The impact of outliers on Cronbach's coefficient [alpha] has not been documented in the psychometric or statistical literature. This is an important gap because coefficient [alpha] is the most widely used measurement statistic in all of the social, educational, and health sciences. The impact of outliers on coefficient [alpha] is investigated for…

A two-stage generalized least squares model is developed for estimating the linear regression of an individual outcome on a group characteristic in studies of multilevel data. Results of this model are compared to the results of analytic methods, and formulas are developed for assessing the accuracy of the traditional approaches. (Author/JAZ)

Determination of statistical power for analysis of variance procedures requires five elements: (1) significance level; (2) effect size; (3) number of means; (4) error variance; and (5) sample size. Significance levels are traditionally chosen to be 0.5, .01, or .001. Effect size is not discussed in this paper. The number of means is determined by…

Some approximate formulas are presented for standard errors of estimated regression coefficients in two-level designs. If the researcher can make a reasonable guess as to parameters occurring in the model, this approximation can be a guide to the choice of sample sizes at either level. (SLD)

Application of the bootstrap method to complex psychological analysis is illustrated using a simulation experiment with two populations with small and large samples. The method provides variance estimates, allows testing of nested competing models, and gives a preliminary idea about parameter variability. (SLD)

An equating design in which two groups of examinees from slightly different populations are administered a different test form with a subset of common items is widely used. This paper presents standard errors and a simulation that verifies the equation for large samples for an equipercentile equating procedure for this design. (Author/BS)

The use of three remedies for errors in the measurement of ability that arise from differences in rater stringency is discussed. Models contrasted are: (1) Conventional; (2) Handicap; and (3) deterministic Rater Response Theory (RRT). General model requirements, power, bias of measures, computing cost, and complexity are contrasted. Contrasts are…

Standard errors and bias of unidimensional and multidimensional ability estimates were compared in a factorial, simulation design with two item response theory (IRT) approaches, two levels of test correlation (0.42 and 0.63), two sample sizes (500 and 1,000), and a hierarchical test content structure. Bias and standard errors of subtest scores…

This Monte Carlo study examined the validity of the chi-square test for model evaluation in different instances of misspecification and sample size. The usefulness of the chi-square difference statistic to compare competing structures and improvement in fit is also addressed. (Author/BS)

Large sample standard errors of linear equating for the single-group design are derived without making the normality assumption. Two general methods based on the delta method of M. Kendall and A. Stuart (1977) are described. One method uses the exact partial derivatives, and the other uses numerical derivatives. Simulation using the beta-binomial…

The relationship between levels of reliability and the power of two bias and differential item functioning (DIF) detection methods is examined. Both methods, the Mantel-Haenszel (MH) procedure of P. W. Holland and D. T. Thayer (1988) and the Simultaneous Item Bias (SIB) procedure of R. Shealy and W. Stout (1991), use examinees' raw scores as a…

The effects of the violation of the assumption of normality coupled with the condition of multicollinearity upon the outcome of testing the hypothesis Beta equals zero in the two-predictor regression equation is investigated. A monte carlo approach was utilized in which three differenct distributions were sampled for two sample sizes over…

One of the most frequently used research methods in education and psychology involves repeated observations on the same individuals. When sample sizes are relatively small and a multivariate analysis lacks power, there are currently two analytical options in testing time effects. One is to assume a time series structure to these observations, and…

This study investigated seven methods for analyzing multivariate group differences. Bonferroni t statistics, multivariate analysis of variance (MANOVA) followed by analysis of variance (ANOVA), and five other methods were studied using Monte Carlo methods. Methods were compared with respect to (1) experimentwise error rate; (2) power; (3) number…

The statistical properties of two methods of estimating gain scores for groups in quasi-experiments are compared: (1) gains in scores standardized separately for each group; and (2) analysis of covariance with estimated true pretest scores. The fan spread hypothesis is assumed for groups but not necessarily assumed for members of the groups.…