2017-10-06

Mike Kutner, PhD, Department of Biostatistics and Bioinformatics, Emory University

"Statistical Tests for Interactions and Main Effects in Two-Factor Unbalanced Cross-Classified Fixed Effects Experiments with No Data in Some Cells"

I will discuss the statistical tests of both the interaction effects and main effects in unbalanced two-factor cross-classified fixed effects experiments with no data in some cells. The testing has been problematic when using standard statistical software packages default options such as SAS, SPSS and BMDP. This lecture will lay out the fundamental reasons why these tests are not routine and will provide you with how one can generate tests for interactions and main effects that unfortunately are dependent upon which cells have no data.  Today, very few textbooks and/or classes in analysis of variance handle this topic particularly well. Furthermore, the default options provided by SAS, SPSS and BMDP are still problematic.

Stay connected TwitterFacebookLinkedInYouTubeInstagram