J. (2011). Dynamic patterns of flow in the workplace:
Characterizing within-individual variability using a complexity science
approach. Journal of Organizational Behavior, 32, 627–651. doi:
10.1002/job.747
Chen, H. (2006). Flow on the net-detecting web users’ positive affects and
their flow states. Computers in Human Behavior, 22, 221–233. doi:
10.1016/j.chb.2004.07.001
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple
regression/correlation analysis for the behavioral sciences (3rd ed.).
Mahwah, NJ: Erlbaum.
Colquhoun, W. P. (1971). Biological rhythms and human performance.
London, England: Academic Press.
Csikszentmihalyi, M. (1990). Flow: The psychology of optimal experience.
New York, NY: Harper Perennial.
Csikszentmihalyi, M. (1996). Creativity: Flow and the psychology of
discovery and invention. New York, NY: Harper Perennial.
Csikszentmihalyi, M. (1999). If we are so rich, why aren’t we happy?
American Psychologist, 54, 821–827. doi:10.1037/0003-066X.54.10
.821
Csikszentmihalyi, M. (2000). Beyond boredom and anxiety: Experiencing
flow in work and play. San Francisco, CA: Jossey-Bass. (Original work
published 1975)
Csikszentmihalyi, M., Abuhamdeh, S., & Nakamura, J. (2005). Flow. In
A. J. Elliot & C. S. Dweck (Eds.), Handbook of competence and
motivation (pp. 598–608). New York, NY: Guilford Press.
Csikszentmihalyi, M., & Figurski, T. J. (1982). Self-awareness and aver-
sive experience in everyday life. Journal of Personality, 50, 15–28.
doi:10.1111/j.1467-6494.1982.tb00742.x
Csikszentmihalyi, M., & Rathunde, K. (1993). The measurement of flow in
everyday life: Toward a theory of emergent motivation. In R. A. Dien-
stbier & J. E. Jacobs (Eds.), Nebraska Symposium on Motivation: Vol.
40.Developmental perspectives on motivation (pp. 57–97). Lincoln:
University of Nebraska.
Daniels, K., & Harris, C. (2005). A daily diary study of coping in the
context of the job demands–control–support model. Journal of Voca-
tional Behavior, 66, 219–237. doi:10.1016/j.jvb.2004.10.004
Demerouti, E. (2006). Job characteristics, flow, and performance: The
moderating role of conscientiousness. Journal of Occupational Health
Psychology, 11, 266–280. doi:10.1037/1076-8998.11.3.266
Demerouti, E., Bakker, A. B., Sonnentag, S., & Fullagar, C. (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.
doi:10.1002/job.760
Edgar, D. M., Dement, W. C., & Fuller, C. A. (1993). Effect of SCN
lesions on sleep in squirrel monkeys: Evidence for opponent processes in
sleep–wake regulation. Journal of Neuroscience, 13, 1065–1079.
Eisenberger, R., Jones, J. R., Stinglhamber, F., Shanock, L., & Randall,
A. T. (2005). Flow experiences at work: For high need achievers alone?
Journal of Organizational Behavior, 26, 755–775. doi:10.1002/job.337
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
720 DEBUS, SONNENTAG, DEUTSCH, AND NUSSBECK
Engeser, S., & Rheinberg, F. (2008). Flow, performance and moderators of
challenge–skill–balance. Motivation and Emotion, 32, 158–172. doi:
10.1007/s11031-008-9102-4
Fullagar, C. J., & Kelloway, E. K. (2009). “Flow” at work: An experience
sampling approach. Journal of Occupational and Organizational Psy-
chology, 82, 595–615. doi:10.1348/096317908X357903
Goltz, H. H., & Smith, M. L. (2010). Yule–Simpson’s paradox in research.
Practical Assessment, Research & Evaluation, 15(15). Retrieved from
http://pareonline.net/pdf/v15n15.pdf
Grech, M. R., Neal, A., Yeo, G., Humphreys, M., & Smith, S. (2009). An
examination of the relationship between workload and fatigue within
and across consecutive days of work: Is the relationship static or dy-
namic? Journal of Occupational Health Psychology, 14, 231–242. doi:
10.1037/a0014952
Hildebrandt, G., Rohmert, W., & Rutenfranz, J. (1974). 12 and 24h
rhythms in error frequency of locomotive drivers and the influence of
tiredness. International Journal of Chronobiology, 2, 175–180.
Hobfoll, S. E. (1989). Conservation of resources: A new attempt at con-
ceptualizing stress. American Psychologist, 44, 513–524. doi:10.1037/
0003-066X.44.3.513
Hormuth, S. (1986). The sampling of experiences in situ. Journal of
Personality, 54, 262–293. doi:10.1111/j.1467-6494.1986.tb00395.x
Hox, J. J. (2010). Multilevel analysis: Techniques and applications. Hove,
England: Routledge.
Ilies, R., Scott, B. A., & Judge, T. A. (2006). The interactive effects of
personal traits and experienced states on intraindividual patterns of
citizenship behavior. Academy of Management Journal, 49, 561–575.
doi:10.5465/AMJ.2006.21794672
Karnani, M. M., Apergis-Schoute, J., Adamantidis, A., Jensen, L. T., de
Lecea, L., Fugger, L., & Burdakov, D. (2011). Activation of central
orexin/hypocretin neurons by dietary amino acids. Neuron, 72, 616–
629. doi:10.1016/j.neuron.2011.08.027
Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data analysis.
New York, NY: Guilford Press.
Klein, D. C., Moore, R. Y., & Reppert, S. M. (1991). Suprachiasmatic
nucleus: The mind’s clock. New York, NY: Oxford University Press.
Kuo, T.-H., & Ho, L.-A. (2010). Individual difference and job perfor-
mance: The relationships among personal factors, job characteristics,
flow experience, and service quality. Social Behavior and Personality,
38, 531–552. doi:10.2224/sbp.2010.38.4.531
Lack, L. C., & Lushington, K. (1996). The rhythms of human sleep
propensity and core body temperature. Journal of Sleep Research, 5,
1–11. doi:10.1046/j.1365-2869.1996.00005.x
LaHuis, D. M., & Ferguson, M. W. (2009). The accuracy of significance
tests for slope variance components in multilevel random coefficient
models. Organizational Research Methods, 12, 418–435. doi:10.1177/
1094428107308984
LeBreton, J. M., & Senter, J. L. (2008). Answers to 20 questions about
interrater reliability and interrater agreement. Organizational Research
Methods, 11, 815–852. doi:10.1177/1094428106296642
McFadden, D. (1974). Conditional logit analysis of qualitative choice
behavior. In P. Zarembka (Ed.), Frontiers in econometrics (pp. 105–
142). New York, NY: Academic Press.
Mehl, M. R., & Conner, T. S. (Eds.). (2012). Handbook of research
methods for studying daily life. New York, NY: Guilford Press.
Meijman, T. F., & Mulder, G. (1998). Psychological aspects of workload.
In P. J. D. Drenth, H. Thierry, & C. J. de Wolff (Eds.), Handbook of
work and organizational psychology: Vol. 2.Work psychology (pp.
5–33). Hove, England: Psychology Press.
Monk, T. H. (2005). The post-lunch dip in performance. Clinics in Sports
Medicine, 24, e15–e23. doi:10.1016/j.csm.2004.12.002
Nakamura, J., & Csikszentmihalyi, M. (2002). The concept of flow. In
C. R. Snyder & S. J. Lopez (Eds.), Handbook of positive psychology (pp.
89–105). New York, NY: Oxford University Press.
Nielsen, K., & Cleal, B. (2010). Predicting flow at work: Investigating the
activities and job characteristics that predict flow states at work. Journal
of Occupational Health Psychology, 15, 180–190. doi:10.1037/
a0018893
Ohly, S., Sonnentag, S., Niessen, C., & Zapf, D. (2010). Diary studies in
organizational research: An introduction and some practical recommen-
dations. Journal of Personnel Psychology, 9, 79–93. doi:10.1027/1866-
5888/a000009
Petrou, P., Demerouti, E., Peeters, M. C. W., Schaufeli, W. B., & Hetland,
J. (2012). Crafting a job on a daily basis: Contextual correlates and the
link to work engagement. Journal of Organizational Behavior, 33,
1120–1141. doi:10.1002/job.1783
Pierro, A., Kruglanski, A. W., & Higgins, E. T. (2006). Progress takes
work: Effects of the locomotion dimension on job involvement, effort
investment and task performance in organizations. Journal of Applied
Social Psychology, 36, 1723–1743. doi:10.1111/j.0021-9029.2006
.00078.x
Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003).
Common method biases in behavioral research: A critical review of the
literature and recommended remedies. Journal of Applied Psychology,
88, 879–903. doi:10.1037/0021-9010.88.5.879
Raudenbush, S. W., Bryk, A. S., Cheong, Y. F., Congdon, R., & du Toit,
M. (2009). HLM6: Hierarchical linear and nonlinear modeling. Lincol-
nwood, IL: Scientific Software International.
Rheinberg, F. (2008). Intrinsic motivation and flow. In H. Heckhausen &
J. Heckhausen (Eds.), Motivation and action (2nd ed., pp. 323–348).
Cambridge, England: Cambridge University Press. doi:10.1017/
CBO9780511499821.014
Rheinberg, F., Vollmeyer, R., & Engeser, S. (2003). Die Erfassung des
Flow-Erlebens [The measurement of flow experiences]. In J.
Stiensmeier-Pelster & F. Rheinberg (Eds.), Diagnostik von Motivation
und Selbstkonzept (Tests und Trends N. F. 2) (pp. 261–279). Göttingen,
Germany: Hogrefe.
Rodríguez-Sánchez, A. M., Schaufeli, W., Salanova, M., Cifre, E., &
Sonnenschein, M. (2011). Enjoyment and absorption: An electronic
diary study on daily flow patterns. Work & Stress, 25, 75–92. doi:
10.1080/02678373.2011.565619
Roque, C. (2011, January 28). From choke to flow: Breaking down your
creative blocks [Blog post]. Retrieved from http://gigaom.com/2011/01/
28/from-choke-to-flow-breaking-down-your-creative-blocks/
Salanova, M., Bakker, A., & Llorens, S. (2006). Flow at work: Evidence
for an upward spiral of personal and organizational resources. Journal of
Happiness Studies, 7, 1–22. doi:10.1007/s10902-005-8854-8
Schüler, J., & Brunner, S. (2009). The rewarding effect of flow experience
on performance in a marathon race. Psychology of Sport and Exercise,
10, 168–174. doi:10.1016/j.psychsport.2008.07.001
Semmer, N. (1984). Stressbezogene Tätigkeitsanalyse [Stress-oriented task
analysis]. Weinheim, Germany: Beltz.
Smith, A. P., & Maben, A. (1993). Effects of sleep deprivation, lunch, and
personality on performance, mood, and cardiovascular function. Physi-
ology & Behavior, 54, 967–972. doi:10.1016/0031-9384(93)90310-C
Snijders, T. A. B., & Bosker, R. J. (2012). Multilevel analysis: An intro-
duction to basic and advanced multilevel modeling (2nd ed.). London,
England: Sage.
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. doi:10.1037/0021-9010
.93.3.674
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 Organi-
zational Psychology, 15, 197–217. doi:10.1080/13594320500513939
Sonnentag, S., Mojza, E. J., Demerouti, E., & Bakker, A. B. (2012).
Reciprocal relations between recovery and work engagement: The mod-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
721
RECOVERY AND FLOW AT WORK
erating role of job stressors. Journal of Applied Psychology, 97, 842–
853. doi:10.1037/a0028292
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. doi:10.1037/0021-9010
.91.2.330
Triemer, A. (2001). Ambulantes psychophysiologisches 24-Stunden-
Monitoring zur Erfassung von arbeitsbezogenen Stimmungen und Emo-
tionen [Ambulant psychophysiologic 24 hours monitoring for the as-
sessment of work-related moods and emotions] (Unpublished doctoral
dissertation). TU Dresden, Germany.
Trougakos, J., Hideg, I., Cheng, B., & Beal, D. (2013). Lunch breaks
unpacked: The role of autonomy as a moderator of recovery during
lunch. Academy of Management Journal. Advance online publication.
doi:10.5465/amj.2011.1072
Van Dongen, H. P. A., & Dinges, D. F. (2000). Circadian rhythms in
fatigue, alertness and performance. In M. H. Kryger, T. Roth, & W. C.
Dement (Eds.), Principles and practice of sleep medicine (3rd ed., pp.
391–399). Philadelphia, PA: Saunders.
Weiss, H. M., & Rupp, D. E. (2011). Experiencing work: An essay on a
person-centric work psychology. Industrial and Organizational Psy-
chology, 4, 83–97. doi:10.1111/j.1754-9434.2010.01302.x
Zapf, D. (1993). Stress-oriented analysis of computerized office work.
European Work and Organizational Psychologist, 3, 85–100. doi:
10.1080/09602009308408580
Received May 31, 2012
Revision received December 13, 2013
Accepted January 2, 2014 䡲
Members of Underrepresented Groups:
Reviewers for Journal Manuscripts Wanted
If you are interested in reviewing manuscripts for APA journals, the APA Publications and
Communications Board would like to invite your participation. Manuscript reviewers are vital to the
publications process. As a reviewer, you would gain valuable experience in publishing. The P&C
Board is particularly interested in encouraging members of underrepresented groups to participate
more in this process.
If you are interested in reviewing manuscripts, please write APA Journals at Reviewers@apa.org.
Please note the following important points:
•To be selected as a reviewer, you must have published articles in peer-reviewed journals. The
experience of publishing provides a reviewer with the basis for preparing a thorough, objective
review.
•To be selected, it is critical to be a regular reader of the five to six empirical journals that are most
central to the area or journal for which you would like to review. Current knowledge of recently
published research provides a reviewer with the knowledge base to evaluate a new submission
within the context of existing research.
•To select the appropriate reviewers for each manuscript, the editor needs detailed information.
Please include with your letter your vita. In the letter, please identify which APA journal(s) you
are interested in, and describe your area of expertise. Be as specific as possible. For example,
“social psychology” is not sufficient—you would need to specify “social cognition” or “attitude
change” as well.
•Reviewing a manuscript takes time (1–4 hours per manuscript reviewed). If you are selected to
review a manuscript, be prepared to invest the necessary time to evaluate the manuscript
thoroughly.
APA now has an online video course that provides guidance in reviewing manuscripts. To learn
more about the course and to access the video, visit http://www.apa.org/pubs/authors/review-
manuscript-ce-video.aspx.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
722 DEBUS, SONNENTAG, DEUTSCH, AND NUSSBECK
A preview of this full-text is provided by American Psychological Association.
Content available from Journal of Applied Psychology
This content is subject to copyright. Terms and conditions apply.