Journal of Adolescence 37 (2014) 145–154

Contents lists available at ScienceDirect

Journal of Adolescence

journal homepage: www.elsevier .com/locate/ jado

Associations among sleep, daily experiences, and lonelinessin adolescence: Evidence of moderating and bidirectionalpathways

Leah D. Doane*, Emily C. ThurstonDepartment of Psychology, Arizona State University, P.O. Box 871104, Tempe, AZ 85287-1104, USA

Keywords:ActigraphySleepStressAffectLonelinessAdolescenceDiary studies

Abbreviations: NA, Negative Affect; PA, Positive* Corresponding author. Tel.: þ1 480 965 5289; f

E-mail address: Leah.Doane@asu.edu (L.D. Doan

0140-1971/$ – see front matter � 2013 The Foundahttp://dx.doi.org/10.1016/j.adolescence.2013.11.009

a b s t r a c t

The present study examined the dynamic associations among daily stress levels, affect, andobjective sleep quality in adolescence. We also explored loneliness as a potential moder-ator of these associations. Seventy-eight adolescents participated over three days. Theycompleted diary reports of stressful experiences and affect five times a day while wearingan actigraph to obtain objective measurement of sleep. They also provided self-reports ofloneliness. High daily stress was associated with shorter sleep duration. Models testingbidirectional associations indicated that prior day stress was associated with shorter sleepduration, but poor sleep duration and sleep efficiency were also associated with greaterstress the next day. Loneliness was a significant moderator of the associations betweendaily stress and sleep duration and latency such that lonely individuals had shorter sleepdurations and sleep latencies after particularly stressful days. Results suggest daily dy-namic associations among loneliness, daily stress, and objective measures of adolescentsleep.� 2013 The Foundation for Professionals in Services for Adolescents. Published by Elsevier

Ltd. All rights reserved.

An accumulation of evidence has demonstrated that adolescents do not get enough sleep (Carskadon, 1990a, 1990b;Morrison, McGe, & Stanton, 1992; National Sleep Foundation, 2011a; Wolfson & Carskadon, 1998). Both biological (e.g.,changes in circadian rhythms) and contextual (e.g., earlier school start times) factors are associated with changes in sleeppatterns during adolescence (Carskadon, 2011, 1999; Dahl & Lewin, 2002). Not surprisingly, adolescents’ lack of sleep hashealth consequences; poor sleep has been associated with obesity (Gupta, Mueller, Chan, & Meininger, 2002), risk for suicide(Choquet, Kovess, & Poutignat, 1993; Liu, 2004), anxiety and depression (Alfano, Zakem, Costa, Taylor, & Weems, 2009),reduced academic performance (Wolfson & Carskadon, 2003), and increased stress levels and negative mood (Lund, Reider,Whiting, & Prichard, 2010). Although prior research has consistently linked poor sleep with adolescent health determinants,fewer studies have focused on predictors of poor sleep that include day-to-day fluctuations in stress or affect, and howmorestable individual characteristics, like loneliness, interact with these daily experiences to influence sleep patterns. Much priorresearch has also been limited by the reliance on single subjective or self-report measures of sleep. Given that affect andstress levels may bias self-reports of sleep behaviors, it appears increasingly important to examine these associations usingobjective indicators of sleep. In an attempt to address these limitations, the current study examined: a) the day-to-dayassociations among negative and positive affect (NA; PA), stress levels and objective sleep quantity and quality; and b)

Affect.ax: þ1 480 965 8544.e).

tion for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved.

L.D. Doane, E.C. Thurston / Journal of Adolescence 37 (2014) 145–154146

whether loneliness moderated the associations among affect, stress, and sleep in an adolescent sample transitioning intocollege.

Daily experiences of stress, affect, and sleep

Stressful experiences, ranging from shift work and daily hassles to childhood trauma, have been linked with sleep dis-turbances (Chung & Cheung, 2008; Greenfield, Lee, Friedman, & Springer, 2011; Lund et al., 2010; Van Reeth et al., 2000). Forinstance, Lund et al. (2010) found that tension and stress were themost important contributors to poor sleep in a large sampleof college students. Additionally, daily self-reports of perceived stress and worry have been shown to predict poorer sub-jective sleep and objective sleep efficiency (Åkerstedt Kecklund, & Axelsson, 2007; Galambos, Howard, & Maggs, 2010;Tworoger, Davis, Vitiello, & McTiernan, 2005). A few studies found no effects of perceived stress on sleep, but thesestudies focused specifically on clinical samples or stress following traumatic events (Lavie, Carmeli, Mevorach, & Liberman,1991; Pillar, Malhotra, & Lavie, 2000).

Experimental sleep deprivation designs have been used to illustrate the linkages between NA and PA in relation to objectivemeasures of sleep. In a combined sample of adolescents and adults, less PAwas seen in sleep-deprivedparticipants as compared tothosewhoreceivedadequate sleep (Talbot,McGlinchey,Kaplan,Dahl,&Harvey, 2010). Similarly, in a sampleof adolescents, lessPAand a lower PA toNA ratiowere seen in sleep-deprived youthwhen compared to thosewhowere rested (Dagys et al., 2012). Thereis less consistent evidence of the associations between NA and PA and subjective and objective measures of sleep in naturalisticenvironments. Although two early studies of healthy adolescents found small (Price, Coates, Thoresen, & Grinstead, 1978) or noassociations (Clark&Watson,1988) between subjective sleep andnegative emotional states,most studies have been conducted inadultorclinicalpopulations (Bower, Bylsma,Morris,&Rottenberg, 2010;Cousinset al., 2011;Hamiltonet al., 2008). Forexample, ina diary study of older adults, subjective, but not objective, assessments of sleep were significantly correlated with less PA andgreater NA the next day (McCrae et al., 2008). In regards to trait measures of affect, Norlander, Johansson, and Bood (2005) foundthat individuals with high trait PA and low trait NA had the greatest subjective sleep quality when compared to individuals withlow PA and high NA. Thus, an association between daily or trait affect and subjective assessments of sleep has been established inadults, but less is known about the associations in healthy adolescents in naturalistic environments.

Associations among affect, stress, and sleep are particularly important during adolescence, as this developmental period ischaracterized by high variability and increases in affect and stress (Arnett, 1999; Colton & Gore, 1991; Larson,Csikszentmihalyi, & Graef, 1980; Larson & Richards, 1994). Adolescents exhibit more extreme levels of emotion comparedto children or adults in daily life and in response to stress (Larson & Richards, 1994). Therefore, beyond biological shifts insleep (e.g. homeostatic pressure; Jenni, Achermann, & Carskadon, 2005), it is possible that adolescents may experience dailyor chronic changes in sleep due to the high variability and intensity of their emotions, coupled with the interpersonal andenvironmental transitions that characterize this developmental period (Graber & Brooks-Gunn, 1996; Smetana, Campione-Barr, & Metzger, 2006).

Daily reciprocal relations among sleep and affect or stress levels

A second consideration for research on daily experiences and sleep is the need to disambiguate whether daytime affect orstress predict sleep that night or whether sleep one evening predicts the next day’s affect or stress levels. One strategy toaddress thepotential for bidirectional effects is to use dailydiarymeasures of stress, affect, and objectivemeasures of sleepoverconsecutive days and nights. Two recent studies of adolescents and young adults have examined the day-to-day associationsamong stress levels or affect and self-reported sleep duration and quality. In a daily experience study of ninth graders, greaterstress during the day predicted less sleep at night, which in turn, was associated with higher anxiety, depressed mood andfatigue the next day (Fuligni &Hardway, 2006). Similar dynamic associationswere found in a two-weekdiary study offirst yearuniversity students;Galambos,Dalton, andMaggs (2009) found that higher PA significantly predictedgreater sleepquality, andin turn, greater sleep quality predicted higher PA and lower NA. As such, a second aim of the present studywas to examine theday-to-day reciprocal associations among stress and affect and objective sleep since, to our knowledge, no daily diary studieshave examined these reciprocal relations in a normative sample of adolescents with objective measures of sleep.

Evidence of moderation: loneliness

Previous studies have shown that the associations between stress and sleep are only significant under specific conditionsor for particular groups. For example, Hanson and Chen (2010) found that the effect of daily stressors on sleep was moderatedby childhood adversity, such that young adults with worse childhood environments slept fewer minutes on the days theyencountered more stressors. Another study found that coping style moderated the effects of stress on sleep (Sadeh, Keinan, &Daon, 2004). College students who utilized more emotion-focused coping styles experienced a shift towards worse sleepquality between low and high stress periods. These studies suggest that individual traits or past experiences may moderatethe associations among daily stressors or affect and sleep, particularly in adolescents or college students.

An individual trait that may be of particular importance for adolescents is loneliness. Although one of the primarydevelopmental tasks of adolescence is to establish intimate relationships (Buhrmester, 1990), studies have shown thatloneliness is prevalent during this developmental period (Heinrich & Gullone, 2006). Loneliness is the pain and distress

L.D. Doane, E.C. Thurston / Journal of Adolescence 37 (2014) 145–154 147

experienced when individuals perceive a lack of quantity and quality in their social relationships (Hawkley & Cacioppo, 2010;Peplau & Perlman, 1982). No empirical research to our knowledge has examined the moderating role of loneliness applied tothe study of daily experiences and sleep. We might expect that lonely individuals will demonstrate stronger relations be-tween daily levels of stress or NA and poor sleep given research showing that lonely individuals view daily events as morestressful and have less confidence in their ability to manage stress (Hawkley, Burleson, Berntson, & Cacioppo, 2003; Hawkley,Preacher, & Cacioppo, 2007). As such, a third research aimwas to examine whether stressful experiences and affect might bedifferentially related to sleep in lonely versus non-lonely individuals.

The present study

We first examined the concurrent associations between daily affect, stress levels, and sleep quantity and quality. We hy-pothesized that low levels of PA, high levels of NA, and high levels of stress would be associated with poor sleep (lower sleepduration and time in bed, lower sleep efficiency, higher sleep latency, and later bedtime) that night. Second, in order to un-derstandwhether therewere daily reciprocal relations between sleep and stress/affect, wepredicted sleep fromboth prior daymeasures of stress and affect while covarying for next day measures of stress levels and affect. We hypothesized that higherpriordaystress levels andgreaterNAwouldpredictworse sleepandworse sleepwouldpredict highernextdaystress levels andgreaterNA. Finally,we investigatedwhether lonelinessmoderated the associations betweendaily affect or stress and sleep.Wehypothesized that we would find stronger associations among daily affect or stress and sleep in lonely individuals. A betterunderstanding of the associations among daily experiences and sleep among lonely and non-lonely adolescents may provideevidence for maladaptive health pathways common to loneliness and poor sleep during this developmental period.

Method

Participants

A total of 82 adolescents participated in this study. Participants were recruited as part of a longitudinal study of thetransition from high school into college. They were contacted through orientation activities for the psychology department ata large southwestern university and email communication. Participants were required to live within 35 miles of the uni-versity, be a senior in high school, and endorse that they were coming to the university in the subsequent fall.

Youth were excluded if they did not provide at least three nights of actigraph sleep data and/or had values outside of theexpected norms (n ¼ 3) or did not provide questionnaire data (n ¼ 1). The final analytic sample consisted of 78 youth (23%male), aged 17–18 years (M¼ 18.05, SD¼ .41). Participants were racially and ethnically diverse: 54% non-Hispanic White, 23%Latino/Hispanic descent, 13% multiracial, 5% African American and 5% Asian American/Pacific Islander. Youth also came fromvarying socioeconomic backgrounds as measured by their parents’ level of education: 11% reported that their parentscompleted some high school, 26.8% of parents had a high school diploma, 22% of parents had some college, 12.2% of parentsreceived an associate’s degree, 18.3% of parents received a bachelor’s degree, and 9.8% of parents received a graduate degree.

Procedures

The University Institutional Review Board approved all procedures. Participants signed consent forms upon delivery ofproject materials; parental consents were collected for participants who were under the age of 18. Participants who con-sented to the study participated by wearing an actigraph and providing diary reports of situations and emotions five times aday for 4 nights and 3 days. They were asked to select three typical consecutive weekdays to participate. All materials neededfor the study were brought to the participant’s home directly by project personnel, who explained all procedures and pro-vided the participant with an e-mail address and phone number where they could reach study staff at any time. Study staffthen called participants the night when they were supposed to begin the protocol. Project personnel picked up completedstudymaterials and paid participants $40 for completion of the protocol. Participants were also sent a summary of basic sleepstatistics across the 4 nights of participation.

Participants were asked to provide paper and pencil diary reports upon waking in the morning, 30 min after waking, tworandomly selected times throughout the day1 and just before bed. In the diary entries, participants reported on their location,what they were doing, thinking, and feeling, and on the presence and nature of any stressful events that had occurred in thelast hour. In total, participants were required to fill out fifteen diary entries (M ¼ 14.18, SD ¼ 1.27).

Measures

Objective sleepDuring the four nights and three days of data collection, individuals wore the Actiwatch Score (Phillips Respironics, Inc.), a

wrist-based accelerometer placed on the non-dominant hand that quantifies movement across the waking day and during

1 These times were chosen for purposes of other study protocol including saliva sampling for cortisol which is not reported in this manuscript.

L.D. Doane, E.C. Thurston / Journal of Adolescence 37 (2014) 145–154148

sleep. Participants were instructed towear the actigraph over the entire study period. They were told to press a button on thewatch to indicate when they were in bed trying to go to sleep and when they woke up, which inserted a marker into theactigraph record. To score data, the Actiware-Sleep software (version 3.4) validated algorithm was used (Oakley, 1997). Ac-tivity counts within each epoch were calculated based on activity levels during the adjacent 2 min period.2 The threshold wasset to 40, with a range of 20–80. Utilizing 1 min epochs and based on significant movement after at least 10 min of inactivity,this algorithm calculates a variety of sleep parameters. We focused on five parameters: sleep duration (actual sleep minutes),sleep efficiency (the proportion of time spent in bed actually sleeping), sleep latency (number of minutes in bed before fallingasleep), total time in bed, and bedtime. Objective indicators of sleep were validated with daily self-reports of bed and waketimes to identify significant outliers and equipment malfunction. Outliers (greater than 2 SD) for sleep parameters or days inwhich there was significant discordance between self-reports and objective measurement were not included in analyses(N ¼ 12 days). Actigraph sleep estimates have been validated against concurrent polysomnography (Sadeh, Hauri, Kripke, &Lavie, 1995).

Daily affect and stress levelsIn each diary entry, participants were asked to rate their levels of affect using the Positive and Negative Affect Schedule

(PANAS; Watson, Clark, & Tellegen, 1988). Example items ask participants to “indicate to what extent you have felt distressedin the last hour ” or “indicate towhat extent you have felt enthusiastic in the last hour.” Participants responded on a scale from0 (very slightly or not at all) to 4 (extremely). We calculated PA and NA scores by averaging scores across ten positive and tennegative items at each time point (Cronbach’s alpha for PA: .92; NA: .88). Participants were also asked to indicate the moststressful event they encountered in the last hour and how stressful that event was. They could respond that the event was notat all stressful (0) to very stressful (3). PA, NA and stress levels were aggregated and standardized across the fivemoments in aday to represent average daily experiences of PA, NA, and stress.

LonelinessThe UCLA Loneliness Scale (Version 3; Russell, 1996) is a measure of a “trait” or global feeling of loneliness and satisfaction

with one’s social network. Research has demonstrated that loneliness is stable and heritability estimates range from 45% inchildren to 48% in adults (Bartels, Cacioppo, Hudziak, & Boomsma, 2008; Boomsma, Willemsen, Dolan, Hawkley, & Cacioppo,2005). Furthermore, recent work has demonstrated that, although related to depression, loneliness is a distinct construct thatlongitudinally predicts depressive symptoms rather than the reverse (Cacioppo, Hawkley, & Thisted, 2010). The UCLALoneliness Scale (Version 3) is highly reliable across several populations with test-retest reliability of .73 and strong internalconsistencies ranging from .89 to .94. Examples items are, “Howoften do you feel that you are no longer close to anyone?” and“How often do you feel you can find companionship when you want it?”. Items are rated on a scale of 1 (never) to 4 (always).Positively worded items were reverse scored and a global loneliness score was calculated after summing across all items. TheCronbach’s alpha was .91 in this sample.

Analytic plan

All sleep outcomes were treated as continuous variables and only unstandardized coefficients are reported. All multi-variate analyses included covariates of gender, race/ethnicity (white vs. non-white), parent’s educational status, age, anddepressive symptoms due to prior research indicating associations with sleep or loneliness (Arber, Bote, & Meadows, 2008;Cacioppo, Hughes, Waite, Hawkley, & Thisted, 2006; Mezick et al., 2008). Hierarchical linear models were utilized to accountfor the nested nature of the data (days nested within person; Raudenbush & Bryk, 2002). At Level 1, we modeled the sleepparameters as well as the daily affect and stress level variables. At Level 2, person-specific parameters including lonelinessand demographic covariates were modeled. As recommended (Enders & Tofighi, 2007), Level 1 variables were centered at thegroup mean and Level 2 variables were centered at the grand mean. Intercepts in all models were estimated as randomlyvarying across persons, and the slopes of all time-varying covariates were fixed if they did not vary significantly across in-dividuals or were covariates in themodels. Within-person and between-person pseudo R2 and total variance explained termswere calculated to reflect the proportion of within-person, between-person, and total variation explained by each model(Singer & Willett, 2003).

To address our first research question (Model 1), we conducted awithin-person analysis.We estimated the average level ofsleep (Level 1 intercept, b0), between-person effects on the intercept (covariates), and time-varying or within-person (Level 1)affect (PA: b1; NA: b2) and stress level (b3) predictors. Next, to examine the reciprocal relations among sleep, daily affect, andstress (Model 2), we predicted sleep from time-varying within-personmeasures of prior day stress and affect while covaryingfor next day levels of affect (b4 and b5) and stress levels (b6). This technique allowed us to uncover the temporal changes instress/affect and sleep over time following analytic techniques from other studies (e.g., Adam, Hawkley, Kudielka, & Cacioppo,2006; Zeiders, Doane, & Adam, 2011). For example, if both prior day and next day measures of stress were associated withsleep quality, this would suggest that stress was associatedwith subsequent sleep, and that sleep was also associated with thenext day’s stress level, even when controlling for the stress level present on the prior day (i.e., capturing within-person

2 The following algorithm was used where A denotes activity counts and E denotes epoch : A ¼ E � 2(1/25) þ E – 1(1/5) þ E þ E þ 1(1/5) þ E þ 2(1/25).

L.D. Doane, E.C. Thurston / Journal of Adolescence 37 (2014) 145–154 149

changes in stress). Secondary analyses predicting affect and stress from previous nights’ sleep were utilized to confirmfindings found within the reciprocal relations modeling.

To examine our third research question whether loneliness was a moderator of the affect or stress and sleep associations(Model 3), loneliness (g06) and depressive symptoms (g05) were included as between-person predictors of the intercept(average level of sleep), and loneliness was entered as a moderator of the associations between the Level 1 predictors (g11, g21,g31) and sleep. Significant two-way interactions between loneliness and stress level, NA and PAwere probed using the simpleslopes technique for hierarchical linearmodeling as outlined in Preacher, Curran, and Bauer (2006) for cross level interactions.We estimated simple slopes and range of significance for associations among stress levels, NA, or PA and sleep at differinglevels of loneliness.

Results

Table 1 presents descriptive and bivariate associations. Participants slept, on average, 5.88 (SD ¼ 1.06) hours per night,however spent 7.16 (SD ¼ 1.14) hours in bed. Participants’ actual time asleep was substantially lower than the National SleepFoundation’s recommendations of 8.5–9.25 h for adolescents and 7–9 h for adults (2011b). On average, participants’ sleepefficiency was 78.91% (SD ¼ 9.53) and they spent 11.4 min (SD ¼ 9.7) falling asleep each night. Finally, the mean bedtimeacross the sample was 11:45 PM (SD ¼ 1.33). All dependent variables were assessed for violations of normality. Sleep latencywas positively skewed (skewness: 3.75; kurtosis: 19.54) and therefore was log transformed for all subsequent analyses.

Averagedaily stresswas significantlyassociatedwith shorter sleepdurations (r¼�.26,p< .05), less sleepefficiency (r¼�.36,p< .01), and shorter sleep latencies (r¼�.16, p< .05) but notwith total time in bed (r¼�.01, ns) or bedtime (r¼�.19, ns). Therewere no significant bivariate correlations between average levels of NA and PA and sleep. Correlations amongNA, PA, and stresslevels ranged from r ¼ �.49 to r ¼ .50. Loneliness was not correlated with sleep but was associated with NA (r ¼ .31, p < .01).

Several covariates were associated with our study variables of interest. Males spent less time in bed (F¼ 9.36, p< .01), hadgreater sleep efficiency (F ¼ 3.50, p < .07), reported lower stress levels (F ¼ 5.43, p < .05), and reported greater levels of PA(F ¼ 5.65, p < .01) than females. White youth reported greater levels of loneliness than non-white youth (F ¼ 5.26, p < .05).Age was correlated with sleep duration (r ¼ �.22, p < .05), latency (r ¼ �.24, p < .05), and bedtime (r ¼ .26, p < .05). Finally,depressive symptoms was significantly correlated with NA (r ¼ .44, p < .01), stress levels (r ¼ .25, p < .05), and loneliness(r ¼ .69, p < .05), but not with any of the objective sleep parameters.

Daily stress, affect, and sleep quality

Multivariate analyses were conducted with sleep duration, efficiency and latency as dependent variables based onpresence of significant bivariate associations with the independent variables of interest. Unconditional models containing nopredictors or covariates at Level 1 or Level 2 were calculated for each sleep variable to indicate the proportion of variance thatwas attributable to within-person (or variability in sleep across the week for an individual) and between-person factors. Forsleep duration, 49% of the variation was due to within-person factors (51% between-person). Regarding sleep efficiency, 44%of variation was due to within-person factors (56% between-person). Finally, for sleep latency, 90% of variation was due towithin-person factors (10% between-person). All of these indicated a great deal of day-to-day fluctuation in sleep.

Table 1Descriptive statistics and zero order correlations of objective sleep quality and quantity, loneliness, affect and stress levels (N ¼ 78).

1 2 3 4 5 6 7 8 9 10 11 12 13 14 Mean (SD) Range

1. Sleep duration(minutes)

1.00 352.86 (63.67) 182.68–485.44

2. Sleep efficiency .56** 1.00 78.91 (9.53) 53.24–92.323. Sleep latency .11 �.11 1.00 11.40 (9.70) .00–40.504. Total time in bed

(minutes).70** �.14 .08 1.00 429.57 (68.29) 254.75–581.5

5. Bed time �.01 .20þ �.34** .06 1.00 23.75 (1.33) 21.42–27.486. PANAS Positive

Affecta�.17 �.05 �.02 �.12 .05 1.00 .86 (.55) .13–2.88

7. PANAS NegativeAffecta

.17 .11 .01 .13 .06 �.49** 1.00 .29 (.27) .01–1.49

8. Stress level �.26* �.36** �.16* �.01 �.19 �.17 .50** 1.00 1.16 (.46) .00–2.469. Loneliness �.08 �.07 �.03 �.01 .19 �.01 .31** .17 1.00 39.98 (9.23) 24.00–61.0010. White race/

ethnicity�.05 �.01 .12 �.01 �.08 �.16 �.03 .03 .26* 1.00 .54 (.50)

11. Male �.12 .20þ �.09 �.33** .17 .26* �.18 �.26* �.08 �.04 1.00 .23 (.42)12. Age .22* .21þ �.24* .18 .26* .01 .17 �.10 .11 .29** .03 1.00 18.04 (.41) 16.75–18.6713. Parent education �.06 .03 .07 �.11 .01 �.05 .06 .05 .05 .25* .05 .04 1.00 3.42 (1.50) 1.00–6.0014. Depressive

symptoms.04 �.02 �.01 .11 .15 �.04 .44** .25* .69* .13 �.09 .10 �.03 1.00 14.80 (8.61) .00–41.00

þp < .10, *p < .05, **p < .01.a PANAS is the Positive and Negative Affect Schedule.

Table 2Sleep duration, efficiency and latency predicted from daily affect, daily stress levels and loneliness (N ¼ 78).

Variable Sleep durationa Sleep efficiencyb Sleep latencyc

Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

Level 1 Intercept b0 359.50** 357.28** 355.13** 80.12** 80.14* 80.04** 1.57** 1.52** 1.51**Between-person effects on interceptWhite race/ethnicity, g01 �18.27 �10.92 �.06 �2.49 �.12 .68 .27 .29 .33Male, g02 �16.82 �46.21* �54.55* 3.75þ 5.58** 5.64** �.29 �.40 �.39Age, g03 43.35 56.55* 53.05* 5.72þ 4.13 4.26 �.63** �.26 �.24Parent education, g04 3.04 6.43 4.08 .88 .86 .84 .08 .11 .10Depressive symptoms, g05 23.30 �.59 .26þ

Loneliness, g06 �23.24* �1.48 �.17Prior day Positive Affect, b1 21.61 5.19 14.65 3.85 2.02 1.94 �.16 �.44 �.01Loneliness, g11 �2.82 3.66 �.15

Prior day Negative Affect, b2 7.21 12.75 16.04 �.68 �2.43 �2.40 �.22 �.03 .12Loneliness, g21 �31.15 .38 �.73

Prior day stress level, b3 �10.63* �15.25* �19.22** �1.14þ �.79 �1.47þ �.17þ �.17 �.20þ

Loneliness, g31 �21.76* �1.29 �.30*Next day Positive Affect, b4 �12.49 17.37 1.90 4.36 �.07 .322Next day Negative Affect, b5 �9.54 �17.03 2.75 1.76 .95 .83Next day stress level, b6 �19.38** �17.18** �2.64** �2.30** .03 .01Within-person pseudo R2 .09 .09 .10 .14 .33 .36 .01 .01 .03Between-person pseudo R2 .06 .08 .08 .11 .12 .09 .04 .07 .07Total variance explained .07 .08 .09 .12 .20 .21 .02 .02 .04

Model 1 included only prior day affect and stress levels. Model 2 included both prior day and next day affect and stress levels.Model 3 included prior day and next day affect and stress levels at Level 1 and levels of loneliness and depressive symptoms at Level 2.þp < .10, *p < .05, **p < .01. Unstandardized coefficients are reported for all models. All independent variables have been standardized.

a Measured in minutesb Measured in percentagec Measured in minutes and log transformed to correct for skewness.

L.D. Doane, E.C. Thurston / Journal of Adolescence 37 (2014) 145–154150

Results frommultivariate analyses are presented in Table 2. Model 1 included the average level of sleep (Level 1 intercept,b0), between-person predictors of the intercept (covariates), and time-varying (Level 1) prior day affect (PA: b1; NA: b2) andstress level (b3) predictors. Next day affect (PA: b4; NA: b5) and stress levels (b6) were added in Model 2 as time-varyingpredictors. Finally, in Model 3, depressive symptoms (g05) and loneliness (g06) were added as between-person predictorsof the intercept. Loneliness was also included as a between-person predictor of prior day affect (PA: g11; NA: g21) and stressslopes (g31) to examine moderating pathways.

In Model 1, higher daily stress was significantly associated with shorter sleep duration that night (b¼�10.63, p< .01) andwas marginally associated with less sleep efficiency (b ¼ �1.14, p < .10) and shorter sleep latency (b ¼ �.17, p < .10). Prior dayPA and NA were not associated with sleep duration, efficiency, or latency. Bidirectional associations were tested by enteringboth prior day and next day stress, NA and PA as predictors (see Table 2, Model 2). Greater prior day and next day stress wereassociated with shorter sleep durations (prior: b ¼ �15.25, p < .05; next: b ¼ �19.38, p < .01). Next day rather than prior daystress levels were associated with reduced sleep efficiency (b ¼ �2.64, p < .01). To confirm that the simultaneous modelingproduced reliable estimates, we conducted a series of analyses predicting next day stress levels from sleep duration andefficiency with the same set of covariates.3 Sleep duration and efficiency were significant predictors of next day levels of stress(duration: b ¼ �.01, p < .01; efficiency: b ¼ �.02, p < .01).

Evidence for moderation: loneliness?

Loneliness was entered as a cross level moderator of daily stress, NA, and PA associations with sleep (Table 2, Model 3).There was a main effect of loneliness on sleep duration such that loneliness was associated with lower sleep duration acrossthe three nights of sleep (g ¼ �23.24, p < .05). Loneliness significantly moderated the associations between stress and sleepduration (g ¼ �21.76, p < .01) and sleep latency (g ¼ �.30, p < .05). Probing of significant interactions revealed a significantassociation between prior day stress levels and sleep duration at high (b ¼�40.97, p < .01) but not low levels (b ¼ 2.55, ns) ofloneliness. The relationship between prior day stress levels and sleep duration was statistically significant at values ofloneliness greater than �.31 (53.8% of the sample). The simple slopes calculations for sleep latency indicated significantassociations between prior day stress levels at high (b ¼ �.50, p < .01) but not at low levels of loneliness (b ¼ .10, ns). Therelationship between stress levels and sleep latency was significant at values of loneliness greater than .10 (41% of thesample). An illustration of the simple slopes for relationships between prior day stress levels and sleep duration and latency atdiffering levels of loneliness can be seen in Figs. 1 and 2.

3 Models were run separately for each sleep parameter because they were highly correlated. Next day PA and NA were also included as covariates.

Fig. 1. Simple slopes plots for sleep duration by daily stress levels for low and high levels of loneliness.

L.D. Doane, E.C. Thurston / Journal of Adolescence 37 (2014) 145–154 151

Discussion

The findings from this study illustrated daily and bidirectional associations among stress and objective sleep, as well assignificant moderation by loneliness, in a sample of adolescents transitioning from high school into college. Youth who re-ported more stressful days were more likely to exhibit shorter sleep durations that night. Furthermore, we found evidence ofbidirectionality such that nights characterized by low sleep duration or low sleep efficiency were associated with greaterexperiences of stress the next day. Finally, relations between stress levels and sleep durationwere only present in adolescentswith average or high levels of loneliness. Similarly, shorter sleep latencies after stressful days were only present in lonelyyouth.

We found partial support for our first set of hypotheses. Bivariate and multivariate associations were identified betweendaily stress and objectively measured sleep duration, efficiency and latency, but not total time in bed or bedtime. Thesefindings highlight the differential associations among perceived stress levels and types of sleep quality and quantity. Ourresults indicated that stress was more associated with quality of sleep and actual time sleeping rather than sleep timing (i.e.,how long someone spends in bed or what time they go to bed). The fact that there were no main effects with sleep timingsuggests that uncontrollable environmental constraints (e.g., work or school schedule) may influence timing more thanpsychosocial experiences. In contrast to sleep timing, participants’ ability to sleep well was impeded by increases in perceivedstress. One possible pathway by which perceived stressors were associated with subsequent sleep quality is continuedphysiological activation. Prior research has indicated bidirectional associations among indicators of physiological stress ac-tivity (e.g., cortisol) and subsequent sleep quality (Zeiders et al., 2011). Therefore, it is possible that increased physiologicalactivation due to perceived stress may have led to poor sleep. Another potential pathway between perceived stress and

Fig. 2. Simple slopes plots for sleep latency (log transformed) by daily stress levels for low, and high levels of loneliness.

L.D. Doane, E.C. Thurston / Journal of Adolescence 37 (2014) 145–154152

subsequent sleep quality is coping behavior in response to stress. Based on previous research indicating that ruminationmoderates the associations between stress and sleep (Sadeh et al., 2004), we hypothesize that maladaptive coping behaviormay have led to diminished sleep quality. Althoughwe could not directly test this hypothesis, future research shouldmeasureindicators of physiological stress activity and coping behavior in conjunction with perceived stressors to understand themoderating (or mediating) pathways among stress, coping styles and sleep quality and timing.

Our findings are consistent with previous research in adult and adolescent populations, in which demands or stressorsduring the day were associated with lower self-reported sleep quality or duration (Åkerstedt, Kecklund, & Axelsson, 2007;Fuligni & Hardway, 2006). We did not find associations among daily experiences of PA or NA and subsequent sleep dura-tion or quality in this study. Of the small number of studies that have examined objective sleep quality and daily affect inadolescents, only those in experimental designs or clinical populations have found significant associations among objectivemeasures of sleep and daily affect (e.g. Cousins et al., 2011; Dagys et al., 2012; Haynes, Bootzin, Smith, & Stevens, 2006).Therefore, associations may not be present in healthy populations of adolescents. Further, differential findings may be due tomeasurement (e.g., state or trait indicators) or methodological (e.g., naturalistic setting) variability across studies.

In regard to our second hypothesis, we found bidirectional associations among stress and sleep duration and efficiencyindicating a day-to-day dynamic: participants’ stress levels were related to sleep that night, and importantly, sleep durationand efficiency were associated with the next days’ stress levels. These findings are similar to daily dynamic associationsbetween affect and sleep that have been demonstrated in healthy and clinical populations of adolescents. For example, in alarge adolescent sample, Fuligni and Hardway (2006) found that stressful demands during a particular day were associatedwith self-reported sleep duration that night and that less sleep was associated with increased feelings of anxiety, depression,and fatigue the next day. In another study, Cousins et al. (2011) found that daily PA and PA to NA ratios were associated withmore time in bed in depressed youth and less time in bed in anxious youth. They further found that sleep was associated withthe next days’ mood such that sleep duration was related to more PA the following day for both diagnostic groups. Thesestudies, along with the findings presented here, demonstrate the importance of going beyond measuring and modelingconcurrent associations to exploring the directionality and dynamicmechanisms between daily experiences and sleep in bothhealthy and clinical populations of varying ages.

Our third hypothesis was partially confirmed. We predicted that loneliness would moderate the stress and sleep quality/duration associations. In our sample, individuals high and average on loneliness who experienced particularly stressful dayswent on to have shorter sleep durations (approximately 41 fewer minutes of sleep on nights after a stressful day). Further,lonely individuals additionally experienced shorter sleep latencies after stressful days. Interestingly, in contrast to priorresearch (e.g., Hawkley et al., 2003), lonely individuals did not view daily events as more stressful. This moderation may bealternatively explained by the way lonely individuals cope with stressors as opposed to the actual levels of stress theyperceive. In terms of their social situations, lonely individuals attribute their problems to uncontrollable, trait-like de-ficiencies; this attributional style is then associated with less successful coping efforts (Anderson, 1980; Peplau & Perlman,1982). As such, we also hypothesize that lonely individuals may ineffectively manage stress through rumination, a poorcoping strategy. Research indicates that rumination often occurs when individuals experience a disconnect between theiractual and desired circumstances (e.g. loneliness; Martin & Tesser, 1996; Schoofs, Hermans, & Raes, 2012; Vanhalst, Luyckx,Raes, & Goossens, 2012). Further, other work has shown that rumination and anxiety mediate the direct associations betweenloneliness and sleep quality (Zawadzki, Graham, & Gerin, 2013). Therefore, lonely individuals’ ability to cope with dailystressors may have increased the salience of stress for these individuals and, ultimately, its relations with sleep that night.

Overall, our findings illustrated hypothesized compensatory mechanisms in sleep whereby when either sleep duration orsleep quality is decreased, the body adapts by improving the opposite sleep component (e.g., corresponding decreases insleep duration and sleep latency; Sadeh, Gruber, & Raviv, 2003). Our findings compliment a study of childrenwho underwentexperimental sleep deprivation and subsequently exhibited increased sleep quality (e.g., decreased night awakenings,decreased sleep latency; Sadeh et al., 2003). Compensatory mechanisms were also found in a recent study of adolescents.Astill, Verhoeven, Vijzelaar, and Van Someren (2013) found a 13% decrease in sleep duration during holidays away from schoolas compared to a typical school week. This decreased sleep durationwas coupled with increases in overall sleep efficiency andfewer night awakenings (Astill et al., 2013). Future research should examine a greater number of potential compensatorymechanisms within naturalistic settings to gain a better understanding of sleep quantity and quality interplay.

To our knowledge, this is the first study to examine the constructs of loneliness, daily affect, stress and objective sleepdynamically across multiple days in adolescence. There are several limitations that should be considered. First, these mea-surements occurred in a select sample of youth who were planning on attending a university close to home and who weredisproportionately female. Further, although it was a racially and ethnically diverse sample, we were not able to examine dif-ferencesbyrace/ethnicitybeyondanexaminationofwhiteversusnon-white individualsbecauseof small cell sizes insomeof thegroups. These limitations of our sample limit the generalizability of our findings tomore diverse populations. However, a recentreport found that two-thirds of a national sample of high school graduates expected to attend college in the subsequent fall(Bureauof LaborStatistics, 2013), thereforeourfindingsmaygeneralize toa largepercentageof adolescents. Futureexaminationsshould utilize a larger and more educationally diverse sample of youth with equal representation of males to females.

Additionally, although we identified short-term bidirectional associations between stress and sleep in lonely individuals,we cannot conclude that the feelings of loneliness occurred first in this study. An alternative hypothesis is that individualswho experienced stressful days may have experienced poor sleep quality and shorter sleep duration prior to our study days,which led to participants reporting experiences of disconnection and isolation. Future research should examine both trait and

L.D. Doane, E.C. Thurston / Journal of Adolescence 37 (2014) 145–154 153

state experiences of loneliness to better understand the causal nature of the associations. Third, we were restricted to threedays and four nights of actigraphic sleep and self-reported stress and affect. Self-report bias may have influenced our diaryconstructs and it is possible that some youth were not compliant. Further, our design and models did not allow us to controlfor the influence of one night’s sleep on the subsequent night beyond accounting for the nesting of the days within in-dividuals. Future research should measure these dynamic constructs over a greater number of days with the inclusion ofcompliance measures for daily diary reports.

In sum, our findings suggest that daily stress levels were associated with objective sleep quality and duration inadolescence, an understudied and unique developmental period in which changes in sleep patterns are especially relevant(Carskadon, 2011). Loneliness played an important role such that the associations among stress and sleep were strongestamong individuals who were lonely. An integral strength of this study was the use of ambulatory assessment to measurethese constructs across multiple days and within naturalistic settings (Trull & Ebner-Priemer, 2013). Future research shouldbuild off of our findings and methodologies to examine how day-to-day processes of stress and sleep in relation to additionaltrait characteristics are predictive of later health and functioning across the transition to university and beyond into adult-hood. Further, our research may have implications for future prevention and intervention programs that include strategies tohelp youth cope effectively with stress as well as implement effective sleep strategies. Our findings suggest that one cannot bedealt with effectively without an improvement in the other. Additionally, such programs may be even more important forlonely individuals, as this population may require tailored strategies to manage stress and improve their sleep habits.

Acknowledgements

The authors would like to thank the participants of the ASU Transition to College Study for the time and effort theycontributed to this research. This research was partially supported by the Institute for Social Science Research at Arizona StateUniversity.

References

Adam, E. K., Hawkley, L. C., Kudielka, B. M., & Cacioppo, J. T. (2006). Day-to-day dynamics of experience-cortisol associations in a population-based sample ofolder adults. Proceedings of the National Academy of Sciences, 103, 17058–17063.

Åkerstedt, T., Kecklund, G., & Axelsson, J. (2007). Impaired sleep after bedtime stress and worries. Biological Psychology, 76, 170–173.Alfano, C. A., Zakem, A. H., Costa, N. M., Taylor, L. K., & Weems, C. F. (2009). Sleep problems and their relation to cognitive factors, anxiety, and depressive

symptoms in children and adolescents. Depression and Anxiety, 26, 503–512.Anderson, C. A. (1980). Motivational and performance deficits as a function of attributional style. Unpublished doctoral dissertation. Stanford University.Arber, S., Bote, M., & Meadows, R. (2008). Gender and socio-economic patterning of self- reported sleep problems in Britain. Social Science & Medicine, 68,

281–289.Arnett, J. J. (1999). Adolescent storm and stress, reconsidered. American Psychologist, 54, 317–326.Astill, R. G., Verhoeven, D., Vijzelaar, R. L., & Van Someren, E. J. W. (2013). Chronic stress undermines the compensatory sleep efficiency increases in response

to sleep restriction in adolescents. Journal of Sleep Research. http://dx.doi.org/10.1111/jsr.12032.Bartels, M., Cacioppo, J. T., Hudziak, J. J., & Boomsma, D. I. (2008). Genetic and environmental contributions to stability in loneliness throughout childhood.

American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 147B, 385–391.Boomsma, D. I., Willemsen, G., Dolan, C. V., Hawkley, L. C., & Cacioppo, J. T. (2005). Genetic and environmental contributions to loneliness in adults: The

Netherlands twin register study. Behavior Genetics, 35, 745–752.Bower, B., Bylsma, L. M., Morris, B. H., & Rottenberg, J. (2010). Poor reported sleep quality predicts low positive affect in daily life among healthy and mood-

disordered persons. Journal of Sleep Research, 19, 323–332.Buhrmester, D. (1990). Intimacy of friendship, interpersonal competence, and adjustment during preadolescence and adolescence. Child Development, 61,

1101–1111.Bureau of Labor Statistics. (2013). College enrollment and work activity of 2012 high school graduates. Retrieved from http://www.bls.gov/news.release/hsgec.

nr0.htm.Cacioppo, J. T., Hawkley, L. C., & Thisted, R. A. (2010). Perceived social isolation makes me sad: five-year cross-lagged analyses of loneliness and depressive

symptomatology in the Chicago health, aging, and social relations study. Psychology and Aging, 25, 453–463.Cacioppo, J. T., Hughes, M. E., Waite, L. J., Hawkley, L. C., & Thisted, R. A. (2006). Loneliness as a specific risk factor for depressive symptoms: cross-sectional

and longitudinal analyses. Psychology and Aging, 21, 140–151.Carskadon, M. A. (1990a). Adolescent sleepiness: increased risk in a high-risk population. Alcohol, Drugs, and Driving, 5, 317–328.Carskadon, M. A. (1990b). Patterns of sleep and sleepiness in adolescents. Pediatrician, 17, 5–12.Carskadon, M. A. (1999). When worlds collide: adolescent need for sleep versus societal demands. Phi Delta Kappan, 80, 348–353.Carskadon, M. A. (2011). Sleep in adolescents: the perfect storm. Pediatric Clinics of North America, 58, 637–647.Choquet, M., Kovess, V., & Poutignat, N. (1993). Suicidal thoughts among adolescents: an intercultural approach. Adolescence, 28, 649–659.Chung, K. F., & Cheung, M. M. (2008). Sleep–wake patterns and sleep disturbance among Hong Kong Chinese adolescents. Sleep, 31, 185–194.Clark, L. A., & Watson, D. (1988). Mood and the mundane: relations between daily life and self-reported mood. Journal of Personality and Social Psychology,

54, 296–308.Colton, M. E., & Gore, S. (1991). Adolescent stress: Causes and consequences. New York: Aldine De Gruyter.Cousins, J. C., Whalen, D. J., Dahl, R. E., Forbes, E. E., Olino, T. M., Ryan, N. D., et al. (2011). The bidirectional association between daytime affect and nighttime

sleep in youth with anxiety and depression. Journal of Pediatric Psychology, 36, 969–979.Dagys, N., McGlinchey, E. L., Talbot, L. S., Kaplan, K. A., Dahl, R. E., & Harvey, A. G. (2012). Double trouble? The effects of sleep deprivation and chronotype on

adolescent affect. Journal of Child Psychology and Psychiatry, 53, 660–667.Dahl, R. E., & Lewin, D. S. (2002). Pathways to adolescent health: sleep regulation and behavior. Journal of Adolescent Health, 31, 175–184.Enders, C. K., & Tofighi, D. (2007). Centering predictor variables in cross-sectional multilevel models: a new look at an old issue. Psychological Methods, 12,

121–138.Fuligni, A. J., & Hardway, C. (2006). Daily variation in adolescents’ sleep, activities, and psychological well-being. Journal of Research on Adolescence, 16,

353–378.Galambos, N. L., Howard, A. L., & Maggs, J. L. (2010). Rise and fall of sleep quantity and quality with student experience across the first year of university.

Journal of Research on Adolescence, 21, 342–349.

L.D. Doane, E.C. Thurston / Journal of Adolescence 37 (2014) 145–154154

Galambos, N. L., Dalton, A. L., & Maggs, J. L. (2009). Losing sleep over it: daily variation in sleep quantity and quality in Canadian students’ first semester ofuniversity. Journal of Research on Adolescence, 19, 741–761.

Graber, J. A., & Brooks-Gunn, J. (1996). Transitions and turning points: navigating the passage from childhood through adolescence. Developmental Psy-chology, 32, 768–776.

Greenfield, E. A., Lee, C., Friedman, E. L., & Springer, K. W. (2011). Childhood abuse as a risk factor for sleep problems in adulthood: evidence from a U.S.national study. Annals of Behavioral Medicine, 42, 245–256.

Gupta, N. K., Mueller, W. H., Chan, W., & Meininger, J. C. (2002). Is obesity associated with poor sleep quality in adolescents? American Journal of HumanBiology, 14, 762–768.

Hamilton, N. A., Affleck, G., Tennen, H., Karlson, C., Luxton, D., Preacher, K. J., et al. (2008). Fibromyalgia: the role of sleep in affect and in negative eventreactivity and recovery. Health Psychology, 27, 490–494.

Hanson,M. D., & Chen, E. (2010). Daily stress, cortisol, and sleep: themoderating role of childhood psychosocial environments.Health Psychology, 29, 394–402.Hawkley, L. C., Burleson, M. H., Berntson, G. G., & Cacioppo, J. T. (2003). Loneliness in every- day life: cardiovascular activity, psychosocial context, and health

behaviors. Journal of Personality and Social Psychology, 85, 105–120.Hawkley, L. C., & Cacioppo, J. T. (2010). Loneliness matters: a theoretical and empirical review of consequences and mechanisms. Annals of Behavioral

Medicine, 40, 218–227.Hawkley, L. C., Preacher, K. J., & Cacioppo, J. T. (2007). Multilevel modeling of social interactions and mood in lonely and socially connected individuals: the

MacArthur social neuroscience studies. In A. D. Ong, & M. H. M. van Dulmen (Eds.), Oxford handbook of methods in positive psychology (pp. 559–575).New York: Oxford University Press.

Haynes, P. L., Bootzin, R. R., Smith, L., & Stevens, S. (2006). Sleep and aggression in substance – abusing adolescents: results from an integrative behavioralsleep-treatment pilot program. Sleep, 29, 512–520.

Heinrich, L. M., & Gullone, E. (2006). The clinical significance of loneliness: a literature review. Clinical Psychology Review, 26, 695–718.Jenni, O. G., Achermann, P., & Carskadon, M. A. (2005). Homeostatic sleep regulation in adolescents. Sleep, 28(11), 1446–1454.Larson, R., Csikszentmihalyi, M., & Graef, R. (1980). Mood variability and the psychosocial adjustment of adolescents. Journal of Youth and Adolescence, 9, 469–490.Larson, R., & Richards, M. H. (1994). Family emotions: do young adolescents and their parents experience the same states? Journal of Research on

Adolescence, 4, 567–583.Lavie, P., Carmeli, A., Mevorach, L., & Liberman, N. (1991). Sleeping under the threat of the scud: war related environmental insomnia. Israel Journal of

Medical Sciences, 27, 677–680.Liu, X. (2004). Sleep and adolescent suicidal behavior. Pediatrics, 27, 1351–1358.Lund, H. G., Reider, B. D., Whiting, A. B., & Prichard, J. R. (2010). Sleep patterns and predictors of disturbed sleep in a large population of college students.

Journal of Adolescent Health, 46, 124–132.Martin, L. L., & Tesser, A. (1996). Some ruminative thoughts. In R. S. Wyer (Ed.), Ruminative thoughts: Advances in social cognition (Vol. 9); (pp. 1–47). Hillsdale,

NJ: Erlbaum.McCrae, C. S., McNamara, J. P. H., Rowe, M. A., Dzierzewski, J. M., Dirk, J., Marsiske, M., et al. (2008). Sleep and affect in older adults: using multilevel

modeling to examine daily associations. Journal of Sleep Research, 17, 42–53.Mezick, E. J., Matthews, K. A., Hall, M., Strollo, P. J., Buysse, D. J., Kamarck, T. W., et al. (2008). Influence of race and socioeconomic status on sleep: Pittsburgh

sleep score project. Psychosomatic Medicine, 70, 410–416.Morrison, D. N., McGee, R., & Stanton, W. R. (1992). Sleep problems in adolescence. Journal of American Academy of Child & Adolescents Psychiatry, 31, 94–99.National Sleep Foundation. (2011a). Annual sleep in America Poll exploring connections with communications technology use and sleep. Retrieved from http://

www.sleepfoundation.org/article/press-release/annual-sleep-america-poll-exploring-connections-communications-technology-use-.National Sleep Foundation. (2011b). How much sleep do we really need?. Retrieved from http://www.sleepfoundation.org/article/how-sleep-works/how-

much-sleep-do-we-really-need.Norlander, T., Johansson, Å., & Bood, S..Å. (2005). The affective personality: its relation to quality of sleep, well-being and stress. Social Behavior and Per-

sonality, 33, 709–722.Oakley, N. R. (1997). Validation with polysomnography of the sleepwatch sleep/wake scoring algorithm used by the Actiwatch Activity Monitoring System. Bend:

MiniMitter, Cambridge Neurotechnology.Peplau, L., & Perlman, D. (1982). Perspective on loneliness. In L. A. Peplau, & D. Perlman (Eds.), Loneliness: A sourcebook of current theory, research and therapy.

New York: Wiley Interscience.Pillar, G., Malhotra, A., & Lavie, P. (2000). Post-traumatic stress disorder and sleep – what a nightmare! Sleep Medicine Reviews, 4, 183–200.Preacher, K. J., Curran, P. J., & Bauer, D. J. (2006). Computational tools for probing interaction effects in multiple linear regression, multilevel modeling, and

latent curve analysis. Journal of Educational and Behavioral Statistics, 31, 437–448.Price, V. A., Coates, T. J., Thoresen, C. E., & Grinstead, O. A. (1978). Prevalence and correlates of poor sleep among adolescents. Archives of Pediatrics &

Adolescent Medicine, 132, 583–586.Raudenbush, S. W. B., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods. Thousand Oaks, California: Sage Publications.Russell, D. W. (1996). UCLA Loneliness Scale (Version 3): reliability, validity, and factor structure. Journal of Personality Assessment, 66, 20–40.Sadeh, A., Gruber, R., & Raviv, A. (2003). The effects of sleep restriction and extension on school-age children: what a difference an hour makes. Child

Development, 74, 444–455.Sadeh, A., Hauri, P. J., Kripke, D. F., & Lavie, P. (1995). The role of actigraphy in the evaluation of sleep disorders. Sleep, 18, 288–302.Sadeh, A., Keinan, G., & Daon, K. (2004). Effects of stress on sleep: the moderating role of coping style. Health Psychology, 23, 542–545.Schoofs, H., Hermans, D., & Raes, F. (2012). Effect of self-discrepancy on specificity of autobiographical memory retrieval. Memory, 20, 63–72.Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis: Modeling change and event occurrence. USA: Oxford University Press.Smetana, J. G., Campione-Barr, N., & Metzger, A. (2006). Adolescent development in interpersonal and societal contexts. Annual Review of Psychology, 57,

255–284.Talbot, L. S., McGlinchey, E. L., Kaplan, K. A., Dahl, R. E., & Harvey, A. G. (2010). Sleep deprivation in adolescents and adults: changes in affect. Emotion, 10, 831–841.Trull, T. J., & Ebner-Priemer, U. (2013). Ambulatory assessment. Annual Review of Clinical Psychology, 9, 151–176.Tworoger, S., Davis, S., Vitiello, M., Lentz, M., & McTiernan, A. (2005). Factors associated with objective (actigraphic) and subjective sleep quality in young

adult women. Journal of Psychosomatic Research, 59, 11–19.Van Reeth, O., Weibel, L., Spiegel, K., Leproult, R., Dugovic, C., & Maccari, S. (2000). Physiology of sleep (Review) – interactions between stress and sleep:

from basic research to clinical situations. Sleep Medicine Reviews, 4, 201–219.Vanhalst, J., Luyckx, K., Raes, F., & Goossens, L. (2012). Loneliness and depressive symptoms: the mediating and moderating role of uncontrollable rumi-

native thoughts. The Journal of Psychology, 146, 259–276.Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of

Personality and Social Psychology, 54, 1063–1070.Wolfson, A. R., & Carskadon, M. A. (1998). Sleep schedules and daytime functioning in adolescents. Child Development, 69, 875–887.Wolfson, A. R., & Carskadon, M. A. (2003). Understanding adolescents’ sleep patterns and school performance: a critical appraisal. Sleep Medicine Reviews, 7,

491–506.Zawadzki, M., Graham, J. E., & Gerin, W. (2013). Rumination and anxiety mediate the effect of loneliness on depressed mood and sleep quality in college

students. Health Psychology, 32(2), 212–222.Zeiders, K. H., Doane, L. D., & Adam, E. K. (2011). Reciprocal relations between objectively measured sleep patterns and diurnal cortisol rhythms in late

adolescence. Journal of Adolescent Health, 48, 566–571.