There are equal sample sizes but violations of homogeneity of variances and normality. Which post Hoc method should I use in SPSS. After of course getting the Welch t-test or ANOVA first.
Is Dunn's Post Hoc test fine or are there some things I have to consider. For example, I found few sources that suggest: Between Dunnett T3, Games and Howell, and Tamhane T2. Dunnett T3 is implored when sample sizes are small (
Generally, the more complicated tests (raw formulas) are more accurate, and the simpler formulas are based on simplifying presumptions. In addition, the accuracy of the tests depends on the available data. I suggest the section of post-hoc tests by the Montgommery's Design and analysis of experiments. Apart from the t-test and Bonferroni, the tests are introduced by the order of complexity (the first the most complicated). See the following very same question:
https://www.researchgate.net/post/Which_post_hoc_test_is_best
In addition to the above don't forget that all of your tests depend on the regression assumptions. I would fix those first and then follow Babak:s suggestions. Best D. Booth BTW a good way to fix them is with the Box-Cox package in R
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS