Why is it that we sometimes get appointments so spectacularly wrong? Clients often tell me, when we are discussing how to terminate an employee's contract because of poor performance, that the employee interviewed so well. It's been the same story for 30 years, in my experience, and the answer is the same: too many people rely on gut instinct at interview - but our entrails are not reliable thinkers.
We all like to pride ourselves on being good judges of character. I'm afraid we are wrong about that too, which is why skilled salespeople can fool most of us most of the time. They use techniques that can be learned, and so do people who come over well in interviews.
Recent research, quoted in The Guardian, by Jason Dana of Yale University looked at unstructured interviews. Participants were asked to predict the academic performance of college students. Some participants were given data on the students' past achievements; others were given the same data, plus the opportunity to interview them.
Consistently, interviewing led to less accurate predictions. The trouble is, we like to see people face to face and this can mean we get too much information. As Guardian columnist Oliver Burkeman commented: "When bombarded by data, we seek refuge in 'sense-making', clinging to stories that make things clear."
Research also shows we are influenced by the halo effect if a person has done something great or shines at one type of work. This glow spreads out over other areas in which that person is less competent, but we remember the halo. Then the similarity effect works like this - "I like that person's jacket, therefore they are like me - we will get on and they will work well here, like me."
Measure and test
So what should we do to mitigate these forces? Start by writing accurate job descriptions and realistic person specifications that are measurable and testable. Use a structured application form, not a CV. Make a shortlist on the basis of the person-spec criteria and structure your interview likewise. Add a relevant work skills test or two on the day, and actually look at the results when making a decision: know what a good answer looks like.
Have a panel of interviewers scoring candidates as they go along - this minimises the risk of stereotyping. Use a short scoring system, such as "fully met, partly met or not met", rather than scoring out of 10, which allows for too much variation.
Be prepared to challenge each other if you stray into story telling, such as "they wouldn't fit in" or "I have a feeling about them".
Use all these techniques and you stand a much better chance of getting someone who can actually do the job, rather than the best blagger.
Gill Taylor is a sector HR consultant