My normal practice was to get the interviewee to talk about what they have done in the past; bonus points for doing something beyond what was necessary just because they liked doing it. Then gently push them to describe why they did it that way, why not this way, and - with 20:20 hindsight - what they would do differently next time.
I like that question. That reminds me one case, when I was helping a friend judge a student and his semestral project (programming). It was evident it’s not his own project and we squeezed out of him, that the code was written by his brother, a professional programmer. But we gave him a chance, started pointing to random fragments of the code, and asking, what the code does. His responses were spot on. He received a passing score. With minimum number of lab points, so he would still have a hard time on the exam, but we didn’t reject him. The reasoning was simple. A complete newbie analyzing code of another person that efficiently and well, that he could easily outcompete seasoned programmers. Plus his brother passed a lot of practical knowledge, which he understood too.
I first encountered the techniques while an undergraduate. I mentioned I had selected a Doug Self audio preamp, and was pushed as to why I thought it was a good preamp. The interviewer listened to my responses and asked follow up questions. At the end I remember saying something was claimed to be some sort of advantage, but I didn't understand why it was a significant advantage. The interviewer liked that, and moved onto other topics.
From that I also learned that if someone (especially a salesman) is prepared to say "don't know" or "no" when "yes" is the obvious easy answer, then you can trust their answers to other questions.
Overall, that interview at HP was the best I experienced, hence I remembered it. I didn't join HP then for unrelated reasons, but did a decade later.
Then ask them how they would approach a simple abstract problem. Explain that you don't have any "correct" answer in mind, you just want to see how their mind works. Bonus points if they ask questions to narrow down possible answers. Example problem: "a toy manufacturer wants to add some traffic lights so a child can play with push-along cars on the carpet. What do you say?".
This is the kind of a question I’d hate being asked. One may fail simply because of things completely unrelated to their future job or even them themselves. If not that such law would be hard to formulate, I would support banning them. To me it’s playing a psychologist, while not understanding even the basic limitations of human’s mind. Getting a pretense of rationality, while any negative outcome usually reflects shortcomings of the interviewer, not the person answering.
I understand why you say that; purely "psychological" questions irritate me too.
However, for the "traffic lights controller" example I mentioned, too many people jump straight to a
single answer based on their previous experience. I want to see them generate
multiple possibilities, e.g. MCU+software, FPGA, discrete logic. Bonus points if they suggest a purely mechanical solution.
This is very different than observing, how a person deals with an on-topic question. There the interviewee is expected to deal with a problem, because this is exactly the service they are offering to the potential employer. Of course the interviewer may still fail at properly evaluate that thinking process, but this is mostly due to picking up wrong signals or assigning them wrong weights. Not because the process itself is inherently flawed.
Yes and no. The more interesting jobs don't have a single specific technology, and the technology will change over time.
For example, in an R&D consultancy potential clients often state the
solution they want developed, but it usually beneficial to understand the client's
problem. That often requires gradual elucidation of their situation, followed by considering solutions that are completely different to the client's thoughts. While the client might think a microprocessor+software is required, often a solution involving FPGA and analogue processing is optima. And sometimes their problem isn't even technical, even though they are trying to use technology to solve it!
For brevity I didn't mention that I also
- adopt the technical aspects of the questions so it is relevant to the job in hand
- tell the interviewee that I use the flow of questions to illustrate the kind of things that happen during a project, so they can appreciate how the company works. Good candidates pick up on that and then ask their own questions
With off-topic questions and, worse, completely random, abstract and irrelevant questions, this is not the case. It’s unlikely in another person’s brain such a question will map to the same notions and invoke the same references as it did in your head. Quite opposite: it may map to something completely different. From the beginning you are not even on the same page and from there it only gets worse. With on-topic questions, you more or less stay within the bounds of industry practices, norms, and knowledge. With off-topics the way a person approaches the question and how you interpret their approach is a function of personal beliefs and experience, of both of you. You can’t peek into their head, you can’t tell what influenced their approach. You’ll interpret it through lens of your own truths, explaining it within your own reality. If that wasn’t bad enough, the person may not be able to produce a reasonable answer at all. The question map map to void. No need to seek far: to me the example you gave above is producing an empty head. Totally blank. It’s not that I don’t know what to respond… I don’t even imagine what kind of a response you’d expect. I tell you that straight. If that was an interview, at this point the person could simply become overwhelmed. Or try to wiggle out of the situation, answering just anything to not fail. It will be random rambling, not representative of how the person thinks while facing an actual problem to solve.
There we agree. I especially dislike "trick" questions to which there is an answer that you either know or don't know. Classic example is "why are manhole covers round". Such questions allow an interviewer to project superiority, which does nobody any good.
As for "tell me about yourself", the best response is "which part of my CV don't you understand?"
Or "why do you want to work here?" => "that's what I'm trying to find out during the interview".
OTOH getting inside someone's head is valuable, since it can inform how they might fit in a company and job. What might be very valuable characteristics for
that job might be unhelpful for
this job.
Provided it is done sensitively in a way that allows an interviewee to express their better qualities (and doesn't belittle the interviewee) both parties can gain. Interviews really should be a two-way process.