How do you know? I mean, how do you test for false negatives in the hiring process? Perhaps you're rejecting lots of people without side projects, but they go on to be successful anyway?
Sure, I mean I didn't say I actually use this test to make hiring decisions. I can just directly test lots of things like ability to code or problem solve (not perfectly for sure, but I do my best to evaluate directly the skills that can be evaluated directly). I am just making the point that there are often very highly correlated attributes that people can have, and one that is easy to test for can give you a lot of information about the ones that are harder to see.
I agree their might be some correlation. What I'm saying is:
- If you really meant you can get 95% precision in hiring from the answer to one question, I believe you only if you're super-conservative in hiring (i.e. will reject unless you're super-confident). In this case, your recall and false negatives is also going to be really high, so your single-question test isn't really helpful.
- If you actually meant 95% accuracy (i.e. precision in the everyday sense, not the math sense), then I don't believe you, because you probably can't estimate your accuracy, unless you also hire some people who fail the interview process.
There are awesome programmers and technical leads who don't have side projects just because they are so focused on their work and being a good parent.
I'm probably over-analysing your original statement, so I'll stop here.
Right, I understand what you're saying, but even being a parent people might put a side project on hold or not have as much time to spend on it, but that doesn't mean they can't talk about what they have done in the past.
You can never truly get rid of false negatives efficiently. However, false negatives are orders of magnitude better than false positives. So just try to minimize false positives and try not to lose sleep on the false negatives.
Agreed. It's especially important for smaller companies where each hire has a big impact. You're trying to minimize risk and have a good to great hire (technically and culturally), not hire the absolute best.