A Multilevel Investigation of the Influence of Family and School Level Factors on Age of Sexual Initiation: A Focus on Sex Differences

Background/Purpose: Early sexual initiation is associated with unwanted pregnancy, depression, and other health risks, yet national surveys reveal that 13% of American teens have sex before age 15 and that U.S. teen pregnancy rates are the highest in the Western world. Previous literature has identified individual, family characteristics associated with early initiation but few have examined school level factors. In addition, sex differences in predictors are clear but findings are inconsistent. The current study uses bioecological and resilience theories and statistical modeling techniques appropriate for adolescents clustered within schools, to examine the influence of school level factors; and explore the mechanisms through which putative risk and protective factors work (i.e., whether school level variables influence the relationship between previously identified risk and protective factors). Relational theory then guides examination of sex differences in particular predictors of ASI.

Methods: The analytic sample includes 10,596 individuals who were interviewed at both Waves I and IV of the National Longitudinal Study of Adolescent Health (Add Health). Participants were in grades 7-12 at Wave I and aged 24-32 at Wave IV (55% female; 57% white; 20% African-American; 15% Latino). Predictor variables were taken from Wave I interview data. The school level variables were aggregated from the individual to the school level. Predictors include GPA, physical development, attitudes about early sex and pregnancy, perceived likelihood of higher education, alcohol use, delinquency, family structure, parents’ education level, childhood abuse, perception of maternal approval of sex, parental monitoring, parent-child relationship quality, depression; and school level averages on parents’ education level, perception of peer attitudes about sex, and likelihood of higher education; and cross-level interactions. The dependent variable, self-reported age of first heterosexual sexual intercourse, is taken from Wave IV report. All variables are adolescent reported with the exception of parents’ education which is taken from the parent interviews at Wave I. Hierarchical linear modeling, a statistical approach that adjusts for the correlated standard errors among individuals within the same school, was conducted in 8 steps, to estimate the effects of predictors on age of initiation for the entire sample, and then for boys and girls separately. 

Results: Results support hypotheses that both family and school level factors influence age of initiation for girls and boys, and that school mean parental education attainment moderates the influence of individual level factors on age of initiation. Results also show that there are more family and school level factors that predict age of initiation for girls (father-child relationship quality, sexual abuse, parental monitoring, and school mean higher education likelihood) than for boys, and that perception of maternal and peer approval of sexual activity are primary predictors of age of initiation for boys.

Implications: Results highlight the importance of school context on understanding the motivations for early initiation. Findings support the need for policy changes and prevention interventions that include partnerships of adolescents, parents, and schools; and that address each of the predictors of early sexual activity (e.g., paternal involvement, family and peer approval of early sex, and school context).