Objectifying

My PhD was all about image analysis in research dentistry. We’d photograph teeth, and I’d program computers to figure out which pixels were, say, stained and which were clean. It’s surprisingly tricky because clean teeth are neither uniform in colour nor flat, so the same colour can be stain in one place and clean in another. The upshot is that on a pixel level, the computer was a bit rubbish.

It wasn’t great on a tooth level, or even a person level. You would, in short, be a fool to use my software as a diagnostic tool for an individual patient. But that doesn’t make the test useless, because that isn’t really what it’s for. It’s designed to eliminate the subjective and hard-to-reproduce standard measure, which is some guy gaping into your maw and declaring it ‘about a six’.

There are loads of tests like this, including more or less all surrogate outcomes, and they’re really useful. And in all cases people try to use them inappropriately and then whinge that their patients aren’t improving and nobody cites their papers.

I was thinking about this because

Turns out that discussing the Bechdel test at 1am with a bunch of drunk male nerds is a mistake #whoknew

— Vicky (@ViolettaCrisis) April 18, 2014

and I realised the Bechdel test is exactly the same: it’s a crappy measure of how, for want of a better term, feminist a particular film is. I don’t think anyone’s contesting that. If you’re making Superman Versus Batman, and you correctly want every scene to drive forward the plot, then any scene with characters other than Superman or Batman discussing things other than Superman or Batman is likely to get cut, and that’s fine.

(Although SUGGESTION: Batman told from the perspective of the Gotham police, and the reveal at the end is his identity. And there’s totally no reason it should turn out to be Bruce Wayne, because everyone expects that. So why should it even be a man? Come on, this would rule.)

But, the test is objective and repeatable, so it’s very useful for comparing between studios or years. If 25% of one studio’s output passes and 75% of another’s passes, that’s useful data, and it’s fairly reliable even, because the test has been around for ages (so you didn’t make it up post hoc) and the results can’t be faked or skewed (much) by subjective biases. Anyone can watch those films and verify the result, in a way you can’t (directly) with subjective opinions like ‘I think Hollywood is getting more sexist’. You can do science with this stuff.

The only ways I can think of that it could be rendered useless would be if Hollywood found out about it and started studying for the test, or if Hollywood became so awesomely enlightened that films started failing because one of the two characters who would otherwise have passed them was genderqueer and didn’t identify as female.

And neither of those feels like a defeat.