Early adolescent sexual debut: the mediating role of working memory ability, sensation seeking, and impulsivity

Abstract

Although deficits in working memory ability have been implicated in suboptimal decision making and risk taking among adolescents, its influence on early sexual initiation has so far not been examined. Analyzing 2 waves of panel data from a community sample of adolescents (N = 347; Mean age[baseline] = 13.4 years), assessed 1 year apart, the present study tested the hypothesis that weak working memory ability predicts early sexual initiation and explored whether this relationship is mediated by sensation seeking and 2 forms of impulsivity, namely acting-without-thinking and temporal discounting. The 2 forms of impulsivity were expected to be positively associated with early sexual initiation, whereas sensation seeking was hypothesized to be unrelated or to have a protective influence, due to its positive association with working memory. Results obtained from structural equation modeling procedures supported these predictions and in addition showed that the effects of 3 prominent risk factors (Black racial identity, low socioeconomic background, and early pubertal maturation) on early sexual initiation were entirely mediated by working memory and impulsivity. The findings are discussed in regard to their implications for preventing early sexual onset among adolescents.

Figure 1

Hypothesized model testing direct and mediational pathways of influence on early adolescent sexual debut. All variables except sexual initiation were assessed at baseline. The direct effects of exogenous risk factors on endogenous personality variables in the model (i.e., sensation seeking, acting-without-thinking, and temporal discounting) were tested. Similarly, the direct effect of working memory on sexual initiation was examined. All exogenous variables were correlated. These paths are not shown in the figure for the sake of clarity.

Figure 2

Final structural equation modeling (SEM) showing the significant paths (p < .05) with standardized coefficients, unless noted otherwise. Model fit: comparative fit index = 0.95; Tucker-Lewis index = 0.94; root-mean-square error of approximation = 0.03, 90% CI [.02, .04]. WM = working memory; AWT = acting-without-thinking; TD = temporal discounting; SEB = socioeconomic background. Dashed lines represent effects significant at p ≤ .10 level. ^uIn case of dichotomous predictors, the unstandardized coefficients are reported for ease of interpretation. ^fPubertal development's effect on working memory was significant only for females, and the female-specific coefficient is reported here. All exogenous variables were assessed at baseline and are correlated. These paths are not shown for reasons of clarity.