Damned Lies and Statistics, FOSS Sexism and Education

Off the Beat: Bruce Byfield's Blog

Feb 17, 2010 GMT
Bruce Byfield

The trouble with statistics is that they can be easily abused. This insight is hardly new to me, but its truth was reinforced when I read Mark Guzdial's suggestion that free and open source software (FOSS) was not a good match for education because few women or minorities participated in it.

Guzdial makes his comments in an article entitled, "The Impact of Open Source on Computing Education." After talking with Michael Terry, an assistant professor at the University of Waterloo who has studied open source usability, Guzdial suggests a number of reasons why FOSS might not be a good fit for computing education.

Guzdial suggests, for instance, that the myth that FOSS developers work for free might discourage students from entering computing science. He also suggests that the emphasis on code in the open source camp might make it harder to make the case that computing has broader social implications, and that the relative rarity of usability testing makes breaking into FOSS harder than proprietary software. At times, Guzdial seems aware that he is dealing with perceptions as much as reality -- at other times, he doesn't.

These comments are effectively addressed by Fedora's Greg DeKoenigsberg. However, what I want to focus upon is what DeKoenigsberg did not.

At the top of Guzdial's list is the argument that:

"At a time when we are trying to broaden participation in computing, open source development is even more closed and less diverse than commercial software development. Open source is overwhelmingly White or Asian and male. Some estimates suggest that less than 1% of open source developers are female."

Guzdial is referring to the FLOSSPOLS study from 1996. Actually, the figure given by FLOSSPOLS is 1.5%, which is still surprisingly low, but Guzdial's rounding down suggests that, on some level, he is determined to present FOSS in the worst possible light.

Or perhaps he is playing devil's advocate, although that is not the impression created by his comments. For example, his response to challenges to his remarks is that "it is not acceptable to criticize religion, Santa Claus, or open source development."

A case of special pleading

Guzdial is using the FLOSSPOLS figure very differently from most people. When FLOSSPOLS cites the figure, or the Geek Feminism bloggers (or, for that matter, when I do), the intent is not to denigrate FOSS. In each of these cases, the figure is raised by active participants in FOSS, who think that improving the figure would benefit the entire community.

Of course, you might disagree with that reasoning, but the perspecitve and intent is clear. Other citers of the figure see the low participation of women as an internal problem to be solved. By contrast, Guzdial uses it as a reason to avoid FOSS altogether.

This is a peculiar argument, no matter how you view it. For one thing, while advisers in post-secondary education might warn about sexism in a career field, since when is it a reason to steer anybody away from their choice of profession?

The fact that only 15.7% of CEOs of Fortune 500 companies are female is not used as a reason why young women should avoid studying business management. So why would the low percentage of women be a reason to avoid studying FOSS? While the percentage of major female CEOs may be ten times greater than the number of female FOSS participants, the number is still dismal.

The same is true of proprietary software. The same FLOSSPOLS study gives a figure of 28% female involvement in proprietary software -- once again, an improvement over the figure for FOSS, but hardly anything to boast about. Yet Guzdial does not suggest that computing science departments should discourage female students from applying to proprietary software companies, and steer them towards academia or private research instead -- although that would be consistent with his attitude towards FOSS.

FOSS and education

But the strangest part of Guzdial's argument is how peripheral his concern is to the actual goal of education. Advising students on their career paths is a service performed by the faculty and staff of any university department, but their main purpose is to educate.

Guzdial, however, says nothing of the potential benefits of FOSS in education. A couple of years ago, Chris Tyler of Seneca College, whose FOSS programs are a model of cooperation between academics and FOSS companies and projects, suggested to me that even students who do not continue working on FOSS benefit by having it in their curriculum.

According to Tyler, working with FOSS gives students the chance to work with much larger code bases than a classroom project ever could. They also greater opportunities to learn skills they need on the job, such as project management and collaboration -- to say nothing of familiarizing themselves with an increasingly important area of software development. Moreover, if they contribute to a project, they can have samples of their skills to show potential employers.

I would also add that FOSS frees computing departments from spending money on licensing, or expecting students to do so -- a serious consideration considering the years of financial restraint that North American post-secondary institutions have endured. True, students taught FOSS may not know a specific proprietary technology, but they will likely learn half a dozen free ones, and enough flexibility to ramp up faster than students whose experience is limited to proprietary software. Yet Guzdial does not address -- let alone refute -- any of these benefits.

By picking at peripheral concerns and setting inconsistent standards, Guzdial gives the impression of being strongly biased against FOSS. In the end, I can't help wondering if he is more concerned to find another argument against FOSS than in steering female students away from sexism. Although he is writing informally, which excuses many faults, his use of the FLOSSPOLS statistic is far from in the best academic tradition. In fact, it strongly suggests an undisclosed agenda.


comments powered by Disqus

Issue 28: GIMP Handbook 4th Ed./Special Editions

Buy this issue as a PDF

Digital Issue: Price $15.99
(incl. VAT)