Panel: Assessment, Incentives, and Policy Evaluation in K-12 Education

Saturday, November 4, 2017: 3:30 PM-5:00 PM
Gold Coast (Hyatt Regency Chicago)

*Names in bold indicate Presenter

Panel Organizers:  John B. Klopfer, Princeton University
Panel Chairs:  James H Wyckoff, University of Virginia
Discussants:  Andrew McEachin, RAND Corporation and Kenneth A. Shores, University of Pennsylvania

Does Incentivizing Value Added Make It More or Less Meaningful?
Isaac Opper, RAND Corporation and Michael Dinerstein, National Bureau of Economic Research

Achievement testing data are frequently used to evaluate the impacts of K-12 education programs. Meanwhile, federal, state, and local policies set incentives for test performance. How do test-based incentives affect the impacts and evaluation of education policies? Isaac Opper and Michael Dinerstein provide evidence that incentivizing teacher value added in the New York City teacher tenure process raises test scores and grades, but these gains fade out more rapidly, and hence over-state long-run impacts. Vivian Wong, Coady Wing, David Martin, and Anandita Krishnamachari design an improved measure of state discretion to weaken accountability standards set under the No Child Left Behind (NCLB) Act, and provide evidence that more stringent implementation of NCLB by the states improves state-collected measures of school performance, but not outcomes on the un-incentivized National Assessment of Educational Progress (NAEP). John Klopfer studies reforms to state school finance legislation, and provides evidence that the NAEP, which is administered early in the year, under-reports gains from reforms because schools spend new funds to add school days after testing; in addition, schools that are incentivized to extend the school year impose large opportunity costs on students and their families. Andrew Johnston provides evidence on the salary and benefit preferences of K-12 teachers, linking these to value-added measures to study the relationship between compensation structure and teacher quality. Taken together, these results show that linking tests to incentives affects the validity of testing programs, and that test design indirectly affects incentives.

See more of: Education
See more of: Panel