I have run a pre-post-test intervention design.
I have two groups: experimental and control
My dependent variable was measured twice, pre and post intervention.
I want to find out if the intervention was more effective (produced a greater change in scores from pre to post test) for first year students compared to second year students.
What analysis do I need to run? I have established the intervention is effective in general (significant interaction was revealed between group and time) but I also want to find out if year of study affects the interventions success.
I would simply say to add the year as a covariate but make sure you have a justification as to why the year matters.
It's not quite that simple because year of school can affect either the pre-test score and post-test scores proportionally and/or the change in scores pre-post. I am assuming you are keeping each individual's pre & post responses paired, eg by running a panel regression. In that case, include you've already included a dummy for pre versus post and will need to add a second dummy for first vs second year student. You will need to interact the dummies as well as entering them separately. Use a random effects regression.
Hi Mi Louis
I would suggest using ANCOVA. You can re-run the data, using pre-intervention scores as a covariate, and post-intervention scores as the DV. This gives a more straightforward comparison of groups, adjusted for any pre-existing differences.
Good luck!
- Badges
- Science topic