Microaggressions and Diurnal Cortisol: Examining Within-Person Associations Among African-American and Latino Young Adults

ARTICLE IN PRESS Journal of Adolescent Health 000 (2018) 1 7

www.jahonline.org Original article

Microaggressions and Diurnal Cortisol: Examining Within-Person Associations Among African-American and Latino Young Adults D1X XKatharine H. Zeiders, D2X XPh.D.a,*, D3X XAntoinette M. Landor, D4X XPh.D.b, D5X XMelissa Flores, D6X XM.S.a, and D7X XAlaysia BrownD8XbX a b

Norton School of Family and Consumer Sciences, University of Arizona, Tucson, Arizona Human Development and Family Science, University of Missouri, Columbia, Missouri

Article History: Received December 24, 2017; Accepted April 26, 2018 Keywords: Diurnal cortisol; Ethnic-racial discrimination; Microaggressions; HPA axis

TAGEDPA B S T R A C T

Purpose: We examined the relations between African-American and Latino young adults’ microaggressions and subsequent changes in weekly diurnal cortisol parameters (i.e., cortisol awakening responses, overall cortisol output (AUC), and diurnal slopes). Methods: Young adults (N = 53, Mage = 20 years, SD = .90; 72% female) participated in a 4-week diary study in which they reported their weekly experiences of microaggressions and completed 2 days of saliva samples each week. Saliva samples were obtained at waking, 30-minutes after waking, and bedtime on each sampling day (six samples each week; 24 samples total). In line with an idiographic approach to stress, young adults’ increases and decreases in microaggressions (relative to their own average) were linked to changes in cortisol parameters the following week. Results: Increases in microaggressions predicted greater AUC the subsequent week, controlling for gender, race, parental education, prior week’s AUC, and weekly behavioral controls. Follow-up analyses of specific types of microaggressions indicated that experiences centered around criminality and secondclass citizenship also related to increases in young adults’ cortisol awakening responses the subsequent week. Microaggressions were unrelated to changes in diurnal slopes. Conclusions: Microaggressions were linked to subsequent changes in diurnal cortisol among AfricanAmerican and Latino young adults. Given the rigorous within-person design, findings point to the importance and impact of subtle forms of discrimination on young adults’ hypothalamic-pituitary-adrenal axis activity, which is theorized to underlie health and well-being. © 2018 Society for Adolescent Health and Medicine. All rights reserved.

The deleterious effects of discrimination experiences are well documented; discrimination is linked to poor mental and physical health, and academic difficulties [1,2]. Physiological processes are theorized to underlie the long-term consequences of discrimination and growing empirical evidence suggests that the hypothalamic-pituitary-adrenal (HPA) axis may, in part, be one of the stress-response systems at play [3]. Yet, notably missing from our

Conflict of Interest: The authors have no conflicts of interest to disclose. * Address correspondence to: Katharine H. Zeiders, Norton School of Family and Consumer Sciences, University of Arizona, Tucson, AZ 85721. E-mail address: [email protected] (K.H. Zeiders). 1054-139X/© 2018 Society for Adolescent Health and Medicine. All rights reserved. https://doi.org/10.1016/j.jadohealth.2018.04.018

IMPLICATIONS AND CONTRIBUTION

Microaggressions are common experiences for racialethnic minority young adults, although the health ramifications remain unclear. This study finds that microaggressions are linked to young adults’ diurnal cortisol (greater overall cortisol output and cortisol awakening response [CAR]). Such findings are important given that diurnal cortisol is a correlate of individuals’ physical and mental health.

understanding of the effects of discrimination broadly, and the relation to the HPA axis specifically, is the examination of discrimination experiences that are considered more subtle [4]. Utilizing a within-person lagged design, the current study examined the relations between subtle experiences of mistreatment, also termed microaggressions [4], and ethnic-racial minority young adults’ levels of cortisol, the main hormone of the HPA axis [5]. Cortisol levels, which follow a strong diurnal pattern—high at waking, peak 30 40 minutes after waking [e.g., CAR] and decline across the day [e.g., diurnal cortisol slope] [5]—have been linked to adults’ and youths’ physical and mental health [6]. Thus, we focused on young adults’ diurnal patterns of cortisol and linked within-person

ARTICLE IN PRESS 2

K.H. Zeiders et al. / Journal of Adolescent Health 00 (2018) 1 7

weekly changes in microaggressions to individuals’ subsequent diurnal cortisol patterns across a 4-week period. Such a design allowed us to test theoretical tenets of stress processes with statistical and methodological rigor. Further, it provided information about the physiological ramifications of microaggressions to help clarify questions about the overall nature of these experiences. Understanding whether microaggressions predict changes in young adults’ diurnal cortisol provides insight into the physiological costs of mistreatment among ethnic-racial minority youth in the U.S. Research has linked overt experiences of discrimination during adolescence and young adulthood to changes in diurnal functioning of the HPA axis [3]; discrimination has been related to greater overall cortisol output, flatter diurnal cortisol slopes, and greater CARs. Given that these diurnal components underlie health, such findings are theorized to reflect processes of allostaic load, or the biological toll of stress on individuals’ long-term functioning [7]. There remains, however, questions about whether subtler experiences of discrimination, microaggressions, may be linked to diurnal cortisol in the same way. Indeed, microaggressions have been found to relate to health and behavioral outcomes that include greater depressive [8,9], anxiety [10], and somatic symptoms [8], lower levels of self-esteem [11], lower sleep quality and duration of sleep [12], and poorer physical health [13]. Additionally, in a recent experimental study with Asian American young adults, microaggressions by white perpetrators, but not Asian American perpetrators, were associated with increased blood pressure [14]. Despite preliminary evidence that microaggressions carry a psychological and physiological toll, the existing research is limited because of its almost exclusive focus on cross-sectional, between-person associations (for exception see [15,16]). Betweenperson approaches provide information about the relations between microaggressions and well-being from a nomothetic approach [17,18]. They investigate relations across a set of individuals and examine the effects of experiencing high levels of microaggressions relative to the other individuals in the sample. This approach is limited because observed and unobserved stable characteristics of individuals, such as personality, can bias (and partially explain) the association [19,20]. Using intensive diary or longitudinal data, however, allows for the examination of the association from an idiographic approach, an approach that focuses on within-person fluctuations and examines the effects of experiencing high levels of microaggressions in comparison to one’s own average level. Such an approach is a more rigorous test because it aligns with theoretical tenets of stress processes and controls for stable between-person characteristics (e.g., personality) that may be theorized to explain the association. Extended to our understanding of microaggressions, an idiographic approach appears particularly needed given the ambiguous nature of experiences. Indeed, researchers have noted that cross-sectional research is unable to disentangle the effects of personality-related factors in linking microaggressions to outcomes, and called for work to take a more rigorous within-person design [15]. Additionally, such designs allow for lagged effects, which may be particularly relevant when ambiguity is experienced—individuals who are uncertain about the meaning of experiences may be more likely to have a delayed response, not captured in contemporaneous associations. Heeding this call, the current study examined the link between microaggressions and diurnal cortisol using an idiographic approach. Specifically, we utilized young adults’ reports of microaggressions across 4 weeks, and linked weekly fluctuations in these events to subsequent weekly changes in diurnal cortisol

parameters. We focused on three commonly discussed classifications of microaggressions: assumptions of inferiority (e.g., experiences in which others assume that you are less intelligent or less deserving), second-class citizen/assumptions of criminality (experiences that relate to substandard service or stereotypes as deviant/criminal), and microinvalidations (experiences in which others do not acknowledge race/ethnicity or ascribe to color-blind ideology), and examined their overall relation (average score) to three parameters of diurnal cortisol: the CAR, the area under the curve (AUC; overall cortisol output across the day), and the diurnal slope (the rate of decline from waking to bedtime cortisol). Consistent with prior work linking discrimination to cortisol [3], we hypothesized that microaggressions would relate to greater CARs and AUCs, and flatter diurnal slopes. We also explored the distinct associations of the three microaggressions to diurnal cortisol parameters, given that some evidence suggests differential relations between types of microaggressions and young adults’ mental health [21]. Understanding whether or not microaggressions physiologically affect young adults is an important preliminary step in the ongoing debate of whether subtle experiences of discrimination can get under the skin. Method Participants and procedures Data from the current study come from a larger weekly diary project focused on understanding stress, health, and relationships among 145 African-American and Latino young adults attending a Midwestern University. The study was approved by University Institutional Review Board (IRB #2002764). Young adults who self-identified as Black/African-American or Hispanic/Latino or Hispanic/Latino and white and between the ages of 18 25 years old were invited to participate in the study. Upon informed consent, participants completed an initial online assessment (approximately 1.5 hours) and weekly diary assessments for the 4 subsequent weeks (approximately 15 minutes each week). Weekly diary assessments were emailed to participants on each Friday. Upon completion of the weekly diaries, participants completed an end assessment (approximately 1 hour). For the current study, we utilized a subsample of 53 participants (16 Latino, 37 African-American) from the larger project that participated in an additional weekly salivary sampling protocol. This subsample provided salivary samples three times per day, for 2 consecutive days (Thursdays and Fridays), across the 4 weeks (total of 24 samples). On each of the sampling days, participants provided a waking, 30-minute postwaking, and bedtime sample. Participants were primarily female (72%), reflecting gender differences in rates of participation for the overall larger study and, in part, gender differences observed in survey research among young adults [22]. Young adults were, on average, 20.00 years old (SD = .90) and a majority were born in the U.S. (98%). Participants were compensated $75 for participation. A majority of participants (72%) completed all 4 weeks of self-reported diaries; 15.1% completed 3 weeks; 13% completed 2 or less weeks. We examined differences on demographic variables between participant who completed 3 4 weeks of nightly surveys to those who completed less than 3 weeks. No differences emerged on participant race, x2 (1) = 2.78, p = .10, gender, x2 (1) = .28, p = .60, living arrangement (i. e., living on-campus vs. off-campus), x2 (1) = .21, p = .65, age, t (51) = 1.25, p = .22, or income, t(50) = ¡.82, p = .41. As for salvia samples, a majority of participants provided 20 24 samples (76%);

ARTICLE IN PRESS K.H. Zeiders et al. / Journal of Adolescent Health 00 (2018) 1 7

17% of participants provided 15¡19 samples, and only 7% reported less than 15 samples. The lowest number of samples received was 5 (from 2 individuals). We examined differences in the number of saliva samples by participant gender, race/ethnicity, and living arrangements; no differences emerged, t(49) = .24, p = .81, t (51) = ¡.14, p = .89, t(51) = .81, p = .75, respectively. Further, we examined correlations between the number of saliva samples and participants age and income; correlations were not significant, r = ¡.01, p = .92 and r = ¡.15, p = .28, respectively. Measures Microaggressions. Participants completed the Racial and Ethnic Microaggressions Scale [23] each week. The scale, which has been validated among Latino and African-American young adults [23], assesses subtle statements and behaviors that communicate or insinuate negative messages to people of color. We utilized three subscales in the current study: Assumptions of Inferiority (eight items; “Someone acted surprised at my scholastic or professional success”), Second-Class Citizen and Assumptions of Criminality (seven items, “Someone clenched her/his purse or wallet upon seeing me”) and Microinvalidations (nine items, “Someone told me they “don’t see color”). Participants reported whether the event had occurred in the past 7 days (0 = No, 1 = Yes). A mean score was computed for each week. Internal consistency reliabilities for the microaggression subscales was not computed because events are considered independent; experiencing one microaggression does not increase individuals’ likelihood that they will experience another [24]. Participants reported on the source (perpetrator) of the microaggression by answering the following question: When thinking about the events of the past week, who were the people that said or did these things to you? Participants were asked to indicate all that applied: (1) friends, (2) other students you know (but not friends with), (3) other students you do not know, (4) instructors/professors, (5) employers or coworkers, and (6) other. Participants also reported on whether the perpetrator was of the same or different race/ethnicity as their own. Salivary cortisol. Participants expelled saliva through a straw into a 2 mL polypropylene tube and labeled tubes with the time and date. Participants were instructed not to eat, drink, or brush their teeth 30 minutes before sampling. Samples were returned, refrigerated at ¡20°C, and then sent on dry ice by courier to Biochemisches Laboratory, Trier, Germany to be assayed for cortisol. Cortisol levels are stable at room temperature for several weeks and unaffected by the conditions associated with shipping [25]. Assays were conducted in duplicate using a time-resolved immunoassay with fluorometric detection [26]. Intra-assay coefficients of variation (CVs) were between 4.0% and 6.7%, and interassay CVs ranged from 7.1% to 9.0%. Consistent with prior cortisol studies [27], 9 cortisol values (less than 1%) were top coded at 1.81 mg/dL to reduce the effects of outliers on the analysis. For the current analyses, we focused on three specific parameters of the diurnal cortisol: the CAR, AUC, and the diurnal slope. The CAR was calculated by subtracting each days’ Sample 2 (30minutes post waking) from Sample 1 (waking). The AUC (with respect to ground) was calculated using the trapezoid formula [28]. The diurnal slope was calculated by subtracting each day’s Sample 1 from Sample 3 (bedtime). An average across the 2 days was computed for each parameter and reflected each participant’s weekly cortisol parameter.

3

Individual and behavioral covariates. Participants reported on their age, race, gender, and parents’ education level during the initial assessment. Daily behaviors were assessed on each of the cortisol sampling days; participants indicated if they smoked cigarettes, consumed alcohol, or caffeine, exercised, napped, or took medication. Behaviors were averaged across the 2 days. Results Analyses were conducted in R statistical software [29]. Given the nested structure of our data (i.e., weeks nested within individuals), multilevel regression models were estimated whereby the intercept and slope components of model equations were allowed to vary randomly across individuals [19]. Weekly cortisol parameters (i.e., CAR, diurnal slope, and AUC) served as dependent variables and were tested in separate models. Microaggressions, prior levels of cortisol (i.e., prior levels of CAR, diurnal slope, and AUC) and behavioral controls (i.e., caffeine, cigarettes, alcohol, exercise, naps, and medications) were lagged 1 week to test temporal effects of microaggressions on cortisol. All time-varying predictor variables were person-centered to disaggregate within- and betweenperson effects [30]. Individual controls (i.e., gender, race, age, and parental education) were included as time-invariant predictors. Microaggression subscales were tested in separate models to avoid issues of collinearity. For the DV (cortisol), intraclass correlations were calculated, and effects of time (weekly changes) on cortisol were ruled out in preliminary models. Microaggressions were included as random effects in all models and compared to simpler models where the random-effect was removed. Nested models were then compared by assessing differences in the sums of squares of residuals for each model; effects were fixed if the change-statistic was nonsignificant and allowed to vary if changestatistic was significant. Statistical assumptions of normality of Level 1 and 2 residuals, and homoscedasticity of Level 2 randomeffects were examined [19]. Finally, for each model with a significant, substantive predictor, we conducted sensitivity analyses where outliers were identified and removed from each analysis to ascertain whether effects remained. Table 1 represents information about microaggression experiences. Differences by gender and race/ethnicity were examined. No differences emerged across gender; only one difference emerged across race/ethnicity. Specifically, 45.9% of African-American participants reported that “someone avoided eye contact with me” at least once across the 4 weeks; only 6.3% of Latino participants reported this event. Across the 4-week period, 75.5% of young adults reported at least one microaggression. The most common microaggressions included someone avoiding eye contact (African-American students), someone of a different ethnic-racial group stating that there was no difference between races, and someone avoiding sitting close to participants. As for the source (or perpetrator) of microaggressions, a majority of young adults reported a single source across the 4 weeks (65%). Of those who reported a single source, 16% reported it was a student they knew (but not a friend), 19% a friend, 24% a student they did not know, and 37% an “other” source. A majority of participants reported that the perpetrator of the microaggression was of a different race/ethnicity as them (97% Week (W)1, 83% W2, 76% W3, and 90% W4). Total microaggressions Total microaggressions were included as a fixed effect for AUC, CAR, and diurnal slopes model. Microaggressions significantly

ARTICLE IN PRESS 4

K.H. Zeiders et al. / Journal of Adolescent Health 00 (2018) 1 7

Table 1 Proportion of young adults reporting microaggressions across 4-week period Inferiority

%

Someone assumed that I grew up in a particular neighborhood. Someone assumed that I would not be intelligent. Someone acted surprised at my scholastic or professional success. Someone assumed that I would not be educated. Someone told me that I was “articulate” after she/he assumed I would not be. Someone assumed that I would have a lower education. Someone assumed that I held a lower paying job. Someone assumed that I was poor. Inferiority total (% of those who reported at least one microaggression) Second class citizen/Criminality (SCCC) Someone’s body language showed they were scared of me. Someone avoided walking near me on the street. Someone avoided sitting next to me in a public space (e.g., restaurants, movie theaters, classroom, and student center). I received substandard service in stores compared to customers of other racial/ethnic groups. Someone clenched her/his purse or wallet upon seeing me. Someone assumed that I would physically hurt them. Someone avoided eye contact with me. SCCC total (% of those who reported at least one microaggression) Microinvalidations I was told that I should not complain about racial/ethnic issues relating to my race/ethnicity Someone told me that they “don’t see color.” I was told that I complain about race/ethnicity too much. I was told that people of all racial/ethnic groups experience the same obstacles. I was told that people of color do not experience racism anymore. Someone told me that she or he was colorblind. Someone told me that they do not see race/ethnicity. Someone of a different racial/ethnic group has stated that there is no difference between the two of us. Someone told me that people should not think about race or ethnicity anymore. Microinvalidations total (% of those who reported at least one microaggression) Total across all microaggressions (% who reported at least one experience across 4-week period)

24.5 18.9 22.6 17.0 13.2 15.1 13.2 13.2 43.4 % 22.6 18.9 17.0 26.4 9.4 15.1 34.0 60.4 % 24.5 26.4 9.4 20.8 24.5 11.3 17.0 28.3 20.8 52.8 75.5

predicted the following week’s AUC cortisol after accounting for prior week’s AUC, race, gender, age, parental education, and behavioral controls (Table 2); on weeks that individuals reported greater levels of microaggressions (as compared to their overall average), they exhibited elevated levels of AUC cortisol the following week. Microaggressions were not associated with individuals’ CAR or diurnal slopes. We then examined each type of microaggression separately.

SCCC microaggressions were not related to individuals’ diurnal slopes. Microinvalidations Microinvalidations were included as random effects for AUC and CAR models, and a fixed effect for diurnal slopes. Microinvalidations were not significantly associated with subsequent weeks’ CAR, AUC, and diurnal slope.

Inferiority microaggressions Discussion Inferiority microaggressions were included as a fixed effect for AUC, CAR, and diurnal slope models. Inferiority microaggressions significantly predicted the following week’s AUC cortisol after accounting for prior week’s AUC, race, gender, age, parental education, and behavioral controls (Table 2); on weeks that individuals reported greater levels of inferiority microaggressions (as compared to their overall average), they exhibited elevated levels of AUC cortisol the following week. These microaggressions were not associated with individuals’ CAR or diurnal slopes. Second class citizen and criminality (SCCC) microaggressions Second class citizen and criminality (SCCC) microaggressions were included as random effects for CAR and diurnal slope models, and fixed effect for the AUC model. SCCC microaggressions significantly predicted the following week’s CAR and AUC; on weeks that individuals reported greater levels of SCCC microaggressions (as compared to their overall average), they demonstrated elevated CARs and AUCs the next week, accounting for prior cortisol levels, race, gender, age, parental education, and behavioral controls.

Discrimination remains a reality for ethnic-racial minority youth in the U.S. Attention to microaggressions has increased due to the frequency and ambiguous nature of these experiences. A growing debate has emerged about the impact of microaggressions and advocated for the importance of utilizing more rigorous methods to capture the impact of such experiences on people’s lives [15]. We followed recommendations for more rigorous designs and linked microaggressions to young adults’ diurnal cortisol from an idiographic approach—focusing on how individuals’ increases or decreases related to their subsequent changes in weekly diurnal cortisol. Results indicated that microaggressions did predict subsequent changes in young adults’ diurnal cortisol. Our findings provide preliminary evidence that, despite such events being subtle, they may evoke a physiological response among African-American and Latino young adults. Viewed in light of the growing empirical evidence that altered diurnal cortisol patterns have implications for long-term health [6], our findings warrant attention. Microaggressions predicted greater AUC, and this association was largely driven by two types of microaggressions: assumptions

ARTICLE IN PRESS K.H. Zeiders et al. / Journal of Adolescent Health 00 (2018) 1 7

5

Table 2 Multilevel regression estimates linking weekly fluctuations of microaggressions to young adults’ diurnal cortisol

Total microaggressions Gender Race Age Parent education Caffeine Cigarettes Alcohol Exercise Naps Medication

Inferiority microaggressions Gender Race Age Parent education Caffeine Cigarettes Alcohol Exercise Naps Medication

Second class/criminality microaggressions Gender Race Age Parent education Caffeine Cigarettes Alcohol Exercise Naps Medication

Micro-invalidations microaggressions Gender Race Age Parent education Caffeine Cigarettes Alcohol Exercise Naps Medication

CAR b (SE)

AUC b (SE)

Diurnal slope b (SE)

.35 (.29) .09 (.05)* ¡.08 (.05)* .01 (.03) ¡.01 (.01) .02 (.04) .00 (.24) ¡.03 (.04) .04 (.03) ¡.02 (.08) ¡.04 (.12) CAR b (SE) .22 (.22) .09 (.05)* ¡.08 (.05) .01 (.03) ¡.01 (.01) .02 (.04) ¡.02 (.25) ¡.03 (.04) .04 (.03) ¡.02 (.08) ¡.03 (.12) CAR b (SE) 1.06 (.48)** .09 (.05)** ¡.09 (.04)** .00 (.02) ¡.00 (.01) .01 (.04) .07 (.19) ¡.05 (.03) .03 (.03) .02 (.07) ¡.01 (.10) CAR b (SE) .03 (.20) .09 (.05)* ¡.08 (.05)* .01 (.03) ¡.01 (.01) .02 (.04) .03 (.24) ¡.03 (.04) .04 (.03) ¡.02 (.09) ¡.04 (.12)

6.97 (3.41)** .14 (.60) ¡.27 (.58) .10 (.32) ¡.00 (.17) ¡.07 (.52) ¡1.05 (2.90) .01 (.47) .18 (.40) ¡1.52 (1.01) ¡1.62 (1.39) AUC b (SE) 6.42 (2.61)** .14 (.60) ¡.24 (.58) .07 (.32) ¡.02 (.17) ¡.21 (.51) ¡2.01 (2.92) ¡.01 (.47) .17 (.40) ¡1.67 (1.00) ¡1.57 (1.38) AUC b (SE) 7.27 (3.03)** .06 (.60) ¡.29 (.58) .10 (.32) .00 (.17) ¡.05 (.51) ¡.42 (2.87) .08 (.47) .20 (.40) ¡1.71 (1.00) ¡1.30 (1.39) AUC b (SE) ¡.29 (3.03) .25 (.60) ¡.34 (.58) .10 (.32) .05 (.17) ¡.45 (.55) ¡.29 (2.83) ¡.13 (.48) .10 (.42) ¡1.28 (1.03) ¡1.64 (1.38)

.28 (.20) .08 (.04)** ¡.04 (.04) .02 (.02) ¡.01 (.01) ¡.09 (.03)*** ¡.14 (.18) .05 (.03) .00 (.02) ¡.01 (.06) .18 (.08)** Diurnal slope b (SE) .06 (.16) .09 (.04)** ¡.04 (.04) .02 (.02) ¡.01 (.01) ¡.09 (.03)*** ¡.13 (.18) .05 (.03) ¡.00 (.02) ¡.01 (.06) .18 (.08)** Diurnal slope b (SE) .25 (.19) .08 (.04)** ¡.04 (.04) .02 (.02) ¡.01 (.01) ¡.09 (.03)*** ¡.13 (.18) .05 (.03) .00 (.02) ¡.02 (.06) .19 (.08)** Diurnal slope b (SE) .19 (.14) .09 (.04)** ¡.04 (.04) .02 (.02) ¡.01 (.01) ¡.08 (.03)*** ¡.11 (.18) .05 (.03) .00 (.02) ¡.00 (.06) .17 (.08)**

Note: All analyses are adjusted for the previous week's measure of cortisol: these coefficients were all significant at p < .01; Gender is coded 1 = Male, 2 = Female; Race is coded 1 = African-American or Black, and 2 = Latino. Cortisol values reflect µg/dL. AUC = area under the curve; CAR = cortisol awakening response; SE = standard error. * p < .10. ** p < .05. *** p < .01.

of inferiority and second class citizens/criminality. Prior work has demonstrated that both of these types of microaggressions relate to aspects of youths’ health [21]. However, our study was the first to link these microaggressions to aspects of physiological functioning. One explanation as to why these two types of microaggressions emerged as important centers around the context and developmental stage of participants in our study. African-American and Latino young adults were attending a predominantly White university within a college town; thus, they represented the numerical ethnic/racial minority group within classrooms and

outside of the university. These contexts can be particularly challenging for ethnic/racial minority students [31 33]. Assumptions of inferiority microaggressions included experiences in which perpetrators expressed surprise or disbelief that African-American or Latino students were intelligent, successful, and/or capable of college. Given that these young adults were already attending college, hearing others make statements related to their academic abilities could be particularly stressful. Related, second class citizen/criminality experiences included times where perpetrators showed fear, weariness, or substandard

ARTICLE IN PRESS 6

K.H. Zeiders et al. / Journal of Adolescent Health 00 (2018) 1 7

service. Although some of these events are not related specifically to academics, they may still reinforce students feeling unwelcome and thoughts of not belonging [32,33]. Such feelings may evoke a physiological response, resulting in heightened levels of cortisol. Note that these same microaggressions were also related to increases in young adults’ CAR—when students reported an increase in these microaggressions, they reported a heightened CAR the subsequent week. The CAR has been linked to general stress in adults [34] and discrimination among adolescents [35]; however, it has also been linked to anticipatory stress [36]. Recent qualitative studies on microaggressions and health suggest that individuals may experience rumination, anger, and intrusive thoughts after the encounter [37,38]. Thus, it could be that when young adults are experiencing microaggressions that have immediate behavioral ramifications (e.g., someone clenching purse or avoiding sitting next to them in a public space), they then experience a heightened physiological state the following week as a result of repeatedly thinking about the encounter and simultaneously experiencing or re-experiencing it. Future research investigating even shorter, more intense assessments of microaggressions and diurnal cortisol (e.g., daily) are needed to disentangle reactions to and anticipatory feelings toward microaggressions. In sum, the current study is the first empirical examination, to our knowledge, linking within-person changes in microaggressions to young adults’ diurnal cortisol. The idiographic design of the study allowed for the examination of whether or not microaggressions related to changes within individuals—a theoretically sound test of the effects of stress on individuals and information about the physiological effects of subtle forms of discrimination. Understanding the effects of microaggressions on young adults’ HPA-axis functioning is needed given that more overt experiences of discriminations have been consistently linked to numerous aspects of health and well-being [1]. Our findings provide initial evidence that certain types of microaggressions are related to overall cortisol output, and to a lesser degree, the CAR. Despite the contributions of the study, there are limitations worth noting. First, the study focused on a small sample of AfricanAmerican and Latino young adults, primarily female, limiting our examination of potential ethnic-racial and gender differences (and the potential interaction) in the relations between microaggressions and diurnal cortisol. Future research with a larger sample would allow for greater attention to within-ethnic/racial variation, and for the examination of other relevant factors (e.g., ethnic identity, skin-tone) found to play a role in experiences of discrimination [39,40]. Second, the study utilized a within-person analyses; although these analyses control for the influence of stable individual or contextual factors, they can be biased by omitted time-varying factors that may change week-to-week or day-to-day [19]. Future research focused on daily changes, while accounting for fluctuating within-person factors that may be related to microaggressions should be explored. Finally, the study did not assess participant compliance during the salivary sampling protocol. Study personnel maintained close contact with participants, increasing compliance. However, exact compliance of sampling timing could not be ascertained. Future work should include electronic compliance tracking to address this limitation. The study is the first to link microaggressions to AfricanAmerican and Latino young adults’ diurnal cortisol profiles. The within-person design provides information about individual-level responses to microaggressions and suggests that these experiences

are linked to changes in diurnal cortisol. Given the growing awareness that changes in diurnal cortisol patterns have implications for long-term overall health and well-being [6], our findings warrant replication and validation in future research. We encourage researchers to build on the present findings by continuing idiographic examinations of microaggressions alongside longitudinal long-term health follow-ups. Acknowledgments Work on this paper was supported, in part, by a University of Missouri Research Council Grant URC-15-073 and the University of Arizona’s Francis McClelland Institute. We gratefully acknowledge participating young adults and the work of our undergraduate/ graduate students: Deidra Bibbs, Ruth Cardenas, Alexandra Davis, Asantewaa Darkwa, Kathleen Holloway, Jayley Janssen, Symone Lenoir, Maiya Smith, Carlton Slaughter, Kimberly Taylor, and Kristen Turner. References [1] Pascoe EA, Smart Richman L. Perceived discrimination and health: A metaanalytic review. Psychol Bull 2009;135:531–54. ~ a-Taylor AJ, Gonzales-Backen MA, et al. Latino adolescents’ [2] Alfaro EC, Uman academic success: The role of discrimination, academic motivation, and gender. J Adolesc 2009;32:941–62. [3] Busse D, Yim IS, Campos B, et al. Discrimination and the HPA axis: Current evidence and future directions. J Behav Med 2017;40:539–52. [4] Sue DW, Capodilupo CM, Torino GC, et al. Racial microaggressions in everyday life: Implications for clinical practice. Am Psychol 2007;62:271–86. [5] Pruessner JC, Wolf OT, Hellhammer DH, et al. Free cortisol levels after awakening: A reliable biological marker for the assessment of adrenocortical activity. Life Sci 1997;61:2539–49. [6] Adam EK, Quinn ME, Tavernier R, et al. Diurnal cortisol slopes and mental and physical health outcomes: A systematic review and meta-analysis. Psychoneuroendocrinology 2017;83:25–41. [7] Juster RP, McEwen BS, Lupien SJ. Allostatic load biomarkers of chronic stress and impact on health and cognition. Neurosci Biobehav Rev 2010;35:2–16. [8] Huynh VW. Ethnic microaggressions and the depressive and somatic symptoms of Latino and Asian American adolescents. J Youth Adolesc 2012;41:831–46. [9] Torres L, Taknint JT. Ethnic microaggressions, traumatic stress symptoms, and Latino depression: a moderated mediational model. J Couns Psychol 2015;62:393–401. [10] Liao KY-H, Weng C-Y, West LM. Social connectedness and intolerance of uncertainty as moderators between racial microaggressions and anxiety among Black individuals. J Couns Psychol 2016;63:240–6. [11] Nadal KL, Wong Y, Griffin KE, et al. The adverse impact of racial microaggressions on college students’ self-esteem. J Coll Stud Dev 2014;55:461–74. [12] Ong AD, Cerrada C, Lee RA, et al. Stigma consciousness, racial microaggressions, and sleep disturbance among Asian Americans. Asian Am J Psychol 2017;8:72–81. [13] Nadal KL, Griffin KE, Wong Y, et al. The injurious relationship between racial microaggressions and physical health: Implications for social work. J Ethn Cult Divers Soc Work 2017;26:6–17. [14] Wong-Padoongpatt G, Zane N, Okazaki S, et al. Decreases in implicit selfesteem explain the racial impact of microaggressions among Asian Americans. J Couns Psychol 2017;64:574–83. [15] Ong AD, Burrow AL. Microaggressions and daily experience: depicting life as it is lived. Perspect Psychol Sci 2017;12:173–5. [16] Ong AD, Burrow AL, Fuller-Rowell TE, et al. Racial microaggressions and daily well-being among Asian Americans. J Couns Psychol 2013;60:188–99. [17] Hankin BL, Young JF, Abela JRZ, et al. Depression from childhood into late adolescence: Influence of gender, development, genetic susceptibility, and peer stress. J Abnorm Psychol 2015;124:803–16. ~ a-Taylor AJ, Updegraff KA, et al. An idiographic and nomo[18] Zeiders KH, Uman thetic approach to the study of Mexican-origin adolescent mothers’ socio-cultural stressors and adjustment. Prev Sci 2015;16:386–96. [19] Singer JD, Willett JB. Applied longitudinal data analysis: Modeling change and event occurence. New York, NY: Oxford University Press, 2003. [20] Wilt J, Revelle W. A personality perspective on situations. In: Rauthmann JF, Sherman R, Funder DC, editors. Oxford handbook of psychological situations. Oxford University Press; 2017.

ARTICLE IN PRESS K.H. Zeiders et al. / Journal of Adolescent Health 00 (2018) 1 7 [21] Nadal KL, Griffin KE, Wong Y, et al. The impact of racial microaggressions on mental health: Counseling implications for clients of color. J Couns Dev 2014;92:57–66. [22] Porter SR, Umbach PD. Student survey response rates across institutions: Why do they vary? Res High Educ 2006;47:229–47. [23] Nadal KL. The racial and ethnic microaggressions scale (REMS): Construction, reliability, and validity. J Couns Psychol 2011;58:470–80. [24] Bollen K, Lennox R. Conventional wisdom on measurement: A structural equation perspective. Psychol Bull 1991;110:305–14. [25] Clements AD, Richard Parker C. The relationship between salivary cortisol concentrations in frozen versus mailed samples. Psychoneuroendocrinology 1998;23:613–6. € rfer RA, Kirschbaum C, Rohde W, et al. Synthesis of a cortisol-biotin [26] Dressendo conjugate and evaluation as a tracer in an immunoassay for salivary cortisol measurement. J Steroid Biochem Mol Biol 1992;43:683–92. [27] Adam EK, Kumari M. Assessing salivary cortisol in large-scale, epidemiological research. Psychoneuroendocrinology 2009;34:1423–36. [28] Pruessner JC, Kirschbaum C, Meinlschmid G, et al. Two formulas for computation of the area under the curve represent measures of total hormone concentration versus time-dependent change. Psychoneuroendocrinology 2003;28:916–31. [29] Core Team R. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, 2017. [30] Curran PJ, Bauer DJ. The disaggregation of within-person and betweenperson effects in longitudinal models of change. Annu Rev Psychol 2011;62:583–619.

7

[31] Harwood SA, Huntt MB, Mendenhall R, et al. Racial microaggressions in the residence halls: Experiences of students of color at a predominantly White university. J Divers High Educ 2012;5:159–73.  rzano D, Ceja M, Yosso T. Critical race theory, racial microaggressions, and [32] Solo campus racial climate: The experiences of African American college students. J Negro Educ 2000;69:60–73. [33] Yosso T, Smith W, Ceja M, et al. Critical race theory, racial microaggressions, and campus racial climate for Latina/o undergraduates. Harv Educ Rev 2009;79:659–91. [34] Chida Y, Steptoe A. Cortisol awakening response and psychosocial factors: A systematic review and meta-analysis. Biol Psychol 2009;80:265–78. [35] Zeiders KH, Doane LD, Roosa MW. Perceived discrimination and diurnal cortisol: Examining relations among Mexican American adolescents. Horm Behav 2012;61:541–8. [36] Fries E, Dettenborn L, Kirschbaum C. The cortisol awakening response (CAR): Facts and future directions. Int J Psychophysiol 2009;72:67–73. [37] Hall JM, Fields B. “It’s Killing Us!” Narratives of Black adults about microaggression experiences and related health stress. Glob Qual Nurs Res 2015;2:1–14. [38] Hoggard LS, Byrd CM, Sellers RM. Comparison of African American college students’ coping with racially and nonracially stressful events. Cult Divers Ethn Minor Psychol 2012;18:329–39. [39] Landor AM, Barr A. Politics of respectability, colorism, and the terms of social exchange in family research. J Fam Theory Rev 2018;10:330–47. [40] Landor AM, Halpern CT. The enduring significance of skin tone: Linking skin tone, attitudes toward marriage and cohabitation, and sexual behavior. J Youth Adolesc 2016;45:986–1002.