Proposes true score models that can account for correlated errors and their effect on coefficient alpha. These models allow random measurement errors on earlier items to affect directly or indirectly the scores on later items. Conditions under which coefficient alpha may yield spuriously high estimates or reliability are discussed. (SLD)

Evaluated coefficient alpha under violations of two classical test theory assumptions: essential tau-equivalence and uncorrelated errors through simulation. Discusses the interactive effects of both violations with true and error scores. Provides empirical evidence of the derivation of M. Novick and C. Lewis (1993). (SLD)

A Markov chain Monte Carlo (MCMC) method and a bootstrap method were compared in the estimation of standard errors of item response theory (IRT) true score equating. Three test form relationships were examined: parallel, tau-equivalent, and congeneric. Data were simulated based on Reading Comprehension and Vocabulary tests of the Iowa Tests of…

Canonical analysis is frequently used in studies of relationships between sets of variables which are difficult to measure accurately, partly because of the true nature of the data and partly because of errors associated with the measurement instruments. Meredith's solution to the fallible data problem is examined. (Author/BJG)

An equation comparable to Spearman's correction for attenuation, which does not depend upon the assumption that error scores are uncorrelated with true scores and with other sets of scores, is derived. (Editor)

A new formula, expressing relative efficiency solely in terms of the standard errors of measurement and the frequency of distributions of true scores, is developed for the relative efficiency of two tests measuring the same trait. Subtests from the Scholastic Aptitude Test provide a numerical illustration. (Author/RC)

Although partial correlation is a correlation of residuals, the correlation of the true-score components of these residuals is not equivalent to the partial correlation of the true scores themselves. The source of this discrepancy is explained and its implications are briefly discussed. (Author)

This paper provides formulas for expected true-score measures and reliability of binary items as a function of their Rasch difficulty parameters when the trait distribution is normal or logistic. With the proposed formula, one can evaluate the theoretical values of classical reliability indexes for norm-referenced and criterion-referenced…

Purpose of this paper is to provide an alternative formulation which allows for the model parameters to be determined given the structural specification of zero mean error and independence among errors for different items and between errors and true scores. (Authors)

A correction by Bond (TM 504 914) of an error by Tucker, Damarin, and Messick is acknowledged. A formula for the correlation between initial true test scores and true difference scores is presented. Observed score considerations should be replaced by emphasis on true score considerations. (Author/CTM)

Shows that the Harris factors of R have psychometric properties similar to those discussed by Kaiser and Caffrey (1965) and Bentler (1968). Specifically it is shown that the Harris factors of R maximize a lower-bound to the reliability of a composite measure derived by Guttman (1945). (Author/RC)

Examines the effects of four parameters on the accuracy of test equating under a relaxed definition of test form equivalence. The four parameters studied were sample size, test form length, test form reliability, and the correlation between true scores of the test forms to be equated. (CK)

The consequences of the assumption that the expected score is equal to the true score are shown and alternatives discussed. (MS)