Environments characterized by negative social relationships (e.g., hostility, neglect, competition) capture antisocial or adverse contexts of development. These antisocial or adverse environments in the home, school, and community can serve as stressors that challenge adolescents and impact their social development (McBride Murry et al., Reference McBride Murry, Berkel, Gaylord‐Harden, Copeland‐Linder and Nation2011). Adolescents may learn to adapt their behavior to match that seen in their broader environment in order to obtain desirable resources (Leventhal & Brooks-Gunn, Reference Leventhal and Brooks-Gunn2000). In particular, stressors related to socially adaptive outcomes such as popularity, social power, and access to dating partners within the peer group appear to be important for most adolescents as they become increasingly motivated to interact with their peer group (LaFontana & Cillessen, Reference LaFontana and Cillessen2010; Volk et al., Reference Volk, Dane, Marini and Vaillancourt2015). Adolescents may employ various strategies to obtain these social outcomes, with some youth using prosocial strategies and other youth using coercive or antisocial strategies whose success may depend on perceptions of broader environmental contexts (Bronfenbrenner & Morris, Reference Bronfenbrenner, Morris, Damon and Lerner1998; Hawley, Reference Hawley2003; Lenzi et al., Reference Lenzi, Vieno, Perkins, Pastore, Santinello and Mazzardis2012). The conscious and nonconscious developmental “decision” of which strategy to use under adverse conditions may depend, at least in part, on individual differences such as personality (Ellis et al., Reference Ellis, Del Giudice, Dishion, Figueredo, Gray, Griskevicius, Hawley, Jacobs, James, Volk and Wilson2012; Volk et al., Reference Volk, Provenzano, Farrell, Dane and Shulman2021).
More specifically, some researchers have hypothesized that psychopathy may be, at least in part, an evolved adaptive and/or learned response for coercively and deceptively securing socially adaptive outcomes – which may consist of evolutionarily relevant resources – in antisocial or adverse environments (Book et al., Reference Book, Methot-Jones, Blais, Hosker-Field, Volk, Visser, Gauthier, Holden and D’Agata2019; Patch & Figueredo, Reference Patch and Figueredo2017; Ribeiro da Silva et al., Reference Ribeiro da Silva, Rijo and Salekin2015). Such a possibility does not require that psychopathic traits are accompanied by psychological well-being or that they be culturally valued (Jurjako, Reference Jurjako2019). Instead, psychopathic traits can include both harmful societal costs and individual benefits as youth tradeoff the various strategies available to them (Ellis et al., Reference Ellis, Del Giudice, Dishion, Figueredo, Gray, Griskevicius, Hawley, Jacobs, James, Volk and Wilson2012). To date, however, there is no evidence specifically linking adverse environments, psychopathic traits, and adaptive outcomes among adolescents, although there is some evidence that such links might exist in adults (e.g., Međedović, Reference Međedović2019). If psychopathy is indeed associated with coercive strategies to obtain resources during adulthood, then such associations are likely evident earlier in development during adolescence, a heightened time for social competition. Therefore, our goal is to examine whether adolescent psychopathic traits are linked with both perceptions of adverse enviroments and socially adaptive outcomes such as social power and dating outcomes.
Psychopathic traits
Psychopathic traits are characterized by interpersonal (e.g., manipulation, grandiosity), affective (e.g., callousness, unemotionality), and behavioral (e.g., impulsivity, sensation seeking) dimensions or facets that collectively form a cohesive superordinate construct that resembles prototypical psychopathy (Ivanova-Serokhvostova et al., Reference Ivanova-Serokhvostova, Molinuevo, González, Hilterman, Pardo, Pera-Guardiola, Bonillo, Batalla, Torrubia and Forth2022; Neumann et al., Reference Neumann, Hare and Newman2007). Studies often find that the facets differentially correlate with outcomes (Neumann et al., Reference Neumann, Hare and Newman2007; Ojanen & Findley-Van Nostrand, Reference Ojanen and Findley-Van Nostrand2019), although youth who score higher on all facets (i.e., more closely resembling prototypical overall psychopathy) may also present unique societal and mental health challenges (e.g., Andershed et al., Reference Andershed, Colins, Salekin, Lordos, Kyranides and Fanti2018; Bégin et al., Reference Bégin, Fontaine, Vitaro, Boivin, Tremblay and Côté2023; Salekin, Reference Salekin2017). Thus, it is important to examine the facets of psychopathy in addition to overall scores (Salekin et al., Reference Salekin, Andershed, Batky and Bontemps2018). Differences in overall or facet-level psychopathic traits, however, are likely dimensional rather than categorical in nature in both youth and adults as well as community and justice-involved populations (Edens et al., Reference Edens, Marcus and Vaughn2011; Murrie et al., Reference Murrie, Marcus, Douglas, Lee, Salekin and Vincent2007), suggesting that differences are a matter of degree rather than kind. Because of the dimensional nature of psychopathy, examining the construct in different populations can be important for delineating how it is associated with criterion variables. In line with this, psychopathic traits have been shown to be associated with several criterion-related outcomes including violence, antisociality, and unethical behavior in non-justice-involved settings of youth and adults (Andershed et al., Reference Andershed, Colins, Salekin, Lordos, Kyranides and Fanti2018; Coid et al., Reference Coid, Yang, Ullrich, Roberts and Hare2009; Pardini & Byrd, Reference Pardini and Byrd2012).
Adverse environments and adolescent social competition
We define adverse environments as those environments that capture a general negative or antisocial quality to the social relationships in one’s local and broader context, including relationships and climates marked by hostility, neglect, and competition (Merçon-Vargas et al., Reference Merçon-Vargas, Lima, Rosa and Tudge2020). Adverse environments can offer a different set of rewards and punishments for social behavior than do average environments. For example, hostile competition for resources may reward more selfish, antisocial behavior (Daly, Reference Daly2016). Adolescence is a time when individuals are increasingly competing with peers for social resources (Volk et al., Reference Volk, Dane, Marini and Vaillancourt2015), making the environment an important factor in whether that competition is pro- or antisocial in nature. To gain a comprehensive understanding of adolescent competition, it is important to understand the interplay between both broad and local environments (Bronfenbrenner & Morris, Reference Bronfenbrenner, Morris, Damon and Lerner1998; Bronfenbrenner, Reference Bronfenbrenner1979). Social learning theories suggest that individuals match their behavior to their perceived broad and local environments through a combination of observation, imitation, and external rewards and punishments (Bandura & Walters, Reference Bandura and Walters1977; Fox, Reference Fox2017). Evolutionary perspectives of human development propose a similar process of development where what is learned, and what influences learning, is biased by the importance of evolutionarily-relevant goals and outcomes such as social power and mating success (Ellis et al., Reference Ellis, Del Giudice, Dishion, Figueredo, Gray, Griskevicius, Hawley, Jacobs, James, Volk and Wilson2012).
To obtain these goals, individuals can learn to respond to adverse environments by engaging in risky or antisocial behavior themselves as a means of meeting the evolutionary goal of surviving and reproducing (Daly, Reference Daly2016; Del Giudice et al., Reference Del Giudice, Ellis and Shirtcliff2011; Ellis et al., Reference Ellis, Del Giudice, Dishion, Figueredo, Gray, Griskevicius, Hawley, Jacobs, James, Volk and Wilson2012). For example, researchers have found that individuals who experienced harsher and more unpredictable environments during childhood report higher adolescent risk-taking and lower sexual restrictedness (Brumbach et al., Reference Brumbach, Figueredo and Ellis2009; Patch & Figueredo, Reference Patch and Figueredo2017). Further, rates of bullying increase when there are few prosocial neighborhood resources and/or adolescents perceive their futures as compromised (Schmidt et al., Reference Schmidt, Pierce and Stoddard2016). Bullying itself can then lead to a greater number of offspring in the long-term (Kretschmer et al., Reference Kretschmer, la Roi, van der Ploeg and Veenstra2022). At the most extreme end, rates of homicide and violence have been found to be the highest among males from late adolescence to early adulthood in cities with higher levels of income inequality (Daly, Reference Daly2016). Thus, antisocial traits like psychopathy, although presenting a simultaneous risk to self and society, can contribute to adolescents’ learning to behave in a way that maximizes their success in environments that potentially reward antisocial behavior.
Adolescent psychopathic traits and adverse environments
Psychopathy may be one specific collection of antisocial traits that, in part, could be a response to adverse environments to acquire socially adaptive outcomes (Gao et al., Reference Gao, Raine, Chan, Venables and Mednick2010; Kerig et al., Reference Kerig, Bennett, Thompson and Becker2012; Ribeiro da Silva et al., Reference Ribeiro da Silva, Rijo and Salekin2015). However, the literature examining links between psychopathy and adversity is mixed, with some studies suggesting greater adversity (Baglivio et al., Reference Baglivio, Wolff, DeLisi and Jackowski2020; de Ruiter et al., Reference de Ruiter, Burghart, De Silva, Griesbeck Garcia, Mian, Walshe and Zouharova2022) while others show no relation to adversity (Bedwell & Hickman, Reference Bedwell and Hickman2022; Zwaanswijk et al., Reference Zwaanswijk, van Geel and Vedder2018). There are also some conflicting results of the directionality of effects. Although a growing number of studies show that environmental adversity may precede psychopathic traits (Backman et al., Reference Backman, Laajasalo, Jokela and Aronen2021; Brummelman et al., Reference Brummelman, Thomaes, Nelemans, Orobio de Castro, Overbeek and Bushman2015; Zara et al., Reference Zara, Bergstrøm and Farrington2024), some findings also suggest psychopathic traits may initiate adverse environmental responses as well (Salihovic et al., Reference Salihovic, Kerr, Özdemir and Pakalniskiene2012). Many longitudinal studies also suggest there are bidirectional effects (Miron et al., Reference Miron, Satlof‐Bedrick and Waller2020; Trentacosta et al., Reference Trentacosta, Waller, Neiderhiser, Shaw, Natsuaki, Ganiban, Reiss, Leve and Hyde2019; Waller et al., Reference Waller, Gardner, Viding, Shaw, Dishion, Wilson and Hyde2014). There may also be population differences where the psychopathy–adversity link may not be as strong in non-justice-involved compared to justice-involved populations (Bedwell & Hickman, Reference Bedwell and Hickman2022; but see also de Ruiter et al., Reference de Ruiter, Burghart, De Silva, Griesbeck Garcia, Mian, Walshe and Zouharova2022). This warrants further research into which, if any, levels of environmental adversity may be linked with psychopathy in non-justice-involved adolescents. We examine four levels of perceived environment, including parenting, schools, neighborhood, and family socioeconomic status (SES).
Parenting captures one of the most proximate environmental contexts that may be associated with adolescent psychopathy. Higher parental neglect and lower parental warmth have been shown to be associated with psychopathic traits among different adolescent populations (Backman et al., Reference Backman, Laajasalo, Jokela and Aronen2021; Gao et al., Reference Gao, Raine, Chan, Venables and Mednick2010). Neglectful and abusive parenting also appear to be associated with psychopathic traits across both non-justice-involved and justice-involved settings (de Ruiter et al., Reference de Ruiter, Burghart, De Silva, Griesbeck Garcia, Mian, Walshe and Zouharova2022). One possibility is that adverse home environments set the precedent for negative expectations in social relationships that influence adolescents adopting similar patterns of behavior themselves. In contexts such as schools, youth psychopathic traits have been concurrently associated with adverse school climates (Fisher & Brown, Reference Fisher and Brown2018) and lower school connectedness (Fanti et al., Reference Fanti, Colins, Andershed and Sikki2017), as well as longitudinally associated with lower student–teacher affiliation (Baroncelli & Ciucci, Reference Baroncelli and Ciucci2019). Although these findings may suggest a more negative and possibly competitive broader school environment, other studies suggest that adolescents higher in psychopathic traits may not suffer adverse peer relations within those school environments (Muñoz et al., Reference Muñoz, Kerr and Besic2008; Van Zalk & Van Zalk, Reference Van Zalk and Van Zalk2015). Homes and schools represent the inner layers of one’s social environment (Bronfenbrenner & Morris, Reference Bronfenbrenner, Morris, Damon and Lerner1998), whereas neighborhoods and SES may capture adolescents’ broader environmental contexts.
When considering broader environments, higher levels of neighborhood violence have been linked with psychopathic traits in justice-involved youth (Schraft et al., Reference Schraft, Kosson and McBride2013), but other studies examining community youth show the relation may be more complex (Markowitz et al., Reference Markowitz, Ryan and Marsh2015; Meier et al., Reference Meier, Slutske, Arndt and Cadoret2008). Some researchers have also found that lower SES was associated with adolescent psychopathic traits, including as a factor affecting the stability of psychopathy (Bégin et al., Reference Bégin, Fontaine, Vitaro, Boivin, Tremblay and Côté2023; Frick et al., Reference Frick, Kimonis, Dandreaux and Farell2003). In research examining community youth, however, the links between SES and psychopathic traits again appear to be quite mixed, including null, negative, and positive relations with SES (Zwaanswijk et al., Reference Zwaanswijk, van Geel and Vedder2018). Overall, our brief review suggests that there may be an association between different levels of environmental adversity and psychopathic traits, but the links have mostly been found in justice-involved settings, which calls for more research in non-justice-involved settings. It has also yet to be found – like in adults (Međedović, Reference Međedović2019) – whether psychopathic traits may be an evolved and/or learned response in adolescents that may be associated with acquiring socially adaptive outcomes in adverse contexts.
Adolescent psychopathic traits as an adaptive strategy
Some researchers have argued that psychopathic traits may have evolved to allow individuals to engage in deceit, manipulation, and intentionally harmful behavior for self-gain at the expense of others (Book et al., Reference Book, Methot-Jones, Blais, Hosker-Field, Volk, Visser, Gauthier, Holden and D’Agata2019; Harris et al., Reference Harris, Rice, Hilton, Lalumiere and Quinsey2007; Međedović, Reference Međedović2019; Mealey, Reference Mealey1995; Ribeiro da Silva et al., Reference Ribeiro da Silva, Rijo and Salekin2015). This selfish approach can lead to the aggressive and coercive pursuit of socially valued outcomes that may represent evolutionarily relevant resources such as social power and dating. In support of possible evolutionary roots of psychopathic traits, and unlike general antisocial traits, psychopathy does not appear to be correlated with prenatal problems or obstetric complications (Bégin et al., Reference Bégin, Fontaine, Vitaro, Boivin, Tremblay and Côté2023; Lalumière et al., Reference Lalumière, Harris and Rice2001; Zara et al., Reference Zara, Bergstrøm and Farrington2024). Further, individual differences in psychopathy are associated with genetic differences (Schermer & Jones, Reference Schermer and Jones2020) and offenders higher in psychopathy appear to be relatively less likely to violently target genetic kin (Krupp et al., Reference Krupp, Sewall, Lalumière, Sheriff and Harris2012). Combined, these findings offer some support that psychopathy may not be the result of perturbed developmental processes, but rather the result of an evolved developmental response to adverse environments (Glenn, Reference Glenn2019; Ribeiro da Silva et al., Reference Ribeiro da Silva, Rijo and Salekin2015). A remaining question, however, is whether psychopathic traits may help adolescents obtain evolutionarily relevant resources such as social power and dating opportunities in adverse environments.
Some evidence suggests that psychopathic traits correlate with evolutionarily relevant outcomes such as increased fecundity, notably, when adults came from adverse environments (Međedović et al., Reference Međedović, Petrović, Želeskov-Đorić and Savić2017). Psychopathic traits may allow individuals to engage in deception and/or self-promotion as a means to obtain desired social power, material resources, and/or sexual partners (Glenn et al., Reference Glenn, Efferson, Iyer and Graham2017; Monteiro et al., Reference Monteiro, Lopes, Nascimento, Gouveia, Shackelford and Zeigler Hill2017; Steininger & Pietschnig, Reference Steininger and Pietschnig2022). For example, in one study, after observing video clips of dating interactions, undergraduate men higher in psychopathy were rated higher in dating desirability by undergraduate women, despite the women not knowing the men’s ratings of psychopathy (Brazil & Forth, Reference Brazil and Forth2020). On the more antisocial side, individuals higher in psychopathy may also resort to coercive sexual tactics when deception or self-promotion fail (Harris et al., Reference Harris, Rice, Hilton, Lalumiere and Quinsey2007). Outside of reproductive opportunities, adolescents higher in psychopathic traits may also be adept at securing socially adaptive outcomes via displays of social power (Ometto et al., Reference Ometto, de Oliveira, Milioni, dos Santos, Scivoletto, Busatto, Nunes and Cunha2016; Van Zalk & Van Zalk, Reference Van Zalk and Van Zalk2015). Bullying is a form of proactive and harmful behavior that is associated with the pursuit, and achievement, of higher levels of social dominance and power (Pouwels et al., Reference Pouwels, Lansu and Cillessen2016; Volk et al., Reference Volk, Provenzano, Farrell, Dane and Shulman2021). As with reproductive goals, psychopathy has been associated with bullying (Ojanen & Findley-Van Nostrand, Reference Ojanen and Findley-Van Nostrand2019; van Geel et al., Reference van Geel, Toprak, Goemans, Zwaanswijk and Vedder2017). Accordingly, as adolescents place increasing value on peer relationships and begin to engage in dating behavior, dominance over peers and dating behavior may become important proxies for adaptive social outcomes sought by adolescents higher in psychopathic traits. It is critical then to examine whether psychopathy could be an adaptive response to adverse environments in adolescents specifically.
Current study
The goal of our study is to examine whether psychopathic traits indirectly linked self-perceived adverse environments (i.e., family SES, parental neglect, school competition, and neighborhood violence) to evolutionarily relevant social outcomes (i.e., social power, dating behavior). Given developmental and evolutionary explanations of psychopathic traits (Ribeiro da Silva et al., Reference Ribeiro da Silva, Rijo and Salekin2015), we expect that self-perceptions of higher parental neglect, higher school social competition, higher neighborhood violence, and lower SES would be indirectly associated with self-perceptions of higher social power and dating behavior through higher levels of psychopathic traits in adolescents. We examine two models: one using overall psychopathy scores and the other using the psychopathy facet scores. Given the holistic nature of environmental influences (Bronfenbrenner, Reference Bronfenbrenner1979), we also do not have any a priori predictions about the strength of the relationship for different levels (i.e., local versus broad) of environmental factors. Finally, we do not have any a priori predictions regarding age or gender differences as we expected indirect mechanisms would work similarly for most adolescents, although we do control for these factors.
Method
Participants
Adolescent participants (N = 396; 230 girls, M age = 14.64, SD = 1.52) were recruited from various extracurricular activities, including sports teams, youth groups, and dance groups across several medium-sized cities in Southern Ontario, Canada. Recruitment also took place in a diversity of locations, including those that may capture more at-risk youth experiences such as community centers, youth shelters, and homeless youth. The majority of participants were Caucasian (73.7%) and fewer were Asian (6.1%), African-Canadian (1.0%), Indigenous-Canadian (0.5%), Multi-ethnic (4.3%), or Other (4.8%). The remainder of participants did not report ethnicity (9.6%). The majority of youth (64.6%) reported their perceived family SES to be “about the same” as the average Canadian.
Procedure
After receiving consent from supervisors of the extracurricular activities, adolescents were invited during a group meeting to participate in a study on adolescent peer relationships. Interested individuals were provided with an envelope that included a parental consent form, an adolescent assent form, an identification number, and an online link to questionnaires. Both the consent and assent forms were required to be completed and returned for data to be used. Questionnaires were presented in random order online. At a prearranged date approximately one week later, completed consent and assent forms were collected, participation was verified using identification numbers, and participants were compensated with $15.
Measures
Psychopathic traits
To assess psychopathic traits, participants completed the 20-item self-report version of the Antisocial Process Screening Device (APSD; Frick & Hare, Reference Frick and Hare2001), which provides an assessment of the multifaceted psychopathy construct (Salekin, Reference Salekin2017), including impulsivity/daring-impulsive (e.g., “You act without thinking of the consequences”), callous-unemotional (e.g., “You are concerned about the feelings of others” [reverse-coded]), and narcissism/grandiose-manipulative traits (e.g., “You think you are better or more important than other people”). We computed both the mean overall score and facet scores. Items were rated on a three-point scale (0 = not at all true to 2 = definitely true). Higher scores reflected higher psychopathic traits. The scales showed acceptable internal consistency (overall: α = .76, impulsivity: α = .67, callous-unemotional: α = .50, narcissism: α = .70).
Socially adaptive outcomes
Social power
To assess the socially adaptive outcome of social power, six items adapted from various papers on social and material resource control were used (Hawley et al., Reference Hawley, Little and Card2008). The items included: “I am good at being able to get what I want from others.” (Item 1), “I usually get what I need, even if others don’t.” (Item 2), “I am able to get others to do what I say.” (Item 3), “I have a lot of power over others.” (Item 4), “In groups I am usually in charge or in control.” (Item 5), and “I usually get my way when I deal with others.” (Item 6). Each item was rated on a five-point scale (1 = never true to 5 = almost always true).
Dating behavior
To assess developmentally appropriate dating resources, items assessing dating frequency, dating partners, and sexual partners were used to assess dating behavior. Participants answered the following questions: “How often do you go on dates with a girl/boy, just the two of you?”, “How many different people have you gone on dates with, just the two of you?”, and “How many sexual partners have you had a voluntary sexual experience with (i.e., more than kissing or making out) since the age of 12?” Participants responded to the first question on a five-pont scale (1 = never to 5 = very often) and responded with a numeric value for the remaining questions.
Self-perceived adverse environmental factors
Perceived family SES
To assess perceived family SES, participants were asked, “Compared to the average Canadian, do you think your family is…” The item was rated on a five-point scale (1 = a lot less rich, 2 = less rich, 3 = about the same, 4 = more rich, and 5 = a lot more rich).
Parental neglect
To assess parental neglect, participants completed 15 items for each parent (i.e., mothers and fathers) from the parental neglect subscale of the Parental Acceptance-Rejection Questionnaire (Rohner, Reference Rohner1984, Reference Rohner2004). A sample item includes, “Pays no attention to me as long as I do nothing to bother him/her.” Each item was rated on a four-point scale (1 = almost never true to 4 = almost always true). An average was computed for mothers (α = .86) and fathers (α = .85) separately, and then an average of the two (r = .72) was computed for a final neglect score. Higher scores reflected higher perceived parental neglect.
School social competition
To assess school social competition, participants completed six items on the Desire for Social Success subscale of the Social and Academic Competition Scale (SACS; Sutton & Keogh, Reference Sutton and Keogh2000). An example item is, “I don’t try hard in class because people won’t like me if I do.” Each item was rated on a four-point scale (1 = strongly disagree and 4 = strongly agree). An average of all items was used, and higher scores reflected higher perceived school social competition (α = .63).
Neighborhood violence
To assess neighborhood violence, participants completed five items on the Children’s Exposure to Community Violence Scale (Richters & Martinez, Reference Richters and Martinez1993 used by Low & Espelage, Reference Low and Espelage2014). An example of an item is, “I have seen somebody arrested.” Each item was rated on a four-point scale (1 = never and 4 = often). An average of all items was used, and higher scores reflected higher perceived neighborhood violence (α = .74).
Results
Preliminary analyses
SPSS 24 was used for preliminary analyses. The maximum number of missing values for the key variables of interest was 2.0% or less, except for the number of sexual partners variable which had 12.9% missing and the number of dating partners which had 8.8% missing. This was expected given that some adolescents may not have begun dating and could have left the question blank although they had the option to indicate zero or skip this question. However, Little’s MCAR test indicated no significant differences in the pattern of results between participants who were missing and not missing data on study variables (χ 2 (323) = 362.563, p = .064). Therefore, Full Information Maximum Likelihood estimation was used to deal with missing cases. All variables met the assumptions of normality, except for the number of dating partners and number of sexual partners. These two variables also had extreme univariate outliers. Six outliers beyond 3.30 standard deviations were winsorized for each variable to maintain rank ordering but reduce impact on the variables. Winsorizing outliers also reduced skewness and kurtosis values for both variables and were within limits for structural equation modeling (< 10; Kline, Reference Kline2016). The significant correlations ranged from small to moderate in size (see Table 1 for correlations, means, and standard deviations).
Correlations, means, and standard deviations for environments, psychopathic traits, and socially adaptive outcomes
Note. N = 396. SES = Socioeconomic Status. *p < .05, **p < .01.
a Gender coded as 1 = Boy, 2 = Girl.
Primary analyses with overall psychopathy scores
Structural equation modeling (SEM) was conducted on Mplus version 8.7 (Muthén & Muthén, Reference Muthén and Muthén2020) to assess whether perceived adverse environmental factors were significantly associated with psychopathic traits, and in turn, whether psychopathic traits were significantly associated with social power and dating behavior, as two indicators reflecting socially adaptive outcomes. First, measurement models were conducted on the dependent variables (i.e., social power and dating behavior) using confirmatory factor analysis. For each latent variable, one indicator path intercept was set to zero and the factor loading was set to 1.0 to allow the means and variances of latent factors to be estimated. Second, a structural model was conducted by estimating paths from the four adverse environmental variables (i.e., family SES, parental neglect, school social competition, neighborhood violence) to psychopathic traits, and the paths from psychopathic traits to the two socially adaptive outcomes. Given that the adverse environmental variables could have associations with the two outcomes independent from psychopathic traits, we also estimated the direct paths from the environment variables to the two outcomes. We controlled for gender and age by allowing these variables to be correlated with one another and the four environment variables, and by estimating paths from age and gender to psychopathic traits and the two outcomes. Finally, the two outcomes were allowed to covary. We estimated the model using maximum likelihood estimation with robust standard errors (MLR) to account for non-normal distribution of the variables. Once determining the significant direct paths, we tested for significant indirect effects from the environmental variables to the socially adaptive outcomes through psychopathic traits using bootstrapping with 5,000 samples and maximum likelihood estimator. We used 95% confidence intervals that did not cross zero to determine significant indirect effects (Biesanz et al., Reference Biesanz, Falk and Savalei2010). Model fit was assessed with the following criteria: root mean square error of approximation (RMSEA) with values of less than .06, the comparative fit index (CFI; Hu & Bentler, Reference Hu and Bentler1999) with values greater than .95, and a standardized root mean square residual (SRMR) value less than .08 (Kline, Reference Kline2016). Non-significant chi-square values can also be indicative of good model fit, but this test is sensitive to sample sizes.
The confirmatory factor analysis for social power revealed a good fit to the model, χ 2 (9) = 14.233, p = .114; CFI = .990; RMSEA = .038, 90% CI [.000,.074]; SRMR = .025. Standardized factor loadings ranged from .547 to .790. As the latent dating behavior variable had three indicators the model was saturated and indicators of model fit were not informative. Standardized factor loadings ranged from .643 to .941. The overall structural model seen in Figure 1 had adequate fit, χ 2 (75) = 160.182, p < .001; CFI = .934; RMSEA = .054, 90% CI [.042, .065]; SRMR = .041. Although the CFI value was slightly below the cut off, this indicator is likely due to the number of parameters estimated, and the RMSEA and SRMR values demonstrated acceptable fit (Hu & Bentler, Reference Hu and Bentler1999).
Significant direct and indirect paths for perceived environments, psychopathic traits, and socially adaptive outcomes. Note. Solid line reflects significant direct paths whereas dashed line reflects significant direct and indirect paths; control variables (age, gender), disturbances, and errors are not shown for simplicity of presentation. Values represent correlations or standardized path coefficients.
Direct effects
Results of the SEM indicated that parental neglect, school social competition, and neighborhood violence were each significantly positively associated with psychopathic traits (see unstandardized and standardized regression coefficients in Table 2). Parental neglect, school social competition, and neighborhood violence were also significantly positively associated with one another. Hence, three of the four perceived environmental variables showed unique associations with psychopathic traits beyond their shared variance. Psychopathic traits, school social competition, and family SES were significantly positively associated with social power. Psychopathic traits and neighborhood violence were also positively associated with dating behavior. Thus, psychopathic traits were uniquely associated with social power and dating behavior beyond the variance accounted for by the perceived environmental variables. With respect to covariates, being older was significantly positively associated with neighborhood violence, parental neglect, psychopathic traits, and both outcomes. Being a boy was significantly positively associated with neighborhood violence, parental neglect, and psychopathic traits.
Direct paths between environments, psychopathic traits, and socially adaptive outcomes
Note. Values represent unstandardized regression coefficients (B), unstandardized coefficient standard errors (SE), and standardized regression coefficients (β). SES = Socioeconomic Status. *p < .05, **p < .01, ***p < .001.
a Gender coded as 1 = Boy, 2 = Girl.
Indirect effects
Given the significant paths from three of the environmental variables to psychopathic traits and the significant paths from psychopathic traits to the two outcomes, we examined whether there were significant indirect effects from these three environmental variables to the socially adaptive outcomes through psychopathic traits. For social power, there were significant indirect effects through psychopathic traits from parental neglect (b = .097, se = .038, β = .059, 95% CI [.039, .186]), school social competition (b = .123, se = .036, β = .076, 95% CI [.064, .207]), and neighborhood violence (b = .109, se = .032, β = .069, 95% CI [.058, .186]). For dating behavior, there were also significant indirect effects through psychopathic traits from parental neglect (b = .092, se = .050, β = .028, 95% CI [.020, .214]), school social competition (b = .117, se = .054, β = .036, 95% CI [.030, .247]), and neighborhood violence (b = .103, se = .052, β = .032, 95% CI [.023, .230]). Thus, unique indirect paths were observed from each of the three perceived environmental variables (parental neglect, school social competition, and neighborhood violence) to both outcomes of social power and dating behavior through higher levels of psychopathic traits.
Secondary analyses with psychopathy facet scores
The same model structure was repeated but replacing overall psychopathy scores with the three psychopathy facet scores (e.g., impulsivity, callous-unemotional, and narcissism). The overall structural model had adequate fit, χ 2 (89) = 172.48, p < .001; CFI = .934; RMSEA = .049, 90% CI [.038, .059]; SRMR = .040, which resembled the fit for the model using overall scores.
Direct effects
The overall SEM showed that different environments were uniquely associated with different psychopathy facets (Table 3; see also Supplemental Table S1 for bivariate correlations). When controlling for shared variance among the environments, family SES was only significantly positively associated with narcissism, whereas parental neglect was only significantly positively associated with callous-unemotionality. School social competition was significantly positively associated with both callous-unemotionality and narcissism, but neighborhood violence was positively linked with impulsivity and narcissism. For the outcome variables of social power and dating behavior, only narcissism was significantly positively associated with social power.
Direct paths between environments, facets of psychopathy, and socially adaptive outcomes
Note. Values represent unstandardized regression coefficients (B), unstandardized coefficient standard errors (SE), and standardized regression coefficients (β). SES = Socioeconomic Status. *p < .05, **p < .01, ***p < .001.
a Gender coded as 1 = Boy, 2 = Girl.
Indirect effects
We examined the possible indirect effects between perceived environments and socially adaptive outcomes via the facets of psychopathy (see Supplemental Figure S1). The results showed that for social power, there were significant indirect effects through the narcissism facet from family SES (b = .061, se = .024, β = .052, 95% CI [.020, .117]), school social competition (b = .170, se = .044, β = .104, 95% CI [.095, .266]), and neighborhood violence (b = .082, se = .036, β = .051, 95% CI [.022, .164]). Thus, when modeling the facets separately, we found that perceived environmental variables (family SES, school social competition, and neighborhood violence) were significantly and indirectly linked to social power through the effects of the narcissism facet specifically.
We performed post hoc power analyses to examine our sample size adequacy using the Monte Carlo approach (Muthén & Muthén, Reference Muthén and Muthén2002). The estimates from the present structural equation models were used as population parameter values for data generation and coverage in the Monte Carlo simulation using 1000 replications. For the key direct paths of interest, all power in our models were greater than .79.
Discussion
Our data suggest that adolescent psychopathic traits may be associated, at least in part, with acquiring socially adaptive outcomes in adverse environments. More specifically, adolescent psychopathic traits significantly indirectly linked self-perceived adverse environments to the evolutionarily relevant social outcomes of social power and dating behavior. This means that youth perceiving adverse conditions at home, in school, and in their neighborhood may have higher levels of psychopathic traits, potentially as an adaptive, learned response to these environments and as a means to acquire socially valued outcomes that can be classified as evolutionarily relevant (i.e., dating, social power). Thus, in adolescence, psychopathic traits may be adaptive within adverse environments for acquiring socially adaptive outcomes, even despite their societal costs and eventual individual costs into adulthood (Auty et al., Reference Auty, Farrington and Coid2015). In contrast, warmer and less hostile environments may inhibit the learning and/or expression of psychopathic traits as their emergence may be less adaptive in environments that evince more prosocial values, or their expression in such environments may not as readily translate into acquiring socially valued outcomes. We first discuss our findings with overall psychopathic traits and then discuss our findings on the facets of psychopathy.
Overall psychopathy, adverse environments, and socially adaptive outcomes
Beginning with our direct paths (see Table 2), psychopathic traits were positively and uniquely related to self-perceptions of parental neglect, school competiton, and neighborhood violence. These findings are consistent with existing literature that shows links between multiple layers of adverse environments and psychopathy (e.g., Fisher & Brown, Reference Fisher and Brown2018; Schraft et al., Reference Schraft, Kosson and McBride2013; de Ruiter et al., Reference de Ruiter, Burghart, De Silva, Griesbeck Garcia, Mian, Walshe and Zouharova2022). Importantly, our path model controlled for the shared variance among the environmental variables, which suggests that each of the three levels of perceived environmental adversity were incrementally associated with psychopathic traits over and above the effects of each other. Psychopathic traits, however, were not related to self-reported family SES which also fits with the mixed evidence regarding SES and psychopathy presented in the literature (Markowitz et al., Reference Markowitz, Ryan and Marsh2015; Zwaanswijk et al., Reference Zwaanswijk, van Geel and Vedder2018). Adolescent psychopathic traits were positively related to social power and dating behavior, both of which are consistent with previous adult research (Glenn et al., Reference Glenn, Efferson, Iyer and Graham2017; Međedović et al., Reference Međedović, Petrović, Želeskov-Đorić and Savić2017). As mentioned above, our path model also suggested that psychopathic traits may be uniquely associated with both of these outcomes after accounting for the other predictors in the model (e.g., environments, demographics).
Social power was also related to higher SES and school competition. This suggests that adolescents who had access to greater financial resources may be better able and/or more willing to exert social power. In this case, coming from a financially supportive environment could serve as a predictor of social power. The relationship between greater school competition and higher social power could suggest that social power may be most associated with wealthy adolescents who experience competitive schooling. On the other hand, the only environment variable that was associated with more dating behavior was neighborhood violence. These data suggest that a hostile and volatile neighborhood environment might facilitate an advanced dating experience in adolescents, which is consistent with some prior work (Wilson & Daly, Reference Wilson and Daly1997). Overall, our findings suggest that different types of evolutionarily relevant outcomes (e.g., social power and dating behavior) can be differentially associated with higher and lower levels of adverse environments as well as psychopathic traits.
Although only neighborhood violence was directly linked to dating behavior and family SES and school competition were directly associated with social power, we found several indirect links between three of the four environmental variables and socially adaptive outcomes through psychopathic traits. Parental neglect, school social competition, and neighborhood violence were each indirectly linked to both social power and dating behavior through higher psychopathic traits. The findings suggest psychopathic traits may provide a link between perceiving multiple independent adverse environments – including parental, school, and neighborhood levels – and socially adaptive outcomes such as social power and dating behavior. The paths were modeled with their shared variance taken into account, meaning parental neglect (as well as school competition and neighborhood violence) was uniquely associated with higher social power and dating behavior via higher psychopathic traits after accounting for the other environmental variables. Thus, each environmental factor may have a unique and independent influence on socially adaptive outcomes in adolescence via psychopathic traits. In addition to these unique paths, however, it would be informative to examine cumulative or interactive risk among multiple negative environments, psychopathic traits, and socially adaptive outcomes (e.g., Baglivio et al., Reference Baglivio, Wolff, DeLisi and Jackowski2020).
The indirect paths are consistent with the hypothesis that psychopathic traits may serve as a response linking adverse environments to social resources (Book et al., Reference Book, Methot-Jones, Blais, Hosker-Field, Volk, Visser, Gauthier, Holden and D’Agata2019; Patch & Figueredo, Reference Patch and Figueredo2017; Ribeiro da Silva et al., Reference Ribeiro da Silva, Rijo and Salekin2015). Our data replicate the indirect pathway between adverse environments, psychopathy, and sexual partners found in adults (Međedović, Reference Međedović2019) and extend those results to include social power and adolescent data. The fact that these indirect relationships existed for adverse social cues involving parents, school, and neighborhood but not adverse financial/SES cues suggests that cues from social or relational environments could be more relevant for psychopathic traits than cues from adverse financial or material environments and/or that cues from the latter might be mixed with respect to psychopathic traits and adaptive outcomes (Zwaanswijk et al., Reference Zwaanswijk, van Geel and Vedder2018). Further research is necessary to understand whether and how SES might be linked with psychopathy (Bégin et al., Reference Bégin, Fontaine, Vitaro, Boivin, Tremblay and Côté2023; see also discussion below).
What do these data suggest about the developmental factors underlying psychopathic traits? Researchers have previously suggested that environmental factors, particularly family factors, may play a role in the development of psychopathic traits (Farrington & Bergström, Reference Farrington, Bergström and Patrick2018). Building off this perspective with our results, youth who develop psychopathic traits could be influenced by negative social relationships that discourage prosocial behavior and instead facilitate the learning and/or expression of selfish, antisocial behavior, including obtaining power over peers and acquiring multiple dating partners. By adopting an evolutionary perspective, we suggest that this process could represent an example of how individuals (with their strengths and propensities) become matched to their surrounding environments, whether adverse or supportive, to maximize the probability of surviving and reproducing in those environments (Brazil, Reference Brazil2024; Ellis et al., Reference Ellis, Del Giudice, Dishion, Figueredo, Gray, Griskevicius, Hawley, Jacobs, James, Volk and Wilson2012).
Although this perspective can help appreciate some of the logic that links psychopathy to adverse environments, it does not excuse any antisocial behavior associated with psychopathy. We strongly emphasize that an adaptive or learned response is not necessarily morally and socially desirable (Ellis et al., Reference Ellis, Del Giudice, Dishion, Figueredo, Gray, Griskevicius, Hawley, Jacobs, James, Volk and Wilson2012). Thus, while our focus in this paper has been on the potential adaptive and functional benefits of youth psychopathic traits to understand why it exists, we stress that this observation must be viewed in light of the negative outcomes associated with psychopathic traits. Moreover, socially adaptive outcomes such as social power and dating behavior, while ultimately adaptive, can have more immediate costs if social power is a result of coercive, aggressive, or violent interpersonal relationships, or if dating behavior coincides with risky sexual behavior and/or coercive and violent dating relationships (Taquette & Monteiro, Reference Taquette and Monteiro2019). We emphasize the importance of interpretation and perspective. Taking an evolutionary view of psychopathic traits can inform why youth may be developing these traits in specific environments (i.e., for their adaptive benefits), but we recognize that other important perspectives focus on the reality of adverse psychological and social outcomes that arise from developing the traits as well. Both perspectives may be necessary to understand why such traits exist and how to successfully prevent them as early in development as possible.
We also stress that the developmental pathways implied in the indirect effects do not leave out the importance of genetics in the development of psychopathic traits (e.g., Schermer & Jones, Reference Schermer and Jones2020). Rather, our data argue that both genes and environments matter, with the latter possibly “unlocking” the expression of the former to lead to the developmental expression of psychopathic traits (e.g., Fox, Reference Fox2017; Glenn, Reference Glenn2019; Mealey, Reference Mealey1995). Further, social learning processes and adopting the antisocial practices of adults in one’s environment could co-facilitate this development along with individual differences in genetic background (i.e., social facultative adaptations; Rutherford, Reference Rutherford2016), suggesting a need to understand both the social learning and genetic components involved both independently and interactively. We also note that our findings may generalize only to community adolescents and further research should examine similar models in justice-involved adolescents, where rates of psychopathic traits, antisocial behavior, and adversity tend to be higher (Baglivio et al., Reference Baglivio, Wolff, DeLisi and Jackowski2020; Forth et al., Reference Forth, Kosson and Hare2003).
Psychopathy facets, adverse environments, and socially adaptive outcomes
To facilitate comparison with research on the multiple dimensions of psychopathy, we refer to the APSD facets here with the more commonly used terms of daring-impulsive (DI) for the impulsivity facet, grandiose-manipulative (GM) for the narcissism facet, and callous-unemotional (CU; see Salekin, Reference Salekin2017). When controlling for the shared variance among environmental variables, we found unique associations to each of the psychopathy facets. Although overall psychopathy was unrelated to family SES, the single-item SES measure was associated positively with GM, which is consistent with some research showing a similar positive link between GM and SES in community populations of youth (Zwaanswijk et al., Reference Zwaanswijk, van Geel and Vedder2018). Building on theories of adolescents’ abuse of power (Volk et al., Reference Volk, Andrews and Dane2022), this may indicate how excess wealth can facilitate the adoption of selfish or antisocial strategies as wealthier individuals have less need to maintain cooperation with others to get desired resources. When controlling for other environmental variables, parental neglect was uniquely associated with CU traits, but not the other dimensions. Higher neglect has been found to correlate with CU traits through the mediating pathway of lower maternal warmth (Bisby et al., Reference Bisby, Kimonis and Goulter2017). Relatedly, lower maternal warmth is a risk factor for increasing levels of CU traits across adolescence and into adulthood (Ray, Reference Ray2018). The directionality is difficult to decipher with our cross-sectional data though. Higher-CU youth may have compromised social affiliation (Viding & McCrory, Reference Viding and McCrory2019), which may affect how these youth interpret their social relationships – including as more neglectful – compared to lower-scoring youth.
School social competition was uniquely positively associated with both CU and GM traits after controlling for other environments’ variance. Because CU traits have been found to correlate with negative school climate (Fisher & Brown, Reference Fisher and Brown2018) and negative academic performance (Frick & Hare, Reference Frick and Hare2001), one possibility is that higher-CU youth may be compensating for a negative school climate by electing to invest in their social standing, which could snowball into (1) decreased academic interest and performance and (2) increased antisocial behavior toward peers to obtain status (van Geel et al., Reference van Geel, Toprak, Goemans, Zwaanswijk and Vedder2017). It is also not surprising that GM traits correlated with school social competition as GM is strongly associated with motivations for social status and holding socially competitive attitudes (Stellwagen & Kerig, Reference Stellwagen and Kerig2013). The last environment variable, neighborhood violence, was uniquely associated with GM and DI traits, which corresponds to previous research using justice-involved youth (Schraft et al., Reference Schraft, Kosson and McBride2013). The link with GM is also consistent with research showing GM tends to increase across adolescence and young adulthood when youth witness violence in their communities (Ray, Reference Ray2018).
Regarding our socially adaptive outcomes, the facet findings were more limited. GM was uniquely associated with higher social power beyond the effects of the environment and the other psychopathy facets. This accords with research showing GM traits may be linked to leadership, social poise, and persuasiveness (Stellwagen & Kerig, Reference Stellwagen and Kerig2013; Weiss et al., Reference Weiss, Lynam and Miller2018). In contrast, none of the facets alone uniquely associated with dating behavior. The bivariate correlations, however, showed that each of the facets were positively linked with the items that formed the dating behavior latent variable (see Supplemental Table S1). Thus, when controlling for shared variance among the facets, no unique relation remained between the facets and dating. It may therefore be the case that any adaptive benefits of dating associated with adolescent psychopathy require input from all three of the facets in order to be expressed (Brazil & Forth, Reference Brazil and Forth2020).
The final analysis using facets showed that only GM traits significantly mediated any unique links between the environment variables and socially adaptive outcomes. Specifically, higher SES, school social competition, and neighborhood violence were linked to higher social power via higher-GM traits. The findings with the latter two environment variables – school social competition and neighborhood violence – is in line with our social learning/adaptive hypotheses that adverse conditions may facilitate the acquisition of socially adaptive outcomes through the use of more psychopathic traits (GM traits specifically here). However, the indirect link with higher SES is counter to this hypothesis. Instead, this finding seems to suggest that less adversity (i.e., higher SES) was linked to higher social power via higher-GM traits. Because of the subjective nature of our SES item, one possibility is that higher-GM youth artificially inflated their SES because they perceive themselves as better than others. It may also be the case that higher SES is genuinely linked with higher-GM traits, which in turn impacts social power. Power facilitates aggressive acquisition of adolescents’ status and resource goals (Andrews et al., Reference Andrews, Cillessen, Craig, Dane and Volk2023; Volk et al., Reference Volk, Andrews and Dane2022). In our case, it appears that a decision to use one’s wealth (i.e., SES) in order to pursue those goals may be related to a desire to possess those resources (i.e., having higher of GM traits). Modest, honest individuals (lower GM) may be less interested in pursuing those goals and thus do not abuse their economic priviledge for their own gains. This suggests that environmental adversity per se might not fully explain higher psychopathic traits, but rather those traits could be the result of the right mixture of hostile and competitive environments that foster selfish and individualistic motives (e.g., GM) alongside environmental affordances that increase the ability of individuals to successfully employ selfish and individualistic (versus cooperative) strategies in those environments. In this way the environment can be seen as increasing both the motivation and the reward for developing and employing psychopathic traits.
Limitations and future directions
Our data largely support the hypothesis that adolescent psychopathic traits may represent an adaptive response to adverse social environments. However, our current data cannot demonstrate a causal direction or linkage, leaving the possibility that adolescents with more psychopathic traits may either elicit or seek adverse environments as well. We also used self-report measures for each construct, including perceptions of the different levels of the environment, suggesting the possibility that our results could have been influenced by shared method variance and/or common dispositional characteristics that may have influenced how participants responded to the three sets of variables in our model. It could also be the case that youth with psychopathic traits have a general negative perception of their environment (Altikriti & Nedelec, Reference Altikriti and Nedelec2020). Although youth perceptions are important to examine, future research could also examine aspects of the environment and social outcomes using measures beyond youth perceptions, including neighborhood crime rates, parental employment, observations at school and home, and peer nominations of social power and dating outcomes. Our measure of school social competition, in particular, may be limited in how it directly captured adversity in the youth’s school independent of the youth’s perceptions of or motivations for school. The items at least indirectly capture a more competitive and adverse social climate (e.g., “I don’t always try hard in class because people won’t like me if I do” [emphasis added]), but future research should aim to examine more direct and objective measures of school climate (e.g., number of student suspensions). It may also be important to extend the research using a clinical measure of adolescent psychopathy (Forth et al., Reference Forth, Kosson and Hare2003).
Another limitation of our study was that we used a general sample of adolescents from a predominantly White, middle-class Canadian region. We thus lack the full range of environmental stressors and psychopathic traits that may be found in more adverse contexts (e.g., juvenile detention centers), which may have reduced our ability to detect effects. Indeed, effect sizes for our direct and indirect effects were small. Although we expect that such extra variance would strengthen our findings (Ometto et al., Reference Ometto, de Oliveira, Milioni, dos Santos, Scivoletto, Busatto, Nunes and Cunha2016), it remains possible that greater variation may end up causing other patterns to emerge. Further, although we controlled for age and gender, our data do not cover childhood, pre-adolescence, or late adolescence, and therefore do not present an overall picture of development across time or across genders. Lastly, although our indirect effects imply a developmental path, our cross-sectional data precludes such a conclusion, and we again call for longitudinal and/or experimental research across childhood and adolescence to test multiple theoretically based hypotheses more directly, including the use of cross-lagged models (Davis et al., Reference Davis, Brittain, Arnocky and Vaillancourt2022). Moreover, we recognize there are alternative, statistically equivalent models possible with our cross-sectional data (e.g., socially adaptive outcomes indirectly linking social environments to psychopathic traits). Although we selected our model based on theory and prior evidence, longitudinal and/or experimental data may be helpful in further testing these alternative and equivalent models. We were also limited in modeling only some measures as latent variables because of poor model fit and our moderate sample size, so we encourage future work to examine all measures as latent variables with larger sample sizes.
As mentioned above, our findings are consistent with both social learning and evolutionary perspectives. Further research that takes into consideration both perspectives could help tease apart predictions made by each. For example, although our study considered desirable and/or adaptive outcomes (i.e., social power, dating), other outcomes could be assessed as well, including those that are undesirable and/or maladaptive (e.g., disease prevalence, injuries, social rejection). Future work could also consider how these perspectives might be combined to create a comprehensive developmental model of psychopathy and other antisocial traits (Ellis et al., Reference Ellis, Del Giudice, Dishion, Figueredo, Gray, Griskevicius, Hawley, Jacobs, James, Volk and Wilson2012; Frick & Viding, Reference Frick and Viding2009).
Conclusion
Overall, our data suggest that psychopathic traits may serve as an indirect link between adverse environments and adaptive outcomes such as social power and dating behavior. Our data may suggest an explanation for why it can be difficult to treat or alter psychopathic traits (Salekin, Reference Salekin2017); because adolescents may gradually learn that those traits may provide individual benefits for them in environments characterized by negative social relationships. If adolescents are benefiting from psychopathic traits when they are in these adverse environments, interventions will need to consider addressing and reducing those benefits alongside treatments aimed at promoting prosocial behavior and relationships.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0954579424000051.
Data availability statement
Available upon reasonable request from last author.
Funding statement
This work was supported by the Brock University Council (10-2015) for Research in the Social Sciences.
Competing interests
None.
Ethical approval
All procedures in this study were approved by the Brock University Social Science Research Ethics Board.
Consent
Informed written consent was obtained from parents of adolescent participants and informed written assent was obtained from adolescent participants.
Open access
