“…In order to control for student-teacher selection, studies in the USA have made use of experiments with random assignment of students and teachers or natural experiments (Clotfelter et al, 2006;Dee, 2004;Garet et al, 2008Garet et al, , 2010Jacob and Lefgren, 2004;Nye et al, 2004). In the absence of such experiments, in this paper, following studies in USA (Betts et al, 2003;Boyd et al, 2006;Clotfelter et al, 2007;Hanushek et al, 2005;Jepsen, 2005;Rivkin et al, 2005;Rockoff, 2004), we use student fixed effects to control for student heterogeneity and student-teacher matching. Student fixed effects alleviate the bias due to students assignment to teachers based on time-invariant student characteristics (static tracking) (Koedel and Betts, 2011;Rothstein, 2010).…”