Abstract
Student assessment nonparticipation (or opt out) has increased substantially in K-12 schools in states across the country. This increase in opt out has the potential to impact achievement and growth (or value-added) measures used for educator and institutional accountability. In this simulation study, we investigated the extent to which value-added measures of teacher quality are affected as a result of varying degrees of opt out, as well as a result of various types of nonrandom opt out. Results show that the magnitude of opt out and choice of classification scheme has a greater impact on value-added estimates than the type of opt-out patterns simulated in this study. Specifically, root mean square differences in value-added estimates increased as magnitude of opt out increased. In addition, teacher effectiveness classification agreement decreased as opt out magnitude increased. One type of opt out, where the highest achieving students in the highest achieving classrooms opted out, had the largest impact on stability than the other types of opt outs.