Emphasizing the measurement of causal effects to arrive at a better understanding of the causal mechanisms involved in statistical theory, a mathematical model for causal inferences in prospective studies is developed and then applied to retrospective case-control studies. Before developing the model, causal agents are delineated, and causal…

A method for comparing the cross validated classification accuracies of linear and quadratic classification rules is presented under varying data conditions for the k-group classification problem. With this method, separate-group as well as total-group proportions of correct classifications can be compared for the two rules. McNemar's test for…

The applicability of the Analysis of Variance, ANOVA, procedures to the analysis of dichotomous repeated measure data is described. The design models for which data were simulated in this investigation were chosen to represent simple cases of two experimental situations: situation one, in which subjects' responses to a single randomly selected set…

A method requiring minimal computational effort is presented for transforming ordered residuals for purposes of testing the correctness of a regression model. The method maintains the same logical ordering in the transformed residuals as that of the original residuals and is suitable for either correlated or uncorrelated data. (Author/BS)

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)

The purpose of this paper is to demonstrate how multiple linear regression provides a viable statistical methodology for dealing with meta-analysis in general, and specifically with the issues of nonindependence and design complexity, such as multiple treatments. Since the F-test and t-test are special cases of the general linear model,…

Meta-analysis has become an important supplement to traditional methods of research reviewing, although many problems must be addressed by the reviewer who carries out a meta-analysis. These problems include identifying and obtaining appropriate studies, extracting estimates of effect size from the studies, coding or classifying studies, analyzing…

The testing of a priori contrasts, post hoc contrasts, and experimental error rates are discussed. Methods for controlling the experimental error rate for a set of a priori contrasts tested simultaneously have been developed by Dunnett, Dunn, Sidak, and Krishnaiah. Each of these methods is discussed and contrasted as to applicability, power, and…

Since computer programs have been available for estimating and testing linear causal models, these models have been used increasingly in the behavioral sciences. This paper discusses the problem that very different causal structures may fit the same set of data equally well. (Author/LMO)

Statistical methods related to Rasch modeling, social network analysis, and latent variable structural equation modeling for experimental designs are critiqued with regard to three important issues: (1) measuring constructs; (2) modeling processes of change; (3) operationalizing social relations. An argument is made for adequate conceptualization…

This report describes a statistically designed experiment that was carried out in an operational setting to determine the contributions of different sources of variation to the unreliability of scoring. The experiment made novel use of partially balanced incomplete block designs that facilitated the unbiased estimation of certain main effects…

The dynamic linear model or Kalman filtering model provides a useful methodology for predicting the past, present, and future states of a dynamic system, such as an object in motion or an economic or social indicator that is changing systematically with time. Recursive likelihood methods for adaptive Kalman filtering and smoothing are developed.…

A statistical test is described which determines homogeneity of effect size of an experiment series. An overall fit statistic is partitioned into between-class fit statistic and within-class fit statistic. These statistics permit assessment of differences between effect sizes for different classes and homogeneity of effect size within classes.…

This article extends both the logic and the statistical procedures used in a recent analysis of Hyde's data on gender differences in cognitive abilities by Rosenthal and Rubin. The logic of a "model fitting" approach to meta-analysis is described. Relevant statistical procedures and goodness-of-fit tests are illustrated. (Author/BS)