Dear Stata users,

I have a query about longitudinal analysis.

I originally ran growth models with data collection wave as the time indicator (coded as 0, 1, 2 for all participants). This meant participants with missing data at some waves were included in the analysis. Now, I would like to use actual years/months since baseline as the time indicator (as there is some individual variation in the time between data collection waves; e.g. 0, 5.79, 10.9 vs. 0, 5.90, 11.20). This new time variable obviously has missing values for participants who didn’t complete data collection waves, and therefore these participants are now not included in the analysis.

Apologies if this seems obvious, but how can I include these participants when using the more precise time indicator? Do I use imputation to replace the missing time data, or perhaps there is a Stata setting I need to apply before running the growth models?

Any thoughts much appreciated.

PS I have contacted the administration team to change my name (to Stephanie Schrempft)