I am working on a project regarding occupation-educational mismatching in the US. I am looking for the percentage share of STEM jobs/ industries by states/ counties in the US from 2011 to 2018 to create instrument variables for mismatching variables.
I am aware that in IPUMS USA, we have variables for occupations as well as industries. We also have places of work. Are there any ways to construct a percentage share of STEM jobs/ industries by states/ counties in the US from 2011 to 2018 based on those variables?
My approach is that I can count how many occupation codes are available in a particular place of work. Then I will count how may of those occupations codes are STEM. Is this a good approach? Are there any STATA commands for this solution?
Thank you so much for your help in advance.