I've read several posts related to this, but I'm not sure. so I'm posting the question in case it's redundant.
I have been conducting an experiment using a within-subject design, comparing three conditions. The collected data is continuous and ranges from 0 to 240 (approximately). (I used a simulator sickness questionnaire; SSQ)
After performing a normality test, I found that the data does not follow a normal distribution. Consequently, I decided to use the Friedman test for statistical analysis.
However, I am now faced with a question regarding the appropriate method for conducting pairwise comparisons. In the past, I have used the Wilcoxon signed-rank test and applied Bonferroni correction.
Recently, I came across the Durbin-Conover test in Jamovi software, which compares ranks instead of scores/means(?) like the Wilcoxon signed-rank test.
I am considering using the Durbin-Conover test for post hoc analysis after the Friedman test. Yet, I wonder why we use the Wilcoxon signed-rank test. Am I misunderstanding something?
Finally, in my experiment, can I use the Durbin-Conover method for pair-wise comparison?
