80

The aim of this study is to replicate a path analysis conducted with a sample of 11 year-old students (Dierendonck, unpublished) on a sample of 15 year-old students. By linking two existing databases (PISA 2006 and EDD 2006), an explanatory model of reading performance of 15 year-old students in Luxembourg is proposed, discussed and compared to the model defined at age 11. The explanatory models are used to determine the influence of exogenous family variables and variables reflecting schooling of students (grade retention, attended school track) on reading performance, but also to estimate the impact of parental involvement variables (at home and at school). The results show a differential effect of these variables at age 11 and age 15. At age 11, several parental involvement indicators are significantly related to reading performance (eg, parents’ self-efficacy for supervising homework, degree of parental degree of information about school, parent–child degree of communication about school activities and daily time spent by an adult to supervise homework). At age 15, no significant relationship, direct or indirect, appears between parental involvement variables and student reading performance