  Chapter 14: Misconceptions When people read, hear, or prepare research summaries, they sometimes have misconceptions about what is or isn't "sound practice" regarding the collection, analysis, and interpretation of data. Here are some of these common (and dangerous) misconceptions associated with the content of Chapter 14. In a repeated measures ANOVA, it's only fair to administer the treatments in the same order to everyone. The F-test for the source called "Subjects" (or what might be called "Between Subjects") provides as much useful information as any of the other F-tests. In a two-way, three-way, or higher-order repeated measures ANOVA, just one of the MS values serves as the denominator for all F-tests of main effects and interactions. Carry-over effects are eliminated if the treatments are administered in all possible orders to the study's subjects. Because each cell mean is based on the same amount of data, a repeated measures ANOVA is robust to its underlying assumptions. The number of subjects involved in a repeated measures ANOVA can be determined by adding 1 to the df value on the "Total" row of the summary table. A repeated measures ANOVA has the same underlying assumptions as do ANOVAs without repeated measures. A 2x4x6 treatments-by-treatments-by-subjects ANOVA will generate 7 F-values, just as is the case in any "regular" three-way ANOVA. F-values that turn out significant using the Geisser-Greenhouse conservative F-test procedure should be viewed with skepticism. A split-plot factorial can only be used in the field of agriculture. A Lindquist Type I ANOVA is to be avoided because it's highly likely to generate Type I errors. If the groups that form the levels of the between-subjects factor have the same n, the researcher is not obligated to check on the sphericity assumption. A three-way mixed ANOVA always has 2 between-subjects factors and 1 within-subjects factor. The total number of subjects in a mixed ANOVA can be determined by adding 1 to the df value on the "Total" row of the summary table. All of the F-values of a mixed ANOVA have the same amount of statistical power. A split plot factorial 2.22 has 2 factors, with 2 levels in the first factor and 22 levels in the second factor. A two-way mixed ANOVA is like a regular two-way ANOVA in that (1) three F-values will appear in the ANOVA summary table and (2) the same MS serves as the denominator when computing the three F-values. In a mixed ANOVA, tests of simple main effects can be used to compare the different levels of the between-subjects factor(s) but cannot be used to compare the different levels of the within-subjects factor(s).