My design is a pre-post control group design. Due to the small samples size (16+16) and the absence of normality is the pre-test scores, I performed a Wilcoxon Signed Ranks Test to test the differences between pre and post test scores for both groups. However, I found out that post-test data is normally distributed for both groups. My questions is:
In this case, in order to test the difference between groups, should I perform Mann–Whitney U Test or Independent- T test?
Advise please!
Thanks in advance.
If I understood correctly, you have two groups, and for each group you have pre and post data. In this case, the distribution of pre- or post-data is irrelevant. Only the distribution of post minus pre is relevant, and here you should also separate the two groups.
PS: I don't know how you came to the conclusion that some scores are normal distributed and others are not. Often, normal distribution tests are used in this context - what is rather stupid. Such test will sometimes allow you to reject the tested hypothesis (the sample is from a normal distributed RV), sometimers it won't. If it does, you don't know if the difference to the normal distribution is of any relevance, and if if does not, you don't know if there are relevant differences but the sample size is too small to recognize this. It's better to make an informed judgement which hypotheses make sense and which are clearly nonsensical.
Hello Mahfoudh,
What is the hypothesis(es) you wish to test? Not knowing, I can only offer generic suggestions.
On the face of it, two common models for data of this sort (two groups; measured twice) are: (a) a one-between (group: control-experimental), one-within (occasion: pre-post) anova; or (b) one-way analysis of covariance, with the pre-treatment scores used as the covariate and post-treatment. Either has an analogous regression model that may be used as well. Model (b) is a bit simpler to explain, since often the effect of interest in model (a) is an interaction term (group X occasion).
However, do note that assumptions of normality apply to the score residuals ("errors") from the model, not to the distribution of raw scores on the DV.
If you are worried about normality (whether you check it or not), you may always use a bootstrap/resampling method to estimate model parameters without having to make distributional assumptions.
Good luck with your work.
t- test alone is quite enough unless otherwise needed for more confirmation.Regards.
Medhat Elsahookie David Morse Jochen Wilhelm This is a bit confusing to me, but overall thank you very much for your suggestions
- Badges
- Science topic