Off the Beat: Bruce Byfield's Blog

Sometimes, free software feminists can frustrate their supporters. Frequently, they seem to have only the vaguest sense of tactics, promoting pop psychology and working against acceptance of their cause with rash actions and statements. But if you even start to doubt that their basic critique of the free software community is valid, all you need to do to quiet your doubts is look at who is running projects and companies. Counting the number of women active in free software may be impossible, but governing boards and executive teams are listed clearly -- and women are seriously under-represented on almost all of them.

Don't believe me? Then take a random sample as I did:

Despite precautions, I may have missed one or two women in leadership roles, either through miscounting or failing to recognize a name as belonging to a woman (if so, I hope the people carrying those names will forgive my linguistic ignorance). However, these errors are unlikely to affect the general picture. Even by deliberately trying to create a biased sample by choosing projects and companies that are supposed to be relatively woman-friendly, it would be hard to get samples that are very different overall than mine.

Even the type of leadership role does not look as though it makes much difference. Although all the listings in these tables are for leadership positions, they include salaried and volunteer positions, elected and appointed positions, and formal boards and lists of leading contributor, small groups of leadership teams and diffuse ones.

However, the details of the positions do not seem to make much difference. In all these circumstances, women generally have 1-2 positions and often are not represented at all.

Reading the statistics
Out of 204 positions on boards and lists of leaders or major contributors, I count 14 that are women, or just under 7%. This percentage is higher than the 1.5% given in the FLOSSPOLS survey about women in free software that is usually mentioned when the subject of women's participation is raised. However, the FLOSSPOLS survey was based on information collected before 2006, and the subjective impressions of long-time community members -- including mine -- are that the famous figure is somewhat higher in the last few years.

However, whether the number of female leaders reflects the general number of women involved in free software is uncertain. Free software leaders are not always project contributors. Nor are they necessarily recruited from their ranks. In at least some cases in my sample, the leading women were chosen for their business or legal expertise rather than their coding or documentation skills.

In addition, at least some women who contribute to free software are like their male counterparts, and have no interest in taking on leadership roles. The assumption that a constant percentage of female contributors will move into responsible positions is almost certainly unjustified.

If I had to guess, I would say that the figure of 7% represented the high point of the range for women's participation in free software. The actual figure is likely to be above that in the FLOSSPOLS survey, but below that given by my tables.

Another question that arises from my sampling is why so many groups have only one or two women. This is so far from what I would expect in a random distribution that I wonder whether tokenism is involved.

I mean no disrespect to the women involved, because tokenism need not be deliberate. As the science fiction write Samuel R. Delany pointed out, people tend to see a group that consists of one-third women as being equally balanced. If that is true, then one or two women on a five person board could easily be misperceived as being closer to an even split than it actually is.

I can't help noticing, too, that in three cases --Canonical, GNOME, and Fedora -- the sole woman is the CEO or Executive Director. Could having a woman in charge assist the misperception simply because she is more visible and active than the male leaders? Ironically, the more competent a woman in charge might be, the less likely the gender imbalance might be noticed.

Under-represented in plain sight

But such matters are details and interpretations. What matters most is the general picture. If the imbalance was just a matter of random distribution, a few examples of under-representation would only be expected. But almost all of them? Given the variety of organizations, the consistent figures seem especially unlikely to be a coincidence. There may be a handful of groups in which the leadership includes more women than men, but not among the most influential or best-known.

Yes, a few women are in positions of power in free software. But, whatever their exact percentage, they are consistently a minority, just like as the women writing code or promoting a new release. The only difference is that they are visible where others are not, and the only question is what we are going to do about it.