This is useful as long as the sampling strategy doesn't change over time. With omnibus questions it can be a bit variable if the studies have some other motivat

This is useful as long as the sampling strategy doesn't change over time. With omnibus questions it can be a bit variable if the studies have some other motivation and the demographics are not consistent over time. edit to add: if you looked at 40 years of this question, you would definitely want to control for the higher number of people with direct exposure to the higher ed system, since that base rate has been on a positive trend for a long time.