I wonder if coffee consumption is correlated with white collar work.
I could see some compounding variables here. With availability and ability to drink it.
That's because the working class is more likely to be forbidden from having coffee at their job location. Folks with desk jobs just drink it at their desks.