CA AND SPOD FOR THE ANALYSIS OF TESTS COMPRISED OF BINARY ITEMS

A test comprised of binary items has an associated theoretical structu re (TS) T-p(D, & E), which may be paired with any of a number of possi ble quantitative characterizations (QCs). The aim of test analysis is to examine whether a set of test items conforms to an appropriately ch osen QC. In this article, the authors consider one of the most common TSs, T-p(D, R, E), in which D = 1 construct, R = monotone increasing i tem/construct regressions, and E=errors in variables. The authors then consider two QCs appropriate for this TS and review tests for each. S everal examples are provided.