Type I error rate and power for the t test, Wilcoxon-Mann-Whitney test, van der Waerden Normal Scores, and Welch-Aspin-Satterthwaite (W) test are compared for two simulated independent random samples from nonnormal distributions. Conditions under which the t test and W test are best to use are discussed. (SLD)

The standardized group mean difference, Cohen's "d", is among the most commonly used and intuitively appealing effect sizes for group comparisons. However, reporting this point estimate alone does not reflect the extent to which sampling error may have led to an obtained value. A confidence interval expresses the uncertainty that exists between…

The wide availability of large amounts of inexpensive computing power has encouraged statisticians to explore many approaches to a basis for inference. This paper presents one such "computer-intensive" approach: the bootstrap of Bradley Efron. This methodology fits between the cases where it is assumed that the form of the distribution…

Describes an activity designed to introduce concepts of randomness and standard deviation. (CP)

Weighting and variance estimation are two statistical issues involved in survey data analysis for large-scale assessment programs such as the Higher Education Information and Communication Technology (ICT) Literacy Assessment. Because survey data are always acquired by probability sampling, to draw unbiased or almost unbiased inferences for the…

Possible bias due to sampling problems or low response rates has been a troubling "nuisance" variable in empirical research since seminal and classical studies were done on these problems at the beginning of this century. Recent research suggests that: (1) earlier views of the alleged bias problem were misleading; (2) under a variety of fairly…

BRRVAR, which uses the Balanced Repeated Replication approach, was designed for use with the Statistical Analysis System (SAS). It was created for the National Center for Education Statistics, to enlarge their capacity to estimate and analyze sampling errors for statistics generated from educational surveys with complex sampling designs. BRRVAR…

A common problem in the behavioral sciences is to determine if a set of observed variables can be more parsimoniously represented by a smaller set of derived variables. To address this problem, the performance of five methods for determining the number of components to retain (Horn's parallel analysis, Velicer's Minimum Average Partial (MAP),…

This report is based on the results of the Survey of Doctoral Scientists and Engineers and the Nonresponse Bias Survey published by the Commission on Human Resources of the National Research Council in 1975. Possible biases in estimates of variables are examined. Estimates based on data from the 1975 survey respondents, plus the nonresponse bias…

A linear classification rule (used with equal covariance matrices) was contrasted with a quadratic rule (used with unequal covariance matrices) for accuracy of internal and external classification. The comparisons were made for seven situations which resulted from combining three data conditions (equal and unequal covariance matrices, minimal and…

Achievement tests which are specifically linked to an instructional program and have been developed in relation to an objectives base and/or to an item generation rule are considered, as well as student response data. Three types of studies are outlined and the kind of procedures thought useful illustrated. As various methods for examining…

The intent of the study was to determine the extent to which test statistics computed by the unweighted means analysis are F-distributed. Applicability criteria were sought in terms of the number of factor levels and the degree to which cell frequencies differ. The unweighted means analysis, a frequently used approximate analysis, was contrasted…

Described and listed herein with concomitant sample input and output is the Fortran IV program which estimates parameters and standard errors of estimate per parameters for parameters estimated through multiple matrix sampling. The specific program is an improved and expanded version of an earlier version. (Author/BJG)

One solution to the problem of obtaining test scores in order to assess multiple outcomes of educational programs is a basically simple technique called matrix sampling. Matrix sampling is the estimation of total test score statistics by administering random subsets of test items to randomly selected students. This chapter lays out the specific…

An investigation was conducted of the characteristics of two estimation procedures and corresponding test statistics used in the analysis of completely randomized factorial experiments when observations are lost at random. For one estimator, contrast coefficients for cell means did not involve the cell frequencies. For the other, contrast…