I am carrying out a multiple range statistical testing in SAS using Duncan, turkey, bonferroni and Gabriel method. While Turkey, Bonferroni and Gabriel gave exactly the same results, Duncan's result was different. I wish to know on which circumstance to use any of these methods and peraventure if any of them is the best and why. Other methods apart from Duncan seems to pick more difference seemingly hidden. Will this be a good basis for their choice over Duncan?
Thank you.
Hello,
Fisrtly I'm assuming you're talking about the Tukey test and not "Turkey"... Nevermind. ;-)
A short time ago, I ask a question about best test for post hoc analysis. It was for parametric test, but anyway, Jochen Wilhelm gave a exellent answer which, I hope, mays help you to answer your question.
https://www.researchgate.net/post/The_choice_of_post-hoc_test
Regards.
Thank you once again Dr Cyril. Your link will be very helpful. But was surprised you were like me when you first asked the question- you wrote "Tuckey" in that link? Lol!
You're write ! ^^ But anyway... I got the good pronunciation with me. ^^
You may be interested by the answer in that link also :
https://www.researchgate.net/post/Which_post_hoc_test_is_best
Regards.
- Badges
- Science topic