There are many human resources professionals who feel that job interviews are not very good predictors of job success and are at times not worth the time. However, most employers cling to the process, feeling that they want to get a personal feel for candidates.
Many managers use the job interview as an opportunity for a little social contact with the candidate, sometimes to get a sense of whether the candidate is likeable or if they will be pleasant to be around. Research has found that the use of unstructured interviews in hiring is no better than simply picking from among pre-qualified applicants at random.
In fact, the interview has the impact of a selection test, perhaps one of several instruments to be used in the selection process. This implies that the job interview should be reproducible. Interviews should be conducted in the same way for all candidates. The same queries should be made and responses should be catalogued so that the relationship between interview responses and successful performance can be tested. The usefulness of the interview format, retention and successful performance should be regularly reviewed.
In some cases, the interview can be regarded as a kind of job sample test, where some of the skills needed in the job itself are actually directly assessed and scored. For example, where spontaneous speaking skills are important for the job, a specific speaking sample can be elicited during the interview and performance can be directly observed.
Interview coaching is a factor that reduces the effectiveness of interviews as selection tools. Many people in the job market seek and find special training in how to conduct themselves in the interview. Levels of interview training complicate any objective judgment of a candidate whose behavior may well reflect learned interview behavior, not real personal qualities.
In most cases, successful performance on the job is not predicted by spontaneous speaking skills or sociability. For many positions, sociability could be a deterrent to successful performance. Candidates with strong social skills can often be frustrated in positions demanding solitary, concentrated effort. Creativity and ambition, so often viewed as socially desirable traits, can often make continuous reliable performance on repetitive, routine yet vital job functions painfully difficult.
Often the best candidate will be one who admits to a way of working or a preference that is not widely socially desirable but which fits perfectly into the demands of the job.
Study after study suggests that the best predictor of job performance is the observation of actual performance. Sometimes samples of actual job tasks can be incorporated in the interview and performance can be directly observed and graded. In many cases, the best selector of success would be an internship or an on-the-job trial.
Regard the interview as a selection tool. Interviews are often best conducted as part of a series of job sample tests and other selection assessment procedures. The interview should be standardized and best conducted by personnel trained in human resources interview techniques.
The interview should be designed around a functional job description. The actual behavior demanded by the job in question should be detailed and the interview should assess likely responses to those kinds of functions. General questions soliciting less relevant responses can be minimized.
The best functional job descriptions go beyond description of the formal duties of the job, but include more detailed factors such as the kinds of supervision, the kinds of peer relations, even the kind of physical location of the job, the level of contact with peers, and what the job actually entails, beyond its formal function.
Fantasy job samples often provide important insight into the way a candidate would handle the demands of a position. The interviewer can describe realistic job related problems and ask how the candidate would handle them. Responses can be viewed in terms of how familiar the candidate is with the problems described and how credible the response would be.