Browsing by Author "Tervonen, Lassi"

Finnish high schools are different in terms Matriculation Examination results. In some of the schools the results are very good on average. However, schools also differ in terms of student characteristics and some of the schools can be seen as ‘elite’, at least in terms of academic ability. If elite high schools students end up with good results in Matriculation Examination, to what extent the results are due to the school environment and to what extent due to the baseline characteristics of students? The Finnish National Joint Application procedure allocates students to general upper secondary schools based on applicants’ preferences and comprehensive school grade point average (GPA). Schools have limited number of seats, which means that they can admit only a limited number of applicants. This procedure leads to an entrance threshold, which determines a cutoff GPA for every school. This study’s focus is on schools in the Helsinki Capital Region which have very high entrance threshold. Those who have GPA above the cutoff are offered a seat, while those who have GPA below the cutoff are not. Because this cutoff is determined randomly, a regression discontinuity design (RDD) can be used to obtain a quasi-experimental setting. RDD exploits thresholds as a rule assigning applicants just above and just below the threshold to treatment and control groups, respectively. This makes it possible to estimate causal elite high school effects using rich data sources containing information on students’ application preferences, GPA, background characteristics, enrollment, Matriculation Examination results, and the cutoff GPAs. However, due to imperfect compliance, the enrollment to elite high schools is probabilistic, so only the elite high school offers are randomized. These offers are used to estimate intention-to-treat effects. Also, by using the offer as an instrumental variable to elite high school enrollment, the effect of enrollment is estimated. This is done by using fuzzy RDD, which allows for imperfect compliance. These estimates represent local average treatment effects for compliers, i.e., for those whose enrollment is determined by their offer status. An elite high school offer gives a possibility to study with high-achieving peers in terms of comprehensive school GPA. The peer group one gets is also often more homogenous in terms of GPA. In some cases, the peers are also more often female and have more often highly educated parents. Nevertheless, intention-to-treat and local average treatment effect estimates suggest that receiving an offer to an elite high school or enrolling to an elite high school does not have systematic positive or negative effect on learning outcomes, as measured by Matriculation Examination results. These results can be seen as evidence against the importance of peer group in terms of learning outcomes. On the other hand, attending in an elite high school can be beneficial in some other way that is not observed in Matriculation Examination grades.