Study findings released on September 1st claimed that women’s median earnings are higher than men’s in 147 out of 150 American communities. The claim comes with a few caveats, including the fact that the difference is all of 8 percent, and only holds for unmarried, childless women under 30.

By Thursday, a report from IWPR on occupational segregation and the gender wage gap reached the opposite conclusion: women’s labor market gains have stalled in the last 15 years or so and, since 2002, have reversed for young women. What is going on here?

It is not the data. Both studies compared women’s median earnings using data produced by the U.S. Census Bureau.

It is partly education and skills. The report claiming that women are out-earning men, produced by Reach Advisors, compared all women and all men, while the IWPR report compared women and men with equivalent education and skills. If the Reach Advisors report had controlled for education, it would have found women earning less.

In fact, much of the gender wage gap (this is the old-fashioned one that favors men) can be traced to continuing occupational segregation. Among highly-educated, high-skill workers, the IWPR study found that jobs dominated by men (i.e. at least 75 of employees are men) yield earnings of over $1400 per week, compared to just over $900 per week in jobs dominated by women. The absolute difference is smaller, but even in low-skill jobs, male-dominated occupations generate earnings around $550 per week compared to just under $410 per week in female-dominated jobs. Educated or not, working with lots of women is a recipe for lower wages.

What the Reach Advisors study could not conclude, but the IWPR report clearly shows, is that women who want to get ahead should be entering male-dominated occupations. For low-education types, that means construction and factory work and transportation. For high-education types, that means engineering and science, breaking the glass ceiling into upper-level management jobs, or working in construction, but as a manager. Integrated occupations also pay better than female-dominated jobs at all skill levels.

And here is where the Reach study is genuinely misleading: the situation for young women is not improving. Although it may have been unintentional, the Reach study did not analyze changes over time. When the IWPR researchers did that, going back all the way back to 1972 data, they found that women entered male-dominated jobs in large numbers throughout the 1970s and 1980s. By the mid-1990s, progress stalled. For young women, matters proceeded to get worse, with a fifth of the reduction in segregation gained in earlier decades wiped out in the few short years since 2002.

The surprising conclusion of putting the studies side-by-side: women are earning on average more than men because of education. Education has helped millions of women do better in the labor market. But women would earn even more if the labor market were desegregated.

Cross posted from FemChat, blog of the Institute for Women's Policy Research.  Robert Drago, Ph.D., joined IWPR as the Director of Research in 2010. Prior to joining IWPR, Dr. Drago held positions as Senior Economist in the Joint Economic Committee of Congress and Professor at the Pennsylvania State University in the departments of Women’s Studies and Labor Studies.