Brown Bag Seminar: Maoxin Zhang, CEMO

Title: Understanding omission behavior from process data: A joint model of responses, response times, and response actions

Abstract: Omission behavior, seeing an item but not responding, is widely observed in low-stakes, large-scale assessments. Treating omitted items as missing at random or incorrect answers might distort parameter estimates and lead to biased conclusions. In this study, we intend to jointly model omission behavior and response behavior via multidimensional latent variable analysis based on process data. We aim to (a) get a better understanding of omission behavior with the information from response times and response actions, (b) improve the precision of item and person parameter estimates, and (c) investigate how ability, speed, interaction propensity, and omission propensity relate to each other.


Published June 7, 2022 11:49 AM - Last modified June 7, 2022 11:49 AM