The effect of violating the assumption of nonspeededness on ability and item parameter estimates in item response theory was studied through simulation under three speededness conditions. Results indicate that ability estimation was least affected by speededness but that substantial effects on item parameter estimates were found. (SLD)

The consequences of underextracting factors and components within and between the methods of maximum likelihood factor analysis and principal components analysis were examined through computer simulation. The principal components score and the factor score estimate (T. W. Anderson and H. Rubin, 1956) tended to become different with…

The theoretical approaches taken by the LOGIST and BILOG computer programs for estimating the parameters of the three-parameter logistic item response model, and their practical consequences, are compared. Guidelines for choosing between the programs for particular applications are offered. (SLD)

Parameter recovery in the graded-response model was investigated using the MULTILOG computer program under default conditions. Results from 36 simulated data sets suggest that at least 500 examinees are needed to achieve adequate calibration under the graded model. Sample size had little influence on the true ability parameter's recovery. (SLD)

Coefficient alpha was examined through computer simulation as an estimate of test reliability under violation of two assumptions. Coefficient alpha underestimated reliability under violation of the assumption of essential tau-equivalence of subtest scores and overestimated it under violation of the assumption of uncorrelated subtest error scores.…

Using computer simulation data, the effect of a single global outlier in two-group classification analysis was explored in terms of the outcome variables of change in classification results (PCHNG), change in misclassification rate (MISDIF), and change in precision of misclassification rate estimation. The precision of misclassification rate…

Using a Monte Carlo simulation, a bootstrap procedure was evaluated for setting a confidence interval on the unrestricted population correlation (rho) assuming various degrees of incomplete truncation on the predictor. Sample size was the most important factor in determining accuracy and stability. Sample size should be at least 50. (SLD)

The adequacy of multidimensional trait and item parameters using multidimensional item response theory analysis was studied through simulation using different degrees of correlation between two trait dimensions and the existence of a restricted trait on the second dimension. Conditions allowing adequate recovery of trait and item parameters are…

In a simulation study of ability and estimation of differential item functioning (DIF) in computerized adaptive tests, Rasch-based DIF statistics were highly correlated with generating DIF, but DIF statistics tended to be slightly smaller than in the three-parameter logistic model analyses. (SLD)

Correlation coefficients are frequently reported in educational and psychological research. The robustness properties and optimality among practical approximations when phi does not equal 0 with moderate sample sizes are not well documented. Three major approximations and their variations are examined: (1) a normal approximation of Fisher's Z,…

The influence of sample size (N) and model parsimony on a set of 22 goodness of fit indices was investigated, including those typically used in confirmatory factor analysis and some recently developed indices. For sample data simulated from 2 known population data structures, values for 6 of 22 fit indices were reasonably independent of N and were…

An investigation of the effects of randomly missing data in two-predictor regression analyses is described. The differences in the effectiveness of five common treatments of missing data on estimates of R-squared values and each of the two standardized regression weights is also investigated. Bootstrap sample sizes of 50, 100, and 200 were drawn…

A method for estimating the first eigenvalue of random data correlation matrices is reported, and its precision is demonstrated via comparison to the method of S. J. Allen and R. Hubbard (1986). Data generated in Monte Carlo simulations with 10 sample sizes reaching up to 1,000 were used. (SLD)

Satisfactory results for interpersonally incomparable ordinal survey responses can be obtained by assuming that rankings are based upon a set of multivariate normal latent variables that satisfy factor or ideal point models of choice. Two statistical methods based upon those assumptions are described and illustrated via simulations. (TJH)

Two new methods of estimating the mean population correlation (M) and the standard deviation of population correlations (SD) were suggested and tested by Monte Carlo simulations. Results show no consistent advantage to using the Pearson correlation or Fisher's Z in estimating M or SD; estimates from all methods are similar. (SLD)