Analysis of variance via confidence intervals /

Informed by Jason Hsu's 1996 hierarchy of levels of inference on comparisons, this book provides a treatment of fixed-effects analysis of variance (ANOVA). Bird indicates that traditional approaches to ANOVA are not as informative as interval estimates of contrasts on effect parameters. The confidence interval approach lays the foundation for understanding contrasts and effect parameters such as Cohen's f. Some of the topics discussed here are not normally found in ANOVA books; these include the MANOVA-model approach to confidence interval construction with minimally one within-subject factor, nonstandard definitions of families of factorial contrasts, and non-central intervals such as Cohen's d and f. Analyses are not supported by usual statistical packages, but instead by a program called PSY. The latter is explained by Kevin Bird, Dusan Hadzi-Pavlovic, and Andrew Isaac in PSY: A Program for Contrast Analysis . A strong background in statistical inference is required. ^BSumming Up: Recommended. Upper-division undergraduates through faculty and established researchers. D. J. Gougeon University of Scranton