I know, in my case, there isn't a direct 1:1 correlation between hourly rate and hour of commute. The calculation for determining that calculation (at least in the context of the study) would be interesting.