Gradual Epiphany

Bias

JD Maturen recently shared a 2005 paper that provides a deft and piercing examination of gender bias in job interviews. In particular, the study showed that people tend to redefine the criteria for success at a job, based not on the actual strengths/weaknesses of the applicant, but on the gender of the applicant and the cultural gender stereotype of the job. Study participants favored male applicants for a police chief job while women applicants were favored for a professor of women’s studies job. Importantly, the study was able to eliminate stereotyping as the source of bias and instead demonstrated that study participants were actually redefining their notion of ‘what it takes’ based on the idiosyncratic credentials of the person they wanted to hire.

In a nutshell, the participants were redefining “merit” based on their own biases and expectations.

This result alone should be setting off alarm bells for anyone in tech who is a hiring manager. The study suggests that when we hire, we tend to hire those people who fit deep-seated biases for the tech industry (i.e. male) and, even worse, we will redefine and reinterpret the merit of a given candidate based on these biases. In addition, male hiring managers will tend to do this measurably more than female managers, and we already have an industry full of male managers, so this will create a self-reinforcing feedback cycle of bias and selection.

This is no meritocracy.

You might, at this point, be shaking your head and laying claim to the belief that the tech industry strives for objectivity so surely this analysis can’t be correct. If we can just be objective and logical, as we love to believe we are, the best applicants will rise to the top. Here is where the study promptly takes your logical legs and sweeps them out from under you. It demonstrates, quite conclusively, that the participants who judged themselves most objective and free of bias, were in fact the MOST biased in their reinterpretation of merit.

The more objective you believe yourself to be, the less objective you probably are.

This is a devastating result. In an industry that prides itself on being a meritocracy, the truth is that we can not even evaluate candidates for a job without falling all over our internal biases. We are biased.

I am biased. I have told myself, time and again, that I am being objective — I can’t help it if I only get a few candidates that are women! I can’t help it if they are less qualified than everyone else! In retrospect, there probably WERE qualified candidates whose merit I dismissed as I rationalized my own biases under an illusory veil of objectivity.

We can not trust our own ability to be objective when it comes to hiring — we are demonstrably incapable of it. Instead, we must rely on external rules to help us ensure everyone gets a fair review. The study outlines two specific things that have been shown to reduce gender bias. Firstly, we must use a predefined structure for every interview — something I’ve personally been doing since I started hiring, albeit for different reasons. Secondly, we must clearly define the standards of merit for a position prior to the review of ANY candidates. Neither of these are a panacea for the problem, but they are relatively simply and effective ways to move in the right direction.

Want to know why there aren’t more women in tech? It’s because we don’t hire them.