The last time I hired I asked two puzzle-type questions during the interviews: a math (statistics) puzzle and a programming puzzle.
The math puzzle had a definite correct answer that can be arrived at by anyone remotely qualified for the position, but which causes even statistics undergrads to scratch their heads for a minute. The real purpose of this is to see if someone is able, willing, and perhaps even eager to actually think.
The programming test does not have a single correct answer. As far as I know there's always a trade-off depending on the importance of competing priorities. I tried to lead them into discussing the design differences to handle the different scenarios. It's very interesting to observe the thought process. Some people think of one fairly workable design and then they are done. Period. Other people realize that a good design in one scenario is suboptimal in another and try a new design. Some go farther and look for an 80% solution to both.
You can call this screening to some degree. It's also about discovery, to be part of the larger picture of how this prospect would work with our team.
You have 50 red marbles, 50 blue marbles, and 2 buckets. You must put all the marbles into the buckets but you may distribute them any way you like. I randomly pick a bucket, and then randomly select a marble from that bucket. How, if possible at all, can you maximize the probability of me picking a red marble?
My intuitive answer would be to put one red marble in one bucket, and the other marbles in the other bucket, making it 100% likely to get red if you picked bucket 1 and almost 50% likely if you picked bucket 2, for a combined probability of just under 75%.
The expression I would want to maximize would be (r/(r+b))+((50-r)/((50-r)+(50-b))), where r and b are integers between 0 and 50 inclusive (I've forgotten the calculus required for this).
Most people somehow get stuck assuming they must distribute the marbles evenly (always 50 marbles in each bucket), even though it's neither stated nor implied. It's interesting to see how long it takes people to challenge their own assumptions.
I expect I would get stuck wondering if I'd be out on my ear if I answered that you could put all the marbles in one bucket, take them back out, and then put one red marble in each bucket.
The math puzzle had a definite correct answer that can be arrived at by anyone remotely qualified for the position, but which causes even statistics undergrads to scratch their heads for a minute. The real purpose of this is to see if someone is able, willing, and perhaps even eager to actually think.
The programming test does not have a single correct answer. As far as I know there's always a trade-off depending on the importance of competing priorities. I tried to lead them into discussing the design differences to handle the different scenarios. It's very interesting to observe the thought process. Some people think of one fairly workable design and then they are done. Period. Other people realize that a good design in one scenario is suboptimal in another and try a new design. Some go farther and look for an 80% solution to both.
You can call this screening to some degree. It's also about discovery, to be part of the larger picture of how this prospect would work with our team.