Simply professing a belief in equality doesn’t protect you from subliminal bias.

It’s pretty much accepted that sexism is rife in academia, but it’s important to recognise that this doesn’t automatically equate to rampant misogyny – much of the anti-female bias is unconscious. Even champions of gender equality can be guilty of subconsciously believing in, or adhering to, a rather less progressive system than they might publicly profess.

For instance, how often do the following questions get asked:
(to a male colleague, newly become a father):
“How long will you be taking off for childcare? Do you think you’ll be able to manage your work?”
(to a male colleague, with children in daycare):
“But you work full time – are your kids in daycare that long?”
(to a male colleague struggling with a new technique):
“Isn’t it funny how men always have a bit of a problem with technology?”
(to the organiser of a conference session):
“Why are there no male speakers in this session?”
(to the co-authors of a paper):
“Why haven’t we named any male reviewers for the paper?”
(to a funding panel organiser):
“Why are there no men on this panel?”

Almost never, presumably. Women are usually ASSUMED to be exclusively taking time off for childcare, while it’s still a bit of a surprise if men do. Nobody questions a father who works full-time about whether his children spend too long in daycare, or if he spends enough time with them. Nobody implies that a man slow to utilise a new technique does so for biological reasons. Gender equality in conference sessions, peer review processes, or funding panels is often discussed, but lethargically enacted.

But this doesn’t mean that the scientific hierarchy is a brotherhood of women-hating chauvinists. In part, perhaps in large part if we’re being generous, it reflects the gap between the conscious mind and the unconscious presupposition. Even male feminists can be guilty of a weird kind of sexism, the subtext to a lot of their well-meaning utterances being “It takes a man to fight effectively for women’s equality”.

We shouldn’t be too surprised about this, and we actually – provided we react accordingly – shouldn’t be too existentially troubled either. We are bombarded by male dominance images of scientist (venerable male professors, men in white coats), and if there’s a woman in a white coat on TV there’s still good odds she’s explaining the “science part” of a hair-nourishing formula in a shampoo commercial. Consequently, it’s hardly surprising that we, men and women alike, easily end up associating science as a male activity in which women are generously allowed to participate on occasion. (Note that the way things are certainly does not mean they are the way they should be – but in the current system it’s not surprising that we acquire such a bias.)

In the social sciences, the implicit association test (IAT), first developed in 1998, has been an ingenious method for detecting these unconscious biases. The test is simplicity itself – just get the test subject to press two keys on a keyboard as words flash up on a computer screen. First off, the left-hand key is pressed for (say) male terms like “son”, “father”, “grandpa”, “man”, and the right key for corresponding female terms. Then the process is repeated with science terms (“physics”, “mathematics”, “astronomy”, “biology”) on one side, and arts terms (“music”, “poetry”, “literature”, “dance”). Then the terms are added – you have to press the left key if a word is either male- or science-related, and the right key if a word is either female- or arts-related. So far so good?

Then the criteria are switched. The left key must be pressed if the word is either male- or arts-related, and the right key if the word is either female- or science-related. All that needs measuring is the change in response time to determine if the subject finds it harder to associate, say, female terms with science ones. And it is a horrible, horrible, sensation – especially for one who professes equality – to feel that awful hesitation in your own response, as one finger itches to press but is unsure if it’s correct when only seconds earlier the opposite configuration was simplicity itself.

The IAT isn’t perfect by a long shot, but it’s still a very good way of circumventing the social-desirability bias that plagues questionnaire tests, and the results can be alarming. Even people who are avowed equal-rights campaigners can find that they have an implicit bias against a particular racial group/sex/religion/whatever, and that insight can be very upsetting. It’s an uncomfortable truth that applies, and needs remembering, in gender relations. Whatever we may consciously articulate, we may subliminally believe something rather different – and avoiding the impact of that subliminal belief is neither easy, nor automatic.

Truly practising equality requires constant self-monitoring, vigilance, and awareness, ideally in a non-confrontational atmosphere. People need help overcoming their implicit bias, and in the long term it’s important that we also start addressing the causes of those biases to limit the degree to which they are acquired by subsequent generations.

Interested in taking an IAT? Project Implicit has a range of different ones here.