NPR When Women Stopped Coding in 1980’s: As we repeat the same mistakes

The NPR Planet Money segment cited below is excellent.  I’m really glad that they reached out to Jane Margolis and Telle Whitney to get the history right.

The question that they don’t address in the segment is, “Why did the classes get so much harder in the mid-1980’s that only the boys who were playing with PCs could succeed at them?”

In the early 1980’s, interest in Computer Science spiked.  There was more interest than there were seats available in CS classes.  Eric Roberts talked about these times in his keynote at the Future of Computing Education Research workshop in January 2014, which I blogged about here.  What to do with the burgeoning enrollment and no additional resources? Caps were put into place, and classes became harder.  Berkeley raised their cap until you had to have 4.0 in all your pre-requisite CS classes to get accepted to the major.  Eric Roberts was chair of the CS department at Wellesley in the early 1980’s, and he told me about introductory CS classes at MIT with insane workload, where only the boys with lots of prior CS experience and who were fanatical about computing were getting through.  Jane Margolis and Alan Fisher talked about this phenomenon in Unlocking the Clubhouse when they describe how the men and women in the CS classes at CMU had different views of the computer, which influenced how they interacted with it and how much time they were willing to put into their classes (nice summary of this story is on Wikipedia).

The classes may not have been made harder explicitly to deal with overcrowding, i.e., to “weed out.” It may have happened in response to an influx of boys who already knew a lot from playing with their PC toys, compounded with a lack of resources because of the overcrowding. With boys who already knew a lot, CS teachers could start skipping over topics, or covering them lightly, or just assigning programming tasks so that the student “figures it out” on his or her own. If a student can’t learn with this approach, then teacher might decide that the student just “can’t” learn to program. Maybe the student doesn’t have the Geek Gene. Some students do succeed with this approach, because they know a lot from prior experience (or have the Geek Gene).

Now, put this in the setting of high enrollments and tight budgets.  A student with lots of prior experience needs less teacher time to succeed. A student with less experience needs more time and effort in order to succeed in CS classes. In lean times, there are fewer resources for teaching, and those with less experience will not get the resources they need to succeed. Students with more experience will succeed just fine, so we continue to have high-quality CS graduates who get good jobs. Unless we look carefully at who is succeeding and who isn’t, we might not even notice that our program now presumes prior experience in order for the student to succeed.

What’s scary is that we may now be following the exact same path.  Eric has been warning about this for some time (see blog post). Enrollment in CS is exploding nationwide.  Now, the caps are starting to be put into place.  Berkeley now requires a 3.0 in the pre-requisite classes to get in to the major.  Here at Georgia Tech, the College of Computing has just requested to have a grade requirement in pre-requisite CS classes before allowing students to transfer into CS.

It’s still the case that it’s mostly wealthier (middle or upper class), white or Asian males who get access to high school CS. That’s in Barb’s AP analysis that got so much coverage this last year (see blog post here and the media coverage here).  AP CS is the most gender-skewed AP (more male than AP Studio Art is female).  So, even if you’re in a school that can afford AP, women will most likely not be in the CS class. In our AP analysis SIGCSE paper last year, we showed how wealth in a state has a strong relationship with AP CS offerings in the state.  We’re now starting to show the relationship continues to the district level as appeared in this blog a few weeks ago.

These kinds of caps have two effects which limit access by women and under-represented minorities (the second of which was pointed out to me by Eric):

• First, the students who succeed the most in intro CS are the ones with prior experience.
• Second, creating these kinds of caps creates a perception of CS as a highly competitive field, which is a deterrent to many students.  Those students may not even try to get into CS.

I understand why caps are going into place.  We can’t support all these students, and there are no additional resources coming.  What else can CS departments do?We might think about a lottery or using something beyond CS GPA to get those seats, something that’s more equitable. State budgets for universities have been cut back across the US, and it’s not clear that anyone (companies or the Federal government) could swoop in and cover that shortfall.  In lean budget times, few university administrators (public or private) are willing to invest in CS right now.  There will likely be a push for more MOOCs in the introductory courses — which is exactly where MOOCs are least effective (see my article in Ubiquity.)

It looks likely that we are going to reduce the diversity in CS, again.  While on our watch.

Mark Zuckerberg. Bill Gates. Steve Jobs. Most of the big names in technology are men.But a lot of computing pioneers, the ones who programmed the first digital computers, were women. And for decades, the number of women in computer science was growing.But in 1984, something changed. The number of women in computer science flattened, and then plunged.

via Episode 576: When Women Stopped Coding : Planet Money : NPR.

Many in the ideas for this blog post came from discussions with the Diversity Task Force of the ACM Education Council. All the mistakes are mine.