Hi Dear Colleagues
I am doing a longitudinal study. As only a small sample was able to be included as the baseline cohort. I added more people later from the second timepoint. also some participants came in the latter two timepoints of the three-time surveys.
Any suggestion on how to include the participants who joined later especially the baseline cohort is small
Thanks
You could treat this as a dynamic population cohort study.
If the recruits at the later time points had been earlier "exposed" you could incorporate their person years retrospectively.
thanks James for the reply.
Normally the cohort study starts with a big group of participants before attrition happens. When I did the first data collection, a big part of the potential participants were not approachable.
Now I wonder how to treat the participants who filled the latter surveys with the first one hav being missed.
By searching for similar design I cannot find any relevant guidance.
Shall i I drop these people who joined the latter two timepoints or there is method to simulate their first timepoint as missing data which is reversed.
I will check on the dynamic cohort study.
Thanks
Kang
This is thinking aloud, as I have no real experience of this. This is a particular type of missingness and I would be informed by that literature. What would really be problematic is if the new chosen sample is due to the response values that is they are NMAR.
Kang,
Did you find anything that worked using a dynamic cohort study? What did you end up doing? Can you drop the citation(s) or methods here?
- Badges
- Science topic
- Similar topics
- Psychology