The Impact of School Accountability Sanctions on Student Outcomes: Evidence from North Carolina

Abstract of work: We take advantage of structural policy breaks of the dual accountability system in North Carolina (federal NCLB and NC-specific ABC) and use RD analysis to evaluate the effectiveness of accountability on student performance, using administrative data provided by NC Department of Public Instruction. This is the first of a series of papers that we will write, which takes advantage of the two accountability systems and their different approaches. Some of these include: • Positive vs. negative sanctions. Teachers and sd10d personnel may be more inclined to alter \reir behavior to receive an award rather than avoid a punishment. • Rewarding proficiency versus rewarding value added. Teachershavegreaterinfluenceover the amount of leaming a student accomplishes in a current year, relative to the influence they have over a studenfs starting 1XJint.Rewarding teachers for a measure that they have more direct infll.ef1Ceover may be a more cost-effective means of induci1g changes in behavior. • Does choice make a difference? Some preli'ninary reports indcate that the take-up rcie d choire in sd100Is that faato meet AyP is very low. The threat of inIrodLK:ingchoice, the first sanction in the tra:liIionaI NCLB fi'ameINork, may be of littleconsequence in oommunities IM1erea large degree of choice, in the form of chater schoois or IiJeraIstudent transfer p:>IK::iesa,lready exists. PJ. the same time, voocher1hreats have shoM1 some degree of effediveness in other settings. • Are school-level sanctions less powerful than teacher-level sanctions? While none of the sanctions we evaluate are true teacher-level sanctions, we can make some inferences about the strength of incentive dilu1ioneffects by comparing effects across schools of varying size. • Can incentivization be a positive-sum game for students? Some evidence suggests that sa1dions lead to test score gans for some students, but losses for others. North Carolina adopted a policy change in 2005 designed to address this concern, and we can determine whether this alteration had any impact on the distribution of test score improvements within schools.