Sean Landis, author of Agile Hiring, discusses how to ask interview questions that "break through the castle walls" so you can get useful information from candidates.
Google “most popular interview questions,” and this one will pop up on most of the lists. If you Google “worst interview questions,” this question is on those lists too. Interviewees ought to study the first set of search results. Interviewers ought to understand why this question appears in the second set, and what they should do about it.
One of the “worst interview questions” articles, The Ten Worst Job Interview Questions Ever, had an interesting way of labeling the negative characteristics of interview questions. I borrowed that to create my own labeling. Here is my summary of why this is such a bad interview question:
Clearly, this is a bad question, but certainly the people who ask it do so sincerely wanting useful information. What information are they hoping this question will reveal? Here are a few things one might be looking for with this question:
What better questions can get the desired information? By taking the negatives about this question and flipping them to the positive, you can create some decent guidelines for evaluating the quality of a question. An interview question should be:
Behavioral questions are trying to get at how the candidate will respond in your work environment. These questions usually take one of two forms: “what you did,” and “what you would do.” “What you did” questions are usually better because they are rooted in reality. The challenge is to locate the candidate’s historic examples that address the questions you want answered. The problem with “what you would do” questions is that they are hypothetical. A moderately clever candidate will tell you what you want to hear, but there is only a weak relationship to how she would act.
“What you did” questions can be tightly or loosely coupled to job history. A tightly coupled line of questioning would start with the discovery of a situation in the candidate’s past, such as being a team lead. You then ask specific questions about that situation: “What was the size and make up of the team?” “What were your primary responsibilities?” “What percentage of your time would you say you designed or coded?” “What did you like and dislike most about that role?” “Is team leadership and management something you want to continue to pursue?” The last question may be the one you really want answered, but the lead-in questions reveal much useful information as well, possibly supporting or conflicting with the final answer. Notice how the first questions were more factual, while the later ones were more personal. The flow builds rapport, creates some constraints for the later questions, and warms the candidate into being honest and accurate.
A loosely coupled line of questioning would be quite different. “Do you want to continue to move toward leadership or remain technical?” (This is assuming you observed that the candidate has leadership experience). “Give me an example where you mentored an individual. Tell me about a time you tried to convince your team to change its collective mind.” These questions get directly to what you think you want to hear about, but allow the candidate more control of what to tell you. You may hear what you want to hear before you hear the truth.
Examples of “What you would do” questions go like this: “If you had a team member that refused to pair program, and your culture encouraged pair programming, how would you handle it?” This is a hypothetical question. “In your ideal leadership role, what percentage of time would you want to spend on leadership and management versus technical tasks?” This question is encroaching on useless and lose-lose.
The tightly coupled “what you did,” questions usually pass the good question tests. They are win-win because you get what you want and they make it easy for the candidate to speak the truth, even if the outcome was not ideal. People tend to be more candid about their mistakes when discussing specifics. You can make it even easier, and get additional useful information, by asking the candidate what she learned from a failure she described.
The loosely coupled “what you did” questions give permission for embellishment and cherry picking. You still may get good answers but, because you’ve lost context, the candidate has more room to tell you what you want to hear.
The “what would you do” questions waive context and win-win, and may encroach on being hackneyed and useless too. I like to reserve these questions for situations where I want to learn something, but the other lines of questioning have failed to reveal it. They may be useful when the candidate has little or no experience in an area, but you still wish to explore potential.
It is good practice, when resorting to “what would you do” questions to see if you can tie it back to reality with a loosely coupled “what you did” question like, “That sounds like a good idea. Can you recall any time you were in a situation like that, and tell me how it played out?” Once back in reality, you have more confidence about the answers and probably can follow with some tightly coupled “what you did” questions.
Questioning a candidate sometimes feels like laying siege to a castle full of treasure. Weaker questions, such as “what would you do” questions, may break through the outer defenses, but it requires a strong line of questioning to break into to the inner workings of the castle. Once you are in, you have more freedom of movement. Many candidates have walls that the interviewer must break through, but once in, the questioning often gets easier.
Focus on questions based in reality that do not lend themselves to canned responses. Before you ask a question, think about whether likely answers will be useful for decision-making. Good questions are easy for the candidate to answer honestly and accurately. Be sure there is sufficient context for the candidate to answer you in a useful way; focusing on the candidate’s past actions often best does this.
The really good candidates will be impressed by quality questions because they reflect intelligence and competence. Poor questions are a turn off. Experienced, quality individuals want to surround themselves by other quality people and will be attracted by your good questions. Work hard to develop the ability to ask good questions and avoid the bad ones.
Do you think these guidelines will help you come up with good questions for your next interview? Which of these guidelines do you like best? Share your opinion in the discussion form for this article, Where Do You Want to Be in Five Years?
If you are convinced that hiring is a mission-critical activity, and you are committed to becoming great at hiring, I suggest you read my book, Agile Hiring. It is strikingly different from what has been written before on hiring. Much of the hiring wisdom is still there, but many of the ideas are new. These new ideas have been developed and tested in companies that are committed to hiring great software professionals. They work. I describe the key technical parts of hiring: resume reviews, phone interviews, on-site interviews, and making the offer.
Sean Landis is author of Agile Hiring, which
is available at:|
The Ten Worst Job Interview Questions Ever is at
Worst interview questions to ask job candidates is at
Sean Landis, a software architect with over twenty years of experience hiring software professionals, has hired in companies with less than ten developers and ones with thousands. He successfully retooled hiring practices at three companies, leading to significant improvements in the quality and quantity of new hires. Sean is a practicing software professional who today practices agile development and has innovatively applied agile principles to hiring.