The proposed paper analyzes tracking effects using the evidence from PISA studies conducted in Poland. The reform of the schooling structure in Poland, implemented in 2000, raised the age of selection into secondary school tracks by one year and separated primary and lower secondary levels of schooling. The reform coincided with the first OECD PISA study and concerned mainly 15 years olds assessed in PISA. Therefore, the PISA study provides rare opportunity to evaluate systematically the results of reform. While the cohort of PISA 2000 students were tracked in the stratified structure of the secondary schools, the students in PISA 2003 and 2006 were in the final year of newly founded general lower-secondary schools and were assessed shortly before the selection into upper-secondary schools. International PISA 2006 study was accompanied by additional grade-based national study of students in 180 schools conducted by the Polish PISA team, in which students of the upper-secondary schools (basically 16 and 17 years olds) were assessed using international PISA instruments. This allows for comprehensive analysis of performance and family background data across the types of schools before and after schooling reform. What is notable in the PISA results in Poland is substantial decrease in the between-school variance of performance between PISA 2000 and PISA 2006. Moreover, significant, and internationally unique improvement in performance was noted. In reading, 15 years old improved their performance significantly, both between PISA 2000 and PISA 2003 and between PISA 2003 and 2006. The improvement, as we are going to argue, can be attributed mainly to the schooling reform. PISA data suggests, raising the age of selection, did not affect negatively, as could be expected from the tracking literature, best performing students. Still, while the transition to the upper secondary schools is now postponed, which is beneficial for student performance, number of problems of upper secondary schools tracked into general and vocational schools were not solved, and are evidenced by the national part of the PISA study.