How to select a post hoc test

In this video, we will discuss how to select an appropriate post hoc test (Fisher's LSD, Bonferroni, Holm, Tukey, Dunnett). We will also compare these tests based on simulation studies.
1. Compare the tests based on simulation studies
2. How to select an appropriate test (08:55)

[@elenapopova4569]

Thank you very much about this video! I would like to add: Games-Howell test after Welch's ANOVA, when variances are not equal in the groups. And also after Kruskal-Wallis the most common Post-Hoc tests are Dunn's or Nemenyi. Best!

[@tilestats]

Thanks for sharing your knowledge!

[@Tepico]

Great thank you. Would you please make also a video on Duncans MRT post-hoc test?

[@alirasteqar243]

Man, you are the best one!!!!!!!!!!!!!!!
Really thanks

[@tilestats]

Thank you!

[@hajarzakzouk7458]

Very good explanation.. I have a question if you please, how can I get the subsets if I have an independent variable containing only two groups ?

[@tilestats]

If you only have two groups there is no need for a post hoc test. You can then simply use a t-test.

[@katibsareeh9056]

Thanks for this great explanation. Please, I performed Fisher's exact test on a 3*2 table which represents 141 patients and I got a significant difference (P value 0.010). What post hoc test should I use to determine which cell or pair of cells is responsible for the significance?

[@tilestats]

You can use Fisher's exact test again to do pairwise comparison. If you do many comparisons, you can adjust the p-values by Bonferroni or Holm.

[@katibsareeh9056]

Does this mean to convert the 3*2 table into two 2*2 tables or what? could you please explain more? and if I again found a significance in any 2*2 table, then how to determine the source of significance?

[@abidhussainwani3028]

Thank you!