The use of confirmatory factor analysis by the LISREL program is demonstrated as an assumption-testing method when computing reliability coefficients under different model assumptions. Results indicate that reliability estimates are robust against departure from the assumption of parallelism of test items. (SLD)

Six stepwise multiple-comparison procedures for repeated-measures means were compared for their overall familywise rates of Type I error when multisample sphericity and multivariate normality were not satisfied. Robust stepwise procedures were identified by Keselman, Keselman, and Shaffer (1991) with respect to three definitions of power. (SLD)

Using Monte Carlo methods, the applicability of large sample theory to maximum likelihood estimates of total indirect effects in sample sizes of 50, 100, 200, 400, and 800 was studied. Samples of at least 200 and 400 are required for the recursive and nonrecursive models, respectively, that were assessed. (TJH)

A review of the literature indicates that the two best procedures for testing variances are one that was proposed by O'Brien (1981) and another that was proposed by Brown and Forsythe (1974). An examination of these procedures for a variety of populations confirms their robustness and indicates how optimal power can usually be obtained. (SLD)

Data on 2,600 children over 21 years were examined for high school graduation, years of schooling completed, teen childbearing, and neighborhood variables (household incomes, race, high school dropouts, female-headed families, unemployment rate). Although neighborhood characteristics may affect outcomes, reliability of estimates of their impact…

Determined the robustness of several types of scoring (number-right; one-, two-, and three-parameter item response; finite-state, and partial-credit) with respect to the violation of the equally classifiable options and option independence made in finite-state scoring using analysis of test responses of 1,232 high school seniors. (SLD)

"How New York City's Charter Schools Affect Achievement" estimates the effects on student achievement of attending a New York City charter school rather than a traditional public school and investigates the characteristics of charter schools associated with the most positive effects on achievement. Because the report relies on an…

This report presents findings from a review of the best evidence on narrowing the gap in outcomes across the five Every Child Matters (ECM) areas for vulnerable groups in the context of improving outcomes for all. The review was commissioned to prepare the ground for work on "Narrowing the Gap" with participating local authorities (LAs).…

This article explores the effect on assessment of "translating" paper and pencil test items into their computer equivalents. Computer versions of a set of mathematics questions derived from the paper-based end of key stage 2 and 3 assessments in England were administered to age appropriate pupil samples, and the outcomes compared.…

Large scale surveys usually employ a complex sampling design and as a consequence, no standard methods for estimation of the standard errors associated with the estimates of population means are available. Resampling methods, such as jackknife or bootstrap, are often used, with reference to their properties of robustness and reduction of bias. A…

Bayesian network models offer a large degree of flexibility for modeling dependence among observables (item outcome variables) from the same task that may be dependent. This paper explores four design patterns for modeling locally dependent observations from the same task: (1) No context--Ignore dependence among observables; (2) Compensatory…

A Monte Carlo study was conducted to investigate the robustness of the assumed error distribution in maximum likelihood estimation models for multidimensional scaling. Results show that violations of the assumed error distribution have virtually no effect on the estimated distance parameters. (SLD)

Robustness of the school-level item response theoretic (IRT) model to violations of distributional assumptions was studied in a computer simulation. In situations where school-level precision might be acceptable for real school comparisons, expected a posteriori estimates of school ability were robust over a range of violations and conditions.…

A maximum likelihood approach to correlational outlier identification is introduced and compared to the Mahalanobis D squared and Comrey D statistics through Monte Carlo simulation. Identification performance depends on the nature of correlational outliers and the measure used, but the maximum likelihood approach is the most robust performance…

Implications of metanalytic results from Monte Carlo studies of the robustness of the F test in the one- and two-factor fixed effects analysis of variance (ANOVA) models and Monte Carlo results for the B. L. Welch (1947) and Kruskal-Wallis (1952) tests are discussed. (SLD)