We are conducting a race, gender, and emplyoment project. We are using data from 1980-current. I am wondering what weights to use and how to apply them if we are just using tabulations for things such as race*gender for certain occupations.
Since you are interested in person-level analyses you would want to use the weight variable PERWT. While 1980 is a “flat” sample, it is still safe to use PERWT because all of the values in the 1980 sample are the same. Weights are simply multipliers, meaning that instead of counting Person A once you would count them 4.65 times because that is how many people they represent in the total population. Major statistical software packages have built in functions for applying weights to frequencies and crosstabulations so you will just need to learn the method for the program you are using.
I hope this helps.