Skip to main content
SearchLoginLogin or Signup

Social Support Moderates the Link Between Chronic Peer Victimization in School and Later Cortisol Secretion

Published onJul 10, 2024
Social Support Moderates the Link Between Chronic Peer Victimization in School and Later Cortisol Secretion
·

Funding

Funding for this study was provided by the Canadian Institutes of Health Research (MOP142350). IOM holds the Canada Research Chair in the Developmental Origins of Vulnerability and Resilience and MB holds the Canada Research Chair in Child Development.

Corresponding Author

Mara Brendgen, Ph.D., Department of Psychology, University of Quebec at Montreal, C.P. 8888 succursale Centre-ville, Montreal, Quebec, Canada, H3C 3P8, email: [email protected].

Keywords

Peer Victimization, Hair Cortisol, Social Support, Adolescence, Emerging Adulthood, Genetic and Environmental Influences

Introduction

Peer victimization, defined as being the recipient of physical, verbal, or indirect aggression, may have important repercussions for mental and physical well-being that often last into adulthood (Moore et al., 2017; Schacter, 2021). Peer victimization may impact mental or physical health by altering the functioning of neurophysiological stress systems, particularly the hypothalamic-pituitary-adrenal (HPA) axis (Schacter, 2021; Vaillancourt et al., 2013). Of the hormones secreted by the HPA axis, cortisol has arguably received most attention, because it can be activated by a broad array of mental and physical stressors, and it is involved in the regulation of many systems critical for well-being. Thus, cortisol is crucial in the central nervous system, contributing to learning, memory, and emotion; in the metabolic system, overseeing glucose storage and utilization; and in the immune system, managing the extent and duration of inflammatory responses and the maturation of lymphocytes (Miller et al., 2007). Cortisol has also been identified as a key factor connecting stress to mental and physical health problems (Miller et al., 2007). Specifically, while cortisol hypersecretion is a risk factor for impaired mental and physical wellbeing (Knezevic et al., 2023), blunted cortisol reactivity is also related to many health problems, including major depression, psychosis, obesity, and cardiovascular morbidity (Brindle et al., 2022).

Studies examining the association between peer victimization and HPA axis functioning have produced equivocal results, however (Kliewer et al., 2019). Part of this inconsistency may be due to differences in cortisol measures. Another source of variability may lie in individuals’ long-term trends of peer victimization. The link between peer victimization and cortisol secretion may also vary depending on the level of perceived social support. To address these issues, this study examined the predictive association of adolescents’ peer victimization trajectories from ages 12 to 17 with later cortisol secretion at age 19 and the potential moderating effect of social support in this regard. Because hair cortisol secretion is partly influenced by genetic factors and is also associated with recent and ongoing victimisation experiences in emerging adulthood (Brendgen et al., 2023), these variables were also included in the analyses.

Peer Victimization and Cortisol Secretion

Cortisol secretion ordinarily adheres to a daily rhythm, peaking shortly after waking, followed by a gradual decline throughout the day (Stone et al., 2001). During acute stress, cortisol surges beyond typical levels, facilitating the release of energy stores in reaction to the threat (Gunnar & Quevedo, 2007). However, prolonged elevation of cortisol due to chronic stress may disrupt the HPA system, posing risks to both mental and physical well-being (Miller et al., 2007). Studies on peer victimization and HPA axis functioning have often focused on saliva-based cortisol measures, either in response to stress or assessing diurnal secretion patterns. A review of 20 studies (Kliewer et al., 2019) suggests that peer victimization is linked to either no or heightened cortisol reactivity in children, whereas victimized adolescents show reduced cortisol reactivity to acute social stress. Results for diurnal cortisol patterns were more inconsistent, with some studies reporting flattened slopes in victimized youth and others finding no association. Similarly mixed findings were observed for the cortisol awakening response and total daily cortisol levels.

Saliva-based cortisol measures are highly responsive to situational factors (e.g., time of day, exercise, mood or physical health) and to non-adherence to sampling protocols (Stalder et al., 2017). More recently, researchers have therefore examined hair-based cortisol measures, which are less affected by situational factors and reflect more chronic cortisol secretion patterns (Stalder et al., 2017). Findings from these studies are also inconclusive, however. Thus, a study of 11-year-olds found that—compared to non-involved youth—being a bully-victim (but not being a "pure" victim or a "pure" bully) was significantly associated with hair cortisol concentration (HCC) (Babarro et al., 2023), with bully-victims showing heightened cortisol levels. In another sample of 7 to 14 year-olds, only victimization from multiple sources—but not victimization by peers in particular—was associated with higher HCC (de Azeredo, 2020). Sex moderation effects were either absent or equivocal. Methodological variations regarding victimization measures may have contributed to the observed inconsistencies. However, the peer victimization-cortisol link may also vary depending on the chronicity of such experiences or the level of perceived social support.

The Potential Role of Chronicity of Peer Victimization

While peer victimization generally decreases from childhood to adolescence (Troop-Gordon & Ladd, 2005), considerable inter-individual differences exist. Several studies employed growth-mixture modeling to identify victimization trajectories, with some focusing on primary school (e.g., Biggs et al., 2010) or secondary school (e.g., Barker et al., 2008) and others covering both school periods (e.g., Oncioiu et al., 2020). Among those with non-stable trajectories, more children experience increasing victimization in primary school, whereas the reverse is observed in secondary school. However, most youth experience consistently low or moderate levels of peer victimization, with only a minority (5-15%) enduring chronic high victimization. Persistently high levels of peer victimization are a particular cause for concern, as it could lead to lasting alterations of the stress-response system. Specifically, it has been suggested that chronic stress initially leads to HPA axis hyper-reactivity, gradually transforming into a hypo-reactive system over continued periods of stress (Miller et al., 2007). In line with this notion, a meta-analysis of 83 studies revealed that a high number of adverse life events before age 18 years correlates with blunted cortisol reactivity later in life (Brindle et al., 2022).

Evidence for the role of peer victimization chronicity in inducing altered cortisol secretion comes from a study of 5th and 6th graders. Using daily diaries, participants reported on peer victimization experiences and provided saliva samples five times per day over four days (Adams et al., 2021). Victimization was linked to blunted cortisol secretion, but only in youth indicating very high rates of peer victimization. To explain this non-linear association, the authors highlight the relative stability of high levels of peer victimization and the expected downregulation of the HPA axis in that context. Only one study, to our knowledge, has examined the association between distinct trajectories of peer victimization and subsequent cortisol secretion. Here, self-reported peer victimization from age 6 to 15 years was assessed to predict HCC at age 17 (Ouellet-Morin et al., 2021). Growth-mixture modeling uncovered three victimization trajectories: A lower, gradually decreasing group (32%), a moderate, more rapidly decreasing group (52.7%) and a third group that remained at higher levels of victimization throughout primary school, with a slight decrease thereafter (15.3%). Further analyses revealed higher HCC among boys experiencing either high or low (but not moderate) levels of peer victimization during their school years. No association was evident among girls. While they also reflect a nonlinear association, these findings contrast with those observed by Adams and colleagues (2021). Methodological differences, including cortisol measurement, may again partly explain this discrepancy. Still, more research is needed to understand the role of victimization duration in the link with cortisol.

The Potential Role of Social Support

Social support is thought to buffer against stress-induced alterations of the HPA axis by increasing oxytocin release in the hypothalamus and activating prefrontal cortex areas regulating affects and behaviors (Gunnar & Hostinar, 2015). Evidence for this moderating effect comes from a study showing that frequent peer victimization in college is linked to higher concurrent cortisol secretion, but only under conditions of low maternal support (Brendgen et al., 2023). In another study, examining persistent adverse experiences in participants from low-income neighborhoods, cumulative exposure to violence during adolescence—whether as a victim or witness—correlated with blunted levels of basal salivary cortisol in young adulthood, but moderate to high levels of paternal support mitigated this link (Aiyer et al., 2014).

To date, no study has investigated whether social support moderates the association between chronic peer victimization and cortisol secretion later in life. Examining the distinct buffering effect of parents and friends is especially important in this context. Although parents continue to exert a major influence on their offspring during adolescence (Collins & Laursen, 2004), friends also become an important source of support (Bokhorst et al., 2010). What’s more, peer victimization often occurs in the presence of others (Bauman, 2020), increasing the likelihood that friends witness these events and can potentially intervene or offer other forms of help. There is indeed evidence that parental support may protect (i.e., buffer) victimized youth from negative outcomes of peer victimization such as internalizing problems (e.g., Bowes et al., 2010), which have been associated with altered cortisol secretion in adolescents and young adults (Zajkowska et al., 2022). Still, other studies found either no moderating role of parental support, or a moderating role only for one or the other sex, although some observed that parental support reduces the deleterious impact of peer victimization via a main effect on mental health (e.g., Bilsky et al., 2013; Burke et al., 2017). A review of 37 studies (Schacter et al., 2021) found similarly inconsistent evidence that friends can protect against the negative outcomes of peer victimization on internalizing symptoms. Subjectively reported measures of mental health are not directly comparable to physiological stress-related measures such as cortisol, however. A formal test of a potential buffering effect of support from parents or friends against altered cortisol secretion in victims of peer bullying is thus warranted. Examining this question in relation to peer victimization experienced during the school years is particularly important, because the developing brain structures and functions related to stress regulation in childhood and adolescence are more vulnerable to the effects of stress compared to the mature brain structures of adults (Pervanidou & Chrousos, 2012).

The Present Study

To address these issues, this study examined whether peer victimization during adolescence (i.e., age 12 to 17 years) predicts altered cortisol secretion in emerging adulthood (i.e., age 19) and whether the chronicity of these experiences and the perceived level of social support from parents or friends moderates this association. We concentrated on hair cortisol, as it captures individuals' cortisol secretion in basal and stress contexts over an extended period (Stalder et al., 2017). It was expected that only youth enduring chronically high levels of peer victimization—but not those experiencing moderate or inconsistent victimization—would show altered (i.e., reduced) levels of cortisol secretion at age 19. We also expected a buffering effect of social support from parents and friends. Potential sex moderation was also examined, although inconsistent support for sex differences with respect to the peer victimization-cortisol link or the buffering effect of social support was found in previous studies.

The hypotheses were tested using a longitudinal sample of twins raised together, which allowed accounting for genetic influences in the analyses. Indeed, empirical findings have suggested a notable genetic effect on HCC in children, adolescents and emerging adults, with heritability estimates varying from 33% to 65% and environmental factors unique to each individual explaining the remaining variance (Brendgen et al., 2023; Tucker-Drob et al., 2017). Moreover, because HCC is believed to indicate cumulative stress experienced recently—while also being malleable by earlier stress experiences—, overall stress effects from concurrent victimization at work, in college, or in romantic relationships at age 19 were controlled, as such experiences have also been associated with altered cortisol secretion (Alhalal & Falatah, 2020; Brendgen et al., 2023; Hogh et al., 2012).

Methods

Participants

Participants were part of a sample of 662 monozygotic and dizygotic twin pairs recruited at birth from the Quebec Newborn Twin Registry, which included all twin births in the Province of Quebec, Canada, between 1995 and 1998. All parents of newborns in the registry living in the Greater Montreal area (n = 989 families) were invited and 662 families agreed to participate. Twins were first seen at 5 months of age and then prospectively assessed thereafter. Participants’ sociodemographic characteristics (Supplement Table 1) closely resembled those of families with a 5-month-old infant in a representative population-based sample of single births assessed in 1997/1998 (Jetté & Des Groseilliers, 2000). At the time of recruitment, the twin sample was representative of the population in the Greater Montreal area in terms of ethnicity. Zygosity was assessed using 8–10 highly polymorphous genetic markers. Twin pairs concordant on all genetic markers were considered monozygotic. When genetic material was insufficient or unavailable (43% of cases), zygosity was determined based on physical resemblance questionnaires at 18 months and again at age 9 (Spitz et al., 1996). Comparing both methods in a subsample of 237 same-sex pairs showed a 94% correspondence rate. Participants with valid peer victimization data for at least one out of the five measurement points between ages 12 and 17 years were included to estimate the peer victimization trajectories (n=495 twin pairs). To be included in the main analyses, participants also needed to have valid cortisol data at age 19, yielding a sample of 375 twin pairs (93 female monozygotic (MZ) pairs, 55 male MZ pairs, 60 female dizygotic (DZ) pairs, 49 male DZ pairs and 118 mixed-sex DZ pairs). A comparison of participants included in this study with those who were excluded or dropped out of the study revealed that study participants came from higher income families, were less likely to grow up in a single parent household and were also less likely to have a parent without a high school diploma (all ps<.01).

Procedure

A web-based questionnaire, available in English or French (the two official languages of Canada), was used to collect data on peer victimization and social support. Most twins participated in hair data collection at the lab, but 11% opted for a home visit and 4% self-collected and mailed their hair samples in a prepaid envelope. Hair samples of at least 3 cm by 1 cm were taken from the posterior vertex area by trained assistants or by self-collectors following an instruction guide. No mean differences were found, and a robust correlation (r=0.91, p<.001) existed between cortisol levels from hair collected at home versus in the lab. Data were collected in the spring.

Measures

Peer Victimization

This was assessed at the end of grades 6, 7, 8, 9 and 11 via self-reports on items based on the Social Experiences Questionnaire (Crick & Bigbee, 1998). Two items reflected verbal victimization (i.e., During this school year, how many times has a student at your school .… “called you names or said mean things to you?”, “made fun of you”) and three items reflected relational victimization (“said mean things about you to other children?”, “stopped you from joining a group when you wanted to join?”, “ignored or pretended not to know you?”). To represent physical victimization, we used two SBQ-based items (“pushed, hit or kicked you?”, “Threatened to hurt you?”) along with a third item to measure face-to-face theft/mugging (“force you to give him or her something that belonged to you?”). An item representing cybervictimization (“… threatened you or said mean things about you via e-mail, chat room, or cell phone?”) was also included in the questionnaire. Responses were given on a three-point scale ranging from 0 (never) to 2 (often). Item scores were averaged to a global peer victimization score at each time point (Grade 6 M=.48, SD=.36; Grade 7 M=.36, SD=.33; Grade 8 M=.24, SD=.27; Grade 9 M=.20, SD=.24; Grade 11 M=.17, SD=.23). Observed values ranged from a minimum of 0 to a maximum of 2, skewness ranged from .91 to 1.91, and kurtosis ranged from -0.64 to 4.39 across the different time points. Alphas ranged from .69 to .79. Longitudinal Means and Covariance Structure (LMACS) analysis showed partial metric and scalar invariance of the victimization scale from age 12 through age 17, with only minor and theoretically explicable deviations from invariance (see Supplement Tables 2 through 4). Given that metric and scalar invariance was observed for most indicators, a meaningful identification of differential developmental trends of victimization experiences via Latent Class Growth Mixture Modeling (see below) was considered feasible (Little, 2013).

Social Support

At the end of grades 7, 8, 9 and 11, these measures were assessed—separately regarding support from the mother, father, and best friend—via self-reports using 6 items from the Network of Relationships Inventory (NRI; Furman & Buhrmester, 1992) that measure emotional validation and instrumental support (e.g., “How often do you turn to this person for support with personal problems”, “How often do you depend on this person for help, advice, or sympathy”, “How often does this person praise you for the kind of person you are?”. The NRI has shown excellent reliability for different age groups, including adolescents (Ackermann et al., 2018). Responses were given on a 5-point Likert-type scale ranging from 0 “never” to 4 “always”, and scores were averaged separately for each source of support and each assessment time (alphas ranged from .82 to .89). Additional Hierarchical Confirmatory Factor analysis—performed separately for support from mother, father, and friend at each time point and including a global factor and two subfactors (emotional validation and instrumental support) with three items each—also showed good fit to the present data (RMSEA varied between .024 and .071, CFI between .985 and .998, TLI between .967 and .997 across the different social agents and the different time points). Temporal stability from one measurement time to the next ranged from r=.41 to .65. The scales were averaged across all time points to compute global social support scales (mother M=2.66, SD=.69, Min=.33, Max=4.00, Skew=-.41, Kurtosis=-.01; father M=2.34, SD=.82, Min=0, Max=4.00, Skew=-.44, Kurtosis=.03; friend M=2.59, SD=.76, Min=0, Max=4.00, Skew=-.22, Kurtosis=-.53).

Hair Cortisol (HCC)

Collected at age 19, samples were washed and steroids were extracted according to a previously validated protocol (Ouellet-Morin et al., 2016), with all samples assayed in duplicates and averaged. Nine participants had extremely high scores (the highest outlier was more than 14 SD) and their data were discarded. Remaining outliers (1.6%, n=13) were winsorized to 3 SD and a logarithmic transformation was applied to the cortisol variable (M=.07 µg/dl, SD=.05). Participants were also questioned about potential confounders (i.e., hair care: washing frequency, treatments, coloring; health-related characteristics: cold, flu and allergies, body mass index, sleeping habits, drug and medication use; health problems: diabetes, head injuries, cardiovascular problems). To control for these variables, we computed standardized residuals of HCC to serve as the dependent variables in subsequent analyses (M=.01, SD=1.00, Min=-2.95, Max=3.12, Skew=0.51, Kurtosis=0.52).

Control Variables

At age 19 years, victimization in college, at work and in romantic relationships in the last 12 months was assessed via self-reports. To this end, the previously used SEQ-based items were adapted to reflect victimization experienced in college (9 items) or at the workplace (9 items). For instance, participants were asked how many times during the past 12 months someone in college “Put you down in front of others”, “Said bad things about you or threatened you using email, chat room, cell phone, or social media”, “Physically threatened you”? These items concord with those used in other studies on victimization experiences in college (Chapell et al., 2006) or at work (Brendgen et al., 2019). Responses were given on a three-point scale from 0 (never) to 2 (often). Victimization in romantic relationships was assessed using 14 items derived from the Revised Conflict Tactics Scale (Straus, 2004), which measure psychological and physical violence experiences (e.g., “How many times has your partner…Insulted or sworn at you; Slapped you; Kicked, hit or bit you”). Responses were given on a 5-point Likert scale from 0 (never) to 2 (often). Individual item scores were averaged (alphas ranging from .71 to .89). A global victimization score at age 19 was calculated by averaging the three scale scores (M=.34, SD=.45, Min=0, Max=3.29, Skewness=2.09, Kurtosis=5.87). For those with valid data for victimization in only two contexts (44.9%) or one context (12.7%), the respective valid data were used for the global victimization scale at age 19.

Analyses

Bivariate correlations of the main study variables are presented in Supplement Table 5. Prior to testing the main hypotheses, Latent Class Growth Mixture Modeling (LCGMM) was conducted with peer victimization measures from ages 12 through 17. This method estimates participants’ individual level (or intercept) and change (or slope) of peer victimization over time and tests whether a better fit to the data is obtained when estimating distinct latent groups. LCGMM aims to identify the fewest number of groups while minimizing intra-group differences and maximizing inter-group differences with respect to the intercept and slope. LCGMM also calculates a posterior probability ranging from 0 to 1 for each participant that represents the probability of belonging to each of the identified groups. One- to eight-group models were estimated and compared. Following recommended guidelines (Van De Schoot et al., 2017), selection of the optimal model was based on the Akaike Information Criterion (AIC), the Bayesian Information Criterion (BIC), the sample-size adjusted BIC (aBIC), Entropy, the Vuong–Lo–Mendell–Rubin (VLMR-LRT) and the Lo–Mendell–Rubin adjusted likelihood ratio test (LMRLRT). For AIC, BIC and aBIC, values typically decline with each additional group and the point at which values plateau indicates the optimal number of groups. For the VLMR-LRT and the LMRLRT, a significant p-value indicates that a given model with k classes fits the data better than the model with k-1 classes. The average posterior class probabilities and Entropy were also examined. These two fit indices can vary between 0 and 1, with higher values indicating better classification precision of the model. Final selection criteria also included the proportion of participants in the smallest group, as well the substantive and theoretically meaningful interpretation of the identified profiles. A maximum likelihood estimator was utilized, and missing data (21.6% of data points) were handled with full information maximum-likelihood estimation (FIML). For 375 out of the 495 twin pairs, valid peer victimization data were available from age 12 to 17 years for at least 3 out of the 5 assessment times. In these analyses, each twin was treated as a separate individual and the Robust Huber-White sandwich estimator was used to account for the non-independence of twin observations.

The main hypotheses were tested using biometric modeling (Neale & Cardon, 2013). Based on a comparison of within-pair correlations in MZ twins (who share 100% of their segregating genes by descent) and those in DZ twins (who share approximately only half of their segregating genes), the relative effects of three latent sources of variance in a measured variable such as HCC can be estimated: 1) additive genetic factors (A), 2) shared environmental factors (C), which affect the two twins of a pair in a similar way, and 3) non shared environmental factors (E), which affect the two twins of a pair differently. The most basic analysis involves fitting a two-group biometric model to the data (Supplement Figure 1), with intra-pair correlations of the latent additive genetic factor (A) fixed to 1.0 for MZ twins and to 0.5 for DZ twins. Within-pair correlations of the latent shared environmental factor (C) are fixed to 1.0 and within-pair correlations of the latent nonshared environmental factor (E) are fixed to 0.0 for both MZ and DZ twins. The factor loadings a, c, and e are fixed to be equal across the two twins in a pair and across MZ and DZ twins and estimate the relative effects of A, C, and E on the measured variable (with E also including measurement error). A higher within-pair correlation in MZ than DZ twins suggests at least some genetic effects. Model fit was assessed using fit indices such as AIC, BIC, CFI, Root Mean Square Error of Approximation (RMSEA) and χ2. More parsimonious follow-up models were also run and compared to the full ACE model using nested χ2 difference tests.

Next, the best fitting biometric model was extended to include the measured predictors (Engelhardt et al., 2019). Model 1 tested only main effects and included victimization at age 19, two dummy variables representing the different peer victimization trajectory groups identified in the LCGMM (using the lowest group as reference group) and the three social support variables. Subsequent sub-steps (Models 2a, 2b, and 2c) added the interaction terms between social support (either from the mother, the father, or the best friend) and each of the peer victimization dummy variables. Interactions were tested in separate models to avoid multicollinearity. We also examined a) the combined portion of the total variance of HCC explained by the predictors, and b) the reduction of relative variance explained by the remaining latent environmental factors in the best fitting ACE model with measured predictors, compared to the relative variance explained by latent environmental factors in the ACE model without predictors. These comparisons gauge to what extent the effects of measured predictors represent “true” environmental effects on HCC.

All biometric analyses were run as five-group models that separated female MZ pairs, male MZ pairs, female DZ pairs, male DZ pairs, and mixed-sex DZ pairs, with means and variances free to vary between the sexes (Neale & Cardon, 2013). Potential sex moderation with respect to the parameters described above was examined using chi-square difference tests between models with and without equality constraints across the sexes. Robust full information maximum likelihood (FIML-R) was used to handle missing data in the social support variables (4.6% of overall data points). Little’s test showed that missingness was likely completely at random, χ2(42) = 32.99, p = .119. All analyses were conducted using Mplus Version 8 (Muthén & Muthén, 1998-2017).

Results

Latent Class Growth Mixture Modeling

The different model fit indices and the model-estimated trajectories of peer victimization between ages 12 and 17 from the final retained model solution (i.e., three-group model) are depicted in Figure 1. The three-group model fit the data better than a model with fewer groups and as well as models with more groups. Inspection of the scree-plot (Supplement Figure 2) revealed that all ICs showed a steep decrease up to the three-group model, with only minimal further decrease thereafter. In addition, models with four or more groups included at least one group with less than 5% of participants. While entropy for the three-group model was just below the .70 threshold (i.e., .69), the average posterior assignment probability for each of the groups ranged from .84 to.87, indicating acceptable classification quality (Nagin, 2010).

As can be seen in Figure 1, peer victimization declined over time for all three groups. However, the groups differed on their level of peer victimization at age 12 (grade 6) and these differences were essentially maintained after the school transition and until age 17 (end of secondary school). The Low group (33.0%) experienced very little peer victimization during adolescence (Intercept=0.20, p<.001; Linear Slope=-2.08, p=.001; Quadratic Slope=2.49, p<.001). However, most participants (55.6%) belonged to the Moderate group, who experienced somewhat more peer victimization at age 12 than the Low group, with a decline to similarly low levels in adolescence (Intercept=0.51, p<.001; Linear Slope=-1.58, p=.001; Quadratic Slope=1.68, p<.001). The High group (11.4%) experienced the highest level of peer victimization already at age 12, and although the frequency declined thereafter, their victimization levels remained higher until age 17 (Intercept=1.03, p<.001; Linear Slope=-1.88, p=.001; Quadratic Slope=1.27, p=.184).

Biometric Modeling

The HCC intra-pair correlation was higher for MZ twins than DZ twins, suggesting at least some genetic influence on HCC (rs = MZ Girls .38 / DZ Girls .03; MZ Boys .45 / DZ Boys .26; Mixed-sex DZ .02). The biometric model of HCC without additional predictors (Table 1) showed that an AE model best fit the data for both sexes, with genetic factors explaining 34% of HCC levels in girls and 40% in boys and the remaining variance explained by unique environmental influences (66% in girls and 60% in boys). These relative genetic and nonshared environmental influences on HCC did not differ between the sexes, as indicated by the superior fit indices of models with (Model 2 and 4) than models without (Models 1 and 3) scalar correspondence of the respective factor loadings. However, to allow testing potential sex moderation in the subsequent extended biometric models, the AE model without cross-sex equality constraints (model 3) was used.

The results of the extended biometric model without interaction terms (Table 2, Model 1) revealed that the High or Moderate versus the Low peer victimization trajectory were not associated with HCC at age 19. The level of social support from the mother, father or best friend was also unrelated to HCC at age 19. No sex moderation of the regression coefficients in Model 1 was observed (interaction ps ranging from .099 to .941). However, the addition of interaction terms with social support (Table 2, Models 2a–c) revealed that the predictive effect of being in the High versus Low peer victimization trajectory on later HCC was significantly moderated by the level of support from a best friend during adolescence (b=0.35, p=.034, Bootstrap 95% CI 0.01/0.64). Examining the region of significance showed that, compared to the Low peer victimization trajectory, the High peer victimization trajectory was associated with blunted (i.e., lower) HCC levels only for youth who reported low support from friends (i.e., -1 SD: b=-.34, 95% CI -0.01/-0.68). At higher levels of support, this association was no longer significant. Support from the mother or the father did not moderate the victimization-HCC link. No sex moderation of the coefficients in Models 2a-c was observed (interaction ps ranging from .114 to .983).

Inspection of the total variance and the remaining latent A and E variances in Model 2c revealed that, together, the predictors explained 15% of the HCC variance in girls and 19.8% in boys. By adding these predictors to the model, the portion of the total variance of HCC explained by latent unique environmental factors E was reduced to 57% in girls and 47% in boys, compared to 66% in girls and 60% in boys in the biometric model without predictors. This reduction suggests that the observed predictive associations represent, at least in part, “true” environmental effects on HCC.

Discussion

This study examined whether distinct trajectories of peer victimization from ages 12 to 17 are differentially associated with HCC at age 19 and whether social support from parents or friends moderates this association. Of specific interest were youth experiencing persistently high levels of peer victimization, as chronic stress may lead to lasting alterations of the HPA axis (Miller et al., 2007). These questions were examined while controlling for genetic effects and effects associated with concurrent victimization in college, at work and in romantic relationships at age 19.

Concordant with previous research (e.g., Tucker-Drob et al., 2017), genetic factors explained a sizable portion of interindividual differences in HCC, with the remainder explained by environmental experiences unique to each individual. The results further indicate that part of these environmental influences may stem from being chronically bullied by peers during the school years. Thus, growth mixture modeling revealed three victimization trajectories—labeled low, moderate and high—which are largely consistent with adolescent victimization patterns found in other samples (e.g., Ouellet-Morin et al., 2021). In line with expectations, results indicated that the High trajectory group—but not the Moderate trajectory group—showed altered HPA axis functioning compared to the Low trajectory group. Specifically, their cortisol secretion showed a blunted pattern, in line with a downward regulation of the neurophysiological stress response system under prolonged exposure to stress (Miller et al., 2007). However, for both female and male victims, this blunted pattern only emerged when friends offered very little help and vanished when more social support by a friend was reported. Adapting to chronic stress with reduced cortisol secretion may provide short-term benefits, because sustained elevated cortisol concentrations may lead to tissue damage and dysregulation of biological systems (Brindle et al., 2022; Miller et al., 2007). Nevertheless, this adaptation may turn maladaptive in the long term, as cortisol hyposecretion has been associated with cognitive decline, obesity, cardiovascular issues and major depression as well as the development of various diseases such as rheumatoid arthritis, chronic fatigue syndrome, and posttraumatic stress disorder (PTSD) (Brindle et al., 2022). In adolescents experiencing moderate levels of peer victimization, periods of intense stress may not be frequent enough to induce lasting alterations in cortisol secretion. It is also possible that youth following a moderate trajectory of peer victimization have greater access to sources of support—notably friends—that can mitigate long-term sequelae than chronic victims. Indeed, compared to their agemates, highly victimized adolescents have fewer reciprocal friends and are also less often nominated as desired friends by their peers (Scholte et al., 2009). Because friends may vary in their ability to offer useful advice and protection against further victimization, having a larger peer network to choose from as potential sources of support may thus be advantageous.

The observed buffering effect of friends’ social support highlights the importance of friends in fostering resilience against stress related to peer victimization. This finding concords with studies examining other outcomes of peer victimization, such as depressive symptoms, academic problems or substance use (Schacter et al., 2021). Interestingly, very small levels of perceived help by a friend may be sufficient, as only victims with extremely low levels of support (i.e., at least 1 SD below the mean) during adolescence showed blunted cortisol levels in emerging adulthood. Their friends might lack the necessary skills or social standing to provide sufficient emotional support or to successfully intervene during bullying episodes. Indeed, friends of bullying victims are often victimized themselves (Scholte et al., 2009). Rather than providing constructive support, some of these friends may engage in co-rumination about their common plight and thus aggravate their stress even further (Guarneri-White et al., 2015). There is also evidence that some adolescents who are frequently bullied by their peers become the target of further harassment by their close friends (Vucetic et al., 2021). In either case, chronically victimized youth whose friends cannot or will not come to their aid might need to seek support from other sources.

Despite the rising importance of friends, parents generally continue to be the main providers of instrumental and emotional support (Bokhorst et al., 2010). Nevertheless, the present results suggest that many parents are not able to shield their offspring from potential negative consequences of chronic peer victimization, at least regarding HPA axis functioning. One explanation why mitigating effects were only found for friends may lie in the fact that bullying incidences often take place where other peers can witness the event (e.g., in the school yard) (Bauman, 2020). In contrast to parents, friends are thus likely to observe many of these occurrences and thus provide immediate assistance. A study of 1266 bullied children and adolescents also found that only 34% told someone at home and the likelihood of disclosure decreased in older compared to younger students (Blomqvist et al., 2020). Victims may hesitate to share their plight with parents because they worry that parents’ actions exacerbate the situation (Newman & Murray, 2005). Unfortunately, this reluctance may leave some bullying victims without any help—especially if they do not receive any support from friends—, with significant consequences for their health-related adjustment. However, research suggests that some chronically victimized youth are open to discussing their problems with school personnel, albeit not necessarily with their teachers (Blomqvist et al., 2020). School social workers or school psychologists may thus play a crucial role especially for bullying victims lacking other sources of support.

Strengths and Limitations

Strengths of this study include its genetically informed and longitudinal design as well as the inclusion of different sources of support. Nevertheless, certain limitations need to be acknowledged. First, no earlier hair cortisol measures were available to directly measure changes in HPA axis functioning. Second, the self-reports of peer victimization and social support may have been subject to perceptual bias and have influenced results due to shared method variance. Still, it has been argued that it is individuals' interpretation of events that affects their cognitive, emotional and—as a result—physiological responses (Cole et al., 2010). Third, although we controlled for illness- and lifestyle-related confounders as well as for concurrent victimization in different contexts, other concurrent stressors that may affect HCC were not assessed, e.g., academic stress or negative life events such as parental divorce (Karlén et al., 2011; Stetler & Guinn, 2020).

Regarding the generalizability of results, the reliance on a predominantly White sample and the loss of more socially disadvantaged participants due to attrition may have limited the variance of victimization and of social support, potentially underestimating the observed associations. Indeed, racial/ethnic minority adolescents are more frequently victimized than others (Fuentes Cabrera et al., 2019). Moreover, non-Whites and lower-SES adults tend to receive less advice and help from others than Whites and higher-SES individuals (Schafer & Vargas, 2016), which may affect their ability to offer support to their victimized children. Future studies should thus replicate the present findings while including measures of other potential stressors and a more socio-ethnically diverse sample.

Conclusion

The present study shows that chronic peer victimization during adolescence may lead to a blunted pattern of cortisol secretion in emerging adulthood, but that this can be offset by even a modest level of social support from a friend. This finding is important because reduced cortisol secretion may signal cumulative “wear‐and‐tear” on the body, possibly causing health problems in the long run. The protective effect of friendship support is encouraging, indicating potential avenues for remediation when prevention of victimization fails. By the same token, the finding of parents’ inefficacy in this regard is worrisome. Parents may need to seek help from professionals such as school psychologists or social workers for additional support and prevent potential long-term health complications in victims of peer bullying.

References

Ackermann, K., Martinelli, A., Bernhard, A., Ueno, K., Freitag, C. M., Büttner, G., Schmiedek, F., & Schwenck, C. (2018). Validation of the Network of Relationship Inventory in Female and Male Adolescents. European Journal of Psychological Assessment, 36(2), 220-228. doi.org/10.1027/1015-5759/a000508

Adams, R. E., Santo, J. B., & Bukowski, W. M. (2021). Indirect effects of HPA axis dysregulation in the association between peer victimization and depressed affect during early adolescence. Psychoneuroendocrinology, 132, 105356. doi.org/10.1016/j.psyneuen.2021.105356

Aiyer, S. M., Heinze, J. E., Miller, A. L., Stoddard, S. A., & Zimmerman, M. A. (2014). Exposure to Violence Predicting Cortisol Response During Adolescence and Early Adulthood: Understanding Moderating Factors. Journal of Youth and Adolescence, 43(7), 1066-1079. doi.org/10.1007/s10964-014-0097-8

Alhalal, E., & Falatah, R. (2020). Intimate partner violence and hair cortisol concentration: A biomarker for HPA axis function. Psychoneuroendocrinology, 122, 104897. doi.org/10.1016/j.psyneuen.2020.104897

Babarro, I., Ibarluzea, J., Theodorsson, E., Fano, E., Lebeña, A., Guxens, M., Sunyer, J., & Andiarena, A. (2023). Hair cortisol as a biomarker of chronic stress in preadolescents: influence of school context and bullying. Child Neuropsychology, 29(5), 742-759. doi.org/10.1080/09297049.2022.2115991

Barker, E. D., Arseneault, L., Brendgen, M., Fontaine, N., & Maughan, B. (2008). Joint Development of Bullying and Victimization in Adolescence: Relations to Delinquency and Self-Harm. Journal of the American Academy of Child & Adolescent Psychiatry, 47(9), 1030-1038. dx.doi.org/10.1097/CHI.ObO13e31817eec98

Bauman, S. (2020). Experiences of adolescent witnesses to peer victimization: The bystander effect. Journal of School Psychology, 80, 1-14. doi.org/10.1016/j.jsp.2020.03.002

Biggs, B. K., Vernberg, E., Little, T. D., Dill, E. J., Fonagy, P., & Twemlow, S. W. (2010). Peer victimization trajectories and their association with children’s affect in late elementary school. International Journal of Behavioral Development, 34(2), 136-146. doi.org/10.1177/0165025409348560

Bilsky, S. A., Cole, D. A., Dukewich, T. L., Martin, N. C., Sinclair, K. R., Tran, C. V., Roeder, K. M., Felton, J. W., Tilghman-Osborne, C., Weitlauf, A. S., & Maxwell, M. A. (2013). Does supportive parenting mitigate the longitudinal effects of peer victimization on depressive thoughts and symptoms in children? Journal of Abnormal Psychology, 122, 406-419. doi.org/10.1037/a0032501

Blomqvist, K., Saarento-Zaprudin, S., & Salmivalli, C. (2020). Telling adults about one's plight as a victim of bullying: Student- and context-related factors predicting disclosure. Scandinavian Journal of Psychology, 61(1), 151-159. doi.org/10.1111/sjop.12521

Bokhorst, C. L., Sumter, S. R., & Westenberg, P. M. (2010). Social Support from Parents, Friends, Classmates, and Teachers in Children and Adolescents Aged 9 to 18 Years: Who Is Perceived as Most Supportive? [https://doi.org/10.1111/j.1467-9507.2009.00540.x]. Social Development, 19(2), 417-426. doi.org/10.1111/j.1467-9507.2009.00540.x

Bowes, L., Maughan, B., Caspi, A., Moffitt, T. E., & Arseneault, L. (2010). Families promote emotional and behavioural resilience to bullying: evidence of an environmental effect. Journal of Child Psychology and Psychiatry, 51(7), 809-817. doi.org/10.1111/j.1469-7610.2010.02216.x

Brendgen, M., Ouellet-Morin, I., Cantave, C. Y., Vitaro, F., Dionne, G., & Boivin, M. (2023). Link Between Peer Victimization in College and Cortisol Secretion: Roles of Genetic Vulnerabilities and Social Support. Journal of Youth and Adolescence, 52(1), 76-90. doi.org/10.1007/s10964-022-01687-1

Brendgen, M., Poulin, F., & Denault, A.-S. (2019). Peer victimization in school and mental and physical health problems in young adulthood: Examining the role of revictimization at the workplace [doi:10.1037/dev0000771]. Developmental Psychology, 55.(10), 2219–2230. doi.org/10.1037/dev0000771

Brindle, R. C., Pearson, A., & Ginty, A. T. (2022). Adverse childhood experiences (ACEs) relate to blunted cardiovascular and cortisol reactivity to acute laboratory stress: A systematic review and meta-analysis. Neuroscience and Biobehavioral Reviews, 134, 104530. doi.org/10.1016/j.neubiorev.2022.104530

Burke, T., Sticca, F., & Perren, S. (2017). Everything’s gonna be alright! The longitudinal interplay among social support, peer victimization, and depressive symptoms. Journal of Youth and Adolescence, 46. doi/10.1007/s10964-017-0653-0

Chapell, M. S., Hasselman, S. L., Kitchin, T., & Lomon, S. N. (2006). Bullying in elementary school, high school, and college. Adolescence, 41(164), 633-648.

Cole, D. A., Maxwell, M. A., Dukewich, T. L., & Yosick, R. (2010). Targeted Peer Victimization and the Construction of Positive and Negative Self-Cognitions: Connections to Depressive Symptoms in Children. Journal of Clinical Child & Adolescent Psychology, 39(3), 421-435. doi.org/10.1080/15374411003691776

Collins, W. A., & Laursen, B. (2004). Parent-adolescent relationships and influences. In R. M. Lerner & L. D. Steinberg (Eds.), Handbook of adolescent psychology (2 ed., pp. 331–361). Wiley.

Crick, N. R., & Bigbee, M. A. (1998). Relational and overt forms of peer victimization: A multiinformant approach. Journal of Consulting and Clinical Psychology, 66(2), 337-347. http://www.scopus.com/inward/record.url?eid=2-s2.0-0031959225&partnerID=40&md5=6120b2fcbfad06a3d53971557548439a

de Azeredo, L. A. (2020). Hair cortisol levels and mental health problems in children and adolescents exposed to victimization. Stress, 23(5), 546–555. doi.org/10.1080/10253890.2019.1690448

Engelhardt, L. E., Church, J. A., Paige Harden, K., & Tucker-Drob, E. M. (2019). Accounting for the shared environment in cognitive abilities and academic achievement with measured socioecological contexts. Developmental Science, 22(1), e12699. doi.org/10.1111/desc.12699

Furman, W., & Buhrmester, D. (1992). Age and Sex Differences in Perceptions of Networks of Personal Relationships. Child Development, 63, 103-115. doi.org/10.1111/j.1467-8624.1992.tb03599.x

Guarneri-White, M. E., Jensen-Campbell, L. A., & Knack, J. M. (2015). Is co-ruminating with friends related to health problems in victimized adolescents? Journal of Adolescence, 39, 15-26. doi.org/10.1016/j.adolescence.2014.11.004

Gunnar, M., & Quevedo, K. (2007). The Neurobiology of Stress and Development. Annual review of psychology, 58(1), 145-173. doi:10.1146/annurev.psych.58.110405.085605

Gunnar, M. R., & Hostinar, C. E. (2015). The social buffering of the hypothalamic–pituitary–adrenocortical axis in humans: Developmental and experiential determinants. Social Neuroscience, 10(5), 479-488. doi.org/10.1080/17470919.2015.1070747

Hogh, A., Hansen, Å. M., Mikkelsen, E. G., & Persson, R. (2012). Exposure to negative acts at work, psychological stress reactions and physiological stress response. Journal of Psychosomatic Research, 73(1), 47-52. doi.org/10.1016/j.jpsychores.2012.04.004

Jetté, M., & Des Groseilliers, L. (2000). Longitudinal study of child development in Québec (ELDEQ 1998–2000). Survey description and methodology, 1(1). Retrieved Retrieved December 1, 2005, from http://www.stat.gouv.qc.ca/publications/sante/bebe_no1_pdf_an.htm

Karlén, J., Ludvigsson, J., Frostell, A., Theodorsson, E., & Faresjö, T. (2011). Cortisol in hair measured in young adults - a biomarker of major life stressors? BMC Clinical Pathology, 11(1), 12. doi.org/10.1186/1472-6890-11-12

Kliewer, W., Sosnowski, D. W., Noh, H., McGuire, K., & Wright, A. W. (2019). Peer victimization and cortisol production in children and adolescents: A systematic review. Journal of Applied Biobehavioral Research, 24(4). doi.org/10.1111/jabr.12172

Knezevic, E., Nenic, K., Milanovic, V., & Knezevic, N. N. (2023). The Role of Cortisol in Chronic Stress, Neurodegenerative Diseases, and Psychological Disorders. Cells, 12(23). doi.org/10.3390/cells12232726

Miller, G. E., Chen, E., & Zhou, E. S. (2007). If it goes up, must it come down? Chronic stress and the hypothalamic-pituitary-adrenocortical axis in humans. Psychological Bulletin, 133(1), 25-45. doi.org/10.1037/0033-2909.133.1.25

Moore, S. E., Norman, R. E., Suetani, S., Thomas, H. J., Sly, P. D., & Scott, J. G. (2017). Consequences of bullying victimization in childhood and adolescence: A systematic review and meta-analysis. World Journal of Psychiatry, 7(1), 60-76. doi.org/10.5498/wjp.v7.i1.60

Muthén, L. K., & Muthén, B. O. (1998-2017). Mplus user’s guide. Version 8. Muthén & Muthén.

Nagin, D. S. (2010). Group-Based Trajectory Modeling: An Overview. In A. R. Piquero & D. Weisburd (Eds.), Handbook of Quantitative Criminology (pp. 53-67). Springer New York. doi.org/10.1007/978-0-387-77650-7_4

Neale, M. C., & Cardon, L. R. (2013). Methodology for genetic studies of twins and families (Vol. 67). Springer Science & Business Media.

Newman, R. S., & Murray, B. J. (2005). How Students and Teachers View the Seriousness of Peer Harassment: When Is It Appropriate to Seek Help? Journal of Educational Psychology, 97(3), 347-365. doi/10.1037/0022-0663.97.3.347

Oncioiu, S. I., Orri, M., Boivin, M., Geoffroy, M.-C., Arseneault, L., Brendgen, M., Vitaro, F., Navarro, M. C., Galéra, C., Tremblay, R. E., & Côté, S. M. (2020). Early Childhood Factors Associated With Peer Victimization Trajectories From 6 to 17 Years of Age. Pediatrics, e20192654. doi.org/10.1542/peds.2019-2654

Ouellet-Morin, I., Cantave, C., Paquin, S., Geoffroy, M.-C., Brendgen, M., Vitaro, F., Tremblay, R., Boivin, M., Lupien, S., & Côté, S. (2021). Associations between developmental trajectories of peer victimization, hair cortisol, and depressive symptoms: a longitudinal study. Journal of Child Psychology and Psychiatry, 62(1), 19-27. doi.org/10.1111/jcpp.13228

Ouellet-Morin, I., Laurin, M., Robitaille, M.-P., Brendgen, M., Lupien, S. J., Boivin, M., & Vitaro, F. (2016). Validation of an adapted procedure to collect hair for cortisol determination in adolescents. Psychoneuroendocrinology, 70, 58-62. doi.org/10.1016/j.psyneuen.2016.05.002

Schacter, H. L. (2021). Effects of Peer Victimization on Child and Adolescent Physical Health. Pediatrics, 147(1). doi.org/10.1542/peds.2020-003434

Schacter, H. L., Lessard, L. M., Kiperman, S., Bakth, F., Ehrhardt, A., & Uganski, J. (2021). Can friendships protect against the health consequences of peer victimization in adolescence? A systematic review. School Mental Health, 1-24. doi.org/10.1007/s12310-021-09417-x

Scholte, R. H. J., Overbeek, G., ten Brink, G., Rommes, E., de Kemp, R. A. T., Goossens, L., & Engels, R. C. M. E. (2009). The Significance of Reciprocal and Unilateral Friendships for Peer Victimization in Adolescence. Journal of Youth and Adolescence, 38(1), 89-100. doi.org/10.1007/s10964-008-9287-6

Spitz, E., Moutier, R., Reed, T., Busnel, M., Marchaland, C., Roubertoux, P., & Carlier, M. (1996). Comparative diagnoses of twin zygosity by SSLP variant analysis, questionnaire, and dermatoglyphic analysis. Behavior Genetics, 26(1), 55-63. doi.org/10.1007/BF02361159

Stalder, T., Steudte-Schmiedgen, S., Alexander, N., Klucken, T., Vater, A., Wichmann, S., Kirschbaum, C., & Miller, R. (2017). Stress-related and basic determinants of hair cortisol in humans: A meta-analysis. Psychoneuroendocrinology, 77, 261-274. doi.org/10.1016/j.psyneuen.2016.12.017

Stetler, C. A., & Guinn, V. (2020). Cumulative cortisol exposure increases during the academic term: Links to performance-related and social-evaluative stressors. Psychoneuroendocrinology, 114, 104584. doi.org/10.1016/j.psyneuen.2020.104584

Straus, M. A. (2004). Cross-Cultural Reliability and Validity of the Revised Conflict Tactics Scales: A Study of University Student Dating Couples in 17 Nations. Cross-Cultural Research, 38(4), 407-432. doi.org/10.1177/1069397104269543

Troop-Gordon, W., & Ladd, G. W. (2005). Trajectories of Peer Victimization and Perceptions of the Self and Schoolmates: Precursors to Internalizing and Externalizing Problems. Child Development, 76(5), 1072-1091. doi.org/10.1111/j.1467-8624.2005.00898.x

Tucker-Drob, E. M., Grotzinger, A. D., Briley, D. A., Engelhardt, L. E., Mann, F. D., Patterson, M., Kirschbaum, C., Adam, E. K., Church, J. A., Tackett, J. L., & Harden, K. P. (2017). Genetic influences on hormonal markers of chronic hypothalamic–pituitary–adrenal function in human hair. Psychological Medicine, 47(8), 1389-1401. doi.org/10.1017/s0033291716003068

Vaillancourt, T., Hymel, S., & McDougall, P. (2013). The Biological Underpinnings of Peer Victimization: Understanding Why and How the Effects of Bullying Can Last a Lifetime. Theory Into Practice, 52(4), 241-248. doi.org/10.1080/00405841.2013.829726

Van De Schoot, R., Sijbrandij, M., Winter, S. D., Depaoli, S., & Vermunt, J. K. (2017). The GRoLTS-checklist: guidelines for reporting on latent trajectory studies. Structural Equation Modeling: A Multidisciplinary Journal, 24(3), 451-467. doi.org/10.1080/10705511.2016.1247646

Vucetic, M., Brendgen, M., Vitaro, F., Dionne, G., & Boivin, M. (2021). Victimization by Friends and Victimization by Other Peers: Common Risk Factors or Mutual Influence? Journal of Youth and Adolescence, 50(3), 563-578. doi.org/10.1007/s10964-020-01270-6

Zajkowska, Z., Gullett, N., Walsh, A., Zonca, V., Pedersen, G. A., Souza, L., Kieling, C., Fisher, H. L., Kohrt, B. A., & Mondelli, V. (2022). Cortisol and development of depression in adolescence and young adulthood – a systematic review and meta-analysis. Psychoneuroendocrinology, 136, 105625. doi.org/10.1016/j.psyneuen.2021.105625

TABLES

Table 1: Biometric Models Without Measured Predictors

Fit Indices

LL

χ2(df)

p

AIC

BIC

RMSEA

TLI

1. ACE(G) – ACE(B)

-880.21

11.03 (9)

.274

1776.42

1807.83

0.055

.0.93

2. ACE Scalar

-881.07

12.75 (11)

.310

1774.14

1797.70

0.046

0.95

3. AE(G) – AE(B)

4. AE Scalar

-881.01

-881.07

12.63 (11)

12.75 (12)

.318

.388

1774.02

1772.14

1797.58

1791.77

0.045

0.029

0.95

0.98

Coefficient Estimates

A

[95% CI]

C

[95% CI]

E

[95% CI]

%A

%C

%E

Girls

1. ACE(G) – ACE(B)

0.60

[.31/.76]

-0.20

[-.67/.41]

0.85

[.72/.98]

32

4

65

2. ACE Scalar

0.63

[.47/.72]

0

[.00/.00]

.85

[.74/.97]

35

0

65

3. AE(G) – AE(B)

.62

[.42/.77]

--

.87

[.73/.99]

34

--

66

4. AE Scalar

.63

[.49/.78]

--

.85

[.74/.97]

35

--

65

Boys

1. ACE(G) – ACE(B)

0.41

[-.48/.69]

0.40

[-.11/.68]

0.70

[.55/.83]

21

19

60

2. ACE Scalar

.54

0

.72

35

0

65

3. AE(G) – AE(B)

.57

[36/.72]

--

.70

[56/.85]

40

--

60

4. AE Scalar

.54

--

.72

35

--

65

Note. A= latent genetic effects, C = latent shared environmental effects, E = latent nonshared environmental effects. LL = loglikelihood, df = degrees of freedom, AIC = Akaike Information Criterion, BIC= Baysian Information Criterion, RMSEA = Root Mean Square Error of Approximation, TLI = Tucker-Lewis Index. G = Girls, B = Boys. 95% CI = 95% Bootstrap confidence intervals. Models with (G) and (B) indication have factor loadings freely estimated across sex; Scalar models are constrained to scalar correspondence of factor loadings across the sexes (i.e., resulting in equivalent relative ACE effects across sex). Scalar coefficient is k = .85 [.74/.96] for models 2 and 4. Best fitting model is bolded.

Table 2: Extended Biometric Models: Regression Results

Model

Measured Predictor Variable

Estimate (SE)

95%CI

p

1

Victimization Age 19 (Control)

.12 (.11)

-.09/.32

.244

High vs Low Peer Victimization 12-17

-.18 (.15)

-.47/.11

.216

Moderate vs Low Peer Victimization 12-17

-.03 (.09)

-.20/.14

.756

Mother's Support

.12 (.08)

-.03/.28

.131

Father's Support

-.06 (.07)

-.20/.07

.373

Friend's Support

-.02 (.07)

-.16/.11

.756

2a

Mother’s Support X

.11 (.11)

-.01/.34

.317

Moderate vs Low Peer Victimization 12-17

Mother's Support X
High vs Low Peer Victimization 12-17

.03 (.20)

-.35/.46

.893

2b

Father’s Support X

.01 (.07)

-.14/.15

.902

Moderate vs Low Peer Victimization 12-17

Father's Support X
High vs Low Peer Victimization 12-17

.20 (.14)

-.10/.47

.166

2c

Friend’s Support X

.07 (.12)

-.15/.31

.582

Moderate vs Low Peer Victimization 12-17

Friend's Support X
High vs Low Peer Victimization 12-17

.35 (.16)

.01/.64

.034

Note. SE = Bootstrap Standard error. 95% CI = 95% Bootstrap confidence intervals. Since estimates showed no significant sex moderation, results from models with equality constraints are shown for parsimony. Significant associations are bolded.

Figures

Figure 1: Developmental Trajectories of Peer Victimization Between Ages 12 and 17 Based on the Three-group Model and Comparative Model Fit Statistics

A graph of the number of patients with age Description automatically generated with medium confidence

Class

LL

AIC

BIC

aBIC

VLM-LRT p

LMR-LRT p

Entropy

1

-2156.08

4328.16

4367.40

4341.99

--

--

--

2

-1843.58

3711.17

3770.04

4341.99

0.005

0.006

0.67

3

-1725.82

3483.65

3562.14

3511.32

0.015

0.017

0.69

4

-1693.48

3426.97

3525.08

3461.56

0.058

0.062

0.73

5

-1671.92

3391.85

3509.58

3433.36

0.116

0.124

0.68

6

-1655.61

3367.22

3504.58

3415.65

0.402

0.410

0.71

7

-1642.46

3348.93

3505.91

3404.28

0.026

0.029

0.73

8

-1635.35

3342.69

3519.30

3404.96

0.031

0.033

0.73

Note. Best model bolded. LL log-likelihood, AIC Akaike information criteria, BIC Bayesian information criterion, aBIC sample size-adjusted BIC, VLM-LRT Vuong–Lo–Mendell–Rubin likelihood ratio test, LMR-LRT Lo–Mendell–Rubin adjusted likelihood ratio test.

* p < .05, ** p < .01 (two-tailed).

Comments
0
comment
No comments here
Why not start the discussion?