English

ERIC ED555593: A Comparison of Three Methods for Computing Scale Score Conditional Standard Errors of Measurement. ACT Research Report Series, 2013 (7)

Professional standards for educational testing recommend that both the overall standard error of measurement and the conditional standard error of measurement (CSEM) be computed on the score scale used to report scores to examinees. Several methods have been developed to compute scale score CSEMs. This paper compares three methods, based on classical test theory, item response theory, and the four-parameter beta compound binomial model. The three methods are compared using data from a single form of the ACT® College Readiness Assessment. The results indicate that all three methods produce comparable results. Figures A-1 through A-8 are contained in the appendix.

Sign in with your free account to read, download EPUBs or listen.