I don’t see a 1970 5% sample. It appears from previous discussions that I can combine the state 1% form 1 and form 2 without having concern about overlap. (I just need to divide the weight by 2. Is this correct? Also, why isn’t there a 1970 5% sample?
You’re correct about combining the 1% samples and dividing the weights by 2. Just be aware that many of the variables are only available in one of the two samples. You can find lots of detail about the 1970 microdata sample design at this page. This sample was created by the Census Bureau back in the 1970’s, so I suspect the main reason that 1% samples were released was due to computing and storage limitations.
Stephanie Moller, PhD
Co-Chief Editor, Social Science Research
Professor, Department of Sociology