Interviewing candidates for jobs is the most common form of assessing competence, capability and personality. Virtually everyone interviews candidates before making a hiring decision, but are they as reliable and valid as we assume?
Reliability is a measure of consistency. An interview should measure what it claims to measure and, ideally, no matter how many times a candidates is interviewed by different recruiters the results would be similar. Validity means that the interview measures some form of job success such as competencies or performance.
In my experience we rarely test whether our interviews are reliable or valid, but make the assumption that they will predict success. Yet, we have no objective way of determining whether or not this is true. We also have no way to determine whether the interviewer was biased or used subjective facts such as age, race, or personal appearance as part of the weighting criteria.
The academic research on interviewing is consistent. The normal, unstructured interview, where recruiters ask whatever questions they feel are relevant and where the same questions may not be asked of every candidate, are not very reliable or valid. They score a validity around 0.20 in predicting success on the job which makes them just a tiny bit better than chance. A score of 1.0 would be a perfectly valid.
A structured interview, where questions are focused to one or two areas and are delivered consistently to every candidate by each recruiter, can reach validities as high as 0.50.