Class Size

Author

Tanay Janmanchi

Using data from Tennessee Project STAR, we seek to understand the causal effect of putting students in smaller classrooms on their grade 4 math scores. Students are randomly assigned to a Kindergarten class of size small or regular, then in grade 4, the students take a math test to measure their progress. One challenge we encountered was that the method of evaluating grade 4 math score might not have been consistent over the years. We found that there is little to no effect of class size on math exam scores. We are 95% confident that the causal effect is between -4.5 and 3.9 and predict the causal effect to be -0.36.

\[ score_i = \beta_0 + \beta_1 class\_size \]

Characteristic

Beta

95% CI

1
kinder

    kindersmall -0.36 -4.5, 3.9
1

CI = Credible Interval