I am currently writing my dissertation on whether or not the adjusted mean for the treatment group’s increased outcome test scores for a post-test is significantly higher than the adjusted mean for the control group’s increased outcome test scores.
I will also run a test using the pre-test and post-test scores of the top ten students that spent the most time using the treatment to show how much their outcome scores increased.
There were 70 students that participated in the study and they all took the pre-test, with 35 in the treatment group and 35 in the control group. Six students (2 from the control group and 4 from the treatment group) did not take the post-test.
My question is should I omit the 6 students’ pre-test scores that did not take the post-test before running an analysis or should I use a simulated test that will fill in the 6 blank post-test scores with an estimated value? Should I run it both ways? I found research that stated it was better to run a simulated test when more than 5% of the data is missing. I will be using the ANCOVA test.
Do you think this is a good test to use or would you suggest a different test? I would appreciate any help you can give me.