Solved – Chi-squared or Kruskal-Wallis

If I collect data for both a control and an experiment group, what tests would I use to detect a significant difference between groups in:

1) numbers with wound healed at week 4, week 8 and week 12

2) % reduction in wound size week 4, 8 and 12

3) change in score (0-100)

4) distance can walk in 6 mins (meters)

I have said

1) anova
2) kruskal-wallis
3 and 4) repeated measures two way ANOVA, are these correct? I wont know if data is parametric or not until I collect I assume?

ANY ADVICE would be much appreciated,
very stressed final year PT student writing research proposal

It is not perfectly clear what you need, but this is what I would recommend:

1) Fisher exact test (or better log-rank test using event yes/no and time);

2) repeated measure ANOVA (plug in also wound size at baseline);

3) repeated measure ANOVA (plug in also baseline score);

4) t test or Mann-Whitney U test (if you only have this at follow-up, otherwise repeated measure ANOVA.

Similar Posts:

Rate this post

Leave a Comment