During the past year, we’ve began to use Pair Programming (PP) as part of our interview process. Having interviewed many candidates using this approach, I have begun to prefer this method to the traditional approach of asking and answering (AA) because:
Traditional interviewing skill emphasize on asking behavioral questions. Since it is so, why not take a step further to see the candidate’s behavior (the idea that the beneficial elements of traditional practices are taken to “extreme” levels). One Chinese politician said “practicing is the only way to verify a theory”, but no one ever said “asking practical questions is one way to verify a theory”. On the other hand, many times people don’t know exactly what they did under those circumstances (have you ever tried to count how many candies and snacks a fat fellow, who claimed he was on a diet, has eaten during one day?), you cannot tell he was lying, because he really believed he had eaten nothing at all.
During AA, an interviewee tries to figure out why he is asked this question and how to give a response to satisfy the interviewer. That’s the only goal for him to accomplish. Interviewer and interviewee are just like two hostile hedgehogs rolled up into two spiky balls. How can you imagine your opponent will tell you his weaknesses? In PP the most important goal is changed to solve a problem and deliver something, this time you two become more like two hedgehogs forming one spiky ball with thorns pointing towards outside against a common enemy.
Problems appear continuously and unpredictably in PP. No one can know what problems they will meet, particularly considering most problems are caused by one’s self. Sometimes you know there is a problem but you don’t know what it really is. You need to find it out by yourself. Sometimes there are several problems simultaneously jumping out for you to handle. Compare this to only facing one question at a time when using AA.
Another benefit is that the interviewer doesn’t have to think over which question I should ask next, so interviewer can focus on observing the interviewee during the whole process.
To help interviewees respond to the AA process, there are a lot of “tips” you can find in books and forums to round the questions you don’t know how to answer. And if a interviewer wants to ensure the interviewee’s thorough knowledge of one specific topic, he may take the risk of offending the interviewee. But in PP all the questions the interviewee wants to round will finally be the debt causing him not able to accomplish the task.
PP can make good candidates outstanding, unqualified ones awful, average ones personalized. Following chart is some statistics from my experience so far this year. Among 38 candidates I can easily tell 9 best candidates and 9 worst candidates. I believe the results’ ratio coincided with our common sense. And I can easily remember the guy whom I have pair programmed with several months ago when I go through the feedback I have written. I was very confident to make these evaluations, which juxtaposes greatly with having to write feedback after an AA-style interview when I always suspected whether my assessment is too subjective or not.
The most important thing is that PP makes me excited and full of anticipation. I am always curious about what how the interviewee will perceive and handle the problem that is presented; and after all making a difficult job enjoyable and interesting is also one of the goals are looking to achieve.
