As a consequence of both No Child Left Behind and Race to the Top, states across the country have endeavored to “turn around” low-performing schools.  In this paper, we use a regression discontinuity design to determine the causal effects of school turnaround in North Carolina. We focus on a broad range of outcomes, including teachers’ perceptions of their working conditions and how they use their time.

As part of its Race to the Top grant, North Carolina received funding to turn around or close 118 schools.  The state identified the schools in spring 2010, and implementation began in the 2010-11 school year.  We focus on elementary and middle schools, where assignment to this program was based on a sharp cutoff on state accountability tests.  This discontinuity in assignment, with those just below the cutoff receiving turnaround treatment and those just above the cutoff continuing with business as usual, allows us to examine the effects of the treatment.  In preliminary work, we find no evidence of manipulation or other problems at the cutoff. 
This paper enhances the set of outcome measures such as student test scores or attendance typically used in studies of this type by drawing on results from multiple waves of a statewide survey of teachers and principals.  This survey provides rich information on how teachers perceive the quality of school leadership and teacher working conditions, and how they use their time.  Administered every other year, the survey asks all teachers in North Carolina about their time use (in hours) and various school leadership and working conditions factors (on a one to five Likert scale).  The survey was administered at baseline in 2010 and after implementation in 2012.  We will add 2014 results when data become available. 

Our preliminary results, based on two specifications (one parametric and one non-parametric) indicate that turnaround increased the amount of time teachers spend on collaborative planning by 20 to 44 minutes per week and on professional development by 24 to 29 minutes, outcomes that are consistent with what one might expect given the structure of the turnaround programs. In addition, however, it also increased time teachers spend on completing paperwork by 14 to 29 minutes and on preparation for local, state, and federal tests by 23 to 55 minutes per week.  The turnaround treatment has no statistically significant effect on overall teaching conditions (as perceived by teachers) although a consistently negative effect emerges on several measures.  For example, the teacher-reported quality of professional development in the treatment schools is around 0.15 standard deviations below that in control schools.  There was little change in teacher turnover during the first full implementation year, but by 2013 teacher turnover was 6.7 to 12.8 percentage points higher in turnaround schools. 

For students, turnaround corresponds to 0.3 to 1.0 percentage point lower average daily attendance, but it has no effect on overall test scores.  Specific subgroups may be particularly affected by the program, with female students and those with limited English proficiency showing larger test score declines.