04-15-2014 11:05 AM
Dear SAS Statistical procedure community
I have a question. I do know that Bonferroni is a very conservative method of means comparisons. I intend to submit my research manuscript to a journal, who have indicated that use of Tukey's as a multiple means comparison procedure is prohibited.
I conducted an experiment with ten levels of treatments and three response variables in a completely randomized design. However, there are suggestions from people who have pre-reviewed my draft manuscript that i should change analysis to Bonferroni adjustment. Some of the data i have are individual counts which i have done log transformation. I used Proc GLM for the analysis. However after the Bonferroni adjustments, i cannot find significant differences among my variables (e.g a mean of 20.7 is not different from that of zero mean (0).
I need to find a better way of doing multiple comparisons among my ten levels (varieties) replicated five times. Which is the best adjustment method that is not very conservative and suitable for this kind of analysis?
Your support will be highly appreciated.
04-15-2014 02:30 PM
Some would argue against performing all possible pairwise comparisons, because too many of these are not of direct interest. However, I know the value of these multiple comparisons in many disciplines. The Bonferroni will be way too conservative with 10 treatments. I think simulation-based adjustments are best. Use
lsmeans treatment / pdiff adj=simulate;