It is argued that time generalizability exists independent of population and situation generalizability, although the three may interact. Problems associated with sampling and randomization efficacy over time are discussed. (Author/GK)

Guttman's assumption underlying his definition of "total images" is rejected. Partial images are not generally convergent everywhere. Even divergence everywhere is shown to be possible. The convergence type always found on partial images is convergence in quadratic mean; hence, total images are redefined as quadratic mean-limits.…

This article presents a fairly quick, easy way to decide what size sample is needed for a survey, given a desired level of confidence and degree of accuracy. The method proposed will work on any multiple response instrument. A technical mathematical explanation is also included. (GK)

The distributional characteristics of Kaiser's Measure of Sampling Adequacy (MSA) were investigated in sample matrices generated from multivariate normal populations of specified correlation levels. Systematic variation of sample size and number of variables revealed the overall MSA to be most influenced by the number of variables. (Author/JKS)

Criticizes an article in which S. A. Cohen and J. S. Hyman describe as "suspicious" the high proportion of statistically significant findings reported in published educational research journals and offers alternative explanations for the high statistical power reported. (GC)

This paper surveys available techniques and introduces an explicit statement of a new statistic to test for equality of correlated proportions in a polychotomous response design. A set of guidelines for the potential user of the techniques is provided. (Author/CTM)

A study was conducted using generated data sets that contained specified amounts of error to determine empirically which of two large sample approximations for the coefficient of concordance or weighted average tau was more appropriate for various numbers of judges and numbers of objects. (CTM)

Tests constructed through item-sampling (where several tests are constructed and randomly assigned to testees) were compared to a traditional constant-item procedure for evaluating student and class progress in a community college freshman psychology class. The item-sampling approach did not appear to be superior. (JKS)

Tests constructed through item-sampling (where several tests are constructed and randomly assigned to testees) were compared to a traditional constant-item procedure for evaluating class progress in a community college intermediate algebra class. The item-sampling format appeared to be slightly less valid than the traditional format. (JKS)

Two studies, both using Type 12 sampling, are presented in which the effects of violating the assumption of essential parallelism in setting confidence intervals are studied. Results indicate that as long as data manifest properties of essential parallelism, the two methods studied maintain precise Type I error control. (SLD)

The distinction between Type 1 and Type 12 sampling in connection with measurement data is discussed, and a method is presented for simulating data arising from Type 12 sampling. A Monte Carlo study is described that shows conditions under which precise confidence level control under Type 12 sampling is maintained. (SLD)

Examines the undercount problem in the 1990 census, the ways in which the Census Bureau has dealt with it in the past, and the Bureau's plans to avoid future undercounts. It also discusses the statistical concept of dual systems estimation in census counting as a means of lessening an undercount. (GR)

Studied three procedures for estimating the standard errors of school passing rates using a generalizability theory model and considered the effects of student sample size. Results show that procedures differ in terms of assumptions about the populations from which students were sampled, and student sample size was found to have a large effect on…

The sample invariant properties of five anchor test equating methods are addressed. Equating results across two sampling conditions--representative sampling and new-form matched sampling--are compared for Tucker and Levine equally reliable linear equating, item response theory true-score equating, and two equipercentile methods. (SLD)

An approximate statistical test of the equality of two intraclass reliability coefficients based on the same sample of people is derived. Such a test is needed when a researcher wishes to compare the reliability of two measurement procedures, and both procedures can be applied to results from the same group. (SLD)