A fairly large number of different methods exist for evaluating comparisons or null hypothetical contrasts. Because it is difficult to choose among them, this paper develops a 2 by 2 classification of the methods. The first dimension of the classification is decision-based error rate vs. experimentwise error rate. The second dimension is planned…

When the results of a series of independent studies are combined, it is useful to quantitatively estimate the magnitude of the effects. Several methods for estimating effect size are compared in this paper. Glass' estimator and the uniformly minimum variance unbiased estimator are based on the ratio of the sample mean difference and the pooled…

Log linear models are proposed for the analysis of structural relations among multidimensional developmental contingency tables. Model of quasi-independence are suggested for testing specific hypothesized patterns of development. Transitions in developmental categorizations are described by Markov models applied to successive contingency tables. A…

The National Assessment of Educational Progress; Second Science Assessment No-Show Study assessed the magnitude and causation of nonresponse biases. A No-Show is defined as an individual who was selected as a sample respondent but failed to be present for regular assessment of the 17-year-old group. The procedure whereby a sample of eligible…

Multidimensional scaling (MDS) a highly reliable measurement technique, often requires an overwhelming task of the subject in the data collection procedure. This investigation was designed to determine the loss of precision in solution associated with five degrees of systematic reduction in the data collection task. Data were simulated via Monte…

The basic design for the national evaluation of the Follow Through program is presented, and some of the related issues of analysis are considered. The design, as it now stands, presents many difficulties for analysis. These analysis issues are seen to include the following: (1) What should be the unit of analysis?; (2) How is the effect of a…

A study was made of the problem of representing the expectations of mean squares associated with analysis of variance sources of variation for experimental designs. These designs have a factorial structure over repeated measures or, for some other reason, have variates within a factorial design not all of which are mutually independent. A simple…

Multiple matrix sampling is a psychometric procedure in which a set of test items is subdivided randomly into subtests of items with each subtest administered to different subgroups of examinees selected at random from the examinee population. Although each examinee receives only a proportion of the complete set of items, the statistical model…

To resolve a recent controversy between Klein and Cleary and Levy, a model for dichotomous congeneric items is presented which has mean errors of zero, dichotomous true scores that are uncorrelated with errors, and errors that are mutually uncorrelated. (Author)

Explicit formulas are derived for the asymptotic sampling variances and covariances of the maximum likelihood estimators for factor-analysis parameters in the special case where there is just one common factor. The effect of the number of variables on these variances and covariances is indicated. A formula is given showing to what extent the usual…

The Automatic Interaction Detector (AID) is discussed as to its usefulness in multiple regression analysis. The algorithm of AID-4 is a reversal of the model building process; it starts with the ultimate restricted model, namely, the whole group as a unit. By a unique splitting process maximizing the between sum of squares for the categories of…

A class of tests for normality using the ratio of two estimates of the standard deviation is generalized to provide a class of tests for multivariate normality using a characterization of the multivariate normal. The posers of some of the tests are examined numerically and compared with the power of a recent similar test. (Author)

Games (1973) criticized the Marascuilo and Levin (1970) "nested" or simple effects design as a poor substitute for the usual logic of a factorial analysis of variance. Marascuilo and Levin's use of "nesting" is completely idiosyncratic and contrary to well-established usage, thus confusing the student. Proper usage is…

A special case of Bloxom's version of Tucker's three mode factor analysis model is developed statistically. A goodness of fit test and an empirical example are presented. (Author/JKS)