Optimizing the length of computerized adaptive testing for the Force Concept Inventory

As a method to shorten the test time of Force Concept Inventory (FCI), we suggest use computerized adaptive testing (CAT). CAT is process administering on computer, with items (i.e., questions) selected based upon responses examinee prior items. In so doing, length can be significantly shortened. step develop CAT-based version FCI (FCI-CAT), examined optimal FCI-CAT such that accuracy and precision [which measure in terms bias, standard error, root-mean-square error (RMSE)] Cohen’s d would comparable full for given class size. First, estimated item parameters three-parameter logistic model response theory, which are used algorithm CAT. For this estimation, 2882 Japanese university students. Second, conducted Monte Carlo simulation analyze how RMSE depend length. Third, post hoc examine consistency results what have been obtained using empirical responses. comparison, 86 pairs pre- post- result, found size 40, may reduce 15–19 items, thereby reducing 50%–63%, an accompanying decrease only 5%–10%. The study were consistent, supports adequacy our its relevance real classrooms.1 MoreReceived 29 September 2020Accepted 26 January 2021DOI:https://doi.org/10.1103/PhysRevPhysEducRes.17.010115Published by American Physical Society under Creative Commons Attribution 4.0 International license. Further distribution work must maintain attribution author(s) published article’s title, journal citation, DOI.Published SocietyPhysics Subject Headings (PhySH)Research AreasAssessmentConcepts & principlesTechnologyPhysics Education Research