About What and From Whom? The Effects of Providing School Performance Information to School Choosers and Citizens

• In Event: Promising Scholarship in Education: Dissertation Fellows and Their Research

Abstract

This project examines the effects of providing school choosers and citizens with additional information about school performance. I study the effects of information provision on both attitudes (e.g., opinions of local schools) and behaviors (e.g., school choices). At a time when governments and others are grappling with how to inform the public about school performance, it is important that we understand how the public responds to various types of information. The responses of school choosers and citizens can shape decisions about which schools to choose, where to live, and how to engage in political matters.
Data and Methods
There are two primary components to this research: field experiments in urban areas and online survey experiments with nationally representative samples.
I conducted field experiments in Milwaukee, Washington, DC, and Philadelphia. In Milwaukee and DC, school-choosing families were assigned to either a treatment or control condition as they moved from elementary to middle school or middle to high school. Those assigned to the treatment condition received booklets with profiles of the schools available to them, including ratings of the schools’ academic performance. Many treatment group families also received support from trained staff. Control group families did not receive this additional information and support unless they specifically requested it. Administrative data from each city enabled testing of how treatment assignment affected school choices, enrollment, and subsequent student outcomes. In Philadelphia, parents and students entering a high school fair were randomly assigned to either a treatment or control condition. Those assigned to the treatment group read through a booklet with information about schools before completing a survey about their beliefs, priorities, and plans. Those assigned to the control group completed the same survey without first reading the booklet.
Later, I conducted online survey experiments with a sample of 1,000 American adults from a YouGov panel weighted to be nationally representative. Of primary interest was how “school choosers” (respondents with children 18 or under) and “citizens” (all respondents) updated their opinions of local schools in response to receiving both official, numerical school performance ratings and narrative comments from parents. Respondents saw these ratings and comments alongside one another for two schools in their neighborhoods. They were randomly assigned to see either two positive comments or two negative comments. They then assigned a grade to each school, A+ through F. I use this random assignment to identify the causal effects of seeing positive versus negative comments. To identify causal estimates of the effects of seeing higher numerical ratings, I use regression discontinuity designs that take advantage of the ways the ratings were constructed. Follow-up work with YouGov and MTurk samples enabled a more nuanced look at people’s responses to different types of school performance information.

Author

• Jon Valant, Stanford University