Quantile Regression Methods for Recursive Structural Equation Models

The second essay explores the heterogeneous peer effects in Dutch primary schools by employing a series of percentiles of classmates achievement distribution as the peers outcome measurement and the quantile regression estimation method. It is found that better performance by lower attainment peers has a significant positive effect on performance of all students in both language and math. This effect is most pronounced at the lower conditional quantiles of the performance distribution, but persists even into the upper tail. In contrast, better performance by high-achieving classmates, has a somewhat more modest effect on language performance, while for math, it has little or no effect on weaker students, but does have strong positive effect on students in the upper tail of the conditional distribution. The findings imply that mixing diverse students may be an efficient policy for the improvement of language performance, tracking appears desirable for the improvement of students' math performance.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Economics

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Ma, Lingjie

Contributors dc:contributor

• Koenker, Roger W.

• Economics, General

Language dc:language

eng

Identifier

(MiAaPQ)AAI3130975

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/85544