Should the man pay for the date?

8/22/2013 10:35:00 AM
Matt Yglesias pokes some fun at a study result showing that 84% of men reported paying for dates most of the time, compared to 58% of women reporting that men paid for their dates most of the time. Apparently, a small number of men are dating a large number of women and then refusing to pay.

Now, I don't have a problem with how couples choose to structure their finances between the two of them. But I do think that women should always offer to pay for their half. It works for Holland, apparently, and also the gay community.

Ok, one problem I have with siding with the "women should pay half" side is that it consists mostly of angry privileged men complaining that women have so much privilege in life. I don't think that: the occasional free dinner hardly offsets the exclusion of women from business, the pay gap, rape-culture, unattainable and unhealthy standards on body image, terrible stereotypes and more. It is plainly clear that men, not women, have it easier in life.

However, men paying for dates is a terrible way to level the playing field. First off, it doesn't do that at all--it only perpetuates the system in which men feel it is acceptable to oppress women, and it contributes to the rape culture by implying that, in exchange for buying dinner, men are entitled to sex. Second, and more importantly, it actually commits the same statistical fallacy that often oppresses minorities.*

Here's what I mean by that: statistically, it is true that men earn more money, on average, than women. But, while the difference in the average male and average female earnings is statistically significant, the difference between an individual man's and an individual women's earnings is not statistically significant. That is, just because men average higher incomes is no reason for a particular woman to assume a particular man will have an income higher than hers. She could be wealthier than the average female, or he could be poorer than the average man. This is a common statistical fallacy I see employed all the time, both inside academia and outside.

It was the same statistical fallacy that killed Trayvon Martin.

In a graph, here's what that fallacy might look like:

As you can see, for any given point on the x-axis, it is impossible to reliably predict whether series 1 will be higher or lower than series 2, even though, on average, we know series 2 is 25 percent higher.

In science journalism, I often hear reporters explain the 95 percent confidence interval by claiming that there's a 95 percent chance that the data will fall inside the interval. THIS IS TOTALLY WRONG. It will always be the case a lot more than 5 percent of the data falls outside of the confidence interval. To give an idea of how wrong this is, in the graph above only 3 of the 30 observations fall within the 95 percent confidence interval! The correct interpretation of a confidence interval is that we are 95 percent confident that the true parameter value (usually the mean) is within the interval. If you want a range for the actual values, rather than the mean of the values, you need to use instead a prediction interval, which will give you a much higher upper limit and much lower lower limit. Then you can say you are 95 percent confident that the data falls within the interval (sorry but "probability" is still too strong a term here--frequentists can only talk about "confidence").

*Ok, this post isn't really about dating. You may have noticed that many of my posts have an ulterior, pedagogical motive. This one is really about the statistical fallacy of using averages for individual predictions.
Lord 8/22/2013 02:33:00 PM
More likely, women in relationships pay more but men don't consider that dating.
Matthew Martin 8/22/2013 02:36:00 PM
Interesting. I hadn't considered that.