So a big link making rounds this morning is this one, with the accompanying headline and tweet “Women earned 76.5c for every $1 men did in US in 2012. Gender pay gap barely budged in a decade.”
The WSJ tweeted this out to their 3.3 million followers
Chelsea Clinton joined the chorus of those upset
What’s amazing is that if people were to go ahead and read the article, they’d find out that gender discrimination is not what is primarily responsible for this gap, and realistically, the article just states facts that most people know already.
Specifically, the pay gap largely persists in part because men are more likely to pursue college majors and advanced degrees in fields that lead to higher-paying careers.
“Women are getting graduate degrees, but not necessarily in fields that give the best salaries,” said Sarah Jane Glynn, associate director of women’s economic policy at the Center for American Progress, a liberal group in Washington.
So why all the hubbub? Because of the headline. If the WSJ were to use the headline “women getting graduate degrees in fields that pay less,” the number of clicks & retweets would suffer.
This issue reminds me of a famous conditional probability question from the admissions offices at UC-Berkeley, where in 1973, 44% of men but just 35% of women were admitted, the difference of which was extremely significant. Instead a suspected systematic bias against women, however, the results were pretty easy to justify- women were applying to more difficult departments!
I stole this table from a site at UPenn, but it makes it obvious that, conditioning on one applying to any number of departments (A through F), there was no gender-based discrimination, and in one department (A), it was actually easier for women to get in. However, because women applied more often to the more difficult-to-get-into departments than their counterparts, their overall application success rates were lower.
|Dept||Numb. Applicants||Admitted||Numb. Applicants||Admitted|
How does this example relate to the gender gap in salaries? If one were to condition on certain fields and/or positions of employment, the salary differences would be substantially smaller, perhaps negligible. Those conditional mean salaries are the true outcomes of interest, but aren’t shown in the WSJ article. (Editor’s note: Do not condition on being a professional basketball player!!)