Hello,
I was wanting to know in general if one has many missing data for certain variables that are being used in analysis such as in association and not much missing data in variables used as descriptives if it is customary practice to apply the multiple imputation method to both sets of variables OR if it is okay to apply the multiple imputation method to the variables, in this case of analysis that have more missing data (higher than 10%), and the descriptive variables where data for every single variable is under 10% missing to apply the complete case analysis?
This is data from a large response driven sample
Thank you!
