The unconscious bias in technical-screens

When you get to a certain level of the org-chart in a startup-style tech company, you spend an embarrassing amount of your time doing interviews. Senior individual-contributor people end up doing technical screens because -- in theory -- we're good at judging who is actually good at something and who is faking it. Sometimes we get to do the culture-fit interviews as well, since we tend to be the people with a lot of tenure.

All of which means I'm in more of these than I used to be.

In theory, the technical screen is awesome because:

  • It's a test of skill, not glad-handing (though glad-handing is an important skill for senior and super-senior IC anyway).
  • It gives the reviewer a sense of a prospect's technical abilities by looking at their work-output.
  • It gives a sense of the prospect's adherence to language styles.
  • It's objective.

And it very well may be, but from talking with others who deal with these I've noticed something else creep in. It comes when a candidate is on the edge of close-enough. Not a clear passing grade, but close.

  • Men: A little pairing, and he'll get up to speed fast. Hire.
  • Women: Not quite where we want her to be, we don't have time to train someone up. Pass.

It's the old adage again:

Men are hired because of their potential, women are hired because of their proven ability.

Subjectivity comes in the edge-cases, as always. No matter how brutally pass/fail you make the tech-screen, if literally everything else is saying hire someone, you will be temped to look at that almost pass for the tech-screen and subjectively judge if it is good enough. That's human nature, we exist to add common-sense to ruthless numbers and automation.

So if you find yourself making that is this candidate good enough or too much work? calculation, ask yourself what biases may be at work on you. Be deliberate.