The Selection Interview: It Doesn’t Work, but it Won’t Go Away

I have never been a fan of interviews. Having been both interviewee and interviewer too many times to count, I can honestly say that most of them fall into one of two categories: (1) chats and conversations or (2) interviewer sales jobs. The chats and conversation style of interview is the most common, and consists of having the candidate talk about themselves and their experiences in a rather unguided way. The interviewer may toss in a few questions or ask for a few clarifications, but generally the entire process is given over to free-flow discussion or monologue. The sales job is also common and is the type of interview hiring managers most frequently use. In these, the interviewer does most of the talking, selling the position and explaining how wonderful the company is. In most cases, the interviewer learns very little about the candidate. The interview as we know it is a product of the 20th century. It became very popular in the first decade of the 20th century as a “scientific” way to select people. It was commonly used to select and promote employees in the booming industries of the East and Midwest, yet a study done as early as 1915 showed that interviewers could not consistently select the best people. A paper written in the Journal of Management History (Vol. l6, No. 3, 2000, pp. 113-126) traces the evolution of the selection interview and underlines over and over how ineffective it is as a predictor of success. The article traces how the interview process was enhanced over the decades leading up to World War II. The U.S. Civil Service Commission, after in-depth analysis, concluded that interviews should contain “genuine problem situations” to predict how applicants would react in real work situations. This led to the setting up of training for interviewers and to the establishment of standards and evaluation methods. Practical tests were part of most interviews in the Thirties and Forties, and by the time World War II started the interview was well established, but still not very predictive of success. <*SPONSORMESSAGE*> At the start of World War II, two different tests were conducted by the army air force to predict how well candidates for flight school would do. Each candidate was interviewed following rigorous standards and then a prediction was made on how well they would do. When the predictions were compared to actual performance years later, the prediction rates were no better than chance. None of the interviewing and screening techniques they used could effectively or reliably predict success. After the war, psychologists showed that group interviews and structured interviews did seem to produce slightly better results, but it would probably require that both types be used to fully assess a candidate. The process is time consuming, requires interviewer training and requires a clear focus on the job requirements. So, if the interview is so non-predictive, why is it so popular? Given the fact that probably somewhere around 75% of all interviews fall into one of the two categories I mention above, and given the fact that there is almost no follow-up done to see if job performance matched the predictions from the interviews, why are they still the most commonly used way to choose candidates for jobs? The first reason is that interviews are easy to do and can be done by anyone. Training is frequently given, but most of that training is focused on the legal aspects of questioning, not on raising predictive capability. The second reason is that we all seek some symbol of power or status and being asked to interview someone is one of these symbols. The idea that your opinion about someone counts is a powerful incentive to keep on doing them. I have never seen a manager who doesn’t want to interview a candidate, even if it is a very cursory process. The third reason is that they are inexpensive. Even the smallest firm can afford to interview. Purchasing tests or conducting on-the-job practice events, on the other hand, are expensive and time consuming. And a final reason is that the interview fulfills one of social needs to meet people face-to-face, to touch and talk to people and find common connections. There is nothing at all wrong with this. What I don’t like about the selection interview process is that it is not predictive of much of anything. Even the best and most rigorously performed structured interviews are not much more predictive than chance. There has actually been very little research done since World War II on the selection interview or on developing new techniques that may be more effective. Some of the recent work in profiling candidates and understanding the specific competencies required for a job are a beginning. Many organizations such as PDI (Personnel Development International) and DDI (Development Dimensions International) have begun to broaden our understanding of this increasingly important area. Skills testing has also greatly expanded and is widely available on the Internet from providers such as BrainBench. Profilers are growing in popularity as well, as characterized by FutureStep. While the selection interview is most likely always going to be here, we need to face the fact that it not a highly predictive tool of job performance or success. In order to predict success, we are going to have to find more powerful and simple tools that are easy to administer, cheap to use, and do not intrude on the time or privacy of the shrinking pool of potential talent. Let’s use the interview as our social introduction and as a way of assimilating the candidate into our organizations, but start looking for better ways to select the best person for the job.