This study sought to develop an instrument for the assessment of grief in caregivers of persons with Alzheimer's disease. Study resulted in a 50-item scale containing three factors. Results suggest that caregiver grief is neither a unitary nor static construct and that a scale such as this may be appropriate for use in supportive, clinical, and…

The book answers practical questions about item response theory (IRT) in four sections: (1) an introduction to the concepts; (2) IRT theories and principles; (3) fundamentals of IRT; and (4) applications of IRT models. The book is useful for anyone trying to learn about IRT and how to use it. (SLD)

An overview of four papers on useful future directions for psychometric research is provided. The papers were drawn from American Psychological Association symposia; they cover the nature of general intelligence, item bias and selection, cut scores, equating problems, computer-adaptive testing, and individual and group achievement measurement.…

Properties of dichotomous Guttman-scalable items are described. Both the elements and eigenvalues of the Pearson correlation matrix of such items can be expressed as simple functions of the number of items if the score distribution is uniform and there is an equal number of items at each difficulty level. (SLD)

Examined reliability of United States Employment Service Interest Inventory (USES-II) by twice administering USES-II to 100 vocational rehabilitation clients. Retest reliability coefficients for the 12 scales of the USES-II ranged from .73 to .88 with a median of .83 for this population. Findings support use of USES-II in occupational exploration…

Some of the most commonly used item response theory models for ordered polychotomous data are described using the taxonomy of D. Thissen and L. Steinberg (1986) as a framework. An empirical study found the graded response model and an extension of the partial credit model very similar in a practical application. (SLD)

Reviews Peabody Individual Achievement Test-Revised (PIAT-R), a revision of the 1970 PIAT. Describes administration and scoring criteria, psychometric properties, and new features of test design. Emphasizes implications for the practitioner. Concludes that PIAT-R may be useful screening tool when general information concerning student achievement…

Evaluates the extent to which newer IQ tests represent an improvement over their historical predecessors, the Binet and Wechsler scales, in terms of the potential for improved educational and psychological service delivery. Discusses three conceptually relevant criteria: theoretical foundations, psychometric integrity, and implications for…

Reviews issues, raised in part because of "Educational and Psychological Measurement" (EPM) policies, regarding "test reliability," which is psychometric terminology, and "score reliability," score-centric terminology. Discusses datametrics and provides a critique of T. Vacha-Haase's proposed meta-analytic reliability generalization via…

Responds to criticisms of some "Educational and Psychological Measurement" policies and the reliability generalization meta-analytic methods of T. Vacha-Haase. Explores consequences of misunderstanding score reliability. (SLD)

B. Thompson and T. Vacha-Haase have examined the statement "the reliability of the test" with emphasis on the following three words: (1) the first "the"; (2) "test"; and (3) the second "the." This discussion focuses instead on the word "reliability." (Author)

Describes determining student responses to teaching through their drawings of the class situation. (SLD)

The classification capabilities of the one-dimensional Kohonen neural network (T. Kohonen, 1995) were compared with those of two partitioning and three hierarchical cluster methods in 2,580 data sets with known cluster structure. Overall, the performance of the Kohonen networks was similar to, or better than, that of the others. Implications for…

The accuracy of Kendall's chi square approximation (M. Kendall, 1962) to circular triad distributions was assessed by comparing estimated indices to enumerated values across upper and lower tails of cumulative proportions of circular triad distributions. In general, the approximation was found to be quite good, and its use is discussed. (SLD)

As examples of models that are not based on normality or its approximation, the logistic positive exponent family of models is discussed. These models include the item task complexity as the third parameter, which determines the single principle of ordering individuals on the ability scale. (SLD)