Research Question/Significance. The No Child Left Behind Act of 2002 (NCLB) dramatically altered federal education policy by sanctioning schools based not only on aggregate student achievement, but also on the achievement of student subgroups within schools, including racial minority, limited-English-proficient (LEP), and low-income students. Despite evidence that this subgroup-specific accountability increased the likelihood that certain schools would be sanctioned (Kane & Staiger, 2002, 2003; Sims, 2013), as well as evidence that such sanctions affect teachers’ career paths (Clotfelter, Ladd, Vigdor, & Diaz, 2004; Feng, Figlio, & Sass, 2010), little research has investigated the effects of subgroup-specific accountability on teachers. Given the prevalence of subgroup-specific accountability, as well as the importance of teachers to student achievement (Rivkin, Hanushek, & Kain, 2005; Rockoff, 2004), these effects are particularly important to understand. This study examines the effects of NCLB’s subgroup-specific accountability on teacher turnover and retention.

Identification Strategies. NCLB held schools accountable for the performance of subgroups only if the numbers of students in a subgroup exceeded a minimum subgroup size, which was determined by each individual state. This study uses a difference-in-differences strategy to compare—in both the pre- and post-NCLB periods—teacher turnover and retention in schools that were held accountable or not held accountable for various student subgroups. Analyses utilize either state fixed effects or school-by-time-period fixed effects to control for time-invariant state and school characteristics, respectively. Additional analyses utilize a regression discontinuity (RD) design to compare teacher turnover and retention in schools whose numbers of subgroup students fell within a narrow range on either side of their states’ minimum subgroup size cutoffs.

Data. This study uses nationwide data from the Schools and Staffing Survey and the accompanying Teacher Follow-up Survey from the periods both before and after the nationwide implementation of NCLB (1999-2001 and 2003-05, respectively). This data is matched with school-level data from the Barnard/Columbia NCLB Database (Reback, Rockoff, Schwartz, & Davidson, 2011), which provides information on the subgroups for which schools nationwide were held accountable during the first two years of NCLB. The Public Elementary/Secondary School Universe Survey provides schools’ numbers of subgroup students in tested grades, which serve as the running variables for the RD analyses.

Preliminary Results. NCLB’s subgroup-specific accountability significantly decreased one-year teacher turnover, but did not affect one-year teacher retention. Difference-in-difference estimates show that the effect of a school being held accountable for one additional subgroup was an additional decrease of .6% in one-year teacher turnover in the post-NCLB period (p=0.04). For Black and Hispanic teachers, teaching in a school held accountable for the same-race subgroup had a particularly strong effect, decreasing one-year turnover by 6% (p=0.09). Limiting the comparison to schools in narrow bandwidths on either side of states’ minimum subgroup size cutoffs, as well as adding school-by-time-period fixed effects, increases the magnitude of these estimates, although their precision varies. However, the relatively consistent magnitude of these effects suggests that subgroup-specific accountability had particularly strong negative effects on the turnover of minority teachers.