Increasingly accountability in education has linked student test scores to teacher and school evaluations. The underlying assumption behind this educator based accountability is that the high stakes linked to student test scores will prompt behavioral change, thus improving student learning and education quality. This study conducts a cross policy analysis using pooled data from the 2009 PISA, categorizing participant countries of the 2009 PISA into three national testing policies based on what type of educator based accountability is applied in the country. Results indicate that initial differences between national testing policy categories are not significant once school types and school practices that select on the student are included. This suggests that potential gains from more stringent accountability may be an artifact of schools under pressure engaging in practices that shape their testing pool, such as admitting only relatively high achieving students or transferring out lower achieving students.

JEL classification: I21, I24, I25, I28

Keywords: Education, PISA, accountability, testing, equity