Hi everyone,
I have a group of 10 kids which have been evaluated pre- and post- intervention on various continous scales.
Now, I would like to see if there any signficant differences between pre and post intervetion scores for each subject, as the number of participants is too small to do anything else I think.
I am not quite sure of which statistical test I should run; I was under the impression that a T-test would show me if there are any significant differences between pre and post on each scale for each subject, but I'm not entierly sure if that is a correct procedure.
Thank you for any help with this!
I'm assuming you can pair the data before and after by student? That is, you know for Student A this was the before score, and this was the after score?
In that case, you want to use a paired test like paired t-test or two-sample paired rank-sum test.
Be sure to read all the assumptions of these tests before you use them.
If you use something like the rank-sum test, be sure you understand the interpretation. People sometimes say it is a test of medians, but that isn't necessarily true.
You can read what I wrote about these tests:
Paired t-test:
http://rcompanion.org/handbook/I_04.html
Paired rank sum test:
http://rcompanion.org/handbook/F_06.html
Thank you for your reply and the links, both have helped me greatly and solved my doubts.
as your data is very small, so you can not apply the parametric tests i.e. t test. In this case you can apply non parametric test like Mann-Whitney U test
- Badges
- Science topic