Lower-to-upper secondary school transition: a Bayesian Lasso approach in data modelling

The determinants of transitioning from lower secondary to upper secondary school for Italian and immigrant teenagers (age range: 16–19) were identified by combining data from the European Union Statistics on Income and Living Conditions (EU-SILC) and the Italian Survey on Income and Living Conditions of Families with Immigrants in Italy (IM-SILC) for 2009. A set of individual, family, and contextual characteristics was selected through the Lasso method and a Bayesian approach to explain the decision not to continue on with upper secondary schooling (yes/no). The interruption of this transition revealed a complex pattern. The variables affecting it positively were squared age and almost all the significant first-order interactions, while negative impacts were observed for father’s age, parents’ education level, the amount of optional technological equipment owned, and the occupations of both parents. Other variables entered through the interactions included the individual’s and parents’ self-perceived health conditions, the degree of urbanisation, the type of macro-region, and so on. There were no gender distinctions and differences between Italians and immigrants disappeared when family background and parental characteristics were taken into account.