Generate a dummy to differentiate (non)missing observations

[Guest]

Hi everyone,

I am working with longitudinal data, namely observations from 1958 (independent vars) and 2008 (dependent var). I want to check wether there are significant differences in the 1958 independent vars between the missing/nonmissing values of the 2008 dependet var. Just to make sure that my final sample is not biased. I'd ideally like to generate a dummy identifying them (missing/nonmissing in 2008) and then make a ttest such as ttest "indepvar" by "dummy"

Thank you in advance,

Francesc

[Nick Cox]

That is just

gen missing2008 = missing(response)

for your response or dependent variable.