The difference between what men and women earn in different fields can lead to some skewed results. Schmidt just wrote a blog post more fully explaining this; here's an example of the kind of error he's afraid will be common:
Every newspaper chart or college loan program that doesn't disaggregate by gender is going to make the majors that women choose look worse than the ones that men choose. Think we need more people to major in computer science, engineering, and economics? Think we need fewer sociology, English, and Liberal Arts majors? That's not just saying that high-paying fields are better: it's also saying that the sort of fields women major in more often are less worthwhile.
How important is gender? Very. A male English major probably makes more than a female math major, and a female economics major makes less than a male history major.