A Rising Tide That Lifts All Boats? Teach For America and Grade-Level Outcomes

• In Event: Teach For America: Impacts Beyond the Classroom

Abstract

Teach For America (TFA) typically places teachers in schools with a history of underperformance. It is in these schools that the presence of a new type of teacher may be particularly helpful in changing the status quo. While a number of studies have assessed the impact of TFA teachers on their own students, the larger effect of TFA on placement schools is an important open question. The goal of this study is to determine whether TFA teachers affect their colleagues and, if so, the extent of those effects.

Ideally, one would use student or classroom level data to compare the scores of non-TFA classrooms in the years before and after the introduction of a colleague TFA teacher. Lacking this micro level data, I use elementary school grade level scores in math and English from the California Department of Education along with TFA placement records for the years 2002-2011 to analyze the effect of TFA’s presence on the student achievement of non-TFA grades.

There are a number of mechanisms through which TFA teachers could affect their colleagues, including shared resources, motivation or effort spillovers, and peer learning. In its most basic form, my identification strategy is to use a difference-in-differences estimator. Control grades in non-TFA schools are used to represent what would have happened in treatment grades (non-TFA grades in TFA schools) in the absence of TFA. Looking before and after a school receives at least one TFA teacher, the change in test scores in a treatment grade relative to the change in an unweighted average of the test scores in control grades yields an estimate of the effect of TFA placement.

Identifying the causal effects of TFA presents a challenge as TFA teachers are not randomly assigned to schools or grades. My empirical strategies account for measured time-varying school characteristics that may affect both the hiring of TFA teachers and student achievement. In addition to comparing schools within the same school district, I use a matched sample to help ameliorate concerns of non-random TFA placement to schools. I find no evidence of selective placement of TFA to grades within a school. My estimates suggest that grades adjacent to TFA grades improve as TFA presence increases. Specifically, when TFA constitutes a relatively large percentage of teachers, I find that grades next to TFA grades improve by roughly ten percent of a standard deviation in both math and English, moving students up nearly four percentage points in the grade-year score distribution in both subjects on average.

Author

• Adam Wright, University of California - Santa Barbara