If we are performing bonferroni corrected pairwise t-tests, does it matter whether they are planned (a priori) or not (post hoc)? And can you perform the t-tests without performing an anova test first, if the goal is to determine whether there is a difference between many groups, and if there is them between which ones?
If I have an significant result in anova for variable x. But a pairwise test is telling me, that group x doesnt differ from the controll group (group without effect from variable x), what do I do? Do i say, that there is no effect for variable x on y? And what if i have no significant anova result for variable k, but i have a signifficant difference between two groups manipulated by variable k?
Thank you
If we are performing bonferroni corrected pairwise t-tests, does it matter whether they are planned (a priori) or not (post hoc)? And can you perform the t-tests without performing an anova test first, if the goal is to determine whether there is a difference between many groups, and if there is them between which ones?
If I have an significant result in anova for variable x. But a pairwise test is telling me, that group x doesnt differ from the controll group (group without effect from variable x), what do I do? Do i say, that there is no effect for variable x on y?
And what if i have no significant anova result for variable k, but i have a signifficant difference between two groups manipulated by variable k?
Where did you get the value of q? it is constant ? or what basis to get the value of q? Thanks
A statistical table - I typically give the value to my students, but it is the "Studentized Range Statistic (q)"