Pathways of school engagement in preventing adolescent delinquency and subtance use

Abstract

Delinquency and substance use impacts negatively on adolescent health and various facets of their individual well-being, therefore translating to negative impact on the economy of many governments irrespective of development status. There is an increasing attention on engagement due to its impact on developmental trajectories and academic success with potentially long-term consequences. Whereas many studies have investigated the impact of engagement on academic failure, limited reports explore the role of engagement in the aetiology of delinquency and health-compromising behaviours and little attention has been paid on influencers of engagement. This study explored the hypothesis that personal and contextual factors at individual and school levels may either mediate or moderate the effect of different dimensions of school engagement on delinquency and substance use. Engagement involved emotional, social and cognitive engagement. Personal factors included psychological factors namely social cognition (measured as the locus of control) and self-perception (measured as self-concept). Contextual factors included parental involvement in school activities, parental commitment to educational goals of the children, communication between parent and child, as well as supervision and monitoring of children by the parents. Trends, distribution and univariate tests of association between socio-demographic variables, substance use, delinquency, engagement and contextual factors were analysed. A total of 898 students were selected using a multistage probability sample design from selected municipalities within Gauteng province as the primary sampling unit, schools within the primary sampling units, and students within the schools. Variability due to grade level and school level factors were analysed and used as input for multilevel cumulative link mixed models assessing the influence of parental and peer factor variables on the psychosocial factors. Correlations, exploratory factor analysis, confirmatory factor analysis, and two structural equation models (SEM) were used to test the hypothesized pathways through which personal and contextual factors influence the impact of different dimensions of engagement on the occurrence of delinquency and substance use. A slight majority of the students were female (57%; n=492). The mean student was 16.7 (standard deviation = 2.6) and age was relatively symmetrical with a median age of 16 years and a majority (66.5%; n=560) distributed between the ages 15 and 18. The majority of the students were from grade 10 (33%; n=283) followed by grades 9 (15.7%; n=131), 11 (15.4%; n=128), 8 (13.3%; n=111) and 7 (13.1%; n=109) while respondents from grades 6 and 12 comprised less than 100 students. The most used substances included alcohol (31.3% of the students) followed by smoking (12.7%) and cannabis (11.3%) when compared with than the hard drugs including amphetamine, barbiturates, cocaine, heroin and other substances. Significant changes in the dependent variables (substance use and delinquency) and the psychosocial variables (social cognitions and self-perceptions) with grade level and school-level were reported (p<0.05). There were significance influences of parent and peer factors on psychosocial factors and psychosocial factors on student engagement (p<0.05). Specific facilitators and indicators of engagement classified as psychosocial factors, social cognition and self-perceptions are presented and discussed. The results indicated that besides peer and parental factors, other factors such as behavioural protection may lead to enhanced engagement and psychosocial factors could be protective against substance use. The study of the nomological network of each of the constructs using factor analysis distinguished two dimensions of delinquency, four of engagement, (emotional engagement, social or behavioural or participatory engagement, and cognitive engagement), two of parental involvement (parental monitoring and parental involvement), four of self-concept (self-perception, self-esteem, self-cognition and self-efficacy), two for social cognition (external and internal dimensions) and three for illicit substance use. The statistical significance of many of the direct and indirect effects from the SEM models confirmed the hypothesis that personal and contextual factors impact the occurrence of delinquency and substance use directly and indirectly. Partial mediation was also revealed where the effects of personal and contextual factors are partially mediated by engagement. The pathways revealed in this study involving direct, indirect effects and total effects of these assets indicate that they do not confer benefits singularly but factors such as psychosocial variables including self-concept and locus of control may act as precursors of other assets such as engagement whereas parental factors may mediate to influence peer factors. An extensive repertoire of developmental assets and liabilities including psychosocial variables and outcomes of delinquency and substance use which can be incorporated in prevention efforts are presented. These factors can be used in computing risk behaviour indices which predict the risk to support early intervention.