Achievement and Growth: The Effects of District Performance Data on Public Opinion

• In Event: 63.069 - What Shapes Public Perceptions of Educational Systems?

Abstract

Public schools, school districts, and states increasingly employ measures of average student growth over time for accountability purposes. Compared to average student achievement at a single point in time, student growth is arguably better able to capture schools’ and school systems’ contributions to student learning. Student growth also bears a weaker relationship to the racial and socio-economic composition of the student body, making it easier to identify highly effective schools that serve predominantly disadvantaged students.

However, there is little research on the political consequences of using student growth to measure educational performance. Using new estimates of district level achievement and growth from the Stanford Education Data Archive and a sample recruited via Amazon’s Mechanical Turk service, I conduct an online survey experiment in which a randomly assigned group receives data about their districts’ average growth while the other group receives data about their districts’ average achievement. This approach allows me to identify the average effects of receiving growth data (as opposed to achievement data) on subjects’ confidence in their district, their stated likelihood of enrolling a child in a local public school, and their attitudes towards a range of other education issues.

I pose three research questions: 1) To what extent are individuals able to estimate average achievement and average growth in their districts accurately? 2) What are the average effects of providing district growth data (as opposed to achievement data) on a range of attitudes towards education issues? 3) Are these effects larger for subjects in districts where the average growth percentile is greater than the average achievement percentile (i.e. relatively high growth districts)?

I find that subjects have a modest understanding of average achievement in their districts, but they do not have accurate prior beliefs about average district growth. Among the entire subject pool, the provision of growth data (as opposed to achievement data) has a small positive effect on the grades that individuals assign to their districts. Among subjects in districts where the growth percentile is greater than the achievement percentile, the provision of growth data has moderately large positive effects on district grades and the stated likelihood of enrolling a child in a local public school. Also among this subset of the subject pool, the provision of growth data increases support for higher teacher salaries and decreases support for school choice reforms such as charter schools and vouchers; however, these effects are small (about 0.1 on a 4-point scale) and only marginally significant.

School districts with relatively high growth rates ought to distribute this information widely. Doing so could increase both confidence and enrollment. This is particularly important for districts that serve disadvantaged communities—many of which have acquired poor reputations for performance despite high levels of student growth.

Author

• David Houston, Harvard University