Caitlin Knowles Myers from the Middlebury College Economics Discussion Group reported last September on a study of about 300 transactions at Boston-area coffee shops concluding that female customers wait on average 20 seconds longer than males, even after controlling for all the obvious things like type/size of order (e.g. regular versus fancy soy-lowfat-decaf-whatever), cash/credit, age, appearance (well-dressed vs. sloppy) and more. I looked at the data, and it seems solid and statistically-significant.
The authors conclude the obvious: that women are subtly discriminated against, in spite of the fact that a free market should stamp out such nonsense, since women are free to bring their business elsewhere. Nobel prizewinner Gary Becker provides ample evidence that free markets do in fact make the world more fair, so this seems like counter-evidence.
After reading the paper more carefully, though, I see that the study didn't control for some other potentially-important variables like:
- whether the wait time was driven by the customer. If females happen to chat more with the barista, wouldn't that matter? What if an indecisive customer asks for the order to be redone?
- whether the customer was a regular or not. Maybe some of the people are there every morning, making the staff processing more efficient.
Although the study did note that the apparent discrimination goes away when there is more female staff is involved, the number of data points is too small to be statistically conclusive.
Finally, there is another qualitative problem with this analysis: who says longer wait times are necessarily "unfair"? Seems plausible, but maybe on closer inspection we'd find that longer wait time is a function of something else, like better service.
[see also Tim Hartford's analysis]