jamovi

Hey, everybody.

I'm looking for some help with my analysis. I come from a humanities background, so my knowledge of statistics is limited.

My study involves two groups:

Intervention group: Completed questionnaires at baseline, post-intervention (1 month), 6-month follow-up, and 12-month follow-up.
Waitlist group: Completed questionnaires at baseline and after a 1-month period.

I have 7 dependent variables (DVs) and want to:

Test the efficacy of the intervention compared to the waitlist group.
Assess whether the results are sustained over the follow-up periods.

I initially planned to use repeated measures ANOVA. However, I've read that it doesn't handle missing data well - and I have missing values in my dataset. After some research, I'm thinking the following approach might be better:

- Use Hierarchical Linear Modelling (Mixed Models in Jamovi) for the intervention and waitlist groups, with pre- and post-measures, for each dependent variable.
- Conduct another Hierarchical Linear Modelling (Mixed Models in Jamovi) for the intervention group alone, with pre-, post-, 6-month, and 12-month measures, for each dependent variable.

Does this seem like a sound approach? Do you have any other suggestions?

Additionally, will mixed models handle missing data automatically? I'd prefer to avoid imputation since I’ve read that it can introduce bias.

Thank you so much for your time and help!

Hi. In my opinion, an elegant analysis such as ANOVA or Linear Mixed Effects analysis would be good and appropriate only if you had much less missing data, and if the design had included 6- and 12-month for the wait-list group. As things stand, I think an appropriate set of analyses would be, simply, a set of three, two-group t tests (with "Group" as the grouping factor), calculated on change-from-baseline scores as illustrated below.
.

Thank you so much for taking the time to respond, I will consider these analyses!