Three separate papers are included in this report. The first describes a two-stage procedure for choosing from among several instructional programs the one which maximizes the probability of passing the test. The second gives the exact sample sizes required to determine whether a squared multiple correlation coefficient is above or below a known…

Based on a Monte Carlo simulation, this study is designed to investigate the power of the Kruskal-Wallis's H-test compared to the power of the F-test for three equal moderate sample sizes drawn at random from distributions of common or different shapes but for which the population distributions have equal variances. The distributions are the…

Describes an investigation of the use of document components, such as terms, sentences, and whole documents, for indexing and retrieval. A number of probabilistic similarity measures based on document components are studied, as well as a new method of handling probability estimations involving small sample sizes. (24 references) (Author/CLB)

The use of measures of magnitude of effect has been advocated as a way to go beyond statistical tests of significance and to identify effects of a practical size. They have been used in meta-analysis to combine results of different studies. Describes problems associated with measures of magnitude of effect (particularly study size) and…

Classroom demonstrations can help students gain insights into statistical concepts and phenomena. After discussing four kinds of demonstrations, the authors present three possible approaches for determining how much data are needed for the demonstration to have a reasonable probability for success. (Author/LMO)

Tau, Lambda, and Kappa are measures developed for the analysis of discrete multivariate data of the type represented by stimulus response confusion matrices. The accuracy with which they may be estimated from small sample confusion matrices is investigated by Monte Carlo methods. (Author/LMO)

Response data from 3,000 subjects to the SRA Attitude Survey were analyzed using the one-parameter and two-parameter logistic latent trait models. The effects of varying sample size on the accuracy of person and item parameters were investigated. (Author/BS)

Describes three methods of sample size determination, each having its use in investigation of social science problems: Attribute method; Continuous Variable method; Galtung's Cell Size method. Statistical generalization, benefits of cell size method (ease of use, trivariate analysis and trichotyomized variables), and choice of method are…

The purpose of this study was to examine empirical Type I error rates and rejection rates for three dimensionality assessment procedures with data sets simulated to reflect short tests and small samples. The TESTFACT G superscript 2 difference test suffered from an inflated Type I error rate with unidimensional data sets, while the approximate chi…

Using data from a national sample (n=716) of American business associations we examine the prevalence of workforce development activities among associations. This study examines the relationship between the level of workforce development services provided in the organization, organizational characteristics, economic development and partnerships.…

Many large-scale testing programs routinely pretest new items alongside operational (or scored) items to determine their empirical characteristics. If these pretest items pass certain statistical criteria, they are placed into an operational item pool; otherwise they are edited and re-pretested or simply discarded. In these situations, reliable…

Actual kurtotic and skewed data and varied sample sizes and estimation methods demonstrated that normal theory maximum likelihood and generalized least square estimators were fairly consistent and almost identical. Standard errors tended to underestimate the estimator's true variation but the problem was not serious for large samples. (SLD)

A method is proposed for partially analyzing additive ipsative data (PAID). Transforming the PAID according to a developed equation preserves the density of the transformed data, and maximum likelihood estimation can be carried out as usual. Simulation results show that the original structural parameters can be accurately estimated from PAID. (SLD)

Monte Carlo methods were used to study the accuracy and utility of estimators of overall error and error due to approximation in structural equation modeling. Effects of sample size, indicator reliabilities, and degree of misspecification were examined. The rescaled noncentrality parameter also was examined. Choosing among competing models is…

A strategy for evaluating/modifying covariance structure models (CSMs) is presented. The approach uses recent developments in estimation under nonstandard conditions and unified asymptotic theory related to hypothesis testing, and it determines the extent of sample size sensitivity and specification error effects by relying on existing statistical…