Gender Variation of Individual Assets and Problem Behaviors in at-Risk Adolescents: A Longitudinal Cross-Lagged Analysis

Background and Purpose: Individual assets, such as social competencies and self-esteem, which promote positive youth development and problem behaviors (aggression, delinquency) tend to co-vary. However, neither the temporal relations, nor potential gender differences in the relations, have been well established. 

Methods: Participants were 372 adolescents (174 boys, 198 girls), primarily African-American, of low socioeconomic status, participating in a longitudinal, prospective study from birth.  Adolescents’ individual assets (commitment to learning, positive values, social competencies, positive identity) were assessed at ages 12, 15, and 17 with the Developmental Assets Profile, a 58-item youth self-report using a 4-point Likert scale, with higher scores indicating greater assets. Problem behaviors (aggression, rule-breaking behaviors, inattention) were assessed at ages 12 and 15 using the Youth Self-Report, a 105-item youth self-rating of his or her own behavior in the last 6 months, with higher scores indicating more problem behaviors. Cross-lagged, structural equation modeling (SEM) with full-information maximum likelihood estimation was employed to examine the longitudinal associations between individual assets and problem behaviors. Gender differences were examined using multiple-group SEM, with chi-square difference tests to compare nested models.

Results: The cross-lagged multiple-group SEM indicated a good fit, Chi-square (114)= 164.04, p= .002, TLI= .98, CFI= .98, RMSEA= .034 (90% CI= .022 - .046), with gender equivalence in factor loadings and intercepts.  However, gender differences were noted in cross-lagged effects. While problem behaviors at age 12 were associated with decreased individual assets at age 15 (β= -.22, p = .03) in girls, no such association was found in boys (β= -.11, p > .20).  Instead, greater individual assets at age 12 were associated with more problem behaviors at age 15 (β= .26, p= .005) in boys.  Higher stability between ages 12 and 15 was found in problem behaviors for both genders (β= .60 for boys; β= .70 for girls) than in individual assets (β= .30 for boys; β= .36 for girls; all p’s <.001).  Between ages 15 and 17, higher stability in individual assets was reported in girls (β= .73) than in boys (β= .43; all p’s <.001).  A stronger negative correlation was observed between individual assets and problem behaviors in girls (r= - .60, p <.001) than in boys (r= - .32, p < .001) at 12 years.  The estimated model accounted for 42% of the total variance in problem behaviors in girls and 33% in boys at age 15; and 52% of the total variance in individual assets in girls and 28% in boys at age 17.  

Conclusions and Implications: Individual assets and problem behaviors were related, yet their temporal relationships differ by gender, underscoring a need to understand gender differences in developmental pathways.  Interventions strengthening individual assets building may reduce problem behaviors in adolescent girls and thereby decreasing its related risk behaviors such as substance use and risky sexual behaviors.  However, alternative intervention may be required for adolescent boys to reduce problem behaviors.