There are one independent variable and two dependent variabls. Two intact groups, one is the experimental and the other is the control group. Pre-tets and post-tests are administered in both groups.
So, the data could be analysed by running a one-way MANCOVA, since there are two dependent variables.
It is also possible to run two separate Independent Samples t-tests, since the two depedent variables are not completely related.
The question is that which one is more accurate and should be used, and why?
Thank you
Hi Masoud. These two tests can be used to achieve functionally similar outcomes, but the underlying question will be whether you are most interested in finding out how closely (or not) the variations in each group match one another (MANCOVA), or whether you are most interested in finding out how distinct (or not) the two groups are to one another (t-tests). T-tests are generally quite quick and easy to run, so why not do both and see how they might compliment one another? This would be an informative exercises to help you pick up on the more nuanced differences between these two anyway. Hope this is helpful!
Hi Masoud
first I should mention that MANCOVA is used to control covariates. when you have two pretest and posttest groups, the scores on the pretest are considered as a covariate to control for preexising differences between the groups. If you use test-test, the possible differeces between groups in the prestest cannot be controled; therefore, this can contaminate the outcome of your posttest. Moreover, you should consider the assumaptions of MANCOVA such as normal distribution.
- Badges
- Science topic
- Similar topics
- Correction