The effect of class size in grades K-3 on adult earnings, employment, and disability status: evidence from a...

Early education interventions have been forwarded as a means for reducing social disparities in income and health in adulthood. We explore whether a successful early education intervention, which occurred between 1985 and 1989, improved the employment rates, earnings and health of blacks relative to whites through 2008.

We used data from Project STAR (Student Teacher Achievement Ratio), a four-year multi-center randomized controlled trial of reduced class sizes in Tennessee involving 11,601 students. Students were initially randomized within 79 schools to classes with 22-25 or 13-17 students. We linked subject records to Social Security Administration (SSA) earnings and disability data collected between 1997 and 2008-when the majority of subjects were between the ages of 18 and 28. We focused our analysis on annual, rather than cumulative, measures of earnings and employment because educational attainment after high school might reduce earnings through age 23. We considered three or more years of statistically significant positive (or negative) annual impacts to be a meaningful effect.

Project STAR improved cognition and high school graduation rates. These benefits were primarily realized among low-income and minority students. These early education benefits did not translate into reduced disability claims in adulthood for treated subjects. However, exposure to small class size increased employment for blacks, and increased earnings for black males (p<0.05). Exposure to small classes also led to an increase in earnings for white males. However, white females exposed to small classes experienced a net decline in earnings and employment across the later years of follow up (p<0.05), offsetting any gains by white males.

Exposure to small class size in grades K-3 appears to improve earnings and employment for black males and earnings for white males, while reducing employment and earnings among white females.