Working Paper: NBER ID: w14777
Authors: Philip Babcock; Julian R. Betts
Abstract: Do smaller classes boost achievement mainly by helping teachers impart specific academic skills to students with low academic achievement? Or do they do so primarily by helping teachers engage poorly behaving students? The analysis uses the grade 3 to 4 transition in San Diego Unified School District as a source of exogenous variation in class size (given a California law funding small classes until grade 3). Grade 1 report cards allow separate identification of low-effort and low-achieving students. Results indicate that elicitation of effort or engagement, rather than the teaching of specific skills, may be the dominant channel by which small classes influence disadvantaged students.
Keywords: class size; student achievement; education production function
JEL Codes: I21; I22
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| smaller class sizes (I24) | increased engagement and effort (D29) |
| increased engagement and effort (D29) | improved educational outcomes (I24) |
| larger classes (C55) | drop in achievement for low-effort students (D29) |
| smaller class sizes (I24) | better outcomes for low-effort students (D29) |
| low-effort students (D29) | significant drop in achievement (I21) |
| low-ability students (D29) | no significant difference in achievement drop-off (I21) |