Significant interactions in two-way ANOVA: example on impact on testing for normality
Jesper Rydén and Sven Erick Alm
U.U.D.M. Report 2008:40 ISSN 1101-3591
In Journal of Applied Statistics Vol. 37, No. 10, October 2010, 1695-1701.
A key issue in various applications of analysis of variance is testing for interaction and the interpretation of resulting analysis of variance tables. In this note is demonstrated that for a two-way ANOVA, incorporating interactions or not may have a dramatic influence when considering the usual statistical tests for normality of residuals. The effect of numerical rounding is also discussed.