Missing data based on other characteristics

[Anne Tremblay]

Dear Stata experts,


I would like to describe the missing data in our sample. For example, I want to know if the percentage of missing data varies based on gender or age. I used the code below. I would like to know if there is a better way to do this. Is test that I should conduct? If I see that boys have 3% more missing data than girls, how can I determine whether that is a lot? Is there a rule of thumb? Do you often present this information in manuscripts?

by gender, sort: mdesc variable1 variable 2 variable 3

Thank you
Anne

[Carlo Lazzaro]

Anne:
welcome to this forum.
Please follow the FAQ to post more effectively. Thanks.
That said:
1) the main issues with missing data are the diagnosis of their mechanism and patterns and the way you deal with them in the light of the diagnostic step;
2) during the last 10 years editors and reviewers have become more demanding about the way missing values are described and dealt with in a paper. Take a look at the literature in your research field.