This is a post I’ve put off writing on purpose because it’s not my favorite topic of discussion. That’s not because I feel shy about it, it’s just because people have usually already made up their minds on this particular topic which makes the opportunity cost of the discussion high. I’ve noticed Lisa Crispin and others making valiant efforts to have this discussion in testing. I agree with her the others involved that it is time.
We need to talk about the role of gender and diversity in testing.
I am tired of hearing about how my being a woman is important to the way I test. It’s a poor definition of “woman” that I don’t believe holds up well if it’s really dissected. There are many different ways to be a woman, and I’m not going to highlight all of them here. I’ll just point out one stereotype that needs to go: women have babies. I don’t have babies and I don’t know that I’ll ever have a baby. In fact, I have plenty of women friends who never want to have a baby. Are we still women?
The flip side of this is that each person has differences that make them valuable on a test team. Hopefully these advantages are obvious enough that there’s no need to go through that argument. The problem arises when we start stereotyping individuals into monolithic groups that are actually quite varied.
Lately I’ve been thinking about this in terms of levels of measurement. Maybe that’s because the book I learned this from, Stephen Kan’s Metrics and Models review, really goes for the throat in the examples used to illustrate levels of measurement.
Nominal: Classifying elements into categories. Kan uses the example of religion by saying that, “if the attribute of interest is religion, we may classify the subjects of the study into Catholics, Protestants, Jews, Buddhists, and so on.”
Ordinal: Ranking is introduced. Kan writes that, “we may classify families according to socio-economic status: upper class, middle class, and lower class.”
Interval Scale: At this level, there are exact, standardized differences between points of measurements. Elements can be compared using addition and subtraction and depend on having some standard of measurement. Kan uses a KLOC example to illustrate this one, “assuming products A, B, and C are developed in the same language, if the defect rate of software product A is 5 defects per KLOC and product B’s rate is 3.5 defects per KLOC, then we can say product A’s defect level is 1.5 defects per KLOC higher than product B’s defect level.”
Ratio Scale: This level is differentiated from the interval scale only because there is a zero element present.
Where do humans fit on this scale? The classifications we have for each other are nominal and ordinal categorizations, but I don’t think that the levels of measurement come anywhere close to defining the measure of human experience. Gender is what I get thrown in my face because I happen to have a vagina. Never mind the fact that I am the one earning the money in my family, I don’t have children and I don’t wear pink or even bake.
There is a nasty undercurrent in testing at the moment that tries to define me as “woman tester.” There’s no need to even look that hard if you want to find it. When I see this I will call it out and I will call it out loudly. I call it out because it undermines hard work done every day not just by women who show up for their tech jobs. It’s also undermining respect shown to women by hordes of male geeks who want things to be better. Guys: I hear you. I know you want me to feel happy and comfortable at work. I know you want more diversity in testing and technology. It means the world to me that you feel this way. I hope that we are far enough along with this problem that others, male, female, transgendered, will call it out with me.
So if you are among those who think we all ought to be wearing badges announcing how great it is that we fit some cultural stereotype/straightjacket, I hope you take some time to rethink that stance. It’s a waste of time we could be spending on other problems in testing.