8 days
Lee DSilver badge
Reply Icon
Re: The first one is free
The problem with fancy interview questions is that you are only testing whether people can answer fancy interview questions.

It's a self-fulfilling prophecy - you're selecting candidates based on some arbitrary criteria which is only vaguely related to their job, so the candidate who "does best" may not actually be good at anything vaguely related to the actual job they need to do.

I've always held that job interviews should be little more than "We pay you for a day, you work with us for a day". Anything else is nonsense. All those lateral thinking and logic tests (which, incidentally, I'm amazing at, being a mathematician) will knock out people who would have been great at the job if it wasn't for that test. All those impressive answers someone gives in interviews where they turn a perceived problem with themselves on its head and make you think they're wonderful? Congratulations you've hired a very good BS'er. All those interview that utterly impress the management types with management-level BS, while all the people who are hands-on, in the field, etc. are completely hating them? If you go ahead, you've hired someone good at speaking managementese who's going to be hated by their underlings and co-workers.

The interview is a selection process - and a two-way one at that! You have to consider "What am I selecting for with this question?" but more importantly "What am I selecting for when I'm looking at the answer?". Because a simple technical question may well be useful to "get right", but someone who admits they don't know, asks if they've be allowed to research the answer, takes a decent stab, tells you what they are sure of and what they are not based on their previous knowledge/experience, and can state where they'd go to get a definitive answer... I'd rather have that guy. I want to select for their METHOD and their communication and honesty, rather than that they don't know what menu the button is under to restart a cluster, or whatever.

Interviews are a selection process. Select for what you want to see, not some arbitrary score system of nonsense. Because like natural selection, if you cull perfectly good applications based on random nonsense and/or the highest-scoring person is just someone lucky, then you're going to end up evolving entirely the wrong direction compared to what you actually want to happen.