I need some advice about how to identify which variables are associated with change between time 1 and time 2. The data-set contains:
1) repeated measures at time 1 and time 2. These are two continuous (scale) measures, e.g. a psychometric measure of self-esteem, and some ordinal measures (Likert-scale type questions).
2) demographic (categorical) data
3) programme process data (also categorical), e.g. which stages of the intervention a participant reached, e.g. did the participant engage in preparatory activities?
I need to analyse whether any variables (demographic and/or process) are associated with, or predict, changes in the repeated measures. My first thought was a multivariate regression but I’d be very grateful for some pointers about how to go about this!
You could try a Hotelling’s T-squared test for dependent groups. This is a multivariate test that compares time 1 and time 2 (same participants at both times) based on two continuous measures. If time 1 and time 2 differ based on those 2 variables, then you go to the univariate analysis to see which continuous measures indeed differ between the 2 time periods. You can also include the categorical variable as a between subjects factor. If you have access to SPSS this would be done the following way.
>general linear model
enter within subjects factor name like “time” with number of levels = 2
enter measure names for your 2 continuous measures like “outcome1” and “outcome2”
select and move appropriate continuous measure to the right pane. There should be 4 measures to move over (i.e., outcome1 for time 1, outcome1 for time 2, outcome2 for time 1, and outcome2 for time 2).
enter your categorical variable into the between subjects factor box.