3 ms·
that's actually what you want, though. It's because only successful candidates appear in the statistics. If Google let everyone in, you'd see a good correlatio
by CHY872 12y ago
that's actually what you want, though.
It's because only successful candidates appear in the statistics. If Google let everyone in, you'd see a good correlation between how well developers do in the interviews and how well they do at Google. If you dump the 95% of people who don't perform well in the phone interview, then what you're left with is a bunch of high performers who performed well to different extents (where the purpose of the interview is not to gauge ability but to demonstrate minimum ability), where the extent to which they performed well is largely noise.
Either a positive or negative correlation would be bad in this case.
- Retric 12y agoYou’re assuming everyone did well at their job. Google selects a lot of terrible people suggesting their process is broken.
- CHY872 12y agoNo such assumption made. If you have a positive correlation in your data, you're probably not placing enough weight on your interviews in your hiring process (i.e. you knew people weren't as good at interview but you still hired them). If you have a negative correlation you're similarly placing too much emphasis on interviews; you're likely hiring people who interview well over people who'd be better in the job. You might still hire bad engineers, but the uncorrelated nature of your interview/job performance data won't help you explain that.