I am analyzing data from a sport psychology experiment I conducted on Motor Imagery. My two treatments are movement and perspective, which led to a 2 x 2 design. My goal was to see if, over 4 trials, dart-throwing scores would be different across groups. I opted for a MANCOVA with the four trials as within factor, movement and perspective as between factors, and an initial trial (Assessment Trial) as a covariate, to take into account their ability.
My supervisor suggested to also look at nonparametric ways of looking at my data. However, all the methods I have found only involve differences between two repeated measures (sign test, or Wilcoxon Signed-Ranks test).
Do you have any suggestion as of how I could rethink of my data to attempt nonparametric analyses?
If you use the assessment trials as a baseline you can convert your four trials into a measure of a change from baseline by simply taking the difference from baseline to score as your measure. You could also attempt to match your between subjects groups based on baseline performance.
Dear Matteo,
according to me you can try to decompose your data analysis in two steps. First, you can use a Wilcoxon Signed-Ranks test and then a Spearman correlation between the assessment trial and the others 4 trials with a Bonferroni correction.
Paul, Nicoletta,
thank you and grazie :)
I will explore both possibilities
You could check if your data have a ranking. In terms of levels of measurement, non-parametric methods result in ordinal data.
You should analyze your data with Kruskal-Wallis test.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS