Dear Statalisters,

Thank-you for reading my query.

I am working with a very big dataset (52 variables, 82284 observations) for longitudinal analysis. The dataset is based on information returned from 3 different surveys. I have converted the dataset to long format so currently there are about 3 different observations (in years) for each ID. There are approximately 11,000 unique ID variables.

I am using mi estimate in Stata/SE 16.1 Stata issued the following error message: "an error occurred when mi estimate executed melogit on m=1 r(2000)" when I attempted to run the following code: mi estimate , cmdok: melogit Obesity i.wave_sg i.BMIgp i.CountryBirth i.Marital HYPERTENSION DM i.Alcohol_quintile || idalias:, or

In this dataset,
  • wave_sg indicates wave/year of data,
  • BMIgp is the 4 BMI groups
  • DM indicates Diabetes Mellitus (Type 1 or 2)
I should note that this analysis has run successfully for the following codes:
  • mi estimate , cmdok: melogit Obesity i.wave_sg i.Alcohol_quintile || idalias:, or
  • mi estimate , cmdok: melogit HeartDisease i.wave_sg i.BMIgp i.CountryBirth i.Marital HYPERTENSION DM i.Alcohol_quintile || idalias:, or
Can anyone assist me? the Stata search results indicate that there are no observations but I struggle to understand how this is feasible when the code works for Obesity in crude analysis and other endpoints in multivariate.

Warm regards,
Sarah