deriving critical level of significance for Bonferroni correction in my data?

j482

New Member
#1
We have conducted an experiment where we test ability to detect weights using two different methods using a repeated measures design. Each participant is asked to detect which weight is heavier between two given weights with weight increased by few percentage, so 4 different weight comparisions.

After the weight detection tests, participant's overall cognitive load is measured by using a Task Load work index that measures individual factors - Mental Demand, TEmporal Demand, frustation experienced, Performance, physical Demand and Effort. A FriendMans Anova is performed to find a significant difference among those factors for the first method. A Bonferroni correction for post-hoc test using Wilcoxins test is required to be applied which to my understanding would require to derive critical level of significance which is = alpha/number of comparisions.

Could anyone explain what it really means here with "number of comparisions"? Does it mean all the pairwise comparision possible among the above mention factors which can be more than 15? Or does it mean 3 main comparisions required for hypothesis which would be : Weight detection using method 1, weight detection using method two and comparision between two methods for weight detection?