Proposes a test for change in composite reliability as a result of scale revision or modification. The test, which is useful for examining whether a revision changes the population reliability of a scale undergoing development, is illustrated with two examples. (SLD)

Developed a general framework to estimate total test or testlet reliability with either fixed or random factors that is based on hierarchical factor analysis. Illustrates the framework with an example based on real data. (SLD)

Studied the sample coefficient alpha for each of the five subscales of the Hopkins Symptom Checklist (HSL; L. Derogaitis and others, 1974) in a sample of 419 adults. Findings show that the normal-theory-based distribution has a systematic bias in describing the behavior of the sample coefficient alpha. (SLD)

Illustrates that there is a critical distinction between reliability and accuracy, explaining through a hypothetical example of shoe fitting that high test reliability does not guarantee good accuracy without consideration of percentile rank scoring, complex measurement models, and other technical detail. (SLD)

It is recognized that the typical difference score used by educators is likely to be relatively weak in reliability. Formulas are developed for estimating changes in the reliability of X-Y under the assumption of equal or unequal variances before and after reliabilities are modified. (SLD)

Explains the use of Strahan's Consistency Index to manage the occurrence of ties in Holland's occupational type codes. (SK)

Developed a statistical test for the hypothesis of the equality of extrapolated coefficients when the original values (Cronbach's alpha) are based on the same sample of persons and, therefore, are statistically dependent. Monte Carlo studies show that the test precisely controlled Type I error, even with small numbers of parttest units or raters.…

This article provides a meta-analysis of 38 studies of J. E. Helms and R. T. Carter's (1990) White Racial Identity Attitude Scale. Systematic measurement errors are implicated as influences on the measure's reliability and validity. (Author/MKA)

Contrary to common belief, reliability estimates of number-right multiple-choice tests are not inflated by speededness. Because examinees guess on questions when they run out of time, the responses to these questions generally show less consistency with the responses of other questions, and the reliability of the test will be decreased. The…

Bruce Thompson's edited volume begins with a basic principle, one might call it a basic truth, "reliability is a property that applies to scores, and not immutably across all conceivable uses everywhere of a given measure." (p. 3). The author claims that this principle is little known and-or little understood. While that is an arguable point, the…

The purpose of this study was to examine the relation between academic achievement and parenting styles among Saudi college students. The participants in this study consisted of 349 male and 219 female college students, whose age ranged from 18-25 years. The instruments for the current study were: (a) an Arabic translation of the Revised Parental…

Much research on requests has been carried out among L1 Chinese adults, L1 Chinese children, L1 children, L2 adults, and L2 children, but no studies to date have simultaneously examined Chinese children's requests in Chinese (L1) and English (L2). The aim of this study is to investigate how Taiwanese elementary school children vary requests…