Abstract
ROMANIUK, JEAN G.; LEVIN, JOEL R.; and HUBERT, LAWRENCE J. Hypothesis-Testing Procedures in Repeated-Measures Designs: On the Road Map Not Taken. CHILD DEVELOPMENT, 1977, 48, 1757-1760. 3 common hypothesis-testing procedures often employed in the analysis of repeated-measures data are the regular univariate F test, the Geisser-Greenhouse modification, and the multivariate Hotelling's T2. A popular road map for choosing among these alternatives implicitly assumes a hierarchy in the rejection probabilities associated with a given data set. This note demonstrates that exceptions to this hierarchy (regular univariate > multivariate > Geisser-Greenhouse) exist, and that the nature of such exceptions depends upon the pattern of mean differences in combination with the structure of the variance-covariance matrix. Until
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have