Journal of Adolescence 36 (2013) 685–693

Contents lists available at SciVerse ScienceDirect

Journal of Adolescence

journal homepage: www.elsevier .com/locate/ jado

Secrets from friends and parents: Longitudinal links withdepression and antisocial behavior

Robert D. Laird*, Brittanee J. Bridges, Monica A. MarseeUniversity of New Orleans, Dept. of Psychology, GP 2001, New Orleans, LA 70148, USA

Keywords:SecrecyFriendshipParentingDepressionAntisocial behavior

* Corresponding author. Tel.: þ1 504 280 5454; fE-mail addresses: rlaird@uno.edu (R.D. Laird), bj

0140-1971/$ – see front matter � 2013 The Foundahttp://dx.doi.org/10.1016/j.adolescence.2013.05.001

a b s t r a c t

Keeping secrets from parents is associated with depression and antisocial behavior. Thecurrent study tested whether keeping secrets from best friends is similarly linked tomaladjustment, and whether associations between secrecy and maladjustment aremoderated by the quality of the friendship. Adolescents (N ¼ 181; 51% female, 48% white,non-Hispanic, 45% African American) reported their secrecy from parents and best friends,the quality of their parent–adolescent relationships and best friendships, and theirdepression and antisocial behavior at ages 12 and 13. Keeping more secrets from bestfriends was associated with more depression, but not with more antisocial behavior, whencontrolling for earlier adjustment, secrecy from parents, and the quality of the friendship.For girls associations between maladjustment and secrecy were conditioned by the qualityof the relationships and whether secrets were kept from parents and friends. Discussionargues for expanding the study of secrecy in adolescence beyond the parent–child dyad.� 2013 The Foundation for Professionals in Services for Adolescents. Published by Elsevier

Ltd. All rights reserved.

Keeping secrets is typically expected to be harmful for the secret-keeper (see Finkenauer, Engles, & Meeus, 2002).Although no one is expected to share everything about their lives with everyone they meet, keeping secrets in close sup-portive relationships may be particularly problematic. In adolescence, parent–child relationship and friendships are oftenclose supportive relationships (e.g., de Goede, Branje, Delsing, & Meeus, 2009), yet adolescents are known to keep secretsfrom both parents and friends (e.g., Finkenauer et al., 2002; Frijns, Finkenauer, & Keijsers, 2013; Merten, 1999). Keepingsecrets from parents predicts greater depression and antisocial behavior problems (Frijns, Finkenauer, Vermulst, & Engels,2005; Keijsers, Branje, Frijns, Meeus, & Finkenauer, 2010; Laird & Marrero, 2010), but it is unknown whether keeping se-crets from best friends also predicts maladjustment. The purpose of the current study was to determine whether keepingsecrets from best friends is associated concurrently and longitudinally with depression and antisocial behavior.

In contrast to non-disclosure, which may be a passive failure to share informationwith another (see Bumpus & Hill, 2008),secret-keeping requires the active withholding of information from another person (Frijns & Finkenauer, 2009). As such,keeping secrets may be difficult and stressful and secret-keeping is thought to increase the likelihood of experiencing stress-related or other psychological problems (Pennebaker, 1989, 1997). Although some researchers postulate that withholdingsecrets from parents may promote a sense of autonomy for the adolescent and be a part of their healthy development (e.g.,Finkenauer et al., 2002), the majority of research has emphasized the negative aspects of keeping secrets from parents.Keeping secrets from parents during adolescence is associated with more depression, antisocial behavior, and loneliness(Finkenauer et al., 2002; Frijns et al., 2005; Keijsers et al., 2010; Laird & Marrero, 2010). Although studies have shown that

ax: þ1 504 280 6049.bridge@my.uno.edu (B.J. Bridges), mmarsee@uno.edu (M.A. Marsee).

tion for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved.

R.D. Laird et al. / Journal of Adolescence 36 (2013) 685–693686

adolescents do keep secrets from their friends (Merten,1999; see also Dolgin & Kim,1994), and that having a secret not sharedwith anyone predicts psychosocial problems that are relieved when the secret is shared (Frijns & Finkenauer, 2009; Frijnset al., 2013), we are not aware of any studies that have directly tested whether keeping secrets from best friends is associ-ated with maladjustment. Generalizing from previous studies documenting links between secrecy and maladjustment, ad-olescents who report keeping more secrets from their best friends are expected to experience more depression and antisocialbehavior problems than adolescents who keep fewer secrets from their best friends.

Keeping secrets from parents and best friends

Keeping secrets from parents and best friends may be linked to negative psychosocial outcomes through several differentprocesses and each process suggests a unique pattern of multivariate associations linking secrecy and outcomes. Keepingsecrets from parents and friends may severely restrict adolescents’ access to instrumental and emotional support (e.g., Frijnset al., 2005), particularly in reference to the topic kept secret. To the extent that secret-keeping is linked to negative outcomesthrough the inability to access social support, secrecy from parents and secrecy from best friends are likely to have uniqueadditive or interactive effects on depression and antisocial behavior. Specifically, if parents and best friends are independentsources of social support, secrecy from parents and secrecy from best friends may be additive in their associations withpsychosocial outcomes. However, if parents and best friends can serve as independent sources of social support (Helsen,Vollebergh, & Meeus, 2000), such that an adolescent who does not have access to support through one relationship canrely instead on the other relationship for support, secrecy from parents and secrecy from best friends may interact to predictpsychosocial outcomes such that poor outcomes will only occur when adolescents keep secrets from both parents and bestfriends.

Secret-keeping also may be an indicator of a dispositional tendency to keep secrets (see Larson & Chastain, 1990) or ofdifficulty trusting others. To the extent that keeping secrets from parents and friends originates in an adolescent’s tendency tobe secretive or distrustful of others, secrecy from parents and secrecy from friends are likely to have redundant effects, suchthat secrecy from friends should not be uniquely associated with depression and antisocial behavior after controlling forsecrecy from parents (and vice-versa).

Finally, keeping secrets may be an indicator of involvement in undesirable behavior that the adolescent wishes to hidefrom parents or friends (Bakken & Brown, 2010; Smetana, Villalobos, Rogge, & Taspoulos-Chan, 2010). However, it is likely thatthe nature of the behaviors that an adolescent wishes to hide from parents may differ substantially from the nature of thebehaviors that an adolescent wishes to hide from a best friend. For example, behaviors or interests that may expose one tosocial ridicule may bemore likely to be kept secret from friends than from parents. In contrast, involvement in common formsof adolescent misbehavior, such as smoking or drinking alcohol, may be kept secret from parents but not from best friends. Tothe extent that secrecy is used to hide misbehavior, secrecy from parents but not from best friends may be associated withmaladjustment.

Secrecy and relationship quality

An adolescent’s decision to keep a secret from an individual is likely influenced by the quality of the relationship withthe individual. Adolescents, and adolescent girls in particular (Keijsers et al., 2010), experiencing lower quality parent–childrelationships are more likely to keep secrets from their parents than adolescents experiencing higher quality parent–childrelationships (Laird & Marrero, 2010; Smetana, Metzger, Gettman, & Campione-Barr, 2006). It is also likely that adolescentswith lower quality friendships are more likely to keep secrets from their best friends (Dolgin & Kim, 1994; Merten, 1999). Tothe extent that secret-keeping is a reflection of the quality of the relationship, keeping secrets from parents and bestfriends may not be associated with negative psychosocial outcomes when controlling for the quality of the parent–childrelationship or friendship. However, it is also possible that secrecy is a more proximal contributor to negative psychosocialoutcomes than relationship quality and that secrecy accounts for the link between relationship quality and the outcomes.Furthermore, it is possible that the links between secret-keeping and outcomes are moderated by relationship quality. Forexample, in the context of a high quality relationship, secrecy may be more strongly associated with negative outcomesbecause the secret-keeper experiences additional stress from keeping the secret, or, in contrast, secrecy may be lessstrongly associated with negative outcomes because the high quality relationship compensates for the effect of keeping thesecret.

In summary, the purpose of the current study was to test secrecy from best friends as a concurrent and longitudinalpredictor of depression and antisocial behavior problems during early adolescence. Analyses tested whether keepingsecrets from parents and best friends were redundant, additive, or interactive predictors of depression and antisocialbehavior. For both secrets from parents and secrets from best friends, analyses tested whether keeping secrets was aunique predictor after controlling for relationship quality and also tested whether keeping secrets interacted withrelationship quality to predict depression and antisocial behavior. Because previous studies have documented sex dif-ferences in secret-keeping (e.g., Frijns & Finkenauer, 2009), psychosocial outcomes, or in the links between secret-keeping and psychosocial outcomes (e.g., Keijsers et al., 2010), analyses controlled for sex and tested sex as a moder-ator of all associations.

R.D. Laird et al. / Journal of Adolescence 36 (2013) 685–693 687

Method

Participants

Participants were part of a larger three-wave longitudinal study of 218 families (see Laird & Marrero, 2010). Data from thesecond and third annual waves of the longitudinal study were used in the current study, because secrecy from parents andbest friends was not assessed in the first wave. Thirty-six participants dropped out of the study prior to the second wave, and1 participant reported not having a best friend at age 12 (and thus did not provide friendship data), resulting in a wave 2sample size of 181. Participants were interviewed during the summers following their 6th (n ¼ 181, M age ¼ 12.4, SD ¼ .53)and 7th (n ¼ 148,M age¼ 13.4, SD ¼ .56) grade school years, hereafter referred to as ages 12 and 13, respectively. The samplewas 50.5% female and 74.4% of the adolescents lived in a two-parent home when the data were collected. Adolescents self-identified their ethnicity as white, non-Hispanic (48.4%), African-American (45.1%), or other (6.6%). Mothers’ educational levelvaried with 2.8% not having completed high school, 10.5% having a high school diploma, 37.0% having attended college ortechnical school, 29.8% having a bachelor’s degree, and 19.9% having a graduate degree.

Girls, c2(N ¼ 182, df¼ 1) ¼ 6.72, p ¼ .01, African Americans, and participants of other ethnicities c2(N ¼ 182, df¼ 2) ¼ 5.88,p¼ .053,weremore likely to drop out of the study following the age 12 interviews thanwere boys or white participants. T-testsshowed that the dropouts had marginally higher quality friendships at age 12, t(179)¼ 1.96, p ¼ .052, than did completers butdropouts did not differ from completers on any other of the age 11 or 12 variables included in this report (ps ¼ .27–.90).

Procedure

Participating families were recruited from a public school system serving the citizens of Baton Rouge, LA in the southernUnited States (population about 400,000). Information letters were distributed to students through schools in the late spring.Participants were compensated $70 per family at age 12 and $90 per family at age 13. Families were provided an overview ofthe structured interview procedure before mothers provided informed consent and adolescents provided written assent.Following the consent discussion, the adolescents were interviewed in a private location within the home. To ensure privacyand reduce social desirability bias, adolescents privately recorded their responses to the questions on a forced-choice answersheet. The structured interviews took about 45 min. Most of the items focused on parent–adolescent interactions and re-lationships. One section of the structured interview focused on relationships with best friends. At the start of the best friendsection, the adolescents were told that the next few sets of items asked about their best friend. If adolescents reported thatthey did not have a best friend, they were instructed to skip the questions. If adolescents reported that they hadmultiple bestfriends, they were instructed to think about the ‘‘best of the best friends’’ when answering the questions.

Measures

SecrecyTo assess adolescents’ secrecy from their mothers and best friends, adolescents completed Larson and Chastain’s (1990)

10-item Self-Concealment Scale twice, once in reference to secrecy from mothers and once in reference to secrecy from bestfriendsduring the age 12 and age 13 interviews. Sample items are: “There are lots of things aboutme that I keep frommymother.”and “I’m often afraid I’ll reveal something to my best friend that I don’t want to.” Adolescents reported the extent to which theyagreedwith each statement using a five-point scale (1¼ strongly disagree to 5¼ strongly agree). Secrecy from parents (as¼ .91 &.90, for ages 12 and 13, respectively) and best friends (as ¼ .89 & .90) were scored as the means of the respective 10 items.

Friendship qualityThe quality of the participants’ current best friendship at ages 12 and 13 was reported using selected items from the

Friendship Quality Scale (FQS: Bukowski, Hoza, & Boivin, 1994). To minimize the length of the interview, a shortened versionof the FQS (i.e., three highest loading items from each of the Help, Security, and Closeness subscales per Bukowski et al., 1994)was used to assess the quality of the friendship (e.g., ‘‘If other kids were bothering me, my friend would help me’’). Ado-lescents responded to the questions about their current best friend on a five-point scale (1¼ never to 5¼ always). An index offriendship quality was computed as the mean of the nine items at ages 12 and 13 (as ¼ .91 & .92).

Parent–child relationship qualityParental acceptance served as the index of parent–child relationship quality. Adolescents responded to 14 items from the

Children’s Report of Parental Behavior Inventory (CRPBI; Schaefer, 1965; Schludermann & Schludermann, 1988) during theinterviews at ages 12 and 13. Adolescents reported how much each item (e.g., “My mother tells me she loves me”) reflectedtheir mother’s behavior on a 5-point scale (1¼ not at all like her to 5¼ a lot like her). Parent–child relationship quality scores atages 12 and 13 were computed as the mean of the respective 14 items (both as ¼ .93).

DepressionAdolescents reported their depressed mood at ages 11, 12, and 13 using the six-item Modified Depression Scale (Orpinas,

1993). Each item (e.g. “In the last month, how often were you very sad?”) was scored on a five-point scale (1 ¼ never to

R.D. Laird et al. / Journal of Adolescence 36 (2013) 685–693688

5 ¼ always). The mean of the six items was computed to index depression (as ¼ .75, .74, & .71, at ages 11, 12, & 13,respectively).

Antisocial behaviorAdolescents reported the frequency of engagement in antisocial behavior at ages 11, 12, and 13 using 26 items from the

Problem Behavior Frequency Scale (Farrell, Kung, White, & Valois, 2000). The items (e.g., “In the last month, howmany timesdid you steal from someone?”) assess physical and non-physical aggression, delinquency, and drug use. Each itemwas scoredon a five-point scale from never (scored 0) to seven or more times (scored 4). The mean of the 26 items was computed to indexbehavior problems (as ¼ .90, .93, & .91, at ages 11, 12, and 13, respectively).

Results

Descriptive statistics and bivariate correlations are shown in Table 1. More secrecy from best friends was associated withmore secrecy from parents, lower friendship quality, more depression, and more antisocial behavior. Boys reported moresecrecy from friends than did girls. More secrecy from parents was associated with lower parent–child relationship quality,more depression, and more antisocial behavior.

A set of six polynomial regression equations were fit to test the primary hypotheses in cross-sectional analyses at ages 12and 13 (controlling depression and behavior problems one year earlier) and in longitudinal analyses from age 12 to age 13. Inthe first equation, age 12 depression was regressed on age 11 depression and the age 12 secrecy from friends, secrecy fromparents, friendship quality, and parent–child relationship quality variables. Two sets of interaction terms also were includedin the equation. The first set of three terms tested the secrets from friends� friendship quality, secrets from parents� parent–child relationship quality, and secrets from friends � secrets from parents interactions. The second set of seven terms testedall two-way and three-way interactions with sex. In the second equation, age 13 depression was regressed on age 12depression and the age 13 secrecy and relationship quality variables and interaction terms. The third equation was longi-tudinal and regressed age 13 depression on age 12 depression and the age 12 secrecy and relationship quality variables.Finally, associations with antisocial behavior were tested with an analogous set of three equations in which antisocialbehavior replaced depression.

The equations first were fit with all interaction terms. However, because including non-significant interaction terms in themodel hinders interpretation of lower-order effects (Cohen, Cohen, West, & Aiken, 2003), the equations were re-fit afterremoving all non-significant interaction terms that were not necessary for interpreting significant higher order interactions.Significant two-way and three-way interactions were interpreted by calculating simple slopes as recommended by Cohenet al. (2003) using tools developed by Preacher, Curran, and Bauer (2006).

Depression

In the cross-sectional age 12 model, more secrecy from parents was associated with more depression (Table 2). The as-sociation between secrecy from friends and depression was moderated by sex and friendship quality as shown by the

Table 1Descriptive statistics and correlations among all variables.

Variable M (SD) 1 2 3 4 5 6 7 8

1. Secrecy from friends – 12 1.96 (.76)2. Secrecy from friends – 13 1.80 (.73) .43***3. Secrecy from parents – 12 2.14 (.87) .54*** .31***4. Secrecy from parents – 13 2.35 (.94) .27*** .39*** .60***5. Friendship quality – 12 4.23 (.72) �.46*** �.33*** �.24*** �.046. Friendship quality – 13 4.25 (.76) �.19* �.36*** �.02 .04 .55***7. P–C Relation. Quality – 12 4.14 (.80) �.31*** �.15 �.47*** �.19* .34*** .19*8. P–C Relation. Quality – 13 4.00 (.86) �.17* �.14 �.31*** �.39*** .24** .32*** .61***9. Depression – 11 2.61 (.83) .27*** .18* .34*** .24** �.07 �.04 �.23** �.37***

10. Depression – 12 2.50 (.75) .46*** .19* .55*** .45*** �.16* .01 �.33*** �.22**11. Depression – 13 2.56 (.71) .04 .33*** .36*** .53*** �.02 �.05 �.21* �.29***12. Antisocial behavior – 11 1.26 (.33) .23** .15 .27*** .17* �.10 �.08 �.28*** �.30***13. Antisocial behavior – 12 1.30 (.42) .07 .18* .27*** .25** �.03 �.07 �.23** �.20*14. Antisocial behavior – 13 1.34 (.41) .07 .20* .30*** .37*** .02 �.08 �.31*** �.29***15. Sex .26*** .26*** .07 �.01 �.42*** �.44*** �.17* �.12

Variable 9 10 11 12 13 14

10. Depression – 12 .50***11. Depression – 13 .31*** .44***12. Antisocial behavior – 11 .40*** .31*** .27***13. Antisocial behavior – 12 .24*** .27*** .21* .69***14. Antisocial behavior – 13 .31*** .33*** .34*** .58*** .82***15. Sex .07 .05 �.06 .20** .12 .20*

ns ¼ 146–181. P–C Relation. ¼ Parent–child relationship. Sex is coded 0 ¼ female, 1 ¼ male. *p < .05, **p < .01, ***p < .001.

Table 2Regression coefficients from the depression models.

Variable Concurrent models Longitudinal model

Age 12 depression Age 13 depression Age 13 depression

B SE B* p B SE B* p B SE B* p

Prior depression .27 .06 .30 <.001 .26 .07 .27 .001 .36 .09 .37 <.001Sex �.13 .10 �.09 .17 �.23 .11 �.16 .033 �.23 .13 �.16 .086Parent–child relationship quality (RQ) �.07 .06 �.07 .28 �.08 .07 �.10 .22 �.03 .08 �.03 .71Friendship quality (FQ) .09 .12 .08 .47 �.04 .08 �.04 .62 .01 .09 .01 .89Secrecy from parents (SP) .28 .06 .32 <.001 .23 .07 .31 .001 .06 .12 .08 .61Secrecy from friends (SF) �.07 .11 �.07 .52 .17 .08 .18 .029 �.13 .15 �.13 .40SP � RQSF � FQ .68 .18 .46 <.001SP � SF �.35 .13 �.30 .011Sex � RQSex � FQ .01 .15 .01 .96Sex � SF .37 .14 .27 .008 �.01 .19 �.01 .97Sex � SP .12 .16 .10 .45Sex � SP � RQSex � SF � FQ �.70 .21 �.43 <.001Sex � SP � SF .42 .19 .25 .032R2 .49 .001 .38 <.001 .31 <.001

Note: Depression at age 11 served as prior depression for the age 12 analyses and depression at age 12 served as prior depression for the age 13 analyses. Age12 secrecy and relationship quality variables served as predictors in the age 12 concurrent model and in the age 13 longitudinal model. Age 13 secrecy andrelationship quality variables served as predictors in the age 13 concurrent model. Sex is coded 0 ¼ female, 1 ¼ male.

R.D. Laird et al. / Journal of Adolescence 36 (2013) 685–693 689

significant 3-way sex� secrets from friends� friendship quality interaction (see Fig.1). Girls’ greater secrecy from friendswasassociatedwith less depression at lower levels of friendship quality (b¼�.554, SE¼ .200, p< .007), but withmore depressionat higher levels of friendship quality (b ¼ .418, SE ¼ .131, p < .002). For boys, secrecy from friends was associated with moredepression at both higher and lower levels of friendship quality (bs ¼ .282 & .318, SEs ¼ .142 & .107, ps ¼ .048 & .003). In thecross-sectional age 13 model, more secrecy from parents and best friends were uniquely associated with more depression.

In the longitudinal age 13 model, the association between secrecy from friends was moderated by sex and secrecy fromparents as shown by the significant sex � secrets from parents � secrets from friends interaction (see Fig. 2). For girls, greatersecrecy from friends was associated with less depression at higher levels of secrecy from parents (b ¼ �.429, SE ¼ .139,p < .003) but not at lower levels of secrecy from parents (b ¼ .175, SE ¼ .232, p ¼ .45). For boys, secrecy from friends wasnot associatedwith depression at higher or lower levels of secrecy fromparents (bs¼�.071 &�.198, SEs¼ .158 & .179, ps¼ .66& .27).

Antisocial behavior

As shown in Table 3, there were no significant interactions in the age 12 or age 13 cross-sectional models. At age 12, moresecrecy from parents was associated with more antisocial behavior whereas more secrecy from best friends was associatedwith less antisocial behavior. At age 13, more secrecy from parents was associated with more antisocial behavior but secrecyfrom best friends was not significantly associated with antisocial behavior.

In the longitudinal age 13 model, greater friendship quality was associated with more antisocial behavior. The associationbetween secrets from parents and antisocial behavior was moderated by sex and parental acceptance as shown by the

0

0.5

1

1.5

2

2.5

3

Low High

Ag

e 1

2 D

ep

re

ss

io

n

Age 12 Secrecy from Friends

Lower FQ, Girls

Higher FQ, Girls

Lower FQ, Boys

Higher FQ, Boys

Fig. 1. Age 12 friendship quality (FQ), secrecy from friends, and sex interact to predict age 12 depression.

0

0.5

1

1.5

2

2.5

Low HighA

ge

1

3 D

ep

re

ss

io

n

Age 12 Secrecy from Friends

Lower SP, Girls

Higher SP, Girls

Lower SP, Boys

Higher SP, Boys

Fig. 2. Age 12 secrecy from parents (SP), secrecy from friends, and sex interact to predict age 13 depression.

R.D. Laird et al. / Journal of Adolescence 36 (2013) 685–693690

significant sex � secrecy from parents � parental acceptance interaction (see Fig. 3). For girls, greater secrecy from parentswas associated with less antisocial behavior at lower levels of parental acceptance (b ¼ �.184, SE ¼ .062, p < .004) but withmore antisocial behavior at higher levels of parental acceptance (b ¼ .110, SE ¼ .027, p < .001). For boys, greater secrecy fromparents was associated with more antisocial behavior at lower levels of parental acceptance (b ¼ .121, SE ¼ .032, p< .001) butsecrecy from parents was not associated with antisocial behavior at higher levels of parental acceptance (b ¼ .077, SE ¼ .066,p ¼ .24). Secrecy from best friends was not associated with antisocial behavior longitudinally for boys or girls.

Discussion

The primary goal of this study was to determine whether keeping secrets from friends was associated with moredepression and antisocial behavior. In bivariate analyses, greater secrecy from friends was associated with more depressionand antisocial behavior. However, after controlling for earlier adjustment, secrecy from parents, the quality of the parent–child relationship, and the quality of the friendship, keeping secrets from one’s best friend was associated with moredepression but not with more antisocial behavior. Keeping secrets from one’s best friend prospectively predicted lower levelsof depression only for girls, and only in combinationwith high levels of secrecy from parents. For girls, more so than for boys,links between secrecy from best friends and depression are conditioned by the quality of the friendship and whether secretsare also kept from parents. Results from the current study argue for expanding the study of secrecy in early adolescencebeyond the parent–child dyad.

Keeping secrets is stressful and potentially harmful for the secret-keeper (Pennebaker, 1989, 1997), and keeping secretsfrom parents predicts greater depression and behavior problems (e.g., Frijns et al., 2005; Keijsers et al., 2010; Laird &Marrero,2010). This was the first study to test whether keeping secrets from friends is similarly associated with depression and

Table 3Regression coefficients from the antisocial behavior models.

Variable Concurrent models Longitudinal model

Age 12 antisocial behavior Age 13 antisocial behavior Age 13 antisocial behavior

B SE B* p B SE B* p B SE B* p

Prior antisocial behavior .79 .08 .61 <.001 .72 .05 .76 <.001 .71 .04 .75 <.001Sex .03 .05 .03 .65 .10 .04 .12 .02 .07 .04 .08 .13Parent–child relationship quality (RQ) �.02 .04 �.03 .63 �.03 .03 �.07 .18 �.04 .05 �.08 .35Friendship quality (FQ) .01 .04 .01 .87 .02 .03 .03 .55 .08 .03 .14 .008Secrecy from parents (SP) .09 .04 .19 .011 .07 .03 .16 .004 �.04 .04 �.08 .30Secrecy from friends (SF) �.11 .04 �.19 .01 �.02 .03 �.03 .60 �.05 .03 �.09 .14SP � RQ .18 .05 .32 <.001SF � FQSP � SFSex � RQ �.04 .06 �.06 .49Sex � FQSex � SFSex � SP .14 .05 .20 .01Sex � SP � RQ �.21 .06 �.32 <.001Sex � SF � FQSex � SP � SFR2 .42 <.001 .719 <.001 .76 <.001

Note: Antisocial behavior at age 11 served as prior antisocial behavior for the age 12 analyses and antisocial behavior at age 12 served as prior antisocialbehavior for the age 13 analyses. Age 12 secrecy and relationship quality variables served as predictors in the age 12 concurrent model and in the age 13longitudinal model. Age 13 secrecy and relationship quality variables served as predictors in the age 13 concurrent model. Sex is coded 0¼ female, 1¼male.

0

0.2

0.4

0.6

0.8

1

Low HighA

ge

1

3 A

ntis

oc

ia

l B

eh

av

io

r

Age 12 Secrecy from Parents

Lower RQ, Girls

Higher RQ, Girls

Lower RQ, Boys

Higher RQ Boys

Fig. 3. Age 12 parent–child relationship quality (RQ), secrecy from parents, and sex interact to predict age 13 antisocial behavior.

R.D. Laird et al. / Journal of Adolescence 36 (2013) 685–693 691

antisocial behavior. Although greater secrecy from friends was associated with more depression and antisocial behavior inbivariate analyses, greater secrecy from friends was associated with more depression, but not more antisocial behavior, inmultivariate analyses. Furthermore, in concurrent analyses at age 12, greater secrecy from friends was associated with lessantisocial behavior after controlling for prior antisocial behavior, secrecy from parents, and the quality of the friendship andparent–child relationship. In sum, greater secrecy from friends is uniquely associated with more depression but is notuniquely associated with more antisocial behavior.

The multivariate pattern of associations linking outcomes with secrecy from parents and secrecy from friends was ex-pected to reveal how secrecymay be linkedwith poorer adjustment. Greater secrecy from friends was associatedwith greatersecrecy from parents but secrecy from parents and friends were independent concurrent predictors of depression at ages 12and 13 (even though secrecy from friends wasmoderated by friendship quality and sex at age 12). Although some adolescentsmay be dispositionally secretive or distrustful of others and thus keep secrets from both parents and friends (see Larson &Chastain, 1990), secrecy in the two types of relationships is not functionally redundant. The potential harm of keeping se-crets in one relationship cannot be fully offset by a lack of secrets in the other relationship. Furthermore, longitudinal resultsshow that inconsistency in secrecy across parents and friends may be most problematic for girls. The highest levels ofdepression were found among girls reporting high levels of secrecy from parents, but not from friends, or high levels ofsecrecy from friends, but not from parents. Thus, in contrast to the expectations that a supportive parent could offsetpotentially detrimental effects of secrecy from one’s best friend or that a supportive best friend could offset potentiallydetrimental effects of secrecy from parents, results show that girls’ inconsistency in secret-keeping behavior across parentsand best friends is linked with the highest levels of depression. In contrast, girls who kept secrets from both friends andparents had the lowest subsequent levels of depression. Perhaps girls who keep secrets in both types of relationships aredemonstrating dispositional secrecy rather than a need to conceal specific behaviors or experiences. Furthermore, girls whodo not reveal secrets to parents or friends do not risk their information being shared with other individuals, do not need tocoordinate privacy boundaries with parents (Petronio, 2002) or friends (Kennedy-Lightsey, Martin, Thompson, Himes, &Clingerman, 2012), and thus do not experience additional stress from needing to regulate someone else’s sharing ofinformation.

These results appear to run counter to Frijns et al.’s (2013) finding that secrets were not linked tomaladjustment when thesecrets were shared with at least one other person. However, Frijns et al. (2013) were testing whether a specific item keptsecret from a parent or friend was shared with someone else, not whether adolescents were consistent or inconsistent interms of the amount of secrecy across relationships. A study that compares parents’ and best friends’ secrecy in reference tothe same topics is needed to resolve the apparent discrepancy between the two studies.

An adolescent’s decision to keep a secret from an individual is likely influenced by the quality of the relationship with theindividual (Dolgin & Kim, 1994; Laird & Marrero, 2010; Smetana et al., 2006). As anticipated, results show negative associ-ations between keeping secrets and relationship quality in both best friendships and parent–adolescent relationships.Although both parent–child relationship quality and secrecy from parents were associated with depression and antisocialbehavior in bivariate analyses, only secrecy from parents remained associated in the multivariate models. For friendships,secrecy was more consistently associated with depression and antisocial behavior than friendship quality at the bivariatelevel, and only secrecy remained associated in the multivariate models. This pattern suggests that secrecy is a more proximalcorrelate of maladjustment than relationship quality and that secrecy at least partially accounts for the link between parent–adolescent relationship quality andmaladjustment. However, the quality of the relationship also appears to moderate the linkbetween secrecy and behavioral adjustment, but only for girls. In girls’ low quality friendships, greater secrecy was linkedwith less depression whereas in girls’ high quality friendships (and in boys’ high and low quality friendships) greater secrecywas linked with more age 12 depression. This finding may reflect girls’ high level of investment in relationship-maintaininggoals. Girls worry more about their friendships and the approval of their friends, report experiencing more stress due to peerinteractions, and show a greater fear of losing friendships than do boys (see Rose & Rudolph, 2006, for a review). Rose and

R.D. Laird et al. / Journal of Adolescence 36 (2013) 685–693692

Rudolph speculate that the increased time spent by girls worrying about the status of their friendships likely contributes togreater internalizing problems, and this seems especially likely within high quality friendships that are presumably valuedmore than low quality friendships. For boys, friendship quality did not seem to make a difference in predicting depression,suggesting that it is act of secret-keeping itself that contributes to boys’ internalizing symptoms.

The quality of the parent–child relationship also moderated the association between secrecy and adjustment problems.Specifically, when girls reported lower levels of relationship quality, greater secrecy from parents was linked with lessantisocial behavior whereas when relationship quality was high (and when boys’ relationship quality was low), greater se-crecy was linked with more antisocial behavior. This finding parallels the best friend finding in that it is only in the highquality parent–child relationship that girls’ secrecy is associated with problems. Similar to girls’ connectedness with peers(Rose & Rudolph, 2006), girls’ relationships with their parents are characterized by higher levels of intimacy and dependencethan are boys’ relationships with parents (Geuzaine, Debry, & Liesens, 2000). Thus, girls’ secrecy within high quality parent–child relationships may have a more negative impact on their behavior than girls’ secrecy within low quality parent–childrelationships. This finding also could show that girls who are engaging in antisocial behavior may rely on secrecy to maintaina good relationship with their parents. In contrast, in lower quality parent–child relationships, girls’ secrecy from parents mayjust be a marker of poor relationship quality rather than an indicator of problem behavior (Keijsers et al., 2010). For boys,research indicates that secrecy seems to matter less in terms of predicting relationship problems with parents (Keijsers et al.,2010). Thus, the link between boys’ greater secrecy and antisocial behavior in low quality parent–child relationships may justreflect that boys are keeping secrets about their antisocial behavior itself. More research is needed on the specific types ofsecrets being kept in order to better interpret these associations.

Results of the current study clearly justify the expansion of research on secrecy in adolescence beyond the parent–childdyad (see also Frijns et al., 2013). However, results also highlight a key issue that needs to be addressed in doing so. The factthat keeping secrets from friends was associated with more depression but not more antisocial behavior whereas keepingsecrets from parents was associated with both more depression and more antisocial behavior suggests that what anadolescent keeps secret from parents may differ substantially from what an adolescent keeps secret from his or her bestfriend. Researchers have typically focused on the qualitative nature of the secrets one keeps (e.g., Bakken & Brown, 2010;Merten, 1999), on the psychosocial correlates of a dispositional tendency to keep secrets (e.g. Finkenauer et al., 2002; Frijnset al., 2005), or on the effects of keeping a specific secret for which the topic is unknown to the researchers (e.g., Frijns &Finkenauer, 2009; Frijns et al., 2013). A few studies have reported on the psychosocial correlates of secrecy and otherforms of information management across different topics but the studies have only considered secrecy from parents (e.g.Darling, Cumsille, Caldwell, & Dowdy, 2006; Laird, Marrero, Melching, & Kuhn, 2013; Smetana et al., 2006, 2010), althoughDolgin and Kim (1994) examined disclosure to close and best friends across a wide range of topics. The expansion of secrecyinto the peer relationship realmwould benefit greatly from a better understanding of what adolescents keep secret from theirfriends and why. Previous studies have described different types and functions of secrets in families (e.g., Imber-Black, 1998;Vangelisti, 1994; Vangelisti & Caughlin, 1997) and researchers could begin by comparing the types of functions of secrets fromparents with the types of functions of secrets from friends.

Adolescents likely keep some behaviors, thoughts, and feelings secret from friends but share them with parents. Ado-lescents also likely keep some behaviors, thoughts, and feelings secret from their parents while sharing them with theirfriends. For example, involvement in risk-taking behavior and sexual explorationmight be shared with friends but kept secretfrom parents. Although sharing secrets appears to reduce the psychological cost of secrecy (Frijns et al., 2013), the informationbeing kept secret probably matters greatly. For example, a teenwho tells her best friend about her involvement in risk-takingbehavior probably does not completely negate the effects of keeping such behavior secret from her parents. It is particularlyimportant for researchers interested in the psychosocial correlates of secrecy to begin documenting the topics that are beingkept secret from another to ensure that it is the secret-keeping process, rather than engagement in behavior one wishes tokeep secret that is the most important factor underlying the associations being tested.

Strengths of the current study include parallel measures of secrecy from parents and best friends, controlling for earliermaladjustment in analyses testing both concurrent and longitudinal associations, and testing replication across two time-points during early adolescence. The primary weaknesses of the study are relying on a single informant to provide globalself-reports of secrecy across two different relationships. As noted previously, future research should attempt to documentthe specific secrets that adolescents are keeping from parents and friends. While self-reports can be justified by the nature ofthe construct, researchers cannot be sure that participants are not keeping secrets from them as well. Relying on a singleinformant likely inflates associations between constructs, but such inflation would likely work against finding unique as-sociations for secrecy from friends when controlling for secrecy from parents and relationship qualities. Additional limita-tions include the fact that the samplewas recruited from a single community and although the demographic characteristics ofthe sample generally reflect the community fromwhich it was recruited, well-educated parents and two-parent families areover represented in the sample and the convenience sample is likely biased by the desire to collect data through personalinterviews in the participants’ homes.

In summary, keeping secrets from best friends during early adolescence is uniquely associated with more depression butnot with more antisocial behavior. Links between secrecy and maladjustment are more conditioned by relationship qualityand secrecy in other relationships for girls than for boys. Future research on secrecy in adolescence should consider secrecyfrom parents and friends, but should seek to better describe the nature of the information being kept secret in both types ofrelationships.

R.D. Laird et al. / Journal of Adolescence 36 (2013) 685–693 693

Acknowledgments

We thank the teachers and principals from the East Baton Rouge Parish School System and the families who participated inthis research. This research was supported by grants from the National Science Foundation (BCS 0517980 & 0807650) to R. D.Laird. The funding source had no involvement in the study design, data collection, analyses, or writing of the report. Thisstudy was conducted as an Undergraduate Honors Thesis by B.J.B. under the supervision of R.D.L. and M.A.M.

References

Bakken, J. P., & Brown, B. B. (2010). Adolescent secretive behavior: African American and Hmong adolescents’ strategies and justifications for managingparents’ knowledge about peers. Journal of Research on Adolescence, 20(2), 359–388. http://dx.doi.org/10.1111/j.1532-7795.2010.00642.x.

Bukowski, W. M., Hoza, B., & Boivin, M. (1994). Measuring friendship quality during pre- and early adolescence: the development and psycho-metricproperties of the Friendship Qualities Scale. Journal of Social and Personal Relationships, 11, 471–484. http://dx.doi.org/10.1177/0265407594113011.

Bumpus, M. F., & Hill, L. G. (2008). Secrecy and parent–child communication during middle childhood: associations with parental knowledge and childadjustment. Parenting: Science and Practice, 8(2), 93–116. http://dx.doi.org/10.1080/15295190802058868.

Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation for the behavioral sciences (3rd ed.). Mahwah, NJ: Erlbaum.Darling, N., Cumsille, P., Caldwell, L., & Dowdy, B. (2006). Predictors of adolescents’ disclosure to parents and perceived parental knowledge: between- and

within-person differences. Journal of Youth and Adolescence, 35, 659–670. http://dx.doi.org/10.1007/s10964-006-9058-1.Dolgin, K. G., & Kim, S. (1994). Adolescents’ disclosure to best and good friends: the effects of gender and topic intimacy. Social Development, 3, 146–157.

http://dx.doi.org/10.1111/j.1467-9507.1994.tb00033.x.Farrell, A. D., Kung, E. M., White, K. S., & Valois, R. F. (2000). The structure of self-reported aggression, drug use, and delinquent behaviors during early

adolescence. Journal of Clinical Child Psychology, 29, 282–292. http://dx.doi.org/10.1207/S15374424jccp2902_13.Finkenauer, C., Engels, R. C. M. E., & Meeus, W. (2002). Keeping secrets from parents: advantages and disadvantages of secrecy in adolescence. Journal of

Youth and Adolescence, 31, 123–136. http://dx.doi.org/10.1023/A:1014069926507.Frijns, T., & Finkenauer, C. (2009). Longitudinal associations between keeping a secret and psychosocial adjustment in adolescence. International Journal of

Behavioral Development, 33, 145–154. http://dx.doi.org/10.1177/0165025408098020.Frijns, T., Finkenauer, C., & Keijsers, L. (2013). Shared secrets versus secrets kept private are linked to better adolescent adjustments. Journal of Adolescence,

36, 55–64. http://dx.doi.org/10.1016/j.adolescence.2012.09.005.Frijns, T., Finkenauer, C., Vermulst, A. A., & Engels, R. C. M. E. (2005). Keeping secrets from parents: longitudinal associations of secrecy in adolescence.

Journal of Youth and Adolescence, 34, 137. http://dx.doi.org/10.1007/s10964-005-3212-z.Geuzaine, C., Debry, M., & Liesens, V. (2000). Separation from parents in late adolescence: the same for boys and girls? Journal of Youth and Adolescence, 29,

79–91. http://dx.doi.org/10.1023/A:1005173205791.de Goede, I. H. A., de Branje, S. J. T., de Delsing, M. J. M. H., & Meeus, W. H. J. (2009). Linkages over time between adolescents’ relationships with parents and

friends. Journal of Youth and Adolescence, 38, 1304–1315. http://dx.doi.org/10.1007/s10964-009-9403-2.Helsen, M., Vollebergh, W., & Meeus, W. (2000). Social support from parents and friends and emotional problems in adolescence. Journal of Youth and

Adolescence, 29, 319–335. http://dx.doi.org/10.1023/a:1005147708827.Imber-Black, E. (1998). The secret life of families. New York: Bantam Books.Keijsers, L., Branje, S. J. T., Frijns, T., Finkenauer, C., & Meeus, W. (2010). Gender differences in keeping secrets from parents in adolescence. Developmental

Psychology, 46(1), 293–298. http://dx.doi.org/10.1037/a0018115.Kennedy-Lightsey, C. D., Martin, M. M., Thompson, M., Himes, K. L., & Clingerman, B. Z. (2012). Communication privacy management theory: exploring

coordination and ownership between friends. Communication Quarterly, 60, 665–680. http://dx.doi.org/10.1080/01463373.2012.725004.Laird, R. D., & Marrero, M. D. (2010). Information management and behavior problems: is concealing misbehavior necessarily a sign of trouble? Journal of

Adolescence, 33, 297–308. http://dx.doi.org/10.1016/j.adolescence.2009.05.018.Laird, R. D., Marrero, M. D., Melching, J. A., & Kuhn, E. S. (2013). Information management strategies in early adolescence: developmental change in use and

transactional associations with psychological adjustment. Developmental Psychology, 49, 928–937. http://dx.doi.org/10.1037/a0028845.Larson, D. G., & Chastain, R. L. (1990). Self-concealment: conceptualization, measurement, and health implications. Journal of Social & Clinical Psychology, 9,

439–455. http://dx.doi.org/10.1521/jscp.1990.9.4.439.Merten, D. E. (1999). Enculturation into secrecy among junior high school girls. Journal of Contemporary Ethnography, 28, 107–137. http://dx.doi.org/10.1177/

089124199028002001.Orpinas, P. (1993). Skills training and social influences for violence prevention in middle schools: A curriculum evaluation. Unpublished doctoral dissertation.

Houston, TX: University of Texas Health Science Center.Pennebaker, J. W. (1989). Confession, inhibition, and disease. In L. Berkowitz (Ed.), Advances in experimental social psychology, Vol. 22 (pp. 211–244). New

York: Academic Press.Pennebaker, J. W. (1997). Opening up: The healing power of expressing emotions. New York: Guilford Press.Petronio, S. (2002). The boundaries of privacy: Dialectics of disclosure. New York: State University of New York Press.Preacher, K. J., Curran, P. J., & Bauer, D. J. (2006). Computational tools for probing interaction effects in multiple linear regression, multilevel modeling, and

latent curve analysis. Journal of Educational and Behavioral Statistics, 31, 437–448. http://dx.doi.org/10.2307/4122453.Rose, A. J., & Rudolph, K. D. (2006). A review of sex differences in peer relationship processes: potential trade-offs for the emotional and behavioral

development of girls and boys. Psychological Bulletin, 132, 98–131. http://dx.doi.org/10.1037/0033-2909.132.1.98.Schaefer, E. S. (1965). Children’s reports of parental behavior: an inventory. Child Development, 36, 413–424. http://dx.doi.org/10.2307/1126465.Schludermann, S., & Schludermann, E. (1988). Questionnaire for children and youth (CRPBI-30). Unpublished manuscript. Winnipeg: University of Manitoba.Smetana, J. G., Metzger, A., Gettman, D. C., & Campione-Barr, N. (2006). Disclosure and secrecy in adolescent-parent relationships. Child Development, 77(1),

201–217. http://dx.doi.org/10.1111/j.1467-8624.2006.00865.x.Smetana, J. G., Villalobos, M., Rogge, R. D., & Tasopoulos-Chan, M. (2010). Keeping secrets from parents: daily variations among poor, urban adolescents.

Journal of Adolescence, 33, 321–331. http://dx.doi.org/10.1016/j.adolescence.2009.04.003.Vangelisti, A. L. (1994). Family secrets: forms, functions, and correlates. Journal of Social and Personal Relationships, 11, 113–135. http://dx.doi.org/10.1177/

0265407594111007.Vangelisti, A. L., & Caughlin, J. (1997). Revealing family secrets: the influence of topic function and relationships. Journal of Social and Personal Relationships,

14, 679–705. http://dx.doi.org/10.1177/0265407597145006.