1
$\begingroup$

I'm really confused about which test to use to test the hyposthesis; "There is no difference in the preference of an exercise programme". Each participant (n=17) attends all three different exercise sessions (aqua aerobics, step aerobics, and circuit training). The 'enjoyment score' is given by each participant and is out of 100.

I know it's ordinal level data so need a non-parametric test. But I'm not sure if it's classed as independent samples, as comparing three different groups (Kruskal-Wallis test), or repeated measures (Friedman test) because the same person is involved in each exercise type?!

Please can someone clarify??

$\endgroup$
1
$\begingroup$

If I understand correctly, you have data in 3 columns (aqua aerobics, step aerobics, and circuit training) and each row represents data for one person.

In this case, you should use the Friedman's test, and afterwards paired Wilcoxon signed-rank tests (for post-hoc testing).

$\endgroup$
  • 1
    $\begingroup$ Thank you for your answer. I did think it was Friedman's but wanted to check. $\endgroup$ – caro Feb 26 '15 at 12:21
  • $\begingroup$ Note that the Wilcoxon signed-rank (WSR) test is somewhat inconsistent with the Friedman test. WSR is an efficient test (but only handles a single difference, not 3 columns) whereas Friedman's test is a generalization of the inefficient sign test so is not so powerful. $\endgroup$ – Frank Harrell Jan 16 at 13:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.