Adaptive Testing With Complex and Severe Constraints in Licensure Testing

• In Event: 65.079 - Roundtable Session 23
  In Roundtable Session: 65.079-14 - Issues in Test Assembly

Abstract

Item selection constraint are typically incorporated into computerized adaptive testing algorithms to ensure the final test forms have comparable, if not superior, statistical and nonstatistical qualities with the paper-and-pencil tests. As the constraints become complex and severe, an effective and efficient adaptive testing method is needed to handle the burgeoning computational complexity and intensity. This study investigates the performances of four different adaptive testing methods in the context of licensure testing with complex and severe constraints. Results are subsequently evaluated and discussed in terms of estimation precision, decision accuracy, item exposure rate control, item pool usage rate, constraint violation, and consumed computing time. This study should advise practitioners how to implement an adaptive test with complex and severe constraints.

Authors

• Xiao Luo, National Council of State Boards of Nursing

• Doyoung Kim, National Council of State Boards of Nursing

• Ada Woo, National Council of State Boards of Nursing