Perceived Parental Protectiveness Promotes Positive Friend Influence

Infant and Child Development Inf. Child. Dev. (2014) Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/icd.1885

Perceived Parental Protectiveness Promotes Positive Friend Influence Brett Laursena,*, Rita Žukauskienėb,*, Saulė Raižienėb, Cody Hiatta and Daniel J. Dicksona a

Florida Atlantic University, Fort Lauderdale, FL, USA Mykolas Romeris University, Vilnius, Lithuania

b

Friend influence over prosocial behaviour and delinquent behaviour was examined as a function of relative parental protectiveness in a community sample of Lithuanian high school students (M = 16.5 years old). Participants completed self-reports describing commitment to personal values, delinquent behaviours, prosocial behaviours, and perceived parental protectiveness. Mutual friends (158 male dyads, 241 female dyads) were identified from peer nominations. Distinguishable dyad Actor–Partner Interdependence Model analyses illustrate how parenting promotes positive peer influence. The results indicate that friend influence is greatest in the context of protective parenting: Adolescents who perceived more parental protectiveness were positively influenced by the strength of their friend’s personal values, whereas adolescents who perceived less parental protectiveness were not. Copyright © 2014 John Wiley & Sons, Ltd. Key words: parent; peer; influence; protectiveness It is far simpler to proclaim the significance of parents to adolescent development than it is to demonstrate it. To be sure, there is evidence that parents have a direct hand in shaping adolescent outcomes, but effects, when found, tend to be modest (Laursen & Collins, 2009). Evidence of peer influence, by contrast, is robust, especially during early adolescence (Hafen, Laursen, & DeLay, 2012). Vigorous debates centre on the manner and degree to which parents contribute to adolescent development (Collins, Maccoby, Steinberg, Hetherington, & Bornstein, 2000). Nuanced models have emerged; some suggesting that parents shape outcomes by fostering opportunities for peer influence. The present study was designed to illustrate procedures for testing these opportunity models. Actor–Partner Interdependence Model (APIM) analyses, conducted on friends drawn from a community sample

*Correspondence to: Brett Laursen, Department of Psychology, Florida Atlantic University, 3200 College Avenue, Fort Lauderdale, FL 33314, USA. E-mail: [email protected] Rita Žukauskienė, Institute of Psychology, Mykolas Romeris University, Ateities str. 20, Vilnius LT-08303, Lithuania. E-mail: [email protected] Copyright © 2014 John Wiley & Sons, Ltd.

B. Laursen et al.

of Lithuanian high school students, examined the degree to which one friend’s personal values influenced the other friend’s prosocial behaviour and antisocial behaviour. Friends were distinguished on the basis of parental protectiveness, to determine whether parent restrictiveness and supervision create a climate that is conducive to peer influence. The child’s social world changes dramatically across the transition into adolescence. Less time is spent in the company of parents and more time is spent in the company of peers (Larson, Richards, Moneta, Holmbeck, & Duckett, 1996). As parent supervision wanes, concerns arise about the influence of peers, especially friends, with whom adolescents spend the bulk of their free time. Considerable evidence suggests that affiliation with deviant peers is associated with increases in delinquency, drinking, and substance use (e.g. Dishion & Piehler, 2009; Vitaro, Brendgen, Ladouceur, & Tremblay, 2001). Less is known about positive peer influence, but recent studies suggest that friends can promote socially desirable features, such as prosocial behaviour (e.g. Sebanc, 2003) and school achievement (e.g. Frank et al., 2008). These studies typically share an important limitation in that they indicate only the presence of peer influence, but fail to identify the source, leaving unclear who is influencing whom. Conventional wisdom assumes that increases in peer influence across the adolescent years must necessarily result in declining parent influence. Scholars take a somewhat different view, challenging the assertion that parent and peer relationships always work at cross-purposes (e.g. Laursen & Mooney, 2008). Contemporary models of influence acknowledge that as more time is spent in the company of peers, parents have less direct influence over child outcomes, but the source of parent influence is thought to change, rather than disappear. Instead of directly influencing child outcomes, parent–child relationships become important moderators of external influence, providing a context that facilitates resistance to peer influence or one that promotes dependence and susceptibility to peer pressure (Fulgini & Eccles, 1993). These contextual models assume that parents amplify or inhibit peer influence by shaping the child’s interest in and receptivity to outside forces. In the present study, we examine whether peer influence varies as a function of relative parent protectiveness. Protectiveness differs from other important parenting constructs, such as warmth, psychological control, and monitoring, which describe the affective climate of the relationship, attempts to control the child’s emotional state, and knowledge about the child’s activities and whereabouts (Laursen & Collins, 2009). Protectiveness describes parent restrictions and supervision designed to shield children from adverse outcomes (Kendler, 1996). As such, it represents a form of parental engagement that is distinct, but overlapping with other parenting practices. It is important to note that our conceptual model does not posit parental protectiveness as directly impacting adolescent outcomes. Rather, perceptions of protectiveness are assumed to provide a context that facilitates peer influence. There are several reasons why protectiveness may be a marker of susceptibility to peer influence. Protectiveness may signal close relations with engaged parents, which promote a worldview that is trusting of others and open to persuasion (Kasser, Ryan, Zax, & Sameroff, 1995). Naiveté may be an unintended consequence of parental protectiveness (Eltringham & Aldridge, 2002). Adolescents accustomed to safe or benign environments in which potential sources of harm have been (or will be) mitigated may be more trusting of peers and more open to their influence. Protectiveness may alienate youth, Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

Parental Protectiveness

inadvertently elevating the significance of friends and romantic partners and increasing the attraction of peer conformity and the validation it brings (Steinberg & Silverberg, 1986). Parental protectiveness may inhibit autonomy development. Limits to behavioural independence may signal an inability to advocate for one’s own view and a tendency to succumb to the wishes of others. Protectiveness may signal a lack of maturity in adolescent children. Parents may be vigilant of adolescent children who lack a firm identity, out of fear for their heightened susceptibility. Adolescents who display greater autonomy appear more adult-like in the eyes of their peers and are more apt to be emulated by their less mature counterparts (Moffitt, 1993). Adolescent behaviours are a reflection of personal values and the degree of conviction attached to those values. Prosocial values are associated with higher levels of prosocial behaviour and lower levels of antisocial behaviour (Hart & Fegley, 1995; Padilla-Walker & Carlo, 2007). This effect is strongest for youth who hold the greatest commitment to their values (Simons, Whitbeck, Conger, & Conger, 1991). New evidence suggests that friend values exert a unique influence on behaviour, over and above an adolescent’s own values. One study found that friend moral reasoning was a unique predictor of pre-adolescent aggression (Gasser & Malti, 2012), and another found links to adolescent prosocial behaviour (Barry & Wentzel, 2006). Presumably, commitment to personal values enhances the importance of these values, because greater commitment implies heightened significance. It is reasonable to assume that parental protectiveness moderates the degree to which one adolescent’s values shape another adolescent’s behaviour, and there is evidence that parenting both amplifies positive effects (Mounts & Steinberg, 1995) and inhibits negative effects (Gauze, Bukowski, Aquan-Assee, & Sippola, 1996). Methodological problems loom large in the measurement of friend influence. Statistical obstacles have plagued efforts to identify the magnitude and direction of peer influence (Laursen, 2005). Adolescent friends are often closely bound to one another, with partners reporting interdependent thoughts, feelings, and behaviours. It follows that data collected from two friends cannot be considered independent. Most conventional statistical procedures assume independent data from participants. When parametric statistics are applied to nonindependent data, the results can be biased, sometimes dramatically so (Kenny, 1996). Recent advances in dyadic data analyses overcome these obstacles. The APIM (Kenny, Kashy, & Cook, 2006) partitions variance shared across partners on the same variable from variance that uniquely describes associations within dyads and between partners. In this manner, unique associations between variables can be measured within individuals (i.e. intra-individual influence) and across partners (inter-individual influence). In the present study, we examine the degree to which commitment to personal values is linked to prosocial behaviour and delinquent behaviour. The distinguishable dyad APIM treats each member of the dyad as belonging to a distinct class of participants, creating separate measures of association between variables, for each type of partner. With this tool, investigators can address the question: Within a friendship dyad, who influences whom? In the present study, we divided friends according to perceptions of parental protectiveness, to examine whether adolescents who perceive themselves to be more supervised than their best friend differed from adolescents who perceive themselves to be less supervised in associations between commitment to personal values and prosocial or delinquent behaviour. Covariates can be added to APIM analyses, to control for attributes known to contribute to intra-individual and inter-individual influence. The goal is Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

B. Laursen et al.

to eliminate confounds that may account for associations between commitment to personal values and adolescents outcomes and to separate perceptions of parental protectiveness from other traits that may confer influence or susceptibility to influence. The list of potential confounds is lengthy, so we focused on variables that describe parenting practices (e.g. warmth, psychological control, and monitoring) and variables that may be linked to interpersonal displays of dominance, persuasion, or malleability (e.g. personality traits). Results from APIM analyses produce measures of association that reflect two different sources of influence. The first emphasizes the role of the agent of influence. One friend may emerge as influential, able to entice or coerce the other to do something he or she might not otherwise do. Influential friends typically have characteristics that confer power or status. The second emphasizes the role of the recipient of influence. One partner may emerge as susceptible to influence, open to being enticed or coerced into doing something he or she might not otherwise do. Susceptible friends typically have characteristics that denote compliance or openness. These two possibilities can be teased apart longitudinally (Marion, Laursen, Kiuru, Nurmi, & Salmela-Aro, 2014), but with cross-sectional data, distinctions must be made on conceptual grounds. The present study focuses on the role of parents as facilitators of friend influence. Available evidence suggests that parenting moderates adolescent outcomes arising from peer experiences (e.g. Allen, Chango, Szwedo, Schad, & Marston, 2012). There is little indication that parenting shapes an adolescent’s effectiveness as an agent of influence (Brown & Bakken, 2011). Therefore, we assume that parental protectiveness affects the degree to which an adolescent is susceptible to friend influence but not the degree to which an adolescent is influential. The present study involves a community sample of high school friends. Within each dyad, one friend was more protected by parents, and the other was less protected by parents. APIM analyses address four goals. The first goal is to describe intra-individual associations between commitment to personal values and outcomes. Is more commitment to personal values linked to greater prosocial behaviour and less delinquent behaviour? Replicating the results of previous studies (e.g. Padilla-Walker & Carlo, 2007), we expected personal values to be associated with favourable outcomes. The second goal is to describe inter-individual associations between commitment to personal values and outcomes. Is one friend’s commitment to personal values linked to the prosocial behaviour and delinquent behaviour of the other friend? Friend moral development has been tied to adolescent outcomes (e.g. Gasser & Malti, 2012), so we expected similar findings for personal values. The third goal is to describe intra-individual associations as a function of relative parental protectiveness. Does parental protection alter the tendency of personal values to promote prosocial behaviour and discourage delinquent behaviour? It was not clear that parental protectiveness would play a role in inter-individual associations between values and behaviours, although it might conceivably dampen associations should parent demands conflict with adolescent values. The fourth goal is to describe inter-individual associations as a function of relative parental protectiveness. Does parental protection promote susceptibility to positive influence from friends? Parental protectiveness was expected to enhance susceptibility to friend influence, because adolescents who are more supervised have less experience with autonomous decision making (Steinberg & Silk, 2002). Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

Parental Protectiveness

METHOD Participants The final sample of 798 participants (316 boys and 482 girls) was drawn from the ongoing POSIDEV Project, a study of all students attending five high schools (9th–12th grades) in a small city in Northeastern Lithuania. Participants ranged in age from 14 to 19 years (M = 16.5, SD = 1.24). According to participant reports, 65.9% lived with two parents, 22.8% received free school lunch, and 21.7% had at least one parent who was unemployed. Procedure Data collection took place in the spring of 2013. All five high schools in the administrative region participated. Students were informed about the study during school hours and told that participation was voluntary. Parents were informed about the study through a written letter and asked to contact the school or the investigators if they did not want their child to participate. The participation rate was 98.9%. Trained research assistants administered questionnaires during regular school hours. Students who were absent on the day of data collection were contacted by research assistants the following week to complete surveys. A total of 1729 students successfully completed questionnaires. Of this total, 19.5% (n = 338) were eliminated because their friendship nominations were not reciprocated. An additional 24.8% (n = 429) were eliminated because their friends were members of higher ranked mutual friend dyads. There were no differences on study variables between those included in the analyses and those excluded. Measures Delinquent behaviour Adolescents completed 15 items assessing antisocial behaviour, from the Youth Self-Report (Achenbach & Rescorla, 2001). Sample item: ‘I cut classes or skip school.’ Items were rated on a scale ranging from 0 (not true) to 2 (very true or often true). Items were averaged to create a scale score. Internal reliability was good (alpha = .82). Parental protectiveness Adolescents completed nine items assessing parental restrictions and supervision, from the Memories of Upbringing Scale (EMBU: Gerlsma, Arrindell, van der Veen, & Emmelkamp, 1991). Sample items: ‘My parents forbade me from doing things other children were allowed to do because they were afraid something might happen to me.’ ‘My parents put limits on what I was and was not allowed to do.’ Items were rated on a scale ranging from 1 (no, never) to 4 (almost always). Items were averaged to create a scale score. Internal reliability was adequate (alpha = .74). The EMBU is widely utilized in European studies of parenting. The instrument has welldocumented convergent and divergent validity (Arrindell, Gerlsma, Vandereycken, Hageman, & Daeseleire, 1998), test–retest reliability (Muris, van Brakel, Arntz, & Schouten, 2011), and cross-cultural equivalence (Deković et al., 2006). Peer nomination inventory Adolescents identified up to eight important peers, defined as ‘someone you talk with, hang out with, and do things with’ (Laursen, Hafen, Kerr, & Stattin, 2012). Each Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

B. Laursen et al.

was labelled friend, romantic partner, or sibling. Important peers could be older or younger, boys or girls, from the same school or a different school. Mutual friends were dyads with reciprocated important peer nominations labelled friendships. Mutual friendships were overwhelmingly same sex, so we restricted our study to these relationships. Each participant was limited to one friendship. Some were involved in more than one friendship; nomination rankings were summed across dyads, and the highest ranked mutual friends were selected (ties were resolved by giving preference to partners who would not otherwise be included). Commitment to personal values Adolescents completed five items assessing the importance of adhering to personal standards, from the character subscale of the Positive Youth Development Inventory (Lerner et al., 2005). Sample item: ‘I stand up for what I believe, even when it’s unpopular.’ Items were rated on a scale ranging from 1 (not important) to 5 (extremely important). Items were averaged to create a scale score. Internal reliability was adequate (alpha = .78). Prosocial behaviour Adolescents completed five items assessing empathy, helping, and cooperation, from the Youth Self-Report (Achenbach & Rescorla, 2001). Sample item: ‘I like to help other people when I can.’ Items were rated on a scale ranging from 0 (not true) to 2 (very true or often true). Items were averaged to create a scale score. Internal reliability was adequate (alpha = .75). Control variables Adolescents completed six items describing perceptions of parental warmth (Arrindell et al., 2001). Items were rated on a scale ranging from 1 (no, never) to 4 (almost always). Internal reliability was high (alpha = .85). Adolescents completed eight items describing perceptions of parental control (Barber, 1996). Items were rated on a scale ranging from 1 (not like him/her) to 3 (a lot like him/her). Internal reliability was acceptable (alpha = .79). Adolescents completed eight items describing parental monitoring (Small & Kerns, 1993). Items were rated on a scale ranging from 1 (never) to 5 (always). Internal reliability was high (alpha = .91). Adolescents also completed the NEO Five Factor Personality Inventory (Costa & McCrae, 2010), which consisted of five 12-item subscales that measured Neuroticism, Extraversion, Openness, Agreeableness, and Conscientiousness. Items were averaged to create a scale score. Internal reliability was acceptable (alpha = .65–.85). Nomination rank describes the order in which the mutual friend was listed on the peer nomination inventory (range = 1 to 8). Plan of Analysis The APIM analyses for distinguishable dyads were conducted within a structural equation modelling framework using Mplus 6.11 (Muthén & Muthén, 1998–2010). We used a four-step procedure to examine the effects of personal values on selfand partner reports of delinquent behaviour and prosocial behaviour, within each friend and between friends who differed on parental protectiveness. In the first step, we distinguished friends on the basis of perceived parental protectiveness. Friends were assigned to roles on the basis of self-reports, with one friend in each dyad classified as more protected (M = 2.66, SD = 0.44) and Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

Parental Protectiveness

the other friend classified as less protected (M = 2.02, SD = 0.41) by parents. Dyads in which partners differed by less than .50 standard deviation in parental protectiveness scores were omitted from analyses (n = 82 dyads). An additional 16 dyads were eliminated because one or both friends did not have data on parental protectiveness. Thus, the final sample included 399 mutual friend dyads (n = 158 boys and 241 girls). We tested for distinguishability on the predictor variable (i.e. personal values) to determine whether the data were suitable for distinguishable dyad analyses. Distinguishability is an empirical construct. Same-sex friends can and do fall into meaningfully different categories (e.g. age, peer acceptance), and it is both appropriate to distinguish between them on the basis of these constructs and necessary to do so if scholars are to determine whether one category of friends has influence over another (Laursen, Popp, Burk, Kerr, & Stattin, 2008). The χ 2 test of distinguishability constrained the variances, interpersonal correlations, and intrapersonal correlations to be equal for more protected friends and less protected friends. A significant χ 2 value with poor fit indicates that friends within dyads should be distinguished on the basis of parental protectiveness. A nonsignificant χ 2 value indicates that partners cannot be distinguished on the basis of parental protectiveness and that APIM analyses for indistinguishable dyads should be conducted (Kenny et al., 2006). In the second step, we conducted a modified APIM analysis with two outcome variables. Figure 1 depicts the measurement model. Actor paths represent associations between variables within an individual (i.e. intra-individual influence). Actor paths describe the degree to which the more protected friend’s commitment to personal values predicts his or her own prosocial behaviour (a1) and his or her own delinquent behaviour (a2). Actor paths also describe the degree to which the less protected friend’s commitment to personal values predicts his or her own delinquent behaviour (a3) and his or her prosocial behaviour (a4). Partner paths represent the influence that one friend has over the other (i.e. inter-individual influence). Partner paths describe the degree to which the more protected friend’s commitment to personal values predicts the less protected friend’s delinquent behaviour (p1) and prosocial behaviour (p2). Partner paths also describe the degree to which the less protected friend’s commitment to personal values predicts the more protected friend’s prosocial behaviour (p3) and the delinquent behaviour (p4). Correlations between friends on the predictor variable (i.e. personal values) and between and within friends on outcome variables (i.e. delinquent behaviour and prosocial behaviour) are included to model statistical nonindependence. The within-dyad correlation on the predictor variable (c) represents friend similarity, and the residual correlations between outcome variables (rc1 through rc6) represent similarity that remains after accounting for actor effects and partner effects. It is important to note that estimates of residual similarity are not simply error but a means to control for additional sources of nonindependence from unobserved variables. Fit indices are not reported because the model is fully saturated. To illustrate the results, we plotted the slopes for each path. For example, to depict the partner path for the less protected friend’s commitment to personal values predicting the more protected friend’s delinquent behaviour (p1), we identified the value of the dependent variable (i.e. more protected friend’s delinquent behaviour) at the mean of the predictor variable (i.e. less protected friend’s commitment to personal values), then plotted values at one standard deviation below and above the mean using the beta weight for the association between the predictor and outcome variables. Supplemental analyses included the addition of correlational paths between parental protectiveness and the Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

B. Laursen et al.

More Protected Friend Commitment to Personal Values

More Protected Friend Prosocial Behavior

a1

rc3 a2

p3

More Protected Friend Delinquent Behavior rc5

p4 c

rc1

rc2

p1 rc6 p2 Less Protected Friend Delinquent Behavior Less Protected Friend Commitment to Personal Values

a3 rc4 a4 Less Protected Friend Prosocial Behavior

Figure 1. Conceptual model for distinguishable dyad Actor–Partner Interdependence Model.

predictor variable, to ensure that actor and partner effects were not being driven by extreme cases on the distinguishing variable. In the third step, follow-up moderator analyses tested for sex differences in patterns of association. Separate multiple group APIM analyses examine whether boys and girls differed on the magnitude and direction of actor paths and partner paths. Group differences in the magnitude of a path were tested by constraining the path to be equal in male and female dyads. A significant χ 2 value indicates poor fit, suggesting sex differences in the magnitude of a path. A nonsignificant χ 2 value indicates that the magnitude and direction of the path do not significantly differ for boys and girls. In the fourth step, we added control variables to the model, to rule out the possibility that the results were driven by variables that could account for associations between commitment to personal values and delinquent behaviour or associations between commitment to personal values and prosocial behaviour. The list of control variables included characteristics of the family (i.e. parental warmth, parental control, parental solicitation of information) and the individual (i.e. age and personality traits) that could plausibly alter results obtained from the conceptual model. We included the nomination rank of the friend to rule out the possibility that patterns of influence were a function of liking and parental protectiveness to rule out the possibility that the results were driven by a few extreme cases. These analyses involved the addition of correlational paths between the control variable and the predictor variable and directional paths between the control variable and the outcome variable. An average of 1.4% (range = 0.1% to 2.3%) of the data were missing on commitment to personal values, prosocial behaviour, and delinquent behaviour. Little’s MCAR test indicated that data were missing completely at random, χ 2(5, N = 798) = 2.49, p > .05. Full-Information Maximum Likelihood estimation procedures were used to handle missing values. Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

Parental Protectiveness

RESULTS Preliminary Analyses A χ 2 test of distinguishability indicated that partners should be distinguished on the basis of relative parental protectiveness, both in the final sample χ 2(18, N = 399) = 1465.72, p < .001, and in the sample that included partners who differed by less than .50 SD in parental protectiveness χ 2(18, N = 481) = 1742.93, p < .001. Intraclass correlations (interpreted as r2) indicated that the data were not independent. There were statistically significant (p < .05) within dyad correlations for delinquent behaviour (r = .17), prosocial behaviour (r = .15), and personal values (r = .15). Separate 2 (sex) by 2 (parental protectiveness: more protected partner and less protected partner) ANOVAs were conducted, with commitment to personal values, delinquent behaviour, and prosocial behaviour as dependent variables. For each, there was a main effect of sex, F(1, 776–793) = 20.00– 31.74, p < .001. Compared with boys, girls reported higher levels of commitment to personal values (d = 0.35, CI = 0.20–0.49) and prosocial behaviour (d = 0.30, CI = 0.15–0.44), and lower levels of delinquent behaviour (d = 0.40, CI = 0.26–0.54). There were no statistically significant main effects or interactions involving parental protectiveness. APIM Analyses for Dyads Distinguished on the Basis of Parental Protectiveness Inter-individual influence Results of the APIM analysis for friends distinguished on the basis of parental protectiveness (n = 399 dyads) are depicted in Figure 2. Actor paths indicated that an adolescent’s commitment to personal values predicted his or her own prosocial behaviour. Higher levels of personal values were associated with greater prosocial behaviour among friends who were more protected by parents (β = .26) and among friends who were less protected by parents (β = .31). Actor paths also indicated that an adolescent’s commitment to personal values predicted his or her own delinquent behaviour, but only for more protected friends. Higher levels of commitment to personal values were associated with fewer delinquent behaviours for friends who were more protected by parents (β = .16). The association between commitment to personal values and delinquent behaviour failed to reach conventional levels of statistical significance (β = .07, p = .17). Figure 3 depicts these associations for more and less protected friends. Intra-individual influence Partner paths revealed that the friend who was more protected by parents was positively influenced by the friend who was less protected by parents, but not the reverse. The personal values of the less protected friend were inversely associated with the prosocial behaviour of the more protected friend (β = .11), such that greater commitment to personal values in the friend who was less protected by parents predicted fewer delinquent behaviours in the friend who was more protected by parents. The personal values of the more protected friend were not associated with the delinquent behaviour of the less protected friend (β = .04). Partner paths also indicated that the personal values of the less protected friend were positively associated with the prosocial behaviour of more protected friends (β = .09), such that greater Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

B. Laursen et al.

Figure 2. Friend influence on prosocial behaviour and delinquent behaviour as a function of commitment to personal values for partners distinguished on the basis of relative parental protectiveness. Note: N = 798 adolescents in 399 dyads. Standardized estimates presented with 95% confidence intervals in brackets. *p < .05, **p < .01.

Prosocial Behavior of More Protected Friend Prosocial Behavior of Less Protected Friend

2

Delinquent Behavior of More Protected Friend

Prosocial and Delinquent Behavior

Delinquent Behavior of Less Protected Friend

1.5

1

0.5

0 (-1 SD)

Mean

(+1 SD)

Commitment to Personal Values

Figure 3. Intra-individual influence: Prosocial and delinquent behaviour as a function of commitment to personal values among the friend reporting more parental protectiveness and among the friend reporting less parental protectiveness. Note: N = 798 participants in 399 friend dyads. Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

Parental Protectiveness

commitment to personal values in the friend who was less protected by parents predicted higher prosocial behaviours in the friend who was more protected by parents. The association between the personal values of the less protected friend and the delinquent behaviour of the more protected friend failed to reach conventional levels of statistical significance (β = .08, p = .12). Figure 4 depicts the statistically significant associations between the less protected friend’s commitment to personal values and the more protected friend’s prosocial and delinquent behaviours.

Supplemental Analyses A similar pattern of statistically significant results emerged when dyads whose perceived parental protectiveness differed by less than .50 SD (n = 82) were included in the analyses. Multiple group APIM analyses were conducted using sex as a moderator. There were no statistically significant χ 2 differences, suggesting that the influence of less protected friends and more protected friends did not vary for boys and girls. Additional analyses were conducted to rule out a series of alternative explanations. Parenting practices (warmth, control, solicitation of information), parental protectiveness, adolescent age, adolescent personality (openness, conscientiousness, extraversion, agreeableness, and neuroticism), and friend nomination rank were separately entered into the model as control variables. In each case, the pattern of statistically significant results did not change.

Prosocial Behavior

Prosocial and Delinquent Behavior of More Protected Friend

2

Delinquent Behavior

1.5

1

0.5

0 (-1 SD)

Mean

(+1 SD)

Commitment to Personal Values of Less Protected Friend

Figure 4. Inter-individual influence: Prosocial and delinquent behaviour among the friend reporting more parental protectiveness as a function of the commitment to personal values of the friend reporting less parental protectiveness. Note: N = 798 participants in 399 friend dyads. Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

B. Laursen et al.

DISCUSSION As expected, commitment to personal values was positively associated with adolescent prosocial behaviour and negatively associated with adolescent delinquent behaviour, within and between friends. Unique to this study is the finding that protective parenting promotes positive peer influence. Adolescents who perceived more parental protectiveness were influenced by their friend’s commitment to personal values, whereas adolescents who perceived less parental protectiveness were not. The findings are consistent with the notion that the importance of parents as socialization agents may extend beyond their direct effects on children and into the opportunities they create for peer influence. Restrictions to autonomy are typically assumed to be negative because they increase adolescent susceptibility to adverse peer influence. The present study suggests that parental restrictiveness is not altogether bad because it may also increase susceptibility to positive peer pressure. How does a commitment to personal values shape behaviour? A cognitive developmental perspective emphasizes the role of moral development, implying that moral behaviours are a reflection of moral judgments (Kohlberg, 1984). Within individuals, personal values should promote prosocial behaviour and inhibit antisocial behaviour. However, personal values should also restrict peer influence, as the individual comes to believe that right and wrong are not a matter of opinion. Moral identity has been proposed as a complementary source of influence. In this model, moral behaviour should vary as a function of the degree to which morality and moral behaviour are central to self-concept (Aquino & Reed, 2002). Here too, personal values should be related to individual behaviour but not to friend behaviour, particularly for those who hold moral behaviour to be a central part of their identity. In contrast, social cognitive theories hold that moral agency is susceptible to peer influence (Bandura, 2001). Moral thought and moral behaviour can be shaped by perceptions of collective agency, which are transmitted by others. Findings that friend commitment to personal values influence adolescent prosocial and delinquent behaviour suggest support for some form of collective agency, a conclusion that must be tempered by an awareness of the limitations of our cross-sectional data. Finally, there is the notion of commitment. Greater levels of commitment, to whatever cause, imply a prioritization of that cause. Being committed to personal values suggests being committed to their implementation. Care should be taken in extending our findings to moral influence because commitment to personal values does not necessarily correspond to a particular level of moral development, only to a desire to see that the values are upheld. The findings imply that parenting provides an important backdrop for the interpersonal application of values. Protectiveness may be perceived as overbearing, and it may prevent age-appropriate exploration of risks, both of which may prove detrimental to development. In this sense, protectiveness may be counterproductive, because although it may reduce the risks to which adolescents are exposed, it may also drive a wedge between parents and adolescent children, who see to renegotiate the terms of their relationship (Smetana, 1988). Successful negotiation of parent–child conflict is key to the development of adolescent autonomy (Allen et al., 2006). Thus, the paradox is that parental attempts to reduce child risk through protectiveness may actually increase them because they preclude the necessary exchanges that foster a unique identity and expressions of autonomy necessary to resist peer influence. Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

Parental Protectiveness

A case can also be made for the notion that elevated susceptibility to peers stems from positive features of parental protection. Parent protection may signal a level of involvement that facilitates the development of a self-system that emphasizes communalism (Kasser et al., 1995). Those with an affiliative self-system may be amenable to principled messages from others, particularly if they reinforce an existing worldview. Attachment theorists make a similar point, suggesting that adolescents who are securely attached to parents are well-positioned to meet well-adjusted agemates and profit from constructive peer relationships (Allen & Land, 1999). Both theories imply individual differences in susceptibility to friends, predicated on individual differences in the quality of relationships with parents. In each case, the safety afforded by parental protection provides adolescents with opportunities for open engagement with agemates. Adolescents who perceive that they are protected by parents may be inclined to explore new opportunities, assuming that family members will shield them from unanticipated trouble. Finally, adolescents with engaged parents are likely to select friends who share affiliative values (Mounts, 2000). Thus, the ties between personal values and positive adjustment should be reinforced by friends who share the same inclinations. Is parental protectiveness good or bad? Perhaps both. It is not difficult to imagine that some youth may view protectiveness as simultaneously positive and negative, an unwelcome burden and a display of affection. We found that age did not alter patterns of friend influence. Early and late adolescents with protective parents were equally susceptible to influence from friends. Across the adolescent years, peer influence over risk taking behaviour tends to peak around age 14 years, waning gradually in the years that follow (Steinberg, 2004). Many assume that peer influence declines because adolescents become more secure in their individual identities (Hafen et al., 2012). If so, our findings imply that protectiveness may be increasingly inappropriate across the adolescent years, not just because it interferes with autonomy development but because it may prolong a period of heightened susceptibility to external forces. The developmental processes postulated move beyond simple parent-driven models. Parents certainly play a role in shaping adolescent values and adjustment outcomes, but as children grow up, parent contributions may shift to a more subtle but equally important role in shaping the child’s perceptions of external forces and his or her openness to peer influence. Greater attention should be given to parents as moderators of peer experiences, for parents clearly have the potential to buffer or exacerbate the effects of friends. In statistical terms, the action is in the interaction. But there is more. We know that the quality of the parent–child relationship at the outset of adolescence anticipates the impact that relationship partners have on relationship trajectories (Laursen, DeLay, & Adams, 2010). Thus, protectiveness may signal high quality relationships with engaged parents or poor quality relationships with increasingly alienated children. Future studies of parents as moderators of peer influence must balance absolute levels of a behaviour against relative levels of a behaviour (compared with peers) and against adolescent perceptions of the overall quality of the affiliation. It is a tall order, but necessary if we are to disentangle parenting behaviours from parent–child relationship quality. This investigation is not without limitations. Of primary concern is the crosssectional nature of the data. The direction of influence, from personal values to adolescent behaviour, is assumed, but causality is not. Longitudinal data are required to determine whether personal values anticipate changes in prosocial behaviour and delinquent behaviour or whether problem behaviour may cause one to waiver in commitment to personal values. In the present study, we made the a priori Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

B. Laursen et al.

assumption that our model measured susceptibility to influence. This assumption can (and should) be tested over time, with stable friends, to determine if influence varies according to characteristics of the influencing agent, the influence recipient, or both (e.g. Marion et al., 2014). Friends select one another on the basis of similar attributes, a tendency that is stronger for observable attributes such as delinquent behaviour than for attitudes, values, and motives (Hafen et al., 2011). Similarity on any behaviour at any particular time point should not be conflated with influence, which can only be measured by changes in similarity. Of additional concern is the fact that personal values may be correlated with other factors that predict adolescent outcomes. Supplemental analyses that control for the contribution of potential confounds, such as relative age, personality, and parental warmth and control, ameliorate, but do not eliminate, this concern. We know that influence within a friendship differs as a function of individual characteristics, such as relative acceptance (Laursen et al., 2012) and relative selfconfidence (DeLay et al., 2013). Although these characteristics are unlikely to account for differences between friends distinguished on the basis of parental protectiveness, other attributes correlated with parenting (e.g. birth order) may well prove important. Finally, our outcome variables were derived solely from self-reports. Objective measures of outcomes have the advantage of avoiding problems arising from shared reporter variance. Nonindependent data can pose a significant threat to the validity of research on influence. The APIM strategy is a powerful and flexible tool for examining influence within close relationships, one that permits data from both partners to be included in the same analyses. The APIM)is an important new tool that overcomes many of these limitations. We have demonstrated how to partition influence between partners from influence within partners, removing the inter-individual similarity that introduces bias into conventional analytic procedures. We have also demonstrated how to test for moderated effects, comparing influence paths across groups of participants (e.g. boys and girls). Finally, we have demonstrated how to rule out alternative explanations through the use of control variables. Nonindependent data pose a challenge to family researchers. The challenge is particularly acute when the focus is on parenting outcomes, because the research premise (i.e. the presence of influence) presumes a lack of statistical independence. One goal of the present investigation was to encourage scholars to apply innovative APIM methodology to the study of parent–child relationships. The APIM can be adapted to address different conceptual models. In the present study, we examined how patterns of friend influence varied as a function of parent–child relationships, using parenting as a moderator. Alternatively, APIM analyses can be used to examine influence within the parent–child relationship (e.g. Cook & Kenny, 2005), directly measuring the strength of influence from parent to child and from child to parent. With recent modifications, the APIM can incorporate longitudinal data to measure influence over behavioural change (Popp, Laursen, Kerr, Stattin, & Burk, 2008). We have demonstrated that friend influence varies as a function of relative parental protectiveness, such that protectiveness fosters conditions in which one friend’s commitment to personal values promotes prosocial behaviour and inhibits delinquent behaviour in the other friend. In so doing, we move beyond models that emphasize the direct role that parents play in shaping adolescent outcomes, focusing instead on parenting as a contextual variable that contributes to individual differences in susceptibility to peer influence. It may well be that during the adolescent years, parents have less of a direct role in shaping outcomes but nevertheless remain important, moderating socialization outcomes by shaping Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

Parental Protectiveness

the perceptions and attitudes that frame how adolescents approach an expanding social world.

ACKNOWLEDGEMENTS This study was supported by a grant to Rita Žukauskienė from the European Social Fund under the Global Grant measure (VP1-3.1-ŠMM-07-K-02-008). Brett Laursen received support for the preparation of this manuscript from the US National Institute of Child Health and Human Development (HD068421) and the US National Science Foundation (0909733).

REFERENCES Achenbach, T. M., & Rescorla, L. A. (2001). Manual for the ASEBA School-Age Forms & Profiles. Burlington, VT: University of Vermont. Allen, J. P., Chango, J., Szwedo, D., Schad, M., & Marston, E. (2012). Predictors of susceptibility to peer influence regarding substance use in adolescence. Child Development, 83, 337–350. Allen, J. P., Insabella, G., Porter, M. R., Smith, F. D., Land, D., & Phillips, N. (2006). A socialinteractional model of the development of depressive symptoms in adolescence. Journal of Consulting and Clinical Psychology, 74, 55–65. Allen, J. P., & Land, D. (1999). Attachment in adolescence. In J. Cassidy, & P. R. Shaver (Eds.), Handbook of attachment (pp. 319–335). New York: Guilford. Aquino, K., & Reed, II, A. (2002). The self-importance of moral identity. Journal of Personality and Social Psychology, 83(6), 1423–1440. Arrindell, W. A., Gerlsma, C., Vandereycken, W., Hageman, W. J. J. M., & Daeseleire, T. (1998). Convergent validity of the dimensions underlying the parental bonding instrument (PBI) and the EMBU. Personality and Individual Differences, 24(3), 341–350. Arrindell, W., Richter, J., Eisemann, M., Gärling, T., Rydén, O., Hansson, S., & Gustafsson, M. (2001). The short-EMBU in East-Germany and Sweden: A cross-national factorial validity extension. Scandinavian Journal of Psychology, 2(2), 157–160. Bandura, A. (2001). Social cognitive theory: An agentic perspective. Annual Review of Psychology, 52(1), 1–26. Barber, B. K. (1996). Parental psychological control: Revisiting a neglected construct. Child Development, 67(6), 3296–3319. Barry, C. M., & Wentzel, K. R. (2006). Friend influence on prosocial behavior: The role of motivational factors and friendship characteristics. Developmental Psychology, 42(1), 153–163. Brown, B. B., & Bakken, J. P. (2011). Parenting and peer relationships: Reinvigorating research on family–peer linkages in adolescence. Journal of Research on Adolescence, 21(1), 153–165. Collins, W. A., Maccoby, E. E., Steinberg, L., Hetherington, E. M., & Bornstein, M. H. (2000). Contemporary research on parenting: The case for nature and nurture. American Psychologist, 55(2), 218–232. Cook, W. L., & Kenny, D. A. (2005). The Actor–Partner Interdependence Model: A model of bidirectional effects in developmental studies. International Journal of Behavioral Development, 29(2), 101–109. Costa, Jr., P. T., & McCrae, R. R. (2010). Bridging the gap with the five-factor model. Personality Disorders, 1(2), 127–130. Deković, M., ten Have, M., Vollebergh, W. A. M., Pels, T., Oosterwegel, A., Wissink, I. B., … Ormel, J. (2006). The cross-cultural equivalence of parental rearing measure: EMBU-C. European Journal of Psychological Assessment, 22(2), 85–91. DeLay, D., Hartl, A. C., Laursen, B., Denner, J., Werner, L., Campe, S., & Ortiz, E. (2013). Learning from friends: Measuring influence in a dyadic computer instructional setting. Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

B. Laursen et al.

International Journal of Research & Method in Education. Advance online publication. doi: 10.1080/1743727X.2013.784961 Dishion, T. J., & Piehler, T. F. (2009). Deviant by design: Peer contagion in development, interventions, and schools. In K. H. Rubin, W. M. Bukowski, & B. Laursen (Eds.), Handbook of peer interactions, relationships, and groups (pp. 589–602). New York: Guilford. Eltringham, S., & Aldridge, J. (2002). Parenting on shifting sands: The transfer or responsibility for safely managing danger. Clinical Child Psychology and Psychiatry, 7, 137–145. Frank, K., Muller, C., Schiller, K. S., Riegle-Crumb, C., Mueller, A. S., Crosnoe, R., & Pearson, J. (2008). The social dynamics of mathematics course-taking in high school. American Journal of Sociology, 113, 1645–1696. Fulgini, A., & Eccles, J. S. (1993). Perceived parent–child relationship and early adolescent orientation toward peers. Developmental Psychology, 9, 622–632. Gasser, L., & Malti, T. (2012). Children’s and their friends’ moral reasoning: Relations with aggressive behavior. International Journal of Behavioral Development, 36(5), 358–366. Gauze, C., Bukowski, W. M., Aquan-Assee, J., & Sippola, L. K. (1996). Interactions between family environment and friendship and associations with self-perceived well-being during early adolescence. Child Development, 67(5), 2201–2216. Gerlsma, C., Arrindell, W., Van der Veen, N., & Emmelkamp, P. M. G. (1991). A parental rearing style questionnaire for use with adolescents: Psychometric evaluation of the EMBU-A. Journal of Personality and Individual Differences, 12(12), 1245–1253. Hafen, C. A., Laursen, B., Burk, W. J., Kerr, M., & Stattin, H. (2011). Homophily in stable and unstable adolescent friendships: Similarity breeds constancy. Personality and Individual Differences, 51, 607–612. Hafen, C. A., Laursen, B., & DeLay, D. (2012). Transformations in friend relationships across the transition into adolescence. In B. Laursen, & W. A. Collins (Eds.), Relationships pathways: From adolescence to young adulthood (pp. 69–89). Thousand Oaks, CA: Sage. Hart, D., & Fegley, S. (1995). Prosocial behavior and caring in adolescence: Relations to selfunderstanding and social judgment. Child Development, 66(5), 1346–1359. Kasser, T., Ryan, R. M., Zax, M., & Sameroff, A. J. (1995). The relations of maternal and social environments to late adolescents’ materialistic and prosocial values. Developmental Psychology, 31, 907–914. Kendler, K. S. (1996). Parenting: A genetic–epidemiologic perspective. The American Journal of Psychiatry, 153(1), 11–20. Kenny, D. A. (1996). Model of interdependence in dyadic research. Journal of Social and Personal Relationships, 13, 279–294. Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). The analysis of dyadic data. New York: Guilford. Kohlberg, L. (1984). Essays on moral development: Vol. 2. The psychology of moral development. New York: Harper & Row. Larson, R. W., Richards, M. H., Moneta, G., Holmbeck, G., & Duckett, E. (1996). Changes in adolescents’ daily interactions with their families from ages 10 to 18: Disengagement and transformation. Developmental Psychology, 32(4), 744–754. Laursen, B. (2005). Dyadic and group perspectives on close relationships. International Journal of Behavioral Development, 29, 97–100. Laursen, B., & Collins, W. A. (2009). Parent–child relationships during adolescence. In R. M. Lerner, & L. Steinberg (Eds.), Handbook of adolescent psychology (pp. 3–42). Hoboken, NJ: Wiley. Laursen, B., DeLay, D., & Adams, R. E. (2010). Trajectories of perceived support in motheradolescent relationships: The poor (quality) get poorer. Developmental Psychology, 46(6), 1792–1798. Laursen, B., Hafen, C. A., Kerr, M., & Stattin, H. (2012). Friend influence over adolescent problem behaviors as a function of relative peer acceptance: To be liked is to be emulated. Journal of Abnormal Psychology, 121(1), 88–94. Laursen, B., & Mooney, K. S. (2008). Relationship network quality: Adolescent adjustment and perceptions of relationships with parents and friends. American Journal of Orthopsychiatry, 78(1), 47–53. Laursen, B., Popp, D., Burk, W. J., Kerr, M., & Stattin, H. (2008). Incorporating interdependence into developmental research: examples from the study of homophily and Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd

Parental Protectiveness

homogeneity. In N. A. Card, J. P. Selig, & T. D. Little (Eds.), Modeling dyadic and interdependent data in the developmental and behavioral science (pp. 11–37). New York, NY: Routledge. Lerner, R. M., Lerner, J. V., Almerigi, J. B., Theokas, C., Phelps, E., Gestsdottir, S., … von Eye, A. (2005). Positive youth development, participation in community youth development programs and community contributions of fifth-grade adolescents: Findings from the first wave of the 4-H study of positive youth development. The Journal of Early Adolescence, 34(4), 17–71. Marion, D., Laursen, B., Kiuru, N., Nurmi, J. E., & Salmela-Aro, K. (2014). Maternal affection moderates friend influence on schoolwork engagement. Developmental Psychology, 50, 766–771. Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674–701. Mounts, N. S. (2000). Parental management of adolescent peer relationships: What are its effects on friend selection? In K. A. Kerns, J. M. Contreras, & A. M. Neal-Barnett (Eds.), Family and peers: Linking two social worlds (pp. 169–193). Westport, CT: Praeger. Mounts, N. S., & Steinberg, L. (1995). An ecological analysis of peer influence on adolescent grade point average and drug use. Developmental Psychology, 31(6), 915–922. Muris, P., van Brakel, A. M. L., Arntz, A., & Schouten, E. (2011). Behavioral inhibition as a risk factor for the development of childhood anxiety disorders: A longitudinal study. Journal of Child and Family Studies, 20(2), 157–170. Muthén, L. K., & Muthén, B. O. (1998–2010). Mplus user’s guide, 6. Los Angeles: Muthén & Muthén. Padilla-Walker, L. M., & Carlo, G. (2007). Personal values as a mediator between parent and peer expectations and adolescent behaviors. Journal of Family Psychology, 21(3), 538–541. Popp, D., Laursen, B., Kerr, M., Stattin, H., & Burk, W. J. (2008). Modeling homophily over time with an actor–partner interdependence model. Developmental Psychology, 44, 1028–1039. Sebanc, A. M. (2003). The friendship features of preschool children: Links with prosocial behavior and aggression. Social Development, 12(2), 249–268. Simons, R. L., Whitbeck, L. B., Conger, R. D., & Conger, K. J. (1991). Parenting factors, social skills, and value commitments as precursors to school failure, involvement with deviant peers and delinquent behavior. Journal of Youth and Adolescence, 20(6), 645–664. Small, S. A., & Kerns, D. (1993). Unwanted sexual activity among peers during early and middle adolescence: Incidence and risk factors. Journal of Marriage and the Family, 55(4), 941–952. Smetana, J. (1988). Adolescents’ and parents’ conceptions of parental authority. Child Development, 59, 321–335. Steinberg, L. (2004). Risk taking in adolescence: What changes, and why? Annals of the New York Academy of Sciences, 1021, 51–58. Steinberg, L., & Silk, J. S. (2002). Parenting adolescents. In M. H. Bornstein (Ed.), Handbook of parenting: Vol. 1. Children and parenting (pp. 103–133). Mahwah, NJ: Erlbaum. Steinberg, L., & Silverberg, S. B. (1986). The vicissitudes of autonomy in early adolescence. Child Development, 57, 841–851. Vitaro, F., Brendgen, M., Ladouceur, R., & Tremblay, R. E. (2001). Gambling, delinquency, and drug use during adolescence: Mutual influences and common risk factors. Journal of Gambling Studies, 17(3), 171–190.

Copyright © 2014 John Wiley & Sons, Ltd.

Inf. Child. Dev. (2014) DOI: 10.1002/icd