Female Computer Scientists Make the Same Salary as Their Male Counterparts

This article seems to have fallen into a common misinterpretation of the results of the study. They assume that failing to reject the null hypothesis automatically accepts the null hypothesis.

The article says "The study authors did find that, on average, women in fields like programming earn 6.6 percent less than men (a Bureau of Labor Statistics study showed that women actually earn 11 percent more, however). But that difference is not statistically significant."

But that the difference is not statistically significant does not mean there is no difference. To show there is no difference, you'd need to show that the power used to run the study was sufficiently high to be confident.

The correct interpretation would be "the study failed to find a significant difference in salaries between men and women, but that does not mean there isn't one." Not as exciting of a headline I guess.

Update: I skimmed the study pdf and the study does not make the mistake as the article. In fact, I applaud the study for using two-tailed tests that gave them less power but more neutrality (so they were checking for gender differences in pay rather than if men were paid more than women). The study shows significant differences in pay in several professions but not engineering or math/CS.

This is a really, really bad article, the author clearly didn't bother to read the study.

I just skimmed it, 64 pages is a bit much right now, but from first look:

First of all, the study [1] just checked graduates one year after graduation - i.e., absolute entry-level. This may say some things but doesn't say anything about female computer scientists in general.

Second of all, the study checked many, many differences in pay - some of which are statistically significant, some are not. The author of OP's article just listed one of these comparisons. To quote from the text:

>All gender differences reported in the text and shown in the figures are statistically significant (p < 0.05, two-tailed t-test) unless otherwise indicated

Moneyquote: >Women working full time earned $35,296 on average, while men working full time earned $42,918

and

>In fact, the pay gap exists within nearly every category of institution and level of selectivity. Among public and private college graduates, women earned 81 percent and 86 percent, respectively, of what men earned one year after graduation. Women who graduated from public universities earned 86 percent of what their male peers earned. The pay gap was largest among graduates of private universities, where women earned just 75 percent of what men earned

And yes, that's a statistically significant difference. It's the very first two figures. (Edit: And depending at which paygroup you look, OP's 6.6% is really low - it goes up to 25% difference!)

tl;dr: There is a pay gap and it's statistically significant.

[1] http://www.aauw.org/files/2013/02/graduating-to-a-pay-gap-th...

Edit: The only thing I can find that is not statistically significant is the pooled pay for women in "Biological, physicial sciences, science technology, mathematics and agricultural sciences" and "Business" - i.e., something different than what OP claims, especially since the average of these jobs is likely to be different from the average of each job itself.

Edit2: I finally found OP's 6.6% pay difference, which is a complete misunderstanding of the article they copied from here [2].

>According to the study, there are seven professions with pay equity. When controlled for all factors other than gender, the earnings difference between men and women is about 6.6%, something most people don’t know.

So the 6.6% are not the pay gap in computer science, but the pay gap for _all_ jobs clustered together! Then it's no wonder that the 6.6% are not statistically significant, several different populations with different means are merged, and since t-tests compare the means, you get no significance.

[2] http://qz.com/182977/there-is-no-gender-gap-in-tech-salaries...

The thing is, salary is so arbitrary. People make up the number. How many tech companies out there actually says "all junior level developers make $80K, level 2 makes $85K"? Some people are paid more. If you are a frontend developer, you are likely getting less than a backend developer or a full stack. Sometimes fullstack engineers get the same salary as a frontend/backend but does two people's job. And I was told by people that asking other people's salary in American culture is rude. How can I find out what my co-workers are making. And even if I knew, how would I know my salary, as a male engineer, is too high? How do we judge? When we do an average on salary, we need several categories: companies (you can't compare Google to a local small business IT firm), location (NYC vs Portland), education, job performance (this is also quite arbitrary).

Please stop reposting this.

This is the kind of "science journalism" that I strongly dislike (and is promulgated by both sides in politically tinged debates like this one). Basically the author has a point of view, and sees in the study some numbers that support said point of view.

That aside, I think the biggest problem is not understanding where (the obvious) disparity in tech comes from. It's not discrimination, and least conscious discrimination. And I don't even think it's unconscious discrimination or side effects of so called "privilege".

It's simply that when tech startups are small, for various reasons (I'll get to this), they have a lot more males than females.* If the startup does well, early hires rise in rank by virtue of being there longer. If/when the reach the size where they are recognized by name, it's too late, most of the top roles (both engineering and management) are filled by males. Even if there was zero discrimination (or feelings of isolation), you would get the topsy turvy industry we see now.

Most studies focus on income but the far bigger disparity is wealth (both in this case and generally). The earlier employees in successful startups are phenomenally more wealthy than latecomers. These people are the ones who go on to start and fund more startups, and the cycle continues.

This is why PG's now infamous remarks about there being few female hackers (as opposed to CS students/engineers) are really insightful. Having Sheryl Sandberg or Marissa Mayer as visible leaders in large companies won't really change much (or anything when it comes to tiny startups). When these large companies hire, it's quite likely they do so scrupulously fairly. But even though the disparity is much better when it comes to new hires, the females amongst them (just as the males) start at the bottom.

Having more female founders is a start, perhaps, but it doesn't really solve the problem of too few female hackers. What you need are more female technical founders. I'm not saying that the non-technical founder role is less important (in fact it can often be a lot more important). But the technical founders are responsible for building the technical team from the ground up. Males tend to have more male friends, just as women tend to have more female friends. And the first people you try to hire are friends.

But even more than just technical founders, if you want to reduce the disparity at the source, there need to be more female "Mark Zuckerbergs" who come up with the idea and hack version 0 when everyone else is out partying (and, importantly, before fundraising is even a thought). This is when a founder has most leverage: when he/she has built the kernel of something that people want and everybody else wants in. Think of the scene from The Social Network (admittedly a hollywood movie) where Zuckerberg assigns roles and equity to his co-founders and basically brushes of the girls' offers to help.

This is a very difficult problem to have, if you want to increase the presence of women in this industry. It's difficult because there's nothing you can do to future "Mark Zuckerbergs" to force him to pick a more diverse set of co-founders. There can be no law, because this is almost like choosing your friends or significant others. A not insignificant percentage of startups with female founders in fact have their significant others as co-founders. This could be great for the startup (I know many great teams like this, obviously pg/jessica being one), but it again points to how difficult it is for females to find co-founders the way males do.

The only thing that would really work is for "these girls" to have been hacking since they were children and think that this is the coolest thing they could be doing. And there needs to be a critical mass of them so they can make friends with each other, because if you think that telling adult male developers to "play nice" is hard, try forcing 10 year old boys to invite 10 year old girls into their crew, (cooties and all).

* It would be cool if there was data to track the ratio of male engineers to female engineers as a successful startup grows. My hunch is the ratio improves steadily along with company size.

is it so hard to believe that women don't enjoy stem fields as much as men?

that men and women are actually different and that it implies different behaviour?

that actually you know what... the more freedom women get the /less likely/ they are to work in a stem field?

maybe we should pay them more in our naive quest to make the unequal equal - because it is 'politcally correct' - which ever kind of correct that is because it certainly isn't 'true'

Woo! Let the back patting begin! We ended sexism!

Where "the same" is 6.6% less.