A behavioral approach to adolescent cannabis use: Accounting for nondeliberative, developmental, and temperamental factors

Abstract

Most behavioral models examine adolescent health risk behaviors using a reflective, deliberate social–psychological framework. In this study, adolescent cannabis use is investigated via an expanded social–psychological model of behavioral decision-making: the Theory of Planned Behavior (TPB). The aim was to examine the contribution of nondeliberative (impulsivity), developmental (perceived parenting styles), and temperamental (moral norms, mental health, delinquency) factors in adolescent cannabis use. A longitudinal questionnaire with baseline and follow-up measurement (14-day interval) was used. Participants were Sixth Form College students (n = 199) aged 16–18 (mean age = 16.44, SD = −0.55). At baseline (T1), demographics, TPB variables, and additional socio-psychological variables were measured. Fourteen days later (T2) self-reported cannabis use was measured. Logistic regression analyses indicated that the impulsivity subcomponent of lack of premeditation and moral norms predicted self-reported cannabis use behavior. Perceived parental rejection predicted cannabis use intentions. Adolescent cannabis use can be better understood through the expanding of behavioral models to account for nondeliberative, developmental, and temperamental factors. Drug education interventions should aim at developing self-instruction training programs teaching adolescents effortful thinking while family-based interventions should focus on encouraging open parent–adolescent communication which has shown to influence adolescents’ cannabis use.

Keywords:

• adolescents
• cannabis use
• parenting
• non-deliberative
• impulsivity

Notes

1 EM is a missing data technique that overcomes some of the limitations of other techniques, such as mean substitution or regression substitution (Schafer & Graham, 2002). These alternative techniques generate biased estimates and, specifically, underestimate the standard errors. Due to the cumulative loss of participants that would have occurred due to list-wise deletion biases estimates (Schafer & Graham, 2002), the maximum-likelihood estimation was used so as to include all cases (Dempster, Laird, & Rubin, 1977). Because the proportion of missing values for most variables was small and/or did not appear to be systematic (p > 0.05), the assumption that the values are missing at random was considered plausible (Little & Rubin, 2002). The missing data were replaced using EM, so that all 199 participants’ data were used throughout.

2 Each variable was measured against the basic TPB variable, separately from one another. Several models were run to test the significance of variables according to the order they were entered in the model. No differences were found regarding the significance. Therefore, the final allocation of variables in each step was conducted according to theoretical relevance.