DP11667 Counting Rotten Apples: Student Achievement and Score Manipulation in Italian Elementary Schools

Author(s):	Erich Battistin, Michele De Nadai, Daniela Vuri
Publication Date:	November 2016
Keyword(s):	Measurement error, Non-parametric bounds, Partial identification, Score manipulation
JEL(s):	C14, C31, C81, I21, J24
Programme Areas:	Labour Economics, Public Economics
Link to this Page:	cepr.org/active/publications/discussion_papers/dp.php?dpno=11667

We derive bounds on the distribution of math and language scores of elementary school students in Italy correcting for pervasive manipulation. A natural experiment that randomly assigns external monitors to schools is used to deal with endogeneity of manipulation as well as possible misclassification of the manipulation status. Bounds are obtained from properties of the statistical model used to detect classes with manipulated scores, and from restrictions on the relationship between manipulation and true scores. Our results show that score distributions are heavily affected by manipulating behavior, with regional rankings by academic performance being reversed once manipulation is taken into account.