Rider Income

The first value is calculated by simply grouping the fifteen exit-stations with likelihood of delayed travel p-values less than 0.01, then averaging the incomes associated with riders traveling to those stations. The second value is the same calculation made for the other 76 exit-stations with p-values greater than 0.01.

This of course does not hold much weight as a statistical test, but serves to say there may be more at play than income alone. While we will go on to find other factors important, income does in fact prove to be a significant predictor of one's likelihood to delay travel for off-peak fares, in the context of a formal regression.