Evaluation of Penalties in Scoring Incomplete Adaptive Tests

• In Event: AERA Poster Session 6
  In Poster Session: Division D Section 1 Poster Session 2

Abstract

All rules for scoring adaptive tests are based on the assumption that the examinees respond to each of the items assigned to them. However, in real-world testing, examinees sometimes do not complete their tests. The question addressed in this research is how to score the incomplete response vectors of examinees who end their test prematurely. The majority of the methods currently in use for scoring incomplete vectors involve penalties designed to persuade examinees from stopping too early. Several of them are reviewed and evaluated against non-penalty methods as the baseline. The results of the study will help us have a better understanding of the scoring methods and to select a scoring rule with a penalty of known size.

Authors

• Qi Diao, ETS

• Wim J. van der Linden, University of Twente