A type I error investigation of modified Scheffe-based multiple-comparison procedures in factorial ANOVA, MANOVA, and multiple-regression situations

摘要

The present study extends the previous one-way ANOVA multiple-comparison findings of Meyers and Beretvas (2003) to interaction comparisons in factorial ANOVA designs, one-factor MANOVA comparisons, and tests of partial regression coefficients in multiple regression, in specific regard to the practical utility of modified (sequential) Scheffe-based procedures. Researchers who are concerned with maintaining familywise Type I error rates while increasing statistical power are encouraged to consider these improved multiple-comparison methods.