The idea is to see if parent occupations change after a lottery win.
One concern is that using monthly data, parent occupations might be super noisy. Maybe we want to do parent occupations in the year prior to a win compared to the year after?
Another concern is whether we want to look at fractions of parents in each category, versus ranking each category and looking at the overall ranking. I'm going to do the ranking one first since I think that is less susceptible to the power concern.
christinakent
commented
4 years ago 
The idea is to see if parent occupations change after a lottery win.
One concern is that using monthly data, parent occupations might be super noisy. Maybe we want to do parent occupations in the year prior to a win compared to the year after?
Another concern is whether we want to look at fractions of parents in each category, versus ranking each category and looking at the overall ranking. I'm going to do the ranking one first since I think that is less susceptible to the power concern.