Constructed-Response Items in an Equating Anchor: Important or Not?

• In Event: Roundtable Session 18
  In Roundtable Session: Methodological Advances in Equating

Abstract

This study compared the accuracy of equating a mixed-format test (70% MC, 30% CR) through several different common-item anchors. Some anchors contained CR items; some did not. Equating through anchors that included CR items was done with and without “trend scoring” of the CR anchor-items responses of the reference-form examinees. This study found that including CR items in the anchor with trend scoring of the responses resulted in more accurate equating than using an MC-only anchor. Including CR items in the anchor without trend scoring resulted in less accurate equating than using an MC-only anchor.

Authors

• Chi-Wen Liao, Educational Testing Service

• Andrea Marie Bontya, Educational Testing Service

• Samuel A. Livingston, ETS