For clarification, I am doing research on asynchronous mathematics video collaboration between students of different countries.

In order to measure this, I designed a 20-some question survey using a discrete Likert scale (1-5, I strongly disagree to I strongly agree with the statement).

I administered this survey once to nearly 100 students via Google Forms and received some seemingly valuable data, including that of a control group and experimental group.

I plan on administering this questionnaire once more in the middle of the treatments and after the collaboration is complete to see if there has been a statistically significant increase in the values.

I was thinking of doing a repeated measures ANOVA, but perhaps another method is better? I could always check sphericity with the control group for a high value? But for the experimental group, I was wondering a better way to compare means than run the test with every single question. Having 20-some regressions in my paper might look overwhelming as well. The questions come in clusters.

Are there any insights on how to go about this in a better way? I included the type of questions included.

Thank you so much!