Closed dominic-comerford closed 5 months ago
1) When amputing ‘time’ in the make_missing function – by adding extra patterns (to bring the total to 7 patterns) - it does not seem to ampute time at all based on the resulting list.
2) I then add missingness in time using the create_miss_time function and subsequently interpolate the imputations for time in a next step. I use MICE to multiply impute the sender and receiver columns. Resulting in a list with MIDS objects, where time is of course not multiply imputed in (as this was done with interpolation).
3) When running the model for the results I get the following error. It seems that the risk set with 931680 rows, which is (the number of timepoints x N x N-1) does not match the 799920 rows in the ‘replacement’ dataframe. I don’t understand this error, and tried excluding patterns (mentioned at 1) that have time missing, as time gets excluded later anyway. But still the problem persists, albeit it with a different mismatch than it has now.
Created on 2024-05-20 with reprex v2.1.0