Working Paper: NBER ID: w18277
Authors: Glenn Ellison; Ashley Swanson
Abstract: This paper explores differences in the frequency with which students from different schools reach high levels of math achievement. Data from the American Mathematics Competitions is used to produce counts of high-scoring students from more than two thousand public, coeducational, non-magnet, non-charter U.S. high schools. High-achieving students are found to be very far from evenly distributed. There are strong demographic predictors of high achievement. There are also large differences among seemingly similar schools. The unobserved heterogeneity across schools includes a thick tail of schools that produce many more high-achieving students than the average school. Gender-related differences and other breakdowns are also discussed.
Keywords: high math achievement; American Mathematics Competitions; school effects; demographic predictors
JEL Codes: C25; I21
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| demographics (J11) | high math achievement (C12) |
| parental education (I24) | high math achievement (C12) |
| income (E25) | high math achievement (C12) |
| unobserved school effects (I21) | high math achievement (C12) |
| high math achievement (C12) | female students (I24) |