I've done quite some reading on multiple imputation (MI), but can not seem to figure out the next question:
I have a dataset with missing values, some rows have many missing values, others have less missing values. They are missing at random.
I want to exclude certain rows because there are missing too many values (columns) , though I'm not sure if I should first create a dataset only with the rows I want to analyze in the end, or use the complete dataset to fill in the missing values (with MI) and delete the rows afterwards.
Does the second method help me create thrust worthier imputations?