Journal of Adolescent Health 54 (2014) 536e542

www.jahonline.org

Original article

Contextual Moderators of Momentary Cortisol and Negative Affect inAdolescents’ Daily Lives

Leah D. Doane, Ph.D. a,*, and Katharine H. Zeiders, Ph.D. baDepartment of Psychology, Arizona State University, Tempe, Arizonab T. Denny Sanford School of Social and Family Dynamics, Arizona State University, Tempe, Arizona

Article history: Received March 15, 2013; Accepted October 11, 2013Keywords: Cortisol; Negative affect; Perceived discrimination; Social support; Adolescence

A B S T R A C TIMPLICATIONS AND

Purpose: To use an ecological momentary assessment design to examine the links betweenmomentary negative affect and cortisol in a sample of adolescents preparing to transition tocollege. Guided by a risk and resilience framework, we also explored whether important ecologicalfactors, perceived discrimination and social support, moderated the momentary associations be-tween negative affect and youths’ cortisol.Methods: Adolescents (N ¼ 77) provided salivary samples and diary reports of affect and expe-riences five times a day over 3 days. They also completed self-report questionnaires on perceiveddiscrimination and social support from family and friends.Results: Within-person increases in momentary negative affect were associated with increases incortisol. Perceived discrimination and social support from friends moderated this association.Adolescents who reported average and high levels of perceived discrimination experiencedexaggerated cortisol responses to negative affect, whereas adolescents who reported low levels ofperceived discrimination did not experience significant reactivity to negative affect. In contrast,adolescents who reported high levels of social support from friends experienced attenuatedcortisol responses to negative affect compared with adolescents who reported average or lowlevels of social support from friends.Conclusions: This study contributes to our understanding of youths’ daily socioemotional expe-riences and physiological reactivity by identifying how perceived discrimination and social supportfrom friends amplified and attenuated, respectively, the effects of negative affect on cortisolreactivity. Examining these processes within adolescents’ naturalistic environments advances ourunderstanding of the moderating role of ecological characteristics in adolescents’ everyday lives.

� 2014 Society for Adolescent Health and Medicine. All rights reserved.

* Address correspondence to: Leah D. Doane, Ph.D., Department of Psychology,P.O. Box 871104, Arizona State University, Tempe, AZ 85287-1104.

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

1054-139X/$ e see front matter � 2014 Society for Adolescent Health and Medicine. All rights reserved.http://dx.doi.org/10.1016/j.jadohealth.2013.10.007

CONTRIBUTION

This study identified linksbetween adolescents’emotional experiencesand physiological re-sponses in their everydaylives. Two ecological fac-tors, perceived discrimi-nation and social supportfrom friends, respectivelyamplified and attenuatedthe physiological re-sponses to emotion. Thesefindings illustrate poten-tial pathways by whichphysiological reactivityand everyday emotionscan affect adolescenthealth.

Recent research has highlighted the importance of under-standing the links between youths’ ecological environments andhealth through measurements of stress response systems thatinclude the hypothalamic pituitary adrenal axis (HPA axis) [1].However, few studies of physiological stress activity havefocused on the implications of daily socioemotional experiences

and more stable ecological factors in adolescents’ naturalisticenvironments. A focus on changes in physiology in response todaily experiences can help unravel connections between socio-emotional experiences in youths’ daily lives and later physicaland mental health [2]. In this study, we used an ecologicalmomentary diary assessment [3] and saliva sampling design toexamine the associations between negative affect (NA) andcortisol among adolescents preparing to transition to college.Drawing from an ecological perspective [4] and the risk andresilience framework [5], we also explored whether contextual

L.D. Doane and K.H. Zeiders / Journal of Adolescent Health 54 (2014) 536e542 537

factors, perceived discrimination and social support, moderatedthe association between NA and youths’ cortisol levels. Investi-gating the interplay between daily socioemotional experiences,ecological factors, and physiology in the naturalistic setting in-forms our larger understanding of the processes by which dailyexperiences affect youths’ long-term functioning.

The HPA axis is considered one of the body’s major stressresponse systems. When individuals encounter stress (which caninclude socioemotional experiences), the HPA axis initiates acascade of events, leading to the eventual release of the hormonecortisol [6]. Cortisol affects many biological functions essential todaily life (e.g., metabolism) and follows a typical diurnal rhythm,with levels highest in the morning, peaking 30 minutes afterwaking (the cortisol awakening response [CAR]), and decliningthroughout the day, reaching nadir in the evening [7]. AlthoughHPA activity in response to a stressor is considered adaptive inthe short term, repeated or long-term chronic activation isthought to lead to maladaptive responses (e.g., too high or toolow responsiveness to stressors) [8].

Examining the release of cortisol in response to stressfulexperiences has largely been studied in laboratory settings.Although such studies provide controlled settings to examinecortisol reactivity, the ecological validity of these findings islimited [9]. Recent work has demonstrated the utility of exam-ining cortisol reactivity in the naturalistic setting by examiningdeviations in cortisol levels away from individuals’ typicaldiurnal rhythm in response to naturally occurring stressors andsocioemotional experiences [10,11]. This naturalistic method iscritical in extending our understanding of individual differencesin cortisol reactivity to everyday, familiar experiences [9]. In thisstudy, we used a similar naturalistic method and examinedmomentary cortisol as an indicator of HPA axis reactivity tonaturally occurring socioemotional experiences in adolescents’daily lives.

Studies have consistently documented a positive link be-tween daily NA and cortisol levels among adults [12,13]. Lesswork, however, has investigated these associations among ad-olescents [10]. Such an examination appears critical because weknow that there are important developmental changes in NAand HPA axis activity during adolescence [14,15], and recentreviews have stressed the importance of distinguishing thedevelopmental differences in the link between affective di-mensions and the cortisol response [16]. The limited adolescentstudies have yielded similar findings to the adult literature;increases in momentary NA have been associated withmomentary increases in cortisol [10]. Other studies that haveinvestigated experiences similar to NA (e.g., loneliness) havefound that cortisol reactivity to these experiences was onlypresent in particular subgroups or was amplified under certainecological contexts such as chronic stress [11,17]. Given thelimited studies examining these constructs in adolescence, andevidence that associations between NA and cortisol reactivitycould vary across important ecological characteristics, a greaterexamination is needed.

The ecological perspective [4] and the risk and resilienceframework [5] suggest that development occurs within envi-ronments in which numerous contextual factors are at play.These contextual factors can exacerbate or diminish the asso-ciation between a risk factor (e.g., NA) and an outcome (e.g.,cortisol) [18]. Ecological factors that exacerbate the associationare considered vulnerability factors, whereas factors thatattenuate the association are considered protective factors [18].

One factor that emerges as important during adolescence isyouths’ experiences of perceived discrimination [19]. Theorysuggests that perceptions of discrimination gain salience duringadolescence because of cognitive and contextual changes [20].Furthermore, discrimination has been consistently linked withmaladjustment and physiological pathways are thought tomediate this link [21]. In relation to the HPA axis specifically,discrimination is theorized to be an especially potent activatorbecause of its uncontrollable and socially evaluative character-istics [22]. Although little work has examined the indirect roleof perceived discrimination in youths’ physiological func-tioning, one study found that individuals experienced thegreatest NA from interpersonal stress when they also reportedhigh levels of perceived discrimination [23]. The amplifyingeffect of discrimination in the relation between mood andcortisol, however, has not been examined. The current studyextends prior work by examining perceived discrimination as apotential vulnerability factor in the relation between NA andcortisol.

Whereas perceived discrimination could be conceptualized asa vulnerability factor, social support is likely to emerge as aprotective factor. Research has consistently found that socialsupport protects individuals from a variety of risk factors [24].Supportive relations have also emerged as protective in the linkbetween stressors and physiological responses [25]. However,the protective nature of social support in the relations betweenNA and cortisol has yet to be examined. The current studyfocused on two sources of social support: support from familyand support from friends. Focusing on these sources of supportseparately appears important because shifts occur in where ad-olescents spend their time (e.g., home vs. school) and in whomthey may confide [26].

The current study investigated within-person associationsbetween NA and diurnal cortisol in a sample of adolescentspreparing to transition to college. We hypothesized that youths’reports of NA would relate to greater momentary cortisol levels,accounting for each individual’s typical diurnal cortisol rhythmand daily behaviors. We also examined the moderating role ofyouths’ perceptions of discrimination and social support. Wehypothesized that perceptions of discrimination would act as avulnerability factor, amplifying the association between NA andyouths’ cortisol levels, whereas social support from both familyand friend support would emerge as protective factors, attenu-ating the association between NA and cortisol levels.

Methods

Participants

Data for the current study come from a longitudinal projectfocused on adjustment during the transition to college (N ¼ 82).Participants were contacted through a psychology departmentorientation at a large southwestern university and were requiredto live within 35 miles of the university, be a high school senior,and plan on attending the university in the subsequent fall. In-dividuals were excluded from current analyses if they werediagnosed with fibromyalgia (n ¼ 1), used corticosteroid medi-cations on sampling days (n ¼ 1), were noncompliant withsampling protocol (n ¼ 2), or had insufficient questionnaire data(n ¼ 2). The final analytic sample consisted of 77 youth (23%male), aged 17e18 years (mean, 18.03; standard deviation [SD],.41). Participants were racially and ethnically diverse: 55% white,

1 This variable was constructed by subtracting the wakeup time from each ofthe individual time points (e.g., 0 ¼ wakeup; .5 ¼wakeup plus 30 minutes, 2.5 ¼beep 1, etc.).

L.D. Doane and K.H. Zeiders / Journal of Adolescent Health 54 (2014) 536e542538

23% Latino/Hispanic descent, 12% multiracial, 5% African-American, and 5% Asian/Pacific Islander.

Procedures

The University Institutional Review Board approved all pro-cedures. Participants who consented to the study provided sali-vary samples and diary reports of emotions five times a day for 3typical consecutiveweekdays. Study personnel broughtmaterialsto participants’ homes, explained all procedures, and provided ane-mail address and phone number where participants couldreach study staff at any time. Upon completion of protocol, projectpersonnel picked up study materials and paid participants $40.

Study materials included three diaries, an actigraph (wrist-based accelerometer), an MEMS 6 (Aardax; Aardex Group, Rich-mond, VA) track cap compliance device with 16 straws, 16 vialsfor saliva sampling, and several questionnaires. Prompted by theactigraph, participants provided a salivary sample immediatelyafter waking, 30 minutes later, approximately 3 and 8 hours afterwaking, and at bedtime for 3 consecutive days. In conjunctionwith salivary samples, participants also completed diary entriesdocumenting their mood, stressful events, caffeine, alcohol,medication and nicotine use, food intake, exercise, and sleepingbehavior in the prior hour. In total, participants were requiredto fill out 15 diary entries (mean, 14.18; SD, 1.27). Becausenoncompliance has been shown to influence cortisol estimates[27], two individuals were excluded in analyses because they didnot use the track cap compliance devices during sampling. Forremaining participants, we considered participants compliant if(1) their waking sample was within 15 minutes of their waketime, and (2) their second sample was between 23 and 37 mi-nutes after their wake time sample [28]. Based on these con-straints, 45 waking samples and 67 wake plus 30 samples wereconsidered noncompliant and not used for analyses.

Measures

Cortisol. Salivary cortisol was collected by passive drool. Partici-pants labeled vials with the time and date of sampling. Completedsamples were collected from participants’ homes andwere storedat�20�C. Samples were sent by courier to Biochemisches Labor atthe University of Trier (Trier, Germany) to be assayed. Cortisolvalues are not significantly affected by transport over a period ofseveral days without refrigeration [29]. Samples were assayed induplicate, using a solid phase time-resolved fluorescence immu-noassay with fluorometric end point detection [30]. The intra-assay coefficient of variation ranged between 4.0% and 6.7%, andthe inter-assay coefficients of variation ranged between 7.1% and9.0%. Cortisol values were transformed using the natural logtransformation to correct for a strong positive skew in the dis-tribution and winsorized at 1.81 mg/dL (equivalent to 50 nmol/L).

Momentary affect. Affect was assessed using the Positive andNegative Affect Schedule [31]. Using 10 items, participantsresponded to items (e.g., “Towhat extent you have felt distressedin the past hour?” “To what extent you have felt enthusiastic inthe past hour?”) using a Likert range from 0 (“Very slightly or notat all”) to 4 (“Extremely”). We calculated NA scores by averagingacross 10 negative items at each time point (a ¼ .88).

Perceived everyday discrimination. A 10-item version of theEveryday Discrimination Scale [32] was used to assess

perceptions of discrimination. The scale asked participants,“Over the past year, in your day-to-day life, how often have any ofthe following things happened to you?” Responses varied from0 (“Never”) to 4 (“Four or more times in the past year”). Exampleitems included “being treated with less respect than others” and“being called names or insulted” (a ¼ .84).

Perceived friend and family social support. Perceived social supportwas assessed using the 20-item Perceived Social Support fromFriends scale and the 20-item Perceived Social Support fromFamily scale [33]. Participants responded to items (e.g., “I have adeep sharing relationship with a number of friends (familymembers),” “Most other people are closer to their friends (family)than I am”) by indicating “Yes,” “No,” or “I don’t know.” In linewith prior work [33], scores were coded 0 (“No” or “Don’t know”)and 1 (“Yes”). A total support score was created by reverse scoringnegative items and summing item scores within each scale.

Analytic plan

Hierarchical linear growth curve models were used to assessthe current study’s questions and to account for the nested na-ture of the data (moments nested within days, and days nestedwithin person) [34]. Specifically, a three-level growth model wasestimated to model the momentary-, daily-, and person-levelchanges in cortisol. Because of cortisol’s strong diurnal rhythm[7], we modeled each individual’s diurnal rhythm at Level 1 (L1)by including a time variable (indicating how long after wakingthe sample was provided1) and a dummy variable representingthe CAR sample. Negative affect was also included at L1 and wasgroup-mean centered by subtracting individuals’ average level ofNA from their level at a given time point. At Level 2 (L2), day-levelcontrols were included (i.e., wake time), and at Level 3 (L3),person-specific parameters (i.e., perceived discrimination, socialsupport from friends and family, aggregate of NA) were modeled.All L1 variables (except time and the CAR dummy code) weregroup-mean centered and L2 and L3 variables were grand-meancentered. All analyses included the following covariates: averagelevel of NA across sampling days, gender, caffeine or nicotine inthe prior hour, oral contraceptive use, race/ethnicity, and par-ents’ education [35].

First, we modeled each individuals’ diurnal cortisol rhythmsaccounting for covariates. Next, we examined the within-personand between-person effects of NA on individuals’ cortisol values.Finally, we examined whether perceived discrimination and so-cial support from family or friends moderated the associationbetween NA and cortisol by including cross-level interactionsbetween discrimination/social support and NA. Significant in-teractions were probed using the simple slopes technique forhierarchical linear modeling [36].

Results

Preliminary analyses and main effect of negative affect on cortisol

Table 1 presents descriptions and correlations among studyvariables. Participants showed the expected diurnal pattern of

Table 1Descriptive statistics and intercorrelations of cortisol, negative affect, perceived discrimination, and social support (N ¼ 77)

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

Minimum Maximum

1. Waking levels of cortisol 1.000 .192 (.468) .05 .542. Cortisol awakening

response�.491y 1.000 .651 (.388) �.38 1.38

3. Cortisol diurnal slope .471y .106 1.000 �.114 (.047) �.22 .004. PANAS negative affect .071 �.059 .138 1.000 .291 (.265) .01 1.495. Perceived discrimination .089 �.314y �.050 .320y 1.000 .981 (.627) .00 2.506. Perceived social support

from family�.146 .095 .038 �.155 �.347* 1.000 12.71 (5.83) .00 20.00

7. Perceived social supportfrom friends

�.033 .146 .003 �.261* �.189 .286* 1.000 13.68 (3.85) .00 20.00

8. Gender (male ¼ 1) .186 �.067 �.043 �.078 �.164 .027 �.138 1.000 .231 (.424)9. Non-Hispanic white �.037 .060 �.039 .063 .042 �.174 �.179 .005 1.000 .551 (.501)

10. Parent education level .085 �.128 .034 .014 �.041 .004 �.135 .056 �.284* 1.000 3.455 (1.544) 1.00 6.0011. Oral contraceptive use �.156 �.236* .214* .085 .129 .015 �.032 �.343y �.107 .000 1.000 .280 (.453)12. Caffeine use before

sampling�.172 �.136 .121 .031 .205 .003 .080 �.208 �.068 .100 .261* 1.000 .064 (.104) .00 .53

13. Nicotine use beforesampling

.065 �.110 .142 .227* .172 �.135 �.279* .154 �.145 .065 �.058 �.015 1.000 .011 (.054) .00 .38

PANAS ¼ Positive and Negative Affect Schedule.* p < .05.y p < .01.

L.D. Doane and K.H. Zeiders / Journal of Adolescent Health 54 (2014) 536e542 539

cortisol: high morning levels (g000 ¼ �1.57; p < .001; equivalentto .220 mg/dL), an approximately 88.5% (g100 ¼ .634; p < .01) rise30 minutes after waking (the CAR),2 and a decline in cortisolacross the day (g200 ¼ �.114; p < .01). Next, we included thewithin-person NA effect in the model at L1 while controllingfor the between-person NA effects at L3 (Table 2, Model 1).Within-person momentary NA was significantly associated withmomentary levels of cortisol (g300 ¼ .142; p < .05), which sug-gested that on occasions when individuals reported greater NA(compared with their own cross-time mean), they demonstrateda corresponding 15% increase in cortisol levels.

Perceived discrimination and social support as moderators

Perceived discrimination significantly moderated the asso-ciations between within-person NA and cortisol levels (g305 ¼.191; p < .01) (Table 2, Model 2). Probing of the interaction atlow (1 SD below the mean) and high (1 SD above the mean)values of perceived discrimination revealed a significant asso-ciation between NA and momentary cortisol levels at highlevels (b ¼ .232; p < .01) but not low levels (b ¼ �.007; ns)(Figure 1). The relationship between NA and cortisol was sta-tistically significant at values of perceived discrimination >1.00(50.1% of the sample).

Social support from friends significantly moderated thewithin-person associations between NA and cortisol levels(g307 ¼ �.022; p < .05), whereas family social support did not(g306 ¼ .010; ns). The simple slopes calculations (Figure 2) indi-cated that there was a significant association between NA andmomentary cortisol at low levels of friend social support (b ¼.197; p< .01), but not at high levels of social support from friends(b ¼ .027; ns). This was confirmed via the region of significance

2 Cortisol levels were log transformed; thus, the effect sizes for all variablescan be interpreted as a percent change per 1 unit change in the independentvariable after using the b% change ¼ ((e

ˇ

b)�1).

values: The relationship between NA and cortisol was significantwhen scores of social support from friends were less than 15(61.5% of the sample).

Discussion

Our findings demonstrated significant within-person associ-ations between adolescents’ experiences of NA and momentarylevels of cortisol. In addition, we found that perceived discrimi-nation and social support acted as vulnerability and protectivefactors, respectively. This evidence suggests that youths’ dailyaffective responses have important links with their daily physi-ological activity, and underscores that wemust take into accountthe broader ecological contexts. Furthermore, our findings holdclinical significance by identifying potential processes of risk andresilience in youths’ daily lives.

The first goal of the study was to examine the associationbetween momentary NA and cortisol among adolescents in thenaturalistic setting. Findings suggested that within-person in-creases in NA related to a 12% increase in cortisol, after accountingfor typical diurnal variation and daily behaviors. Our findingsalign with prior empirical work among adults [12,13] and thelimited work with other adolescent samples [10]. From a physi-ological perspective, momentary increases in cortisol are adaptivebecause they prepare an individual to cope with a stressor orsocioemotional experience [8]. Investigating these associationsbroadens our understanding of physiology and affect in adoles-cence and provides information about potential pathways un-derlying the development of disorders and later well-being.

We also examined ecological characteristics that might serveas protective or vulnerability factors. First, we found that per-ceptions of discrimination acted as a vulnerability factor,amplifying the association between NA and cortisol. Our findingsuggests that youths under levels of high discriminationexhibited an exaggerated physiological response to NA,compared with youths who were under lower levels ofdiscrimination. This is consistent with research showing that

Table 2Multilevel growth model regression estimates predicting daily cortisol activityfrom negative affect, perceived discrimination, and social support

Model 1 Model 2

Coefficient SE Coefficient SE

Intercept: waking level, g000 �1.570x .032 �1.570x .032Wake time, g010 .045* .021 .049y .021Male, g001 .126 .101 .111 .107White, g002 .141y .076 .128 .081Parent education, g003 �.006 .023 �.004 .023Oral contraceptive use, g004 �.098 .072 �.084 .078Between-person NA, g005 �.168 .118 �.164 .119Perceived discrimination, g006 �.058 .048Social supportefamily, g007 �.006 .005Social supportefriends, g008 .001 .009

Cortisol awakening response, g100 .634x .046 .636x .048Wake time, g110 �.029 .031 �.031 .032Male, g101 �.096 .159 �.110 .124White, g102 .032 .134 .030 .096Parent education, g103 �.024 .032 �.026 .031Oral contraceptive use, g104 �.348z .144 �.351z .129Between-person NA, g105 .224 .175 .208 .177

Time since waking: slope, g200 �.114x .005 �.113x .005Wake time, g210 �.006y .003 �.006y .003Male, g201 �.008 .013 �.008 .013White, g202 �.018* .011 �.018* .011Parent education, g203 .003 .003 .003 .003Oral contraceptive use, g204 .022y .010 .022y .010Between-person NA, g205 .016 .015 .014 .015

Momentary NA, g300 .142x .050 .112z .041Male, g301 �.051 .102 �.121 .102White, g302 .070 .081 .026 .082Parent education, g303 �.057y .025 �.064z .021Oral contraceptive use, g304 �.056 .087 �.030 .085NA � perceived discrimination, g305 .191z .062NA � social supportefamily, g306 .010 .007NA � social supportefriends, g307 �.022y .010

Momentary caffeine consumption, g400 .095 .122 .083 .121Momentary nicotine use, g500 .508 .414 .470 .402

All cortisol levels reflect log10 mg/dL.NA ¼ negative affect; SE ¼ standard error.

* p < .06.y p < .05.z p < .01.x p < .001.

Figure 1. Simple slope plots of momentary cortisol by momentary levels ofnegative affect at low and high levels of perceived discrimination.

Figure 2. Simple slope plots of momentary cortisol by momentary levels ofnegative affect at low and high levels of social support from friends.

L.D. Doane and K.H. Zeiders / Journal of Adolescent Health 54 (2014) 536e542540

individuals with chronic exposure to discrimination are morelikely to view new stressors as threatening and potentiallyharmful, thus activating a greater, and perhaps sustained, stressresponse [37]. This exaggerated response may result from psy-chological sensitivity to stimuli: Youths with greater exposure todiscrimination may find experiences of negative affect to bemore stressful than do their peers without such exposure.Although a cortisol response is considered adaptive in the shortterm, if these responses occur repeatedly owing to a chronicstressor such as discrimination, this could potentially lead todysregulation of the HPA axis [2]. Our findings align with theo-retical discussions of the indirect effects of discrimination [38];however, an alternative explanation could focus on the fact thatsome youth may be predisposed to having greater cortisol re-sponses to NA and may in turn perceive greater discrimination.Furthermore, personality or temperamental characteristicsassociated with cortisol, NA, and discrimination [39] couldexplain our findings. Although this cannot be disregarded, wecontrolled for between-person differences in NA that likely re-flected temperamental differences in negative emotionality. Ourfindings complement theoretical work [38] and align with thelimited prior empirical work demonstrating that perceptions of

discrimination have the ability to exacerbate the effects ofstressors [23]. Our study represents the first empirical exami-nation of the role of discrimination in youths’ daily experiencesof NA and physiological reactivity.

Second, we considered whether social support from familyand friends might serve as protective factors, attenuating theassociations between NA and cortisol. We found partial supportfor this hypothesis, because only social support from friends wassignificant. One explanation as to why support from friendsemerged as protective and support from family did not could bebecause of the increasing importance of peers in adolescents’lives [26]. When difficult daily experiences arise, adolescentsmight bemore inclined to go to their friends for support, and thissupport might be especially effective at reducing the heightenedphysiological state. Our findings complement work on the directassociations among social support and stress physiology [25],but also extend our knowledge of the importance of friends inadolescents’ physiological response to daily socioemotional

L.D. Doane and K.H. Zeiders / Journal of Adolescent Health 54 (2014) 536e542 541

experiences. Our findings suggest that youths with high levels ofsocial support may perceive increases in NA as less threateninggiven the high availability of coping resources available to them.For example, prior research has found that high levels ofperceived social support and the use of reappraisal coping stra-tegies were associated with attenuated cardiovascular responsesto stress [40]. Indeed, if a physiological response can be elimi-nated or lessened in the short-term through tangible copingresources, this might be preferable and lead to more adaptiveoutcomes.

The current study contributes to our understanding of youths’daily socioemotional experiences and physiological reactivity byidentifying ecological characteristics that amplify and/or atten-uate the effects of NA on cortisol reactivity. Furthermore, itassessed experiences in the naturalistic setting, allowing forthese associations to be tested in realistic ecological settingsexperienced by adolescents. From a prevention scienceperspective, our findings identify malleable protective factors(i.e., friend support) that could be the focus of intervention/prevention efforts.

Despite contributions, several limitations should be consid-ered. First, although the current study collected intensive mea-surements of daily experiences and cortisol activity, we wererestricted to 3 days of measurement and we assessed associa-tions at one time point. Future studies assessing a greaternumber of days and across a greater period of development (e.g.,the transition to high school) could uncover the complexities ofthe physiological processes underlying daily experiences andultimately inform our understanding of the ways in which dailyexperiences might accumulate to influence youths’ adjustment.Furthermore, although we tracked compliance of salivary sam-ples, paper-and-pencil measures of momentary affect were used(rather than electronic devices), which can lead to risk for non-compliant sampling [3]. Third, we employed a sampling strategywhich resulted in a sample that was predominately female,college-attending, and from a small geographic area, and whichlimited the generalizability of the results. Futurework focused onunderstanding these associations in a larger, more representativesample of youth is necessary. Finally, we assessed stable in-dicators of our participants’ ecological contexts. Youths’ experi-ences of discrimination and social support are likely to changefrom day to day or even across a particular day. Furthermore,there are numerous other environmental factors to consider inyouths’ daily lives (e.g., changes in work or school). More work isneeded that specifically addresses contextual changes (daily andacross the day) during adolescence to fully illuminate the role ofcontext, socioemotional experiences, and youths’ physiology.

Despite limitations, the current study provides importantknowledge about adolescents’ daily socioemotional experiencesand physiological responses in their naturalistic settings. It alsocontributes to our understanding of the role of context in theassociation between affect and cortisol. Examining these pro-cesses can advance our knowledge of adolescent adjustment andinform our larger understanding of the moderating role ofecological characteristics.

Acknowledgments

The authors thank the participants of the Arizona State Uni-versity Transition to College Study for the time and effort theycontributed to this research.

Funding Sources

This study was partially supported by the Institute for SocialScience Research at Arizona State University.

References

[1] McHale SM, Blocklin MK, Walter KN, et al. The roles of daily activities inyouths’ stress physiology. J Adolesc Health 2012;51:623e8.

[2] McEwen BS. Stress, adaptation, and disease: Allostasis and allostatic load.Ann N Y Acad Sci 1998;840:33e44.

[3] Stone AA, Shiffman S. Ecological momentary assessment (EMA) in behav-ioral medicine. Annal Behav Med 1994;16:199e202.

[4] Bronfenbrenner U. The ecology of human development. Cambridge (MA):Harvard University Press; 1979.

[5] Masten AS. Ordinary magic: Resilience processes in development. AmPsychol 2001;56:227e38.

[6] Kirschbaum C, Hellhammer D. Salivary cortisol in psychobiologicalresearch: An overview. Neuropsychobiology 1989;22:150e69.

[7] Kirschbaum C, Hellhammer D. Salivary cortisol. In: Fink G, ed. Encyclopediaof stress, Vol. 3. San Diego (CA): Academic Press; 2003. p. 379e83.

[8] Miller GE, Chen E, Zhou ES. If it goes up, must it come down? Chronic stressand the hypothalamic-pituitary-adrenocortical axis in humans. PsycholBull 2007;133:25e45.

[9] Kudielka BM, Gierens A, Hellhammer DH, et al. Salivary cortisol in ambu-latory assessmentdsome dos, some don’ts, and some open questions.Psychosom Med 2012;74:418e31.

[10] Adam EK. Transactions among adolescent trait and state emotion anddiurnal and momentary cortisol activity in naturalistic settings. Psycho-neuroendocrinology 2006;31:664e79.

[11] Doane LD, Adam EK. Loneliness and cortisol: momentary, day-to-day, andtrait associations. Psychoneuroendocrinology 2010;35:430e41.

[12] Jacobs N, Myin-Germeys I, Derom C, et al. A momentary assessment studyof the relationship between affective and adrenocortical stress responses indaily life. Biol Psychol 2007;74:60e6.

[13] Smyth J, Ockenfels MC, Porter L, et al. Stressors and mood measured on amomentary basis are associated with salivary cortisol secretion. Psycho-neuroendocrinology 1998;23:353e70.

[14] Gunnar MR, Wewerka S, Frenn K, et al. Developmental changes in HPAactivity over the transition to adolescence: Normative changes and asso-ciations with puberty. Dev Psychopathol 2009;21:69e85.

[15] Larson RW, Moneta G, Richards MH, et al. Continuity, stability, and changein daily social experience across adolescence. Child Dev 2003;73:1151e65.

[16] Gunnar MR, Adam EK. The hypothalamic-pituitary-adrenocortical systemand emotion: Current wisdom and future directions. Monogr Soc Res ChildDev 2012;77:109e19.

[17] Matias GP, Nicolson N, Freire T. Solitude and cortisol: Associations withstate and trait affect in daily life. Biol Psychol 2011;86:314e9.

[18] Rutter M. Psychosocial resilience and protective mechanisms. In: Rolf AS,Masten D, Cicchetti KH, et al., eds. Risk and Protective Factors in theDevelopment of Psychopathology. New York: Cambridge University Press;1990. p. 181e214.

[19] Fisher CB, Wallace SA, Fenton RE. Discrimination distress during adoles-cence. J Youth Adolesc 2000;29:679e95.

[20] Spears Brown C, Bigler R. Children’s perceptions of discrimination: Adevelopmental model. Child Dev 2005;76:533e53.

[21] Meyers HF. Ethnicity and socio-economic status-related stresses incontext: An interactive review and conceptual model. J Behav Med 2009;32:9e19.

[22] Dickerson SS, Kemeny ME. Acute stressors and cortisol responses: Atheoretical integration and synthesis of laboratory research. Psychol Bull2004;130:355e91.

[23] Bennett GG, Merritt MM, Edwards CL, et al. Perceived racism and affectiveresponses to ambiguous interpersonal interactions among African Amer-ican men. Am Behav Sci 2004;47:963e76.

[24] Masten AS, Tellegen A. Resilience in developmental psychopathology:Contributions of the project competence longitudinal study. Dev Psycho-pathol 2012;24:345e61.

[25] Uchino BN, Garvey TG. The availability of social support reduces cardio-vascular reactivity to acute psychological stress. J Behav Med 1997;20:15e27.

[26] Larson R, Richards MH. Daily companionship in late childhood and earlyadolescence: Changing developmental contexts. Child Dev 1991;62:284e300.

[27] Kudielka BM, Broderick JE, Kirschbaum C. Compliance with saliva samplingprotocols: Electronic monitoring reveals invalid daytime profiles in non-compliant subjects. Psychosom Med 2003;65:313e9.

L.D. Doane and K.H. Zeiders / Journal of Adolescent Health 54 (2014) 536e542542

[28] DeSantis AS, Adam EK, Mendelsohn KA, Doane LD. Concordance betweenself-reported and objective wakeup times in ambulatory cortisol salivarycortisol research. Int J Behav Med 2010;17:74e8.

[29] Clements AD, Parker CR. The relationship between salivary cortisol con-centrations in frozen versus mailed samples. Psychoneuronendrinology1998;23:613e6.

[30] Dressendorfer RA, Kirschbaum C, Rohde W, et al. Synthesis of acortisolebiotin conjugate and evaluation as a tracer in an immunoassay forsalivary cortisol measurement. J Steroid Biochem Mol Biol 1992;43:683e92.

[31] Watson D, Clark LA, Tellegen A. Development and validation of briefmeasures of positive and negative affect: The PANAS scales. J Pers SocPsychol 1988;54:1063e70.

[32] Williams DR, Yu Y, Jackson JS, et al. Racial differences in physical andmental health: Socio-economic status, stress and discrimination. J HealthPsychol 1997;2:335e51.

[33] Procidano ME, Heller K. Measures of perceived social support from friendsand from family: Three validation studies. Am J Commun Psychol 1983;11:1e24.

[34] Raudenbush SW, Bryk AS. Hierarchical linear modeling. Thousand Oaks(CA): Sage; 2002.

[35] Adam EK, Kumari M. Assessing salivary cortisol in large-scale, epidemio-logical research. Psychoneuroendocrinology 2009;34:1423e36.

[36] Preacher KJ, Curran PJ, Bauer DJ. Computational tools for probing in-teractions in multiple linear regression, multilevel modeling, and latentcurve analysis. J Educ Behav Stat 2006;31:437e48.

[37] Broudy R, Brondolo E, Coakley V, et al. Perceived ethnic discrimination inrelation to daily moods and negative social interactions. J Behav Med 2007;30:31e43.

[38] Garcia Coll C, Crnic K, Lamberty G, et al. An integrative model for the studyof developmental competences in minority children. Child Dev 1996;67:1891e914.

[39] Nater UM, Hoppmann C, Klumb PL. Neuroticism and conscientiousness areassociated with cortisol diurnal profiles in adultsdRole of positive andnegative affect. Psychoneuroendocrinology 2010;35:1573e7.

[40] Lee YSC, Suchday S, Wylie-Rosett J. Perceived social support, coping styles,and Chinese immigrants’ cardiovascular responses to stress. Int J BehavMed 2012;19:174e85.