A FORTRAN program for maximizing and cross-validating split-half reliability coefficients is described. Externally computed arrays of item means and covariances are used as input for each of two samples. The user may select a number of subsets from the complete set of items for analysis in a single run. (Author/JKS)

Construct validity patterns of test-criteria and item-criteria correlations are shown to be inconsistent across samples. The results of an investigation of construct validity patterns on two published personality scales is presented. (JKS)

Item analysis is a very important consideration in the test development process. It is a statistical procedure to analyze test items that combines methods used to evaluate the important characteristics of test items, such as difficulty, discrimination, and distractibility of the items in a test. This paper reviews some of the classical methods for…

A schedule was developed for assessing the future outlook of hospitalized psychiatric patients and administered to samples of patients from two different hospitals. A factor analysis was done for each sample. (CK)

The Relative Difficulty Ratio (RDR) was developed as an index of test or item difficulty for use when raw score means or item p-values are not directly comparable because of chance score differences. Computational RDR are described. Applications of the RDR at both the test and item level are illustrated. (Author/BW)

Use of the same term "split-half" for division of an n-item test into two subtests containing equal (Cronbach), and possibly unequal (Guttman), numbers of items sometimes leads to a misunderstanding about the relation between Guttman's maximum split-half bound and Cronbach's coefficient alpha. This distinction is clarified. (Author/JKS)

Test bias detection methods based on item response theory (IRT) are reviewed. Five such methods are commonly used: (1) equality of item parameters; (2) area between item characteristic curves; (3) sums of squares; (4) pseudo-IRT; and (5) one-parameter-IRT. A table compares these and six newer or less tested methods. (SLD)

Item-discrimination indices are numbers calculated from test data that are used in assessing the effectiveness of individual test questions. This article asserts that the indices are so unreliable as to suggest that countless good questions may have been discarded over the years. It considers how the indices, and hence overall test reliability,…

In test assembly, a fundamental difference exists between algorithms that select a test sequentially or simultaneously. Sequential assembly allows us to optimize an objective function at the examinee's ability estimate, such as the test information function in computerized adaptive testing. But it leads to the non-trivial problem of how to realize…