Sara Vargas posted on Wednesday, December 07, 2011 - 10:20 pm
I have two samples of medical patients that received an intervention and are nearly identical in terms of study design, assessments, etc. except for two main differences:
1) the samples receive the intervention at different points during medical treatment. I am interested in exploring this difference as a moderator of growth trajectories in the intervention versus control groups using a two-group design where sample 1 is one group and sample 2 is the other group, but...
2) both samples have 4 time points but the second 2 time points do not match (1 sample is 0, 3, 6 and 12 while the other is 0, 3, 9, and 15). I do not expect the trajectories of change in outcomes to be linear.
Is there any way to analyze such data so that I can actually interpret the moderating effect of sample given that they ALSO differ in terms of follow-up time point and the trajectories are not expected to be linear?