Determinants of the choice for professional teacher education programs: A multinomial multilevel approach

In recent years, concerns have risen regarding the Flemish teacher labour market : there is a fear of a decreasing quality of inflowing students in teacher education programmes. In this paper we investigate the transition from secondary school to higher education in Flanders, more specifically to professional teacher education programms (i.e. Bacherlor in Education). We used merged administrative databases from the Flemish Department of Education in order to construct a longitudinal dataset, enabling to track individual students from secondary to higher education. We obtained data from the entire population of secondary school leavers in the academic year 2004-2005 (n=51.902). These data allow tracking individual students during their higher education career from the academic year 2005-2006 until 2011-2012. The data include individual demographic student characteristics, detailed individual enrolment information in secondary education and higher education and a number of secondary school characteristics. Since pupils are nested in schools and in municipalities multilevel estimation techniques were used to model the transition from secondary education to various potential higher education programmes. Both random intercepts and random slopes models were estimated. Our findings show that the odds of enrolment in teacher education programs are highly influenced by several individual, secondary school and regional characteristics. While the differences with academic bachelors have been found to be more pronounced when considering most of the determinants, the inflow in teacher education was also found to be substantially different from the inflow in other professional programs. However, not all variance that is detected at the level of secondary schools can be explained by secondary school characteristics.