I'm hoping to run a latent profile analysis in order to examine if there are groups of participants who recover differently after having a vaccination, using both behavioural and cardiac data.
I have simultaneously collected video and cardiac data one minute prior and three minutes following a vaccination. I am hoping to look across 4 different 30-s time periods (baseline, immediately post-needle, 1-minute post needle and 2-minutes post needle) for both behaviour and cardiac data.
My questions are as follows:
1) Can I look at two repeated measures (i.e., behaviour/cardiac) simultaneously using LPA?
2) Should I enter the 8 variables into the model (cardiac and behavioural data across the vaccination period), or is there a way to examine the intercept/rate of change across the four time periods for each variable?
1) Yes. You can use either two latent class variables (one for each process using Model c1:, Model c2: ) and let them correlate, or use one for both.
2) For 2-process LPA, having the 8 variables together is good. With rate of change I think you refer to growth modeling which is certainly also possible and appropriate if you think there is a simple growth form (linear, quadratic etc).