The Ninja Economist takes on your attacks over the (lack of a) gender gap in tech salaries

March 10, 2014
March 10, 2014

I recently wrote an article for Quartz in which I challenged the idea of gender wage discrimination in some fields in the tech industry.  A few minutes after the article posted on the website, the World Economic Forum sent the tweet:

But, that’s not actually what I was arguing. Headlines and tweets have the problem of fitting a lot of information into a small amount of space, so nuance is often lost. The central thesis of my article was that new research shows no significant difference in earnings one year after graduation between male and female engineers, who have the same credentials and make the same choices regarding their career. How’s that for a more precise headline?

On Twitter, many readers have questioned the data collection and research methodology. Others wondered how accurately I was portraying the issue of the gender gap. The following are the comments that came up most often in tweets, reader annotations and personal emails, and the responses from Claudia Goldin, a professor of Economics at Harvard University and Catherine Hill, director of research at American Association of University Women, who was co-author of the study, and myself:

Since Quartz is a publication for a general audience, it’s possible some terms were misunderstood when not clearly defined. Confounding typically means “confusing” or “surprising,” but in statistics, it also refers to hidden variables that can lead researchers to think a relationship between two variables exist when they don’t. A popular example is, “Ice cream sales increase swimming deaths.” The two things are positively correlated but there is a more likely explanation than their relationship. In this case, the “season” is the confounding variable since summertime influences both ice cream sales and swimming pool deaths.

In AAUW’s study, the sample was restricted to those under 35 years old, who were earning their first bachelor’s degree. If someone is over 35, he or she is more likely to have more work experience, which would lead to higher wages when compared to a recent graduate with little or no work experience. Also, if someone has two bachelor’s degrees, he or she is also likely to earn more money than someone of the same age, but who holds one degree.  Therefore, removing these two confounding factors makes it easier to study the effects of gender.

Most labor economists agree that the majority of the gender wage gap is explained through choices like education, occupation, or hours worked. Goldin, former president of the American Economic Association, is widely considered one of the top scholars in her field and believes that “the gender gap in wages is a summary statistic for gender differences in work.”

“Women without children have earnings almost equal to those of comparable men,” Goldin told me. “As women have increased their productivity-enhancing characteristics and as they ‘look’ more like men, the human capital part of the wage difference has been squeezed out.”

I don’t make the claim that wages in the first year are representative of lifetime wages. Though Goldin’s research also shows that in some science and technology fields for those under 45, women earn more than men (pdf) when controlled for education and work hours.

It’s true that the gender gap widens as women get older, but more research needs to be done to understand why the pay gap increases more in some occupations than in others. One explanation is the relationship between work hours and earnings, since some mothers reduce the number of hours they work or choose a more flexible schedule—two decisions that affect career path and lifetime earnings.

“The gender gap in pay would be considerably reduced and might even vanish if firms did not have an incentive to disproportionately reward individuals who worked long hours and who worked particular hours,” says Goldin.  “Such change has taken off in various sectors, [including] technology.”

First, median earnings don’t control for factors such as education or occupation, which are controlled for in the AAUW study. Second, women’s earnings rapidly decline to .70 (over her lifetime) and the study I am referencing looks at graduates one year out of college. Comparing raw earnings versus adjusted earnings isn’t analogous and neither is comparing lifetime earnings and young cohort earnings.

Constructing birth cohorts using data from the US Census and ACS, Goldin observes the following (pdf, table 1b): “Men and women begin their employment with earnings that are fairly similar, both for full-time year-round workers and for all workers with controls for hours and weeks. College graduate men and women working full-time, full-year earn in the 90% range when controlled for hours, weeks and education.”

Highlighting pay equality in STEM fields might lead to more young women entering those occupations, a change many people would support. When I asked if Professor Goldin feels there is public perception of wage discrimination, she agreed that “Now, it’s a lot less.” Pay equality may also encourage women to stay in the labor force longer since their participation in the labor market will be stymied if they feel disadvantaged, according to Goldin.

“Regarding our estimate of earnings for engineering majors one year out of college, it is similar to findings by other scholars [pdf],” said Dr. Hill. “The commenter seems to be misunderstanding the difference between the whole engineering workforce and those working one year after college graduation.  Of course, we would expect salaries to rise over time.”

The authors agree with my concluding statement that the two highest paying jobs, according to their study, are computer science professional and engineer. “We don’t have the breakdown on the relative importance of math, computer science and physical sciences,” Hill said.  “In general, computer science and engineering account for the majority of STEM jobs.”

According to a statement from Dice.com, career website that serves information technology and engineering professionals: “With tech workers, the compensation gender gap has disappeared. Average salaries are equal for male and female tech pros, provided we’re comparing equal levels of experience and education and parallel job titles.”

“It is not lazy,” Hill told me, “It’s simply what we can accomplish with the data set we used. Our study is limited to one particular time because we’re trying to get an apples-to-apples comparison. We have a panel of experts that reviews our reports; this information is on our site.”

The gender pay gap is a focal point of AAUW’s research and advocacy work. Taking a closer look at the data, the study finds that women’s choices—college major, occupation, hours at work—do account for part of the pay gap. It’s beneficial to do a study that only captures a portion of a woman’s career because it can help us have a more detailed understanding of the changing factors that affect the gender wage gap. By studying cohorts (age bands), Goldin’s research shows that women and men start their careers very close in earnings, then the gap widens for the first few decades but then narrows as men and women become older.

I asked Hill if my premise and conclusion were supported by their evidence.  She replied that my statements were consistent with their findings, that “one year after college, among those students who go on to work full time in the year after graduation, we found no statistical difference in the earnings of men and women who took jobs in engineering and math, computer sciences or physical sciences.”

“We used the Baccalaureate and Beyond data that is a federal survey for which information is available on the NCES website,” said Hill. “For most of our calculations, we used the online tool that is on the website and students (or anyone) can easily reproduce the numbers. For the regression analysis, we used internal data using a consulting firm called MRI.  MRI is a well respected firm and a leader in educational analysis. The report was based on an earlier analysis that I did with Judy Dey, an economist who now works for the BLS.  It is called Beyond the Pay Gap.  We did a similar analysis although it is not identical.”

Some have said my approach is excessively optimistic. In my article, I do state that this study shows there is a gender pay gap by college major, in most occupations, and overall.  However, I decided to focus on the positive aspect, that there are some professions with no statistical difference in earnings between men and women in the beginning of their career.

My conclusion was this:  The tech industry is unique in its history of being “equal pay for equal work”: A longitudinal study of female engineers in the 1980s, by Laurie Morgan at the University of Michigan, showed a wage penalty of “essentially zero” for younger cohorts. Today, the two highest paying professions with wage equality are in technology (computer scientist and engineer).

I support pay equality as well as a healthy discussion surrounding it. It’s important not to overlook the challenges women have faced in the fight for equal pay while we celebrating the more recent successes.

I asked Goldin if we shouldn’t discuss a narrowing pay gap because it hurts women by minimizing the issues they face. She replied, “Facts are facts. Truth doesn’t hurt.”
Follow Cynthia on Twitter @NinjaEconomics. We welcome your comments at ideas@qz.com

Top News

Powered by WordPress.com VIP
Follow

Get every new post delivered to your Inbox.

Join 23,610 other followers