>Unstructured interviews reliably degrade the decisions of gatekeepers (e.g. hiring and admissions officers, parole boards, etc.). Gatekeepers (and SPRs) make better decisions on the basis of dossiers alone than on the basis of dossiers and unstructured interviews. (Bloom and Brundage 1947, DeVaul et. al. 1957, Oskamp 1965, Milstein et. al. 1981; Hunter & Hunter 1984; Wiesner & Cronshaw 1988). If you're hiring, you're probably better off not doing interviews.
The "make candidate solve puzzles" technical interview sounds like a very poorly designed IQ test. Just replacing it with more objective standardized IQ tests would help a lot.
The link you cited is talking specifically about "unstructured interviews", the sort where you're judged based on handshake firmness and questions like "What is your greatest weakness?" It shouldn't be too surprising that these are poor predictors of job performance.
Programming technical interviews are not supposed to be like that. They're supposed to be some mixture of a relevant-ish skills test and a plausibly-deniable IQ test. (One constraint here is that anything which is overtly an IQ test will make the company's lawyers get twitchy and start nervously trying to remember the details of the Supreme Court decision in Griggs v. Duke Power Co.)
>Unstructured interviews reliably degrade the decisions of gatekeepers (e.g. hiring and admissions officers, parole boards, etc.). Gatekeepers (and SPRs) make better decisions on the basis of dossiers alone than on the basis of dossiers and unstructured interviews. (Bloom and Brundage 1947, DeVaul et. al. 1957, Oskamp 1965, Milstein et. al. 1981; Hunter & Hunter 1984; Wiesner & Cronshaw 1988). If you're hiring, you're probably better off not doing interviews.
The "make candidate solve puzzles" technical interview sounds like a very poorly designed IQ test. Just replacing it with more objective standardized IQ tests would help a lot.