0

This might be a super simple question but I cant seem to find an answer. The issue I am comparing 2 shoes at 3 speeds and their impact on a dependant variable. Therefore lets say shoe1/speed1 = A; shoe2/speed1 = B; shoe1/speed2 = C; shoe2/speed2 = D; shoe1/speed3 = E; shoe2/speed3 =F. Now i am comparing shoe vs shoe at the same speed. That would mean:

A vs B

C vs D


E vs F

Do i use the bonferroni correction here? Or would I only use it if I went like:

A vs B

A vs C


A vs D


and so on.
PlebHawkDown
  • 21
  • Adjustments for multiple comparisons are only necessary (helpful) in some situations and for some types of inferences. You may not need (or wish) to apply one for your study. Why not treat speed as a continuous variable rather than categorical? Plot a graph showing the two shoes speed vs dependent variable curves. You might find that the inference that you would like to make does not involve all of the comparisons. – Michael Lew Sep 08 '23 at 21:40
  • Id expect that looking at speed in a continuous way would be insufficient with only three data points available, or am i missing something in your answer? – PlebHawkDown Sep 08 '23 at 22:05
  • Please specify the hypothesis you are testing or the question that needs to be answered. For example, if the goal is to test if there's an effect of shoe and speed, I would use some ANOVA. If the goal is to compare a certain brand of shoes against competitors, multiple tests with adjustments like in your second option might work. – Slava Sep 09 '23 at 03:02
  • Basically i am testing if the shoes (that only differ in cushioning used) have an effect on ground reaction forces/loading rate produced. So I am looking at does the shoe itself affect the ground reaction force and is one shoe more impacted by speed than the other. I ran a 2 way repeated measures ANOVA and now want to run some pairwise testing/t tests to look at the factors that proved to be significant. – PlebHawkDown Sep 09 '23 at 10:25
  • ANOVA tests for any difference between group means, so a typical post-hoc test would test all group pairs. For 3 groups that would be 6 tests. Here's a relevant textbook with chapters 4 and 5 on contrasts and multiple testing http://users.stat.umn.edu/~gary/Book.html – Slava Sep 09 '23 at 16:57

0 Answers0