The chart below clearly shows that those who study male-dominated majors generally earn more after college than those who study majors that are dominated by women. Each of the squares below represents a college major; the bigger the square, the greater the number of recent graduates. The color of the square indicates the category of study — science, humanities, etc.
Median yearly earnings are shown on the vertical axis: Majors that appear toward the top of the chart tend to earn more, and those toward the bottom earn less. The gender makeup of the major is on the horizontal axis, with majors that are male dominated on the left and female-dominated majors on the right.
See how the points slope downward to the right? That shows that female-dominated majors, like elementary education and psychology, tend to be lower earning, while male-dominated majors, especially engineering, tend to earn much more. "Sure enough, there's a clear and significant negative correlation between a college major's median year earnings and gender ratio," Olson writes.
There are a few interesting outliers. Nursing is dominated by women (it's 90 percent female) but tends to be relatively highly paid, at least in the initial years after graduating, with $48,000 in median earnings. The interesting outlier for the men is transportation science, which is heavily male but not very highly earning.
Olson speculates on what is driving this trend, but the reasons aren't precisely clear. Through analyzing the data, he rules out the explanation that female-dominated majors are earning less because they didn't find a job after college. He finds there isn't a significant correlation between the gender ratio of a subject and how often its grads get jobs, as the chart below shows.
One theory that Olson entertains is that male-dominated majors tend to be focused on quantitative skills, which are in general highly valued. Engineering is both male dominated and highly valued and paid by businesses.
In the chart below, Olson shows that the median yearly earnings of a college major tend to go up with the college kid's quantitative SAT score, suggesting that math skills are indeed important for finding a well-paid job.
As some of Olson's commentators point out, there are some other potential explanations. It could be that society just doesn't value "women's work" as much as "men's work," due to lingering prejudice and discrimination. The difference could be in part due to different salary negotiation strategies used by men and women. Or it could be because men, who traditionally have been familial bread-winners, are more motivated to seek out jobs with higher incomes, while women prioritize other job attributes.
Regardless of the reason, if it's salary that you care about, engineering looks like a solid bet.
Note: This post has been updated with a newer version of the first chart, which Olson updated after publication to remove some outliers in the data.