Hello, I am trying to run a Tukey analysis on four groups of data (A, B, C, D). I get significance or non-significance for all the comparisons, with the exception of comparing group A with C and comparing group B with C. For these comparisons it always says not tested. I can get comparisons between all the groups using Bonferroni or Holm methods. Am I overlooking something silly? Thanks.
Hi,
this behavior is by design. The Tukey test orders the comparison pairs and looks at them one by one. Once one comparison is not significant, subsequent comparisons are skipped: See the http://en.wikipedia.org/wiki/Tukey's_range_test#Order_of_comparisons in the Wikipedia article. If you are curious about the implementation in code, http://sourceforge.net/p/xltoolbox/code/ci/master/tree/cls1WayANOVA.cls#l1011 (Visual Basic code).
The Tukey test is considered more ‘conservative’ than the others, i.e., it may very well be that the other tests report significant differences where the Tukey doesn’t.
Kind regards
Daniel
Ok, great thanks for the explanation.
News! I’ve released the very first alpha version of the XL Toolbox NG (‘next generation’). It has almost no features so far, but a lot of work was done under the hood