Thoughts on the Pipeline

As I outline here, we have a shortage of people in computer science in general, and especially a shortage of women and minorities.  People usually break the shortages down into two parts: pipeline (getting people to enter the field) and retention (getting them to stay).  CS has problems with both parts.

There are many factors in the process of people choosing a career.  It’s not an overnight thing; it’s a lifetime evolution of identity.  We all have different personalities, different balances of creativity vs. analytics, different natural abilities, different upbringing, different genetics, different values.  We are members of a society that pushes on us certain expectations and social norms.

Let me sum up my thinking on this topic so far, though.  Let’s take a page from crime-solving.  To perform any act, a person requires means, motive, and opportunity.  Let’s examine those with respect to computer science.  To enter CS, one must have:

  • Means: access to computing resources and instruction.  This is actually a major problem in our pipeline.
    • The hardware cost alone is enough to ensure an economic imbalance that reduces minority representation in CS.  Solutions would need to include monetary support from both public and private sectors.
    • There is also a severe shortage of instructors, since anyone with software skills can make so much more money in the software industry than in teaching.  The shortage of labor leads to a bigger shortage of labor, and the only solution I can see is what the TEALS program is trying to do: involve professional software engineers in CS education.  I wholeheartedly support this initiative – you can learn more at https://www.tealsk12.org/.
  • Motive: desire to pursue the occupation.  There are many components to this, including:
    • You must believe you can succeed.  I think this is one of the top causes of imbalance in our pipeline.  Stereotypes in our society say that computers are for white male nerds who naturally feel an affinity for those computers.  Anyone else – minorities, females, and even white males who don’t self-identify as nerds – has absorbed a message that this isn’t natural for them.  The term stereotype threat has been coined for the anxiety associated with being in the outgroup, and studies have shown that being reminded of your status as a member of the outgroup can instantly reduce performance.  For example, women do worse when told they’re working on a “math test” instead of a “a problem-solving task,” and the difference goes away when they’re informed about stereotype threat.  There is a difference in how men and women evaluate their own performance, also.  Men are more confident and rate their competence higher than women do.  This is broader than just computer science, but may be especially true when stereotypes are pushing men and women into specific gender roles.  Solutions to ensure that women and minorities feel they can succeed, MUST include breaking down our societal stereotypes.  (Thus my book project!)  We also need to get more role models into the public eye.
    • You must believe you will enjoy the work (enough).  Stereotypes impact this one in the same way as above.
    • You must believe the work has value.  Jane Margolis states, and I believe her to be right, that women tend to be more interested in computer science when you emphasize the valuable things they can do with it (CS-as-a-means), the benefits they can bring to society, rather than just assume people will pursue CS because they enjoy it (CS-as-an-end).  I actually think it is likely this is true of minorities in addition to women, because I believe the gender-difference is rooted in stereotypes.  So we can bring more balance to the CS population by being careful about how we talk about the value of CS careers.
    • You must believe you will make enough money to support the lifestyle you desire.  The shortage of labor in CS makes this very easy to satisfy; here’s a link from the federal Bureau of Labor and Statistics on that subject.  Though I’ve actually encountered perception issues here.  I’ve given career talks to high school students in the past (esp. when I was living in Atlanta) where the students actually believed jobs were declining, due to jobs moving overseas.  They had it backwards: companies were setting up engineering centers overseas because they couldn’t hire enough talent in the U.S.  And since that time, emphasis on these centers has reduced, because working across oceans is hard.  Fixing broken perceptions on the financials should include building programs to discuss CS career opportunities at the high school level.  We should be actively promoting CS early enough that students have time to explore it before making career decisions.
  • Opportunity: In addition to having access to the resources, an individual must have exposure to them, the time to use them, and an environment that makes it safe and natural to explore one’s curiosity.  I don’t have data to back it up, but I believe that historically and likely even now, people’s biases and stereotypes lead them to expose boys to computers more often than girls, and at a younger age.  As parents and educators it is our duty to consciously even the playing field for our sons and daughters here.

A lack of means almost certainly impacts underrepresented minorities, but probably is not a major cause of gender imbalance in software.  Reduced opportunity is probably a factor in both race and gender.  But as I see it, reduced motivation is the #1 factor in the shortage of women in our pipeline: women self-select out.  We can work to change that by changing our social norms, promoting new role models, and changing the way we talk about the value of CS.

The pipeline isn’t our only problem in software diversity.  But it is the start of the problem.  And finally, let me note that if your organization is meeting the pipeline percentages in hiring, and publishing papers, and choosing speakers, and promoting leaders: you’re not leading, you’re just maintaining status quo.

 

Leave a Reply