I am not a fan of the usefulness of interviews. The research on job type interviews (unstructured) suggests they do not work as intended. The research on structured interviews is modestly positive if very well structured. This was struck home to me as I sat through a semi-structured selection process for a networking group. All 8 of us in the group were fascinating (and vouched for by someone in the network), but in practise the group interview only assessed how well we could present ourselves in a narrow, artificial way with bias to those good at presenting, and with some extravert factors. (I say that as someone who does have presenting skills)
I believe it would have been better (or at least equal and less hassle) to use a random process. This was further reinforced in the pub later where the natural flow of conversation between the 8 of us was an order of magnitude better.
Still, the unstructured interview is not going away any time soon, but – gentle reader – if you take one message from this, is to realise interviews very limited effectiveness in judging any thing to do with a job. See below for the academic study on it, plus NY times article:
(Note, from my reading one of the more effective ways of testing someone in a job, is to give them that job for a trial period – although this is impractical in many instances, it’s one of the driver for the value of internships and one of the problems of how to give internships)