Emotion Regulation as a Determinant of Recovery ... · Resources theory (Hobfoll, 1989), we expect that emotional stress during signiﬁcant study-related events negatively impacts

Emotion Regulation as a Determinant of RecoveryExperiences and Well-Being: A Day-Level Study

Eva Maria Schraub, Sarah Turgut, Vera Clavairoly,and Karlheinz SonntagHeidelberg University

This study examined the impact of two emotion regulation strategies,reappraisal and expressive suppression, on recovery experiences andaffective well-being after significant study-related events. In a sample of63 undergraduate students who completed a time-contingent daily diaryover 14 consecutive days (726 diary entries), the assumption that per-ceived emotional stress during study-related events would reduce affec-tive well-being at bedtime ( � �0.28, p � .001) was supported.Multilevel analyses further showed that recovery experiences partiallymediated this negative relationship ( � 0.39, p � .001). As postulated,reappraisal buffered the adverse effects of emotional stress on recoveryexperiences ( � 0.05, p � .01). Unexpectedly, expressive suppressionhad the same buffering effect ( � 0.04, p � .05). We conclude thatan additional, fine-grained focus on context and time would usefullyenhance our knowledge of the effects of emotion regulation on stressresponses.

Keywords: diary study, emotion regulation, well-being, recovery

After stressful events, people need time to recover to restore theirresources (Meijman & Mulder, 1998). Accordingly, recent evidenceshows that recovery experiences are positively related to different mea-sures of psychological well-being (e.g., Geurts & Sonnentag, 2006; Hahn,Binnewies, Sonnentag, & Mojza, 2011; Sonnentag, 2003; Sonnentag &Fritz, 2007). However, while studies indicate that high work demands

Eva Maria Schraub, Sarah Turgut, and Karlheinz Sonntag, Department of Industrial andOrganizational Psychology, Psychological Institute, Heidelberg University, Germany; VeraClavairoly, Department of Psychological Aging Research, Psychological Institute, HeidelbergUniversity, Germany.

Correspondence concerning this article should be addressed to Eva Maria Schraub, Workand Organizational Psychology, Heidelberg University, Hauptstrasse 47-51, 69117 Heidelberg,Germany. E-mail: [email protected]

International Journal of Stress Management © 2013 American Psychological Association2013, Vol. 20, No. 4, 309–335 1072-5245/13/$12.00 DOI: 10.1037/a0034483

309

increase the risk of not being able to relax after work (Cropley & Purvis,2003; Rau, 2006; Sonnentag & Bayer, 2005), the follow-up questionremains largely unanswered (for an exception, see, e.g., Sonnentag &Kruel, 2006): Which determinants impede or facilitate recovery experi-ences after demanding and stressful days? To ground practical advice(e.g., for stress management training design) on empirical evidence, weneed to identify the processes that influence recovery experiences afterstressful days.

One process that may explain recovery from job stress is emotionregulation. The Job Demands-Resources model (Demerouti, Bakker,Nachreiner, & Schaufeli, 2001) predicts that personal resources moderatethe consequences of job demands (Bakker & Demerouti, 2007; Xantho-poulou, Bakker, Demerouti, & Schaufeli, 2007). In line with this predic-tion, research shows that individual differences in emotion regulationaffect the way work-related emotional events relate to individual well-being and performance. For example, studies by Ciarrochi, Dean, andAnderson (2002) and by Giardini and Frese (2006) indicate that emotion-ally competent people who actively address emotional job stressors dohave less adverse health effects later. Raftery and Bizer (2009) found thatnegative feedback enhanced test performance of individuals who habitu-ally reinterpreted (i.e., reappraised) situations (“reappraisers”), but didnot affect test performance of individuals suppressing the expression oftheir feelings (“suppressors”). Schraub, Stegmaier, and Sonntag (2011)showed that changes at their workplace did not affect suppressors’ well-being. Extending this line of individual differences research, we examinethe impact of emotion regulation as a determinant of students’ recoveryfrom emotional stress during study-related events.

Altogether, our aim in this study is to integrate research lines on workstress, recovery, and emotion regulation. First, we examine the role ofrecovery experiences in the relationship between emotional stress duringstudy-related events and later affective well-being. Second, we investi-gate the impact of two different emotion regulation strategies on therelationship between emotional stress and recovery experiences. To ourknowledge, the role of emotion regulation has not yet been analyzed withregard to the recovery process. Third, we further extend research onemotion regulation by analyzing situational emotion regulation behaviorin a diary design. This design allows for detection of the effects ofintrapersonal variation in the use of specific emotion regulation strategies,while controlling for individual differences in emotion regulation. To ourknowledge, situational emotion regulation has so far only been analyzedin the emotional labor context (i.e., a context where emotions need to beregulated as part of one’s job; Hochschild, 1983); for example, in cus-tomer service (e.g., Totterdell & Holman, 2003).

310 Schraub, Turgut, Clavairoly, and Sonntag

THEORETICAL BACKGROUND AND HYPOTHESESDEVELOPMENT

Effects of Emotional Stress on Recovery Experiences and SubsequentAffective Well-Being

Emotional stress is characterized by negative emotional experiences,such as anger or anxiety (Chang, Johnson, & Yang, 2007). According to thePerson–Environment fit perspective (cf. Edwards, Caplan, & Harrison,1998), it indicates a lack of match (i.e., a mis-fit) between environmentaldemands or supplies and a person’s abilities or needs. Emotional stress canlead to a variety of negative consequences for individuals’ well-being,attitudes, and behaviors (cf. Fisher & Ashkanasy, 2000; Fredrickson &Joiner, 2002; Weiss & Cropanzano, 1996). Previous studies have shown thatevents at work that are being perceived as stressful are negatively related toaffective well-being (Elfering et al., 2005). If an emotional event at work isfurthermore being perceived as significant, the likelihood increases that itseffects spill over into leisure time (e.g., Williams & Alliger, 1994). In thepresent study we conceptualize the perceived stressfulness of a study-relatedevent as “an individual’s response to work-related environmental stressors”(Jex, Beehr, & Roberts, 1992, p. 623).

As proposed by Conservation of Resources theory (Hobfoll, 1989),people strive to protect and accumulate resources to deal with their environ-ment. Being a means to regenerate depleted resources (cf. Sonnentag, Per-rewé, & Ganster, 2009), recovery experiences should help people to restoreaffective well-being and associated personal resources such as optimism afterperiods of emotional stress. People experience recovery by refraining fromany activities involving effort expenditure so that resources can be rebuilt(Hobfoll, 1998; Meijman & Mulder, 1998) or by actively engaging inrecovery activities which foster new resources (Geurts & Sonnentag, 2006;Hobfoll, 1998; Sonnentag, 2001). Sonnentag and Fritz (2007) identifieddistinct recovery experiences that promote recovery. One of them is psycho-logical detachment, defined as an “individual’s sense of being away from thework situation” (Etzion, Eden, & Lapidot, 1998, p. 579). Another recoveryexperience is relaxation, which is “characterized by decreased sympatheticactivation and becomes evident in a decrease in heart rate, muscle tension,and other indicators of activation” (Sonnentag, Binnewies, & Mojza, 2008, p.675). Recent research demonstrates that resource-providing activities pro-mote the recovery experiences psychological detachment and relaxationwhich in turn are positively associated with well-being and with feelingrecovered (Ragsdale, Beehr, Grebner, & Han, 2011).

311Emotion Regulation, Recovery, and Well-Being

Several diary studies highlight the importance of adequate recoveryexperiences for well-being (cf. Demerouti, Bakker, Geurts, & Taris, 2009).Nevertheless, these studies also indicate that especially when resources arespent (e.g., because of high job demands), the risk of insufficient recoveryafter work increases (Cropley & Purvis, 2003; Rau, 2006; Sonnentag &Bayer, 2005). One reason for this effect may be prolonged cognitive engage-ment (i.e., lack of psychological detachment), which is a likely reaction toemotional stress (cf. Geurts & Sonnentag, 2006). Thus, it seems that recoveryexperiences are often impeded at precisely the times when they are mostneeded.

According to stress theories such as the Person–Environment Fit per-spective (cf. Edwards, Caplan, & Harrison, 1998) and the Conservation ofResources theory (Hobfoll, 1989), we expect that emotional stress duringsignificant study-related events negatively impacts people’s affective well-being at bedtime, which serves as an indicator of feeling recovered (Son-nentag, 2001). As recovery experiences during afterwork hours restore lostresources and positively affect peoples’ well-being (Demerouti et al., 2009),we further hypothesize that recovery experiences mediate the negative effectsof emotional stress on affective well-being. Thus, we formulated the follow-ing hypotheses:

Hypothesis 1: Emotional stress during a significant study-related eventnegatively affects affective well-being at bedtime.

Hypothesis 2: Recovery experiences mediate the negative relationshipbetween emotional stress during a significant study-related event andaffective well-being at bedtime.

Emotion Regulation as a Moderator of the Effects of Emotional Strain

Emotion regulation refers to people’s efforts to influence the experience,intensity, duration, or expression of activated emotions during emotionallydistressing events (Gross, 1998b). Gross (2001) developed a process-orientedmodel of emotion regulation to classify the strategies people use to regulatetheir emotions, and distinguished between antecedent-focused regulation andresponse-focused regulation. Antecedent-focused regulation (e.g., cognitivereappraisal of the situation) comes early in the emotion-generative process.Response-focused regulation (e.g., expressive suppression) is applied whenemotions are already fully experienced and only modifies the emotionaldisplay, not the experience. Researchers ascribed to and empirically foundquite diverse effects for these different types of emotion regulation.


Expressive suppression, on the one hand, is an ineffective strategy interms of altering emotional experiences. It decreases the emotional expres-sion, but not the intensity of the felt emotion (Gross, 2001; cf. Gross & John,2003). Because of the cognitive load that the suppression of emotionalexpression imposes, expressive suppression is associated with impairedmemory and social functioning. Various experimental studies demonstratedthis effect. Richards and Gross (1999, 2000), for example, found that studyparticipants who were asked to suppress emotional expressions during thepresentation of emotion-eliciting material showed decreased incidental mem-ory for information presented during the suppression period and increasedcardiovascular activation. Schmeichel, Vohs, and Baumeister (2003) re-vealed that after a required suppression period, study participants performeddemanding logical tasks significantly worse than a control group. Johns,Inzlicht, and Schmader (2008) revealed a similar effect for spontaneousexpressive suppression: Study participants who suppressed the expression ofnegative emotions performed worse on tests of cognitive ability afterward.As a study by Butler and colleagues (2003) shows, restricted cognitivecapacity because expressive suppression does not only affect mere cognitivetasks, but also translates to being less responsive in social interactions.Expressive suppression is further associated with negative long-term out-comes. People who frequently suppress their emotional expression appar-ently experience less positive and more negative emotions and have lowercognitive capacity as well as worse cardiovascular functioning (Richards,2004; Srivastava, Tamir, McGonigal, John, & Gross, 2009). A longitudinalstudy showed that suppressors had weaker social connections at the end ofcollege (English, John, Srivastava, & Gross, 2012). They also reported moredepressive symptoms, were less satisfied with their lives and relationships,and were more pessimistic about their future (Gross & John, 2003).

Reappraisal, on the other hand, is an effective emotion regulation strat-egy, because it occurs very early in the emotion generative process and theneven prevents the development of negative emotions (Gross, 2002). Thus, itimposes a much smaller cognitive load than expressive suppression, becauseit does not consume cognitive capacity to monitor one’s feelings and behav-ior later on (Gross, 2002; Richards & Gross, 2000). Accordingly, reappraisaldoes not negatively influence cognitive performance. For example, it did notdecrease memory when people were asked to reappraise the situation whileviewing emotional slides (Richards & Gross, 2000). In contrast to expressivesuppression, reappraisal does not impair interpersonal communication (But-ler et al., 2003). It is associated with positive long-term outcomes on health,memory, and social relationships (English et al., 2012; Gross & John, 2003;Richards & Gross, 2000; Srivastava et al., 2009).

Reviewing the emotion regulation literature, we found that with fewexceptions (e.g., Tschan, Rochat, & Zapf, 2005), most empirical studies


published so far were either experimental (e.g., Gross, 1998a), analyzedindividual differences (e.g., Ciarrochi et al., 2002; Giardini & Frese, 2006;Raftery & Bizer, 2009; Schraub et al., 2011), or focused on emotional labor(e.g., Totterdell & Holman, 2003). However, in environments where displayrules are weaker and more informal than they are in the service context (cf.Bono & Vey, 2005), people may determine for themselves when and how toregulate their emotions. Moreover, theories on interpersonal effects of emo-tion regulation (Côté, 2005; Van Kleef, 2009) and the independence ofemotion regulation styles suggest that people may apply different and some-times concurrent emotion regulation strategies depending on the context. Toboth complement and extend prior studies, we therefore chose to examinesituational regulation efforts instead of individual differences in this diarystudy. We adapted an emotion regulation questionnaire to situational regu-lation to gain insight into short-term consequences of specific emotionregulation behavior.

Concerning the effects of emotion regulation, we argue that the strategyof reappraising the situation might buffer negative effects of emotional stresson recovery experiences because it is applied early in the emotion-generativeprocess and changes peoples’ interpretations of the respective situation(Gross, 2001). Reappraisal changes the emotional experience to being morepositive, making subsequent regulation efforts unnecessary. Thus, resourcesare not bound for further self-control. We hypothesize that:

Hypothesis 3: Reappraisal buffers the negative impact of emotionalstress on recovery experiences.

In contrast, researchers relate expressive suppression to mainly negativeoutcomes because it does not change the experience of emotional stress, butonly suppresses its display, thereby consuming cognitive resources thatotherwise would be available for other tasks (Raftery & Bizer, 2009).Because of this heightened cognitive load, we expect this regulation strategyto interfere with recovery experiences and hypothesize that:

Hypothesis 4: Expressive suppression enhances the negative impact ofemotional stress on recovery experiences.

To summarize, we expect recovery experiences to be an explanatorymechanism for a negative relationship between emotional stress duringstudy-related events and affective well-being at bedtime. Moreover, we deemthe use of reappraisal and expressive suppression to differentially affect therelationship between emotional stress and recovery experiences. The frame-work that integrates the research questions is depicted in Figure 1.


METHOD

Contextual Factors

This diary study was conducted with undergraduate students of a Germanuniversity who belonged to the first year of students in a reorganizedcurriculum (i.e., bachelor/master programs replacing a diploma curriculum).All students had received excellent grades in their final high school exams.Facing so far unstandardized processes, first-year students often reportedinsecurity regarding curriculum requirements and examination procedures.Because they knew that they would eventually have to apply and compete fora place in a master’s program to obtain an academic degree that wouldqualify them for the jobs they desired, they also experienced some pressureconcerning their future careers. In addition, growing international competi-tion and the implementation of tuition in Germany contributed to makestudying a full-time job with a high stress level (Cooke, Bewick, Barkham,Bradley, & Audin, 2006; Obergfell & Schmidt, 2010).

Apart from classes at university, students need to self-determine theirtime and resources for their studies. According to the literature, recoverybecomes even more difficult in an unregulated work–life situation (Ahrent-zen, 1990; Cropley, Dijk, & Stanley, 2006; Sonnentag & Kruel, 2006).Although a student’s life may differ from an employee’s life concerning theamount of work hours, it is important to note that their average level of stressand strain may not necessarily differ from an employee’s level. Recentstudies show that students experience high levels of stressors and strains(e.g., Obergfell & Schmidt, 2010) and as Ragsdale and colleagues (2011) putit, “stress processes with a weekly cycle occur with students as well” (p. 154).In the final questionnaire of the diary, several participants reported that theyhad experienced the diary as a great opportunity to reflect on their study-related and private experiences and to become aware of certain behaviortendencies.

Recovery experiences

Affective well-being

Emotional stress

Expressive suppression Reappraisal

Figure 1. The proposed conceptual scheme.


Sample and Procedure

In return for research participation credits required by their schedule, 67full-time undergraduate students of a German university volunteered toparticipate in the study. All of them completed a paper-and-pencil question-naire containing questions about demographics and personal traits. They thenreceived a structured paper-based diary within which they had to answer a1-page questionnaire each night before going to bed on 14 consecutive days.The research assistants reminded participants of this task each night via shortmessages on their cell phones. They were assured of anonymous datatreatment, and that their cell phone numbers could not be assigned to theirdata. The research assistants also pointed out that they could be contacted incase of any questions or issues. We matched questionnaires by an individual,self-generated code.

Out of the 67 diaries that had been distributed, 65 were returned; thisequals a return rate of 97%. Because two participants had to be excludedbecause of being on holiday while participating in the study, the final sampleconsisted of 63 participants (51 women and 12 men) with an average age of21 years (SD � 2.9 years). All were full-time students, working on studyassignments for between 3 and 12 hr per day, with an average working timeof 4.8 hr per day (SD � 2.1).

Measures

The focus study variables emotional stress, recovery experiences, emo-tion regulation, and affective well-being at bedtime were assessed in thediary, whereas control variables were assessed in the general questionnaire.We instructed participants to refer to their studies when asked for work-related experiences.

Emotional Stress

Analogous to the procedure used by Gable, Reis, Impett, and Asher(2004), we asked participants to recapture their most significant study-relatedevent of the respective day and to briefly describe it. To assess emotionalstress during this event, the questionnaire then provided nine items from atranslated and adapted version of Fisher’s (2000) job emotion scale (Cole,Bruch, & Vogel, 2006). The participants had to rate their experience ofemotions such as “frustrated,” “disappointed,” and “worried” in relation tothe study-related event on a 5-point Likert scale ranging from 1 � “not at all”


to 5 � “very much.” Because this rating reflects a specific and situation basedevaluation of the perceived emotions during the study-related event inquestion, the emotional stress values reflect how stressful the event wasperceived. Cronbach’s alpha indicated a reliability of � � .89.

Recovery Experiences

We assessed two specific recovery experiences in the evening, psycho-logical detachment and relaxation, with items from Sonnentag and colleagues(2008) recovery experience questionnaire in its German version. In total, sixitems asked to what extent the participants detached from their studies andrelaxed. Example items are, “Tonight, I was able to forget about universitywork” (psychological detachment from work) and, “Tonight, I was doingthings during which I was able to relax” (relaxation). Participants could ratethe items on a scale ranging from 1 � “not at all” to 5 � “very much.” Toexamine the factor structure prior to aggregating the items of this scale, weconducted an exploratory factor analysis. Without rotation, all items con-verged on one factor with an eigenvalue greater than one. This factoraccounted for 72.8% of the variance; all item loadings exceeded .82. Cron-bach’s alpha for the composite scale was � � .93.

Emotion Regulation

For the assessment of the participants’ emotion regulation, we adaptedfour reappraisal and two expressive suppression items from the Germanversion (Abler & Kessler, 2009) of Gross and John’s (2003) emotion regu-lation questionnaire to situational emotion regulation. This tool measurescognitive reappraisal and expressive suppression as two uncorrelated stylesof intrapersonal emotion regulation and demonstrates adequate psychometricproperties in terms of validity and reliability (Gross & John, 2003). We askedthe participants to indicate to what extent they reappraised the situation (e.g.,“I controlled my emotions by changing the way I think about the situation Iwas in”) and suppressed the expression of their feelings (e.g., “I kept myemotions to myself”) during the study-related event they had describedbeforehand. Participants answered via a 7-point Likert scale ranging from1 � “not at all” to 7 � “very much.” To assure that reappraisal andexpressive suppression formed two separate factors, we submitted all emo-tion regulation items to a principal components analysis with oblique rota-tion. Corroborating the measures’ discriminant validity, two factors emergedwith eigenvalues greater than 1, accounting for 78.0% of the variance. The


items’ primary loadings on their appropriate factors were greater than .82;cross-loadings were lower than .26. The internal consistency was � � .89 forreappraisal (Cronbach’s Alpha) and r � .80 for expressive suppression(Spearman’s correlation coefficient).

Affective Well-Being

We assessed affective well-being at bedtime with six items of thewell-being scale by Warr, Butcher, and Robertson (2004) that we translatedinto German using the back-translation procedure (Brislin, 1970). Partici-pants rated their general nonevent related well-being with these items (e.g.,“At the moment, I feel happy”) on a 5-point Likert scale ranging from 1 �“not at all” to 5 � “very much.” Because of the employed diary design thismeasure reflects a situational evaluation of the affective well-being. Cron-bach’s alpha for this scale was � � .83.

Controls

To ensure that day-level affective well-being could actually be explainedby the day-level predictors, we controlled for the sociodemographic data ageand gender (assessed with one item each), as well as for dispositionalaffectivity. Positive and negative affectivity significantly influence a person’srecovery, affective well-being, and performance (Connolly & Viswesvaran,2000; Lyubomirsky, King, & Diener, 2005; Marco & Suls, 1993; Watson &Clark, 1984). We measured dispositional affectivity using Krohne, Egloff,Kohlmann, and Tausch’s (1996) validated German version of the Positiveand Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988).Participants rated the extent to which they generally experience 10 positivefeelings (e.g., “I generally feel proud”) and 10 negative feelings (e.g., “Igenerally feel upset”) on a 5-point Likert scale ranging from 1 � “not at all”to 5 � “very much.” Cronbach’s alpha was � � .83 for positive affectivityand � � .89 for negative affectivity.

Data Analyses

With the diary design of this study, we collected repeated measurementdata. The two-level study consisted of day-level data (Level 1) and person-level data (Level 2), with days being nested in persons. For this kind of study,it is necessary to use the multilevel random coefficient modeling method


(MRCM; also called hierarchical linear modeling, HLM) to accurately fit thehierarchical data structure (e.g., Netzlek, Schröder-Abé, & Schütz, 2006;Raudenbush & Bryk, 2002). This method offers the advantage of workingwith different levels of analysis simultaneously, such that interrelations ondifferent levels are statistically independent of each other (Netzlek et al.,2006). In the analyses, each data level is treated as a formally independentsubmodel. Use of the MRCM for analyzing multilevel data are important forseveral reasons: First, the MRCM offers the advantage that any bias instandard errors and statistical tests because of the interdependence of the dataare adjusted for (Krull & MacKinnon, 2001; Snijders & Bosker, 1993).Second, MRCM takes the hierarchical data structure into account and thusavoids aggregation of the day-level variables which would result in a loss ofpower of the statistical tests because of the reduced number of higher levelunits. Thus, if the variability in the data was neglected, misinterpretation ofthe results would likely occur (Hox, 2010).

We used HLM 6.0 (Raudenbush, Bryk, Cheong, Congdon, & Du Toit,2004) for our analyses. We centered the person-level control variablespositive and negative affectivity at the grand mean and all day-level predic-tors at the respective person mean.

RESULTS

Descriptive Results

Participants reported 726 study-related events altogether (M � 11.5;SD � 2.3). Table 1 depicts the means, SDs, and correlations of all studyvariables. The correlation of r � .43 between reappraisal and expressivesuppression on the day-level indicates that these two strategies were oftenapplied in conjunction.

Hypotheses Testing

To test our hypotheses, we first calculated null models (Model 0) thatincluded the intercept as the only predictor. For data evaluation this step isnecessary, as it verifies whether sufficient variance exists in the criterionvariables on the day-level as well as on the person-level to be explained bythe respective predictors. For each hypothesis, we then added the relevantcontrol variables in a second model (Model 1) and afterward conductedanalyses with the predictors (Models 2 and 3). The models integrate everyparticipant by a regression line including an intercept and a slope. For each


Tab

le1.

Mea

ns,

SDs,

and

Inte

rcor

rela

tions

Bet

wee

nSt

udy

Var

iabl

es

Var

iabl

eM

SD1

23

56

78

9

1.E

mot

iona

lst

ress

1.96

0.96

�.2

5**

�.4

0**

.54*

*.4

0**

2.R

ecov

ery

expe

rien

ces

2.78

0.82

�.1

4.5

3**

�.1

0**

�.1

3**

3.A

ffec

tive

wel

l-be

ing

3.32

0.78

�.5

1**

.60*

*�

.16*

*�

.15*

*

5.R

eapp

rais

al2.

371.

52.2

9**

.14

�.0

6.4

3**

6.E

xpre

ssiv

esu

ppre

ssio

n2.

701.

76.3

8**

.04

�.0

9.5

0**

7.G

ende

ra—

—.0

5.1

4.1

6.0

6.1

78.

Age

21.2

42.

91�

.01

.08

.05

.29*

.18

.04

9.Po

sitiv

eaf

fect

ivity

3.57

0.49

�.3

4**

.35*

*.4

7**

.12

�.1

2�

.02

�.1

310

.N

egat

ive

affe

ctiv

ity2.

960.

66.3

6**

�.4

7**

�.4

9**

.03

.18

�.0

7.1

8�

.59*

*

Not

e.B

elow

diag

onal

:pe

rson

-lev

elda

ta(N

�63

),ab

ove

diag

onal

:da

y-le

vel

data

(N�

726)

.a

1�

fem

ale,

2�

mal

e.*

p�

.05.

**

p�

.01.


multilevel model, model fit indices (deviances) indicate the model fit for thedata. Differences of the deviances of two subsequent models follow achi-square distribution and indicate if the additional predictors explain asignificant additional amount of variance.

As Model 0 in Table 2 shows, the variance on both levels was indeedsufficient for both recovery experiences and affective well-being. Further-more, both reappraisal and expressive suppression showed high levels ofintrapersonal variance, indicating that it made sense to study their effects ona daily basis.

Next, we entered the control variables gender, age, and negative as wellas positive affectivity (Level 2) as predictors in Model 1. In Model 2, weadditionally entered emotional stress (Level 1). Finally, in Model 3, weincluded recovery experiences (Level 1). We tested each model for improvedfit over the previous model by calculating differences in the deviances (� �2log likelihood) and submitting them to a chi-square test. Table 3 depicts theresults.

The analysis showed that Model 1 improved significantly over Model 0(� �2 log likelihood � 25.16, df � 7, p � .001). The control variablespositive and negative affectivity were significant predictors in this model. Asproposed in Hypothesis 1, the intensity of emotional stress during a signif-icant study-related event negatively affects affective well-being in the lateevening. To test this hypothesis, we compared the model fit of Model 1 to theone of Model 2 in which the variable emotional stress was entered. As Model2 showed an improved model fit (� �2 log likelihood � 137.35, df � 8,p � .001), emotional stress contributed significantly to the prediction ofaffective well-being, and did so beyond the effects of negative and positiveaffectivity. Thus, data supported Hypothesis 1. The intensity of emotionalstress during a significant study-related event had a negative impact onaffective well-being at bedtime.

In Hypothesis 2, we postulated that recovery experiences mediate thenegative relationship between emotional stress and affective well-being.Comparing the model fit between Model 2 and Model 3, the difference

Table 2. Hierarchical Linear Modeling Estimates of Null Models

Dependent variable �00 �2 00

% of total variance thatis within-person

Affective well-being 3.32 0.46 .15*** 75.41Recovery experiences 2.78 0.56 .11*** 83.58Reappraisal 2.37 1.92 .38*** 83.48Expressive suppression 2.70 2.74 .35*** 88.67

Note. N � 726 occasions, N � 63 participants. �00 � pooled intercept; �2 � within-personvariance; 00 � between-person variance. % of total variance that is within-person wascomputed with the formula �2/(�2 � 00).*** p � .001.


Tab

le3.

Mul

tilev

elE

stim

ates

for

Mod

els

Pred

ictin

gA

ffec

tive

Wel

l-B

eing

:R

ecov

ery

Exp

erie

nces

asM

edia

tor

Mod

el0

Mod

el1

Mod

el2

Mod

el3

Est

imat

eSE

tE

stim

ate

SEt

Est

imat

eSE

tE

stim

ate

SEt

Inte

rcep

t3.

320.

0562

.65*

**

3.32

0.04

76.5

0***

3.32

0.04

76.6

2***

3.32

0.04

76.6

6***

Gen

der

0.15

0.11

1.36

0.15

0.11

1.36

0.15

0.11

1.37

Age

0.02

0.01

1.52

0.02

0.01

1.54

0.02

0.01

1.54

Posi

tive

affe

ctiv

ity0.

260.

122.

27*

0.27

0.11

2.40

*0.

270.

112.

40*

Neg

ativ

eaf

fect

ivity

�0.

210.

07�

2.86

**

�0.

200.

07�

2.80

**

�0.

200.

07�

2.81

**

Em

otio

nal

stre

ss�

0.28

0.03

�8.

52***

�0.

190.

03�

6.71

***

Rec

over

yex

peri

ence

s0.

390.

0411

.11*

**

Dev

ianc

e19

06.0

318

80.8

717

43.5

215

67.4

9�

Dev

ianc

e25

.16*

**

137.

35***

176.

03***

�df

34

11

R2

R2

R2

Lev

el1

(with

in-

pers

on)

Var

ianc

e0.

460.

460.

000.

400.

130.

320.

20L

evel

2(b

etw

een-

pers

on)

Var

ianc

e0.

150.

100.

330.

090.

100.

090.

00

Not

e.N

�72

6oc

casi

ons,

N�

63pa

rtic

ipan

ts.

Uns

tand

ardi

zed

regr

essi

onco

effic

ient

sar

ere

port

ed.

*p

�.0

5.**

p�

.01.

***

p�

.001

.


between the deviances was again significant (� �2 log likelihood � 176.03,df � 9, p � .001). This indicates that recovery experiences contributedsignificantly to the prediction of affective well-being beyond the previousvariables.

According to Baron and Kenny (1986), we tested the data for a mediationeffect in a stepwise procedure (results can be found in Table 3 and 4): First,the analyses showed that a significant relationship between the independentvariable, emotional stress, and the mediator, recovery experiences existed( � �0.22, p � .001). Second, we tested whether the relationship betweenthe mediator, recovery experiences, and the dependent variable, affectivewell-being at bedtime, was significant. The analyses revealed a significantrelationship between recovery experiences and affective well-being ( �0.39, p � .001). Third, we tested the relationship between the independentvariable, emotional stress, and the dependent variable, affective well-being,which proved to be significant ( � �0.28, p � .001). The effect ofemotional stress on affective well-being decreased (from � �0.28 to ��0.19, p � .001) but remained significant after the mediator was included inthe analyses. Because this speaks against full mediation, we tested for apartial mediation effect conducting the Sobel Test (Sobel, 1982) in itsadjusted form for multilevel data (Krull & MacKinnon, 2001). In support ofHypothesis 2, the test revealed that the mediating effect for recovery expe-riences was significant (z � �.57, p � .001). Thus, recovery experiencespartially mediated the negative relationship between emotional stress andaffective well-being.

Hypotheses 3 and 4 postulated different moderating effects of reappraisaland expressive suppression on the negative impact of emotional stress onrecovery experiences. We supposed reappraisal to be buffering (Hypothesis3) and expressive suppression to be enhancing (Hypothesis 4) the negativeimpact of emotional stress on recovery experiences. Again, we computedmodels of multilevel estimates in order to test the prediction of recoveryexperiences. Table 4 depicts the results. As before, Model 1 contained thecontrol variables gender, age, and negative as well as positive affectivity(Level 2) as predictors. The difference of the likelihood ratio between Model0 and Model 1 was significant (� �2 log likelihood � 25.16, df � 7, p �.001). In a next step, we entered emotional stress, reappraisal and expressivesuppression as predictors in Model 2, which was then compared with Model1. Model 2 showed a significantly improved model fit (� �2 log likelihood �146.91, df � 10, p � .01). While emotional stress was negatively related torecovery experiences ( � �0.22, p � .01), reappraisal and expressivesuppression did not prove to be significant predictors of recovery experi-ences. To test the moderation hypotheses (Hypotheses 3 and 4), we includethe interactions between emotional stress and reappraisal and expressivesuppression, respectively, in Model 3. Compared with Model 2, Model 3


Tab

le4.

Mul

tilev

elE

stim

ates

for

Mod

els

Pred

ictin

gR

ecov

ery

Exp

erie

nces

:R

eapp

rais

alan

dE

xpre

ssiv

eSu

ppre

ssio

nas

Mod

erat

ors

Mod

el0

Mod

el1

Mod

el2

Mod

el3

Est

imat

eSE

tE

stim

ate

SEt

Est

imat

eSE

tE

stim

ate

SEt

Inte

rcep

t2.

780.

0556

.93*

**

2.78

0.04

66.6

8***

2.78

0.04

64.9

3***

2.74

0.05

59.0

9***

Gen

der

0.11

0.10

1.12

0.12

0.10

1.21

0.13

0.10

1.26

Age

0.02

0.01

1.66

0.02

0.02

1.34

0.02

0.02

1.26

Posi

tive

affe

ctiv

ity0.

110.

961.

150.

070.

100.

700.

080.

100.

73N

egat

ive

affe

ctiv

ity�

0.24

0.06

�3.

79**

�0.

220.

07�

3.14

**

�0.

220.

07�

3.27

**

Em

otio

nal

stre

ss�

0.22

0.03

�6.

36***

�0.

220.

03�

6.60

***

Rea

ppra

isal

0.01

0.02

0.64

0.00

0.02

�0.

05E

xpre

ssiv

esu

ppre

ssio

n�

0.03

0.02

�1.

61�

0.04

0.02

�1.

98*

Em

otio

nal

stre

ss�

Rea

ppra

isal

0.05

0.02

2.62

**

Em

otio

nal

stre

ss�

Exp

ress

ive

supp

ress

ion

0.04

0.01

2.05

*

Dev

ianc

e19

06.0

318

80.8

717

33.9

617

24.8

4�

Dev

ianc

e25

.16*

**

146.

91***

9.12

***

�df

34

32

R2

R2

R2

Lev

el1

(with

in-

pers

on)

Var

ianc

e0.

460.

460.

000.

400.

130.

390.

03L

evel

2(b

etw

een-

pers

on)

Var

ianc

e0.

140.

090.

360.

070.

220.

070.

00

Not

e.N

�72

6oc

casi

ons,

N�

63pa

rtic

ipan

ts.

Uns

tand

ardi

zed

regr

essi

onco

effic

ient

sar

ere

port

ed.

*p

�.0

5.**

p�

.01.

***

p�

.001

.


showed a significantly smaller likelihood ratio (� �2 log likelihood � 9.12,df � 12, p � .01). The interactions of emotional stress during a study-relatedevent with reappraisal ( � 0.05, p � .01) and expressive suppression ( �0.04, p � .05) showed significant influences on recovery experiences in theevening. Figure 2 displays these interaction effects.

An inspection of the simple slopes revealed that as expected in Hypoth-esis 3, reappraisal buffered the negative impact of emotional stress onrecovery experiences. However, in contrast to Hypothesis 4, expressivesuppression did not enhance the negative impact of emotional stress onrecovery experiences, but had a buffering impact as well. We tested thesimple slopes of both interaction effects using an online HLM calculator(Preacher, Curran, & Bauer, 2006). We selected 1 SD above the mean (highreappraisal and expressive suppression, respectively) and 1 SD below themean (low reappraisal and expressive suppression, respectively) as condi-tional values of the emotion regulation strategies. For low reappraisal andexpressive suppression, the negative simple slopes relating emotional stressto recovery experiences were significant (z � �1.98 for reappraisal and z ��1.96 for expressive suppression, p � .05). For high reappraisal and ex-pressive suppression, the simple slopes were also negative, but nonsignificant(z � �1.83, p � .078 for reappraisal and z � �1.85, p � .066 for expressivesuppression). Thus, only on days when reappraisal and/or expressive sup-pression were low, emotional stress was significantly, negatively related torecovery experiences.

Taken together, the data supported Hypotheses 1–3. Emotional stress hada negative relationship with affective well-being and recovery experiencespartially mediated this relationship. That is, the negative impact of emotionalstress on recovery experiences was weaker when the person reappraised thesituation. In contrast to our expectations in Hypothesis 4, expressive sup-

Figure 2. The moderating effects of reappraisal and expressive suppression on the relationshipbetween emotional stress and recovery experiences.


pression had the same effect as reappraisal; it also buffered the negativeimpact of emotional stress on recovery experiences.

DISCUSSION

The present study integrates different lines of research from work stress,emotion regulation, and recovery, which to our knowledge have not beenconsidered in combination before. The study shows how the situationalapplication of emotion regulation can improve recovery experiences andsubsequent affective well-being after stressful study-related events. It furtherextends the literature by taking an intrapersonal approach on these topics,leading to results that complement the conclusions drawn from individualdifferences research.

The aim of the present study was threefold: Our first goal was to examinethe role of recovery experiences in the relationship between emotional stressand affective well-being. Analyses showed a negative impact of emotionalstress during a significant study-related event on affective well-being atbedtime. Recovery experiences partially mediated this negative relationship.Our second goal was to test whether the application of emotion regulationstrategies as a response to emotional stress impacts recovery experiencesafter stressful events at university or concerning the studies. The use ofreappraisal to regulate one’s emotions buffered the negative impact ofemotional stress on recovery experiences, as did the use of expressivesuppression. Both the mediation and moderation results even hold whencontrolling for trait positive and negative affectivity, strengthening the va-lidity and significance of this study. Our third goal was to shed light on theeffects of situational (instead of habitual) emotion regulation by implement-ing a diary design.

In summary, the study extends previous research in the stress domain inseveral ways. With regard to existing research on recovery experiences, ourresults improve the understanding of the link between job/study demands andhealth outcomes. Mediating effects of recovery experiences have recentlybeen supported in a longitudinal study (Kinnunen, Feldt, Siltaloppi, &Sonnentag, 2011); however, our study is among the first to support theseresults on the day-level. Because there is only little research concerningpredictors of recovery experiences, our findings further expand currentknowledge by revealing emotional stress to be an inhibitor of recoveryexperiences (e.g., Cropley & Purvis, 2003; Sonnentag & Bayer, 2005).Furthermore, with its focus on situational emotion regulation, our studyextends recovery research by pointing out a determinant that has similarbeneficial effects as for example job control (cf. Cropley et al., 2006); both


reappraisal and expressive suppression apparently help in detaching andrelaxing from emotional stress.

These results complement research on individual differences in emotionregulation and on experientially manipulated emotion regulation, whichhighlight reappraisal as a healthy form of emotion regulation (e.g., John &Gross, 2004; Mauss, Cook, Cheng, & Gross, 2007). They imply that, in thenotion of the Job Demands-Resources Model (Demerouti et al., 2001),emotion regulation strategies can be considered a personal resource thatmoderates the consequences of emotional job demands. In accordance to thetheoretical assumption that reappraisal prevents the experience of emotionalstress to grow bigger by reframing the situation (Gross & John, 2003), ourresults show that reappraisal apparently makes further self-control after theemotionally stressful event less necessary, so that resources are freed forpromoting recovery experiences. Unexpectedly, we found that expressivesuppression, which is considered a rather unhealthy way of emotion regula-tion when applied habitually (John & Gross, 2004; Srivastava et al., 2009),also buffered negative effects of emotional stress. Although we argued for aresource-consuming negative effect of expressive suppression, this contra-dictory finding is in line with results of some other studies, giving rise to thequestion of whether expressive suppression should generally be considered adetrimental strategy (e.g., Befahr & Cronin, 2010; Cole, Walter, & Bruch,2008; Schraub et al., 2011). In the present study, the definition and measure-ment of expressive suppression as situational emotion regulation may explainthe unexpected positive effect of expressive suppression. This means that weaccounted for inter- and intrapersonal variation and focused on emotionregulation in a specific situation rather than a habitual regulation strategy.Suppressing one’s emotional expression during the experience of increasedemotional stress, in this case, turned out to be a useful mechanism to copewith the stress. We concur with Befahr and Cronin (2010) in assuming thatinterpersonal effects might play a significant role in explaining beneficialeffects of expressive suppression. In the present study, the situational sup-pression of expressions of emotional stress may have evoked more positiveresponses from interaction partners. These positive responses may have, inturn, distracted from the negative experience. This would be in line withSonnentag and Fritz’ (2007) argumentation that the expression of negativeemotions can undermine recovery experiences by inducing continuing cog-nitive and emotional engagement with the adverse experience.

Our study results underline the importance of considering intrapersonalvariation in emotion regulation; more than 80% of the variance in emotionregulation was attributable to intrapersonal variance. Thus, contextual or stateantecedents seem to be stronger predictors of momentary emotion regulationthan individual differences are. As discussed earlier, such a state focus may


lead to different outcomes than a habitual focus, and it may only be thechronic use of expressive suppression that has detrimental effects.

We consider the diary design of the present study to be its particularstrength. Reducing probability for retrospective biases (Alliger & Williams,1993), the diary method more adequately captures emotional experiences andwell-being than do assessments at only one or two points of time, becauseemotions and well-being change in short intervals. Furthermore, effects ofintrapersonal variance in emotion regulation can only be detected by repeatedtime- or event-contingent measurement, as used in this study. Indeed, thediary design and the respective adaptation of Gross and John’s (2003)emotion regulation questionnaire made it possible to reveal short-term effectsof expressive suppression that diverge from those typically found in cross-sectional studies, or in longitudinal studies merely using individual differencemeasures. In addition, the high intrapersonal variance in affective well-being(about 75%) implies that by analyzing day-level antecedents of affectivewell-being, we gained information that gets lost in studies that conceptualizeaffective states as between-subjects variables (Netzlek et al., 2006).

Limitations and Implications for Future Research

Clearly, the sample of this study limits the generalizability of its results.Findings from examining undergraduate students who reported 4.8 workhours per day on average cannot be directly applied to employees in a worksetting; demographic characteristics like age, family responsibilities andeducation might be important moderators of the consequences of significantemotional events on well-being. However, Chang and colleagues (2007),who compared employee and student samples with regard to the relationshipbetween the experience of emotional stress and organizational citizenshipbehaviors in a meta-analysis, found a stronger effect for the employeesample. Taking these authors’ finding into account, our findings could beeven more pronounced in an employee sample. Although we consider ourresults relevant for the current generation of university students, we recom-mend their replication in another context and with a demographically morediverse sample.

A methodological issue that may be improved in future research is datacollection. A time-contingent assessment with higher frequency (e.g., threetimes per day) or an event-contingent assessment would allow the capture ofevents, emotions and behavior even closer to their occurrence and withhigher internal validity. However, our repeated-measurement design allowedfor control of between-person differences in the focus study variables andthus represents a more adequate assessment for the dynamic constructs wefocused on than a cross-sectional assessment would have been.


For a possible explanation of the moderation effects, especially for theunexpected effect of expressive suppression in the present study, we draw oninteractional theories of emotion regulation (Côté, 2005). These suggest thatinteraction partners’ reactions determine one’s well-being, meaning that (a)one’s (positive) emotional expression is mirrored in (b) (positive) reactionsby the interaction partner, which again (c) influences (enhances) one’swell-being. Research showing that the suppression of verbal and nonverbalexpressions of emotional stress at the workplace buffers negative effects onrecovery and performance (Brown, Westbrook, & Challagalla, 2005; Cole etal., 2008; Sanz-Vergel, Demerouti, Moreno-Jiménez, & Mayo, 2010) sup-ports this line of reasoning. It may well be that expressive suppression haspositive effects in high-power-distance situations (e.g., between student andprofessor), but that it has negative effects in equal-power relationships (e.g.,between student and significant other). Because we did not assess reactionsby interaction partners in this study, we recommend follow-up research toshed light on this possible explanation of beneficial short-term effects ofexpressive suppression.

A related variable that determines the effects of expressive suppressionare cultural values. In experimental settings, Butler, Lee, and Gross (2007,2009) demonstrated that cultural context moderated the consequences ofexpressive suppression on both social and physiological measures duringface-to-face interactions between American university students. Matsumoto,Yoo, and Nakagawa (2008) further showed that cultural values determine theuse and the consequences of certain emotion regulation strategies. We canthus hypothesize that cultural orientations in Germany or, more specifically,in the academic setting, may have influenced our findings. As Schwartz(1999) demonstrated, Germans value intellectual autonomy higher than af-fective autonomy, which means that the pursuit of one’s own intellectualideas is more important than the pursuit of affectively positive experiences.Hofstede (2001) found that Germans (compared with other cultures) have amedium level of uncertainty avoidance, which means they don’t feel verycomfortable with ambiguity, thus taking measures to reduce it and to controlthe future. A hypothesis drawn from this could be that the combination ofrather low Affective Autonomy and rather high uncertainty avoidance mightmake emotion regulation via expressive suppression more probable andaccepted. The assessment of cultural context ranging from participants’values on the individual level to organizational and societal culture (cf. Chao,2000) would help to clarify this issue and extend the study’s external validity.

An additional variable that also relates to cultural values and that shouldbe addressed in further studies is work significance. If work is highlysignificant for a person’s self-concept, negative work-related emotional ex-periences might provide more threat to this person’s self and, consequently,have stronger negative effects.


As shown in this and quite some other studies, recovery experiences arean important resource for affective well-being. Guided by Conservation ofResources theory (Hobfoll, 1989), a next step of research could be toinvestigate what factors help people not only to engage in recovery experi-ences, but to preserve their positive effects.

Practical Implications

The importance of emotion regulation and of daily recovery experiencesin maintaining people’s well-being has been supported in this study. Becausehigh levels of psychological stress and strain have been reported for thecurrent student generation (Obergfell & Schmidt, 2010), we encourage uni-versities to expand training and coaching programs, for example by integrat-ing a preventive module on healthy studying techniques in introductorycourses. In such stress management trainings (for examples, see Roger &Hudson, 1995; Walach et al., 2007), the topics of emotion regulation andrecovery experiences should be addressed (cf. Hahn et al., 2011). Accordingto our findings, participants should be informed about the potential harm ofemotional stress for their well-being, learn how to regulate negative emo-tional experiences, as well as become aware of the importance of activelyseeking recovery experiences. Altogether, students would thereby learn toreflect on their work-life-balance. Doing so in an early stage might not onlyprevent cases of psychologically exhausted students, but might also exertbeneficial effects on people’s stress management in their later careers.

CONCLUSION

As this study demonstrates, recovery experiences depend on the waypeople deal with experiences of emotional stress. On days on which partic-ipants either reappraised the situation or suppressed their emotional expres-sions more than usual, they reported greater recovery experiences. In linewith previous research, we conclude that reappraisal can be recommended asa healthy strategy to regulate one’s emotions. In addition, the suppression ofemotional expressions may, applied in the situation evoking emotional stress,be helpful in overcoming experiences of emotional stress. We conclude thatby learning how to use emotion regulation during emotionally stressful worksituations, people can foster subsequent recovery experiences that restoretheir resources and consequently improve their affective well-being at bed-time.


REFERENCES

Abler, B., & Kessler, H. (2009). Emotion Regulation Questionnaire: Eine deutschsprachigeFassung des ERQ von Gross und John [Emotion Regulation Questionnaire: A Germanversion of the ERQ by Gross and John]. Diagnostica, 55, 144–152.

Ahrentzen, S. B. (1990). Managing conflict by managing boundaries: How professionalhomeworkers cope with multiple roles at home. Environment and Behavior, 22, 723–752.

Alliger, G. M., & Williams, K. J. (1993). Using signal-contingent experience samplingmethodology to study work in the field: A discussion and illustration examining taskperceptions and mood. Personnel Psychology, 46, 525–549.

Bakker, A. B., & Demerouti, E. (2007). The Job Demands-Resources model: State of the art.Journal of Managerial Psychology, 22, 309–328.

Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in socialpsychological research: Conceptual, strategic, and statistical considerations. Journal ofPersonality and Social Psychology, 51, 1173–1182.

Befahr, K., & Cronin, M. A. (2010, August). The efficacy of venting as a function of theresponse by third-party listeners. Paper presented at the Academy of Managementmeeting, Montréal, Canada.

Bono, J. E., & Vey, M. A. (2005). Toward understanding emotional management at work: Aquantitative review of emotional labor research. In C. E. Härtel, W. J. Zerbe, & N. M.Ashkanasy (Eds.), Emotions in organizational behavior (pp. 213–233). Mahwah, NJ:Erlbaum Publishers.

Brislin, R. W. (1970). Back-translation for cross-cultural research. Journal of Cross-CulturalPsychology, 1, 185–216.

Brown, S. P., Westbrook, R. A., & Challagalla, G. (2005). Good cope, Bad cope: Adaptive andmaladaptive coping strategies following a critical negative work event. Journal of AppliedPsychology, 90, 792–798.

Butler, E. A., Egloff, B., Wilhelm, F. H., Smith, N. C., Erickson, E. A., & Gross, J. J. (2003).The social consequences of expressive suppression. Emotion, 3, 48–67.

Butler, E. A., Lee, T. L., & Gross, J. J. (2007). Emotion regulation and culture: Are the socialconsequences of emotion suppression culture-specific? Emotion, 7, 30–48.

Butler, E. A., Lee, T. L., & Gross, J. J. (2009). Does expressing your emotions raise or loweryour blood pressure? Journal of Cross-Cultural Psychology, 40, 510–517.

Chang, C-H., Johnson, R. E., & Yang, L.-Q. (2007). Emotional strain and organizationalcitizenship behaviour: A meta-analysis and review. Work & Stress, 21, 312–332.

Chao, G. T. (2000). Multilevel issues and culture: An integrative view. In J. K. Klein & S. W. J.Kozlowski (Eds.), Multilevel theory, research, and methods in organizations (pp. 308–346). San Francisco, CA: Jossey-Bass.

Ciarrochi, J., Dean, F. P., & Anderson, S. (2002). Emotional intelligence moderates therelationship between stress and mental health. Personality and Individual Differences, 32,197–209.

Cole, M. S., Bruch, H., & Vogel, B. (2006). Emotion as mediators of the relations betweenperceived supervisor support and psychological hardiness on employee cynicism. Journalof Organizational Behavior, 27, 463–484.

Cole, M. S., Walter, F., & Bruch, H. (2008). Affective mechanisms linking dysfunctionalbehavior to performance in work teams: A moderated mediation study. Journal of AppliedPsychology, 93, 945–958.

Connolly, J. J., & Viswesvaran, C. (2000). The role of affectivity in job satisfaction: Ameta-analysis. Personality and Individual Differences, 29, 265–281.

Cooke, R., Bewick, B. M., Barkham, M., Bradley, M., & Audin, K. (2006). Measuring,monitoring and managing the psychological well-being of first year university students.British Journal of Guidance & Counselling, 34, 505–517.

Côté, S. (2005). A social interaction model of the effects of emotion regulation on work strain.Academy of Management Review, 30, 509–530.

Cropley, M., Dijk, D-J., & Stanley, N. (2006). Job strain, work rumination, and sleep in schoolteachers. European Journal of Work and Organizational Psychology, 15, 181–196.


Cropley, M., & Purvis, L. J. M. (2003). Job strain and rumination about work issues duringleisure time: A diary study. European Journal of Work and Organizational Psychology,12, 195–207.

Demerouti, E., Bakker, A. B., Geurts, S. A. E., & Taris, T. W. (2009). Daily recovery fromwork-related effort during non-work time. In S. Sonnentag, P. L. Perrewé, & D. C.Ganster (Eds.), Current perspectives on job-stress recovery (pp. 85–123). Bingley, UnitedKingdom: JAI Press/Emerald Group Publishing.

Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The job demands-resources model of burnout. Journal of Applied Psychology, 86, 499–512.

Edwards, J. R., Caplan, R. D., & Harrison, R. V. (1998). Person-environment fit theory:Conceptual foundations, empirical evidence, and directions for future research. In C. L.Cooper (Ed.), Theories of organizational stress (pp. 28–67). Oxford, United Kingdom:Oxford University Press.

Elfering, A., Grebner, S., Semmer, N. K., Kaiser-Freiburghaus, D., Ponte, S. L-D., & Witschi,I. (2005). Chronic job stressors and job control: Effects on event-related coping successand well-being. Journal of Occupational and Organizational Psychology, 78, 237–252.

English, T., John, O. P., Srivastava, S., & Gross, J. J. (2012). Emotion regulation and peer-ratedsocial functioning: A 4-year longitudinal study. Journal of Research in Personality, 46,780–784.

Etzion, D., Eden, D., & Lapidot, Y. (1998). Relief from job stressors and burnout: Reserveservice as a respite. Journal of Applied Psychology, 83, 577–585.

Fisher, C. D. (2000). Mood and emotions while working: Missing pieces of job satisfaction?Journal of Organizational Behavior, 21, 185–202.

Fisher, C. D., & Ashkanasy, N. M. (2000). The emerging role of emotions in work life: Anintroduction. Journal of Organizational Behavior, 21, 123–129.

Fredrickson, B. L., & Joiner, T. (2002). Positive emotions trigger upward spirals towardemotional well-being. Psychological Science, 13, 172–175.

Gable, S. L., Reis, H. T., Impett, E. A., & Asher, E. R. (2004). What do you do when thingsgo right? The intrapersonal and interpersonal benefits of sharing positive events. Journalof Personality and Social Psychology, 87, 228–245.

Geurts, S. A. E., & Sonnentag, S. (2006). Recovery as an explanatory mechanism in the relationbetween acute stress reactions and chronic health impairment. Scandinavian Journal ofWork, Environment and Health, 32, 482–492.

Giardini, A., & Frese, M. (2006). Reducing the negative effects of emotion work in serviceoccupations: Emotional competence as a psychological resource. Journal of OccupationalHealth Psychology, 11, 63–75.

Gross, J. J. (1998a). Antecedent- and response-focused emotion regulation: Divergent conse-quences for experience, expression, and physiology. Journal of Personality and SocialPsychology, 74, 224–237.

Gross, J. J. (1998b). The emerging field of emotion regulation: An integrative review. Reviewof General Psychology, 2, 271–299.

Gross, J. J. (2001). Emotion regulation in adulthood: Timing is everything. Current Directionsin Psychological Science, 10, 214–219.

Gross, J. J. (2002). Emotion regulation: Affective, cognitive, and social consequences. Psy-chophysiology, 39, 281–291.

Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes:Implications for affect, relationships, and well-being. Journal of Personality and SocialPsychology, 85, 348–362.

Hahn, V. C., Binnewies, C., Sonnentag, S., & Mojza, E. J. (2011). Learning how to recoverfrom job stress: Effects of a recovery training program on recovery, recovery-relatedself-efficacy, and well-being. Journal of Occupational Health Psychology, 16, 202–216.

Hobfoll, S. E. (1989). Conservation of resources: A new attempt at conceptualizing stress.American Psychologist, 44, 513–524.

Hobfoll, S. E. (1998). Stress, culture, and community: The psychology and philosophy of stress.New York, NY: Plenum Press.

Hochschild, A. R. (1983). The managed heart: Commercialization of human feeling. Berkeley,CA: University of California Press.


Hofstede, G. (2001). Cultures consequences: Comparing values, behaviors, institutions andorganizations across nations. Thousand Oaks, CA: Sage.

Hox, J. J. (2010). Multilevel analysis: Techniques and applications (2nd ed.). New York, NY:Routledge/Taylor & Francis Group.

Jex, S. M., Beehr, T. A., & Roberts, C. K. (1992). The meaning of occupational stress items tosurvey respondents. Journal of Applied Psychology, 77, 623–628.

John, O. P., & Gross, J. J. (2004). Healthy and unhealthy emotion regulation: Personalityprocesses, individual differences, and life span development. Journal of Personality, 72,1301–1333.

Johns, M., Inzlicht, M., & Schmader, T. (2008). Stereotype threat and executive resourcedepletion: Examining the influence of emotion regulation. Journal of ExperimentalPsychology-General, 137, 691–705.

Kinnunen, U., Feldt, T., Siltaloppi, M., & Sonnentag, S. (2011). Job demands–resources modelin the context of recovery: Testing recovery experiences as mediators. European Journalof Work & Organizational Psychology, 20, 805–832.

Krohne, H. W., Egloff, B., Kohlmann, C-W., & Tausch, A. (1996). Untersuchungen mit einerdeutschen Version der ‘Positive and Negative Affect Schedule’ (PANAS) [Research witha German version of the PANAS]. Diagnostica, 4, 139–156.

Krull, J. L., & MacKinnon, D. P. (2001). Multilevel modeling of individual and group levelmediated effects. Multivariate Behavioral Research, 36, 249–277.

Lyubomirsky, S., King, L., & Diener, E. (2005). The benefits of frequent positive affect: Doeshappiness lead to success? Psychological Bulletin, 131, 803–855.

Marco, C. A., & Suls, J. (1993). Daily stress and the trajectory of mood: Spillover, responseassimilation, contrast, and chronic negative affectivity. Journal of Personality and SocialPsychology, 64, 1053–1063.

Matsumoto, D., Yoo, S. H., & Nakagawa, S. (2008). Culture, emotion regulation, and adjust-ment. Journal of Personality and Social Psychology, 94, 925–937.

Mauss, I. B., Cook, C. L., Cheng, J. Y. J., & Gross, J. J. (2007). Individual differences incognitive reappraisal: Experiential and physiological responses to an anger provocation.International Journal of Psychophysiology, 66, 116–124.

Meijman, T. E., & Mulder, G. (1998). Psychological aspects of workload. In P. J. D. Drenth,H. Thierry, & C. J. D. Wolff (Eds.), A handbook of work and organizational psychology:Work psychology (pp. 5–33). Hove, United Kingdom: Psychology Press Ltd.

Netzlek, J. B., Schröder-Abé, M., & Schütz, A. (2006). Multilevel analyses in psychologicalresearch. Advantages and potential of multilevel random coefficient modeling. Psycholo-gische Rundschau, 57, 213–223.

Obergfell, J., & Schmidt, L. (2010). Zwangsjacke Bachelor?! Der Einfluss von Anforderungenund Entscheidungsfreiräumen auf das Stressempfinden und die Gesundheit Studierender- Ein Vergleich von Bachelor - und Diplomstudierenden im Rahmen des Demand-Control-Modells von Karasek [Strait jacket bachelor?! The influence of demands and decisionlatitude on students’ stress experience and health: A comparison of bachelor and diplomastudents guided by Karasek’s demand-control model]. Unpublished thesis, University ofHeidelberg, Germany.

Preacher, K. J., Curran, P. J., & Bauer, D. J. (2006). Computational tools for probing interactioneffects in multiple linear regression, multilevel modeling, and latent curve analysis.Journal of Educational and Behavioral Statistics, 31, 437–448.

Raftery, J. N., & Bizer, G. Y. (2009). Negative feedback and performance: The moderatingeffect of emotion regulation. Personality and Individual Differences, 47, 481–486.

Ragsdale, J. M., Beehr, T. A., Grebner, S., & Han, K. (2011). An integrated model of weekdaystress and weekend recovery of students. International Journal of Stress Management, 18,153–180.

Rau, R. (2006). Learning opportunities at work as predictor for recovery and health. EuropeanJournal of Work and Organizational Psychology, 15, 158–180.

Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and dataanalysis methods (2nd ed.). Thousand Oaks, CA: Sage.


Raudenbush, S. W., Bryk, A. S., Cheong, Y., Congdon, R., & Du Toit, M. (2004). HLM 6:Hierarchical linear and nonlinear modeling. Lincolnwood, IL: Scientific Software Inter-national.

Richards, J. M. (2004). The Cognitive Consequences of Concealing Feelings. Current Direc-tions in Psychological Sciences, 15, 131–134.

Richards, J. M., & Gross, J. J. (1999). Composure at any cost? The cognitive consequences ofemotion suppression. Personality and Social Psychology Bulletin, 25, 1033–1044.

Richards, J. M., & Gross, J. J. (2000). Emotion regulation and memory: The cognitive costs ofkeeping one’s cool. Journal of Personality and Social Psychology, 79, 410–424.

Roger, D., & Hudson, C. (1995). The role of emotion control and emotional rumination in stressmanagement training. International Journal of Stress Management, 2, 119–132.

Sanz-Vergel, A. I., Demerouti, E., Moreno-Jiménez, B., & Mayo, M. (2010). Work-familybalance and energy: A day-level study on recovery conditions. Journal of VocationalBehavior, 76, 118–130.

Schmeichel, B. J., Vohs, K. D., & Baumeister, R. F. (2003). Intellectual performance and egodepletion: Role of the self in logical reasoning and other information processing. Journalof Personality and Social Psychology, 85, 33–46.

Schraub, E. M., Stegmaier, R., & Sonntag, K. (2011). The effect of change on adaptiveperformance: Does expressive suppression moderate the indirect effect of strain? Journalof Change Management, 11, 21–44.

Schwartz, S. H. (1999). A theory of cultural values and some implications for work. AppliedPsychology: An International Review, 48, 23–47.

Snijders, T. A., & Bosker, R. J. (1993). Standard errors and sample sizes for two-level research.Journal of Educational Statistics, 18, 237–259.

Sobel, M. E. (1982). Asymptotic intervals for indirect effects in structural equations models. InS. Leinhart (Ed.), Sociological methodology (pp. 290–312). San Francisco, CA: Jossey-Bass.

Sonnentag, S. (2001). Work, recovery activities, and individual well-being: A diary study.Journal of Occupational Health Psychology, 6, 196–210.

Sonnentag, S. (2003). Recovery, work engagement, and proactive behavior: A new look at theinterface between nonwork and work. Journal of Applied Psychology, 88, 518–528.

Sonnentag, S., & Bayer, U.-V. (2005). Switching off mentally: Predictors and consequences ofpsychological detachment from work during off-job time. Journal of Occupational HealthPsychology, 10, 393–414.

Sonnentag, S., Binnewies, C., & Mojza, E. J. (2008). ‘Did you have a nice evening?’ Aday-level study on recovery experiences, sleep, and affect. Journal of Applied Psychology,93, 674–684.

Sonnentag, S., & Fritz, C. (2007). The Recovery Experience Questionnaire: Development andvalidation of a measure for assessing recuperation and unwinding from work. Journal ofOccupational Health Psychology, 12, 204–221.

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. EuropeanJournal of Work and Organizational Psychology, 15, 197–217.

Sonnentag, S., Perrewé, P. L., & Ganster, D. C. (2009). Current perspectives on job-stressrecovery. Bingley, United Kingdom: JAI Press/Emerald Group Publishing.

Srivastava, S., Tamir, M., McGonigal, K. M., John, O. P., & Gross, J. J. (2009). The social costsof emotional suppression: A prospective study of the transition to college. Journal ofPersonality and Social Psychology, 96, 883–897.

Totterdell, P., & Holman, D. (2003). Emotion regulation in customer service roles: Testing amodel of emotional labor. Journal of Occupational Health Psychology, 8, 55–73.

Tschan, F., Rochat, S., & Zapf, D. (2005). It’s not only clients: Studying emotion work withclients and co-workers with an event-sampling approach. Journal of Occupational andOrganizational Psychology, 78, 195–220.

Van Kleef, G. A. (2009). How emotions regulate social life: The emotions as social information(EASI) model. Current Directions in Psychological Science, 18, 184–188.


Walach, H., Nord, E., Zier, C., Dietz-Waschkowski, B., Kersig, S., & Schüpbach, H. (2007).Mindfulness-based stress reduction as a method for personnel development: A pilotevaluation. International Journal of Stress Management, 14, 188–198.

Warr, P., Butcher, V., & Robertson, I. (2004). Activity and psychological well-being in olderpeople. Aging & Mental Health, 8, 172–183.

Watson, D., & Clark, L. A. (1984). Negative affectivity: The disposition to experience aversiveemotional states. Psychological Bulletin, 96, 465–490.

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measuresof positive and negative affect: The PANAS scales. Journal of Personality and SocialPsychology, 54, 1063–1070.

Weiss, H. M., & Cropanzano, R. (1996). Affective Events Theory: A theoretical discussion ofthe structure, causes and consequences of affective experiences at work. In B. M. Staw &L. L. Cummings (Eds.), Research in organizational behavior: An annual series ofanalytical essays and critical reviews (pp. 1–74). Greenwich, CT: JAI Press.

Williams, K. J., & Alliger, G. M. (1994). Role stressors, mood spillover, and perceptions ofwork-family conflict in employed parents. Academy of Management Journal, 37, 837–868.

Xanthopoulou, D., Bakker, A. B., Demerouti, E., & Schaufeli, W. B. (2007). The role ofpersonal resources in the job demands-resources model. International Journal of StressManagement, 14, 121–141.

Received January 30, 2011Revision received August 7, 2013

Accepted August 9, 2013 �

E-Mail Notification of Your Latest Issue Online!

Would you like to know when the next issue of your favorite APA journalwill be available online? This service is now available to you. Sign up athttp://notify.apa.org/ and you will be notified by e-mail when issues of interestto you become available!


Documents

Emotion Regulation as a Determinant of Recovery ... · Resources theory (Hobfoll, 1989), we expect that emotional stress during signiﬁcant study-related events negatively impacts