Although it makes a great deal of sense to attempt to design reading research so as to be able to generalize beyond the particular reading passages used in an experiment, the suggestion that tests of statistical significance are a necessary part of making valid generalizations is nonsense. (HOD)

This paper explains how to combine probabilities from discrete distributions, such as probability distributions for nonparametric tests. (Author/BW)

Estimation is by means of iterative convergence to maximum likelihood estimates, and two approaches to assessing fit of the model to sample data are discussed. Relation of this general probabilistic model to other, more restricted models is explored and three cases of the general model are applied to exemplary data. (Author/RC)

Indicates that the explanation of what it means to obtain a significant F-ratio in discriminant function analysis, presented by K. E. Anderson in an earlier volume of Science Education, is incorrect. (JR)

Remarkability is introduced as a quantifiable attribute of given data and as a basis upon which one may rationally judge its scientific value. Applications of remarkability theory to various research and statistical problems and procedures are discussed. (DG)

Monte Carlo results were used to evaluate procedures for discriminating between groups. A multiple testing version of Hotelling's T-squared and the Bonferroni procedure were most powerful for detecting at least one true difference, depending on conditions examined. A multiple Bonferroni procedure was superior in power for detecting all true…

Practical formulas for several analysis of variance (ANOVA) designs and models are presented which make it possible for readers to compute strength of association measures without the use of complete ANOVA tables. (Author/PN)

Points out that 108 studies in the first 25 volumes of AV Communication Review used univariate analysis of variance to test experimental hypotheses, and suggests that researchers include omega-squared values in their reports to permit estimation of the relationship between x and y variables, and correlation data for comparing studies. (MER)

Suggests six analytic areas for dual-career research (marriage type, domestic responsibility, satisfaction, self-image, career salience, and career line); provides questions for a large-scale survey research project; and analyzes scale types. The scales provide useful direction for dual-career family research. (Author)

Three nonparametric alternatives to the parametric Bartlett test are presented for handling the K-sample equality of variance problem. The two-sample Siegel-Tukey test, Mood test, and Klotz test are extended to the multisample situation by Puri's methods. These K-sample scale tests are illustrated and compared. (Author/GDC)

Hawkin's procedure for testing a sequence of observations for a shift in location could have applicability for assessing change within a single subject. Monte Carlo results suggest that Hawkins' procedure is robust with respect to moderate violations of its underlying assumptions of homogeneity of variance and normality. (Author/GDC)

Errors of misclassification and their effects on categorical data analysis are discussed. The chi-square test for equality of two proportions is examined in the context of errorful categorical data. The effects of such errors are illustrated. A correction procedure is developed and discussed. (Author/MH)

A Monte Carlo approach was used to examine bias in the estimation of indirect effects and their associated standard errors. Results illustrate the adverse effects of nonnormality on the accuracy of significance tests in latent variable models estimated using normal theory maximum likelihood statistics. (SLD)