3 ms·
Surprising and disappointing that the best methods were only barely over 50% correlation with performance on the job. What does the algorithmic whiteboard codi
by wffurr 13y ago
Surprising and disappointing that the best methods were only barely over 50% correlation with performance on the job.
What does the algorithmic whiteboard coding tech interview count as? Does it qualify as work sample and brain teaser combined with interview? I suspect it's too far removed from what software engineering is actually like to be a good work sample, and the prior knowledge required to solve algorithm problems doesn't make it a great brain teaser either.
- tokenadult 13y agoSurprising and disappointing that the best methods were only barely over 50% correlation with performance on the job. That's probably about the best we could expect, as job-seeking behavior when a candidate desires a job is different from job-doing behavior over the long term after someone is hired and knows more about the company. The appalling thing is that companies still use procedures (personal unstructured interviews, "personality" tests, and in Europe even handwriting analysis) that are demonstrably much worse than that, rather than the work-sample tests and general cognitive ability tests that at least give companies their best chance to hire workers who will do well on the job.