Google looking for an algorithm for keeping women

November 12, 2012 at 4:08 pm 21 comments

Glad to hear that Google is aware that they’re losing women, and that they’re trying to study themselves to figure out where.  I hope that they’ll be successful.

A big part of the problem is what they’re not doing and not seeing.  As one quoted former Google executive said, “I don’t think there’s a gender bias per se, but I think the c-suite at Google is going to belong to product owners, not business people. People witness it as a demotion of women. I don’t view it as that. I view it as a demotion of business.” Do the folks doing these analyses see that kind of distinction, between product owning and business? Or consider the nice example from the NYTimes piece, quoted below. If you assume that having accomplishments and bragging about accomplishments go hand-in-hand, you might not see that you can have one without the other. Clearly, Google is now seeing what happened (that women weren’t interviewed because they weren’t bragging), and that gives us hope.

Meanwhile, there is the very Google-y approach of gathering data on precisely when the company loses women, then digging deeper to figure out what is happening and to try to fix it.

The results, Mr. Bock said, have been noticeable — at least outside the senior levels of the company. One-third of Google’s 34,300 employees are women. He would not say what percentage of technical employees were women, but he said it was better than the national average of about 25 percent.

Google’s spreadsheets, for example, showed that some women who applied for jobs did not make it past the phone interview. The reason was that the women did not flaunt their achievements, so interviewers judged them unaccomplished.

via In Google’s Inner Circle, a Falling Number of Women – NYTimes.com.

About these ads

Entry filed under: Uncategorized. Tags: , , , .

Where are the graduate CS Education programs? £3m investment in Computer Science and Digital Literacy in Wales

21 Comments Add your own

  • 1. Alfred Thompson  |  November 12, 2012 at 5:21 pm

    A failure to brag is not an issue for women only. There are men who have the same issue. In general modern management expects employees to toot their own horns. If an employee doesn’t do a lot of self promotion then managers ignore them or assume they are not doing anything. I see it as laziness on the part of managers but apparently keeping track of what employees are doing is not seen as part of their jobs these days. Ultimately being able to sell oneself to management is the most important skill an employee can develop regardless of gender.

    Reply
    • 2. Erik Engbrecht  |  November 13, 2012 at 9:11 am

      I think it’s more nuanced than that. Take, for example, organizations that do merit planning (determining the size of raises) by ranking. The ranking is often done several layers up in the organization. In some ways this is good, because it makes the sample size large enough that you should have something that looks like a normal distribution of employees.

      But it also means that in order for someone to be ranked particularly far above average there needs to be general agreement among the managers. In order for a first-line manager to push a high performing employee up in the rankings, especially a more senior one, that person needs to have a reputation with the other managers, and ideally with directors and possibly VPs. So a couple layers of management may have a clear idea what their employees are doing, but eventually the pool gets too big. So now in order to be meaningfully (with meaning being the size of your raise) labeled a high performer you need to short circuit the normal communication processes up and across the management chain.

      My point is that I think in many organizations, especially for more senior employees, raises and promotions and choice assignments require a level of agreement among management and senior leadership that is hard to attain if they can’t put a face to the name and connect it with some project about which they have at least some direct knowledge.

      Reply
    • 3. Julia  |  November 19, 2012 at 4:42 pm

      Speaking of bragging, if you’re a woman, there is not only the floor, but also a ceiling to it. If you don’t talk about your achievements loudly enough, you’re seen incompetent. But if you have quite a bit to talk about (not even bragging – just enthusiastically talking about something you have a great interest in), and the interviewer/manager is not very competent in it, then you injure his ego, and you’re labeled as pushy and arrogant. It’s like you need to fit exactly in between being seen competent enough, but still less competent then your interviewer/manager/team members. Everyone of your interviewers, and everyone of your team members.

      Reply
      • 4. Mark Miller  |  November 19, 2012 at 5:57 pm

        I have a little experience with this. Managers can be very insecure about their own competence, but they probably have very good ways of hiding it. Just because they have an executive position doesn’t mean they earned it through competence. I worked for such a manager once, and it was not pleasant. I didn’t figure out what was going on until I’d worked with him for about 6 months. He was constantly throwing me off guard, making me, and many other employees he managed–men and women–look like uncoordinated goofs. I got so upset by this I could feel a nervous breakdown coming on… I left my position several months later. It was not an easy thing to do. I had worked at the company for 3 years before he became my manager, and I had a good working relationship with most people there.

        I had the chance to read up on this, and it turns out managers like him kind of do this on purpose. They know they’re less competent their employees, but they can’t let that show. Part of that is sabotaging their underlings, making them look less competent. The only way to cope with a boss like this was to diminish one’s own accomplishments, and compliment your boss’s, even though you’re not that impressed with them. You can do that, transfer out to another division, or leave. Those are your options in that situation.

        I saw a comment about this earlier in this discussion, basically, count your blessings. If interviewers are turning you down because they’re intimidated by you, you’d be better off working for someone better than them. Believe me, dealing with a boss who’s threatened by your competence Is not worth the trouble.

        Reply
  • 5. Valerie Barr  |  November 12, 2012 at 7:56 pm

    Similarly, companies that go look for applicant github accounts and other public code repositories will miss out on applicants who don’t care to display their work publicly (a form of bragging).

    Reply
    • 6. Erik Engbrecht  |  November 13, 2012 at 8:36 am

      Projects on GitHub provide concrete evidence of technical skills, and if the projects are collaborative they also provide concrete evidence of how a person interacts with teammates.

      Also, since in many cases these projects must be done in free time, they indicate a level of passion about software development. They also indicate the direction of that passion, because if I person is going to choose to work during their free time, they are probably going to do it in an area that interests them.

      It’s not uncommon in creative fields for people to have a portfolio, and for prospective employers or clients to expect to see it prior to making a hiring decision.

      Yes, GitHub often projects involve a certain amount of attention seeking. But they also provide significant amount of concrete information that is normally lacking during the hiring process.

      Reply
      • 7. Mark Guzdial  |  November 13, 2012 at 9:06 am

        Erik, I don’t think you and Valerie are disagreeing. Sure, GitHub provides concrete evidence of skills, but NOT being involved in GitHub is not evidence of a LACK of those skills. If a company were only to hire people who were involved in GitHub, then they would be getting people for whom they had evidence of performance, but they might be missing others who can perform but chose not to participate in GitHub.

        Reply
        • 8. Erik Engbrecht  |  November 13, 2012 at 10:14 am

          Maybe, I mostly objected to the term “bragging,” and I think labeling it as such is counter productive.

          Hiring and promotion decisions should be made based on the presence of evidence, not the lack. If a business considers GitHub projects to be solid evidence, but also is concerned about passing over people who simply choose not to participate in GitHub, then it needs an alternative means of acquiring equivalently strong evidence.

          Reply
          • 9. Mark Guzdial  |  November 13, 2012 at 11:30 am

            I doubt if the alternatives will allow potential employers to acquire “equivalently strong evidence.” But that doesn’t mean that the potential employee isn’t equally strong. In a tight computing labor market, the onus is on the employer to find employees.

            Reply
      • 10. gaudetetheology  |  November 18, 2012 at 10:30 am

        Also, since in many cases these projects must be done in free time, they indicate a level of passion about software development. . . . because if I person is going to choose to work during their free time, . . .

        A person has to have free time before they can choose to work during their free time. Since women are still significantly more likely to have a “second shift” of childcare, family, or housework responsibilities, relying on open source contributions structurally disadvantages women. It can also disadvantage people who can’t afford a computer and internet at home, or who need to work two jobs, etc.

        Reply
        • 11. Mark Guzdial  |  November 18, 2012 at 4:13 pm

          That’s a great point. Caroline Simard’s research at the Anita-Borg Institute bore this out. Men are more likely to be promoted into mid-level management than women, because mid-level (vs. executive) management still requires technical expertise, and men (with fewer “second shift” responsibilities) have more time to update and maintain their technical skills. As Caroline put it, “Saying high tech is a meritocracy doesn’t make it so.”

          Reply
          • 12. Erik Engbrecht  |  November 19, 2012 at 9:53 am

            “Men are more likely to be promoted into mid-level management than women, because mid-level (vs. executive) management still requires technical expertise,”

            Figure of merit: Technical expertise

            “and men (with fewer “second shift” responsibilities) have more time to update and maintain their technical skills.”

            Source of gender bias: Women on average have less time outside work to maintain technical skills, and are thus less likely to maintain them

            “Men are more likely to be promoted into mid-level management than women”

            Result: Application of gender biased figure of merit results in gender biased promotions.

            The meritocracy is working, in that it is promoting people who do well on its figure of merit. What I really think you are arguing is that the figures of merit used by the meritocracy are wrong.

            So let’s add a couple hypothetical figures of merit for mid-level management:
            1. Ability to multitask, particularly switching rapidly between technical and non-technical responsibilities
            2. Effectiveness at delegating tasks

            I think #1 is demonstrably biased towards women, because the female brain is more wired for multitasking than the male brain is. Speaking anecdotally, I think #2 is also female biased.

            It’s not that high tech isn’t a meritocracy, it’s that high tech often defines its figures of merit somewhat myopically, and probably fails to adjust them effectively for positions of broader authority and more limited direct technical involvement.

            Reply
            • 13. Mark Guzdial  |  November 19, 2012 at 10:05 am

              I disagree with your analysis. Women could have comparable technical expertise given comparable opportunities to develop that expertise.

              Reply
              • 14. Mark Guzdial  |  November 19, 2012 at 11:27 am

                I may have misinterpreted you, Erik — let me try again. If the figure of merit is “Has technical expertise,” then the SV hiring is a meritocracy. It just seems incredibly short-sighted, and counter to my discussions with recruiters, to base hiring based on current expertise. Recruiters I’ve spoken to say that they want “technical capability,” which I interpret as, “the ability to develop technical expertise, especially for new technologies.” If “capability” is the figure of merit, the current system does not adequately test and hire based on that merit.

                Reply
    • 15. Julia  |  November 17, 2012 at 4:21 am

      I rather noticed that the great majority of companies in SV don’t bother looking at my real projects/code. Like, they would even ask for links and code samples but as I submit my resume with the links right on the first page or in a cover letter, they start interviewing without ever looking at my technical blog, portfolio, code samples. I know as I track visits to my links. Some others just click on and out. May be about 1% or less actually look at something, I suspect that 99% don’t even read my resume. Is it because they expect pure bragging, with no evidence? Or because my name is female, so don’t bother?

      Reply
      • 16. Erik Engbrecht  |  November 17, 2012 at 6:21 pm

        My guess would be some combination of laziness, hubris, and stupidity.

        I know a few time I’ve been contacted by recruiters saying “I saw your projects on GitHub and thought you might…” and it was clear all they did some keyword match that turned out to be completely wrong. Using GitHub (or similar) as a recruitment or evaluation mechanism takes time because you have to evaluate code and enough expertise to evaluate it. Both of these traits tend to be lacking, especially in early stages of recruitment.

        Or perhaps they work primarily of off recommendations coming from a relatively small network of trusted people. I’m not in SV but I know a lot of people like that, especially non-technical people or technical people looking for people outside their own area of expertise.

        Or perhaps the projects you link to don’t look relevant to them.

        And some probably do see your feminine name and dismiss you. Such people will always exist. No amount of legislation or corporate policy or economic advantages conferred to those with more rational hiring policies will create a universal barrier against bigots occasionally working their way into a position of influence. Do you really want to work with/for them?

        I think poor recruiting practices are the rule, not the exception. I also think many companies make high-minded statements about their practices that they don’t actually follow, or don’t follow during initial screening (resume review, initial phone interview(s), etc).

        But the fact that many (probably most) companies don’t choose to use the objective evidence available to them doesn’t mean it’s a bad idea to use it or a bad idea for a potential recruit to provide it.

        Reply
        • 17. Julia  |  November 19, 2012 at 4:29 pm

          My case is: they don’t even click on my links, not even to see if the link exists! Usually, combined with expressing a lot of interest. I don’t know how more relevant the links could be – they are presented as “technical blog”, “portfolio”, “code samples” – not “my photo site”. I also had cases when people asked for my resume in professional groups, but never clicked on my links either. Then the common pattern is: long interviews on the phone (sometimes, a series), telling how impressed they were, will schedule further steps/onsite interview and then I never hear back from them. It’s funny that it’s not just 1 or 2 companies, but more like a rule for everybody not to ever let me step into the office (and if I do, there is a noticeable reluctance to show me around – with nothing to be ashamed of with my appearance, hygiene or dress). Like all of them share the same HR department. There is also an interesting correlation – if nobody clicks on my links, I have a very little chance to see them in person (unless, I see them in a different place), no matter how much interest they express.

          After awhile the candidate stops trusting her own feeling about employer expressions of interest, and even starts missing calls for further interviewing (because promises to call already have a very little meaning). I wish somebody made a research on it, it’s like a very distorted reality.

          My guess is, people are more interested in women/minorities as tokens, for PR reasons (see, we interview women/minorities, we are meritocracy! really, it’s better to be rejected upfront then to be somebody’s PR tool).

          Of course nobody wants to work with bigots. But what should we do if they are more widespread then anybody can imagine (like 9 out of 10), just hiding? “Hiding” is saying one thing, doing another, which doesn’t sound subconscious.

          Reply
  • 18. Mylène  |  November 13, 2012 at 2:10 am

    I am always amazed that policies that disproportionately (not absolutely — disproportionately) promote men and disproportionately fail to promote women are seen as “not gender bias, per se.” It makes me wonder what people think gender bias is, exactly? Some kind of violent hatred or paternalistic contempt?

    It’s time we redefined bias — it is not a feeling or an intention. It is an effect — usually unintentional. It doesn’t make me a bad person if I discover I am having that effect. It makes me irresponsible if I don’t do anything about it, though. According to the article, Google has identified some of the problem points and made substantive changes, with corresponding substantive improvements in retention and recruiting. If their data collection continues, it may turn out to be useful beyond their walls.

    Reply
    • 19. Julia  |  November 17, 2012 at 4:30 am

      It doesn’t need to be redefined – discrimination is defined by the government not only as a direct hostility but also as “impact based” – policies and practices failing to promote certain groups exactly fit.

      Reply
  • 20. Mark Miller  |  November 14, 2012 at 12:57 am

    Seeing the article and discussion here has helped me see this issue in a new way. I’ve had discussions with others re. gender discrimination in IT, or within the CS discipline over the years, but discrimination to me connotes an active effort to push out people who don’t fit what’s considered acceptable, or there’s a hostile work environment for people who don’t “fit in.” Mylène’s comment is interesting, since perhaps in many cases women are not being screened out deliberately, but rather because their behavior and responses do not communicate what are considered desirable traits. So perhaps some introspection about “what’s considered desirable” is in order.

    I’ve heard about this same sort of thing with IT job interviews, that good people get screened out because they don’t toot their own horn. It affects male applicants pretty often, from what I’ve heard.

    Reply
  • [...] Google looking for an algorithm for keeping women: “Glad to hear that Google is aware that they’re losing women, and that they’re trying to study themselves to figure out where.  I hope that they’ll be successful. A big part of the problem is what they’re not doing and not seeing.” [...]

    Reply

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Trackback this post  |  Subscribe to the comments via RSS Feed


Recent Posts

November 2012
M T W T F S S
« Oct   Dec »
 1234
567891011
12131415161718
19202122232425
2627282930  

Feeds

Blog Stats

  • 966,078 hits

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 3,135 other followers

CS Teaching Tips


Follow

Get every new post delivered to your Inbox.

Join 3,135 other followers

%d bloggers like this: