Friday, June 12, 2009
Fast Company disses interviews
Those of you who know about research in personnel selection know that while interviews have been shown to be predictive of job success, several other types of selection mechanisms often out-perform them. Cognitive ability is often mentioned as the holy grail of predictors, but in terms of overall utility and defensibility, I recommend work sample exams. So do the authors of a recent article in Fast Company.
As the authors (who also penned Made to Stick) point out, interviewers are often snowed by candidate interview skills. Often only when you make them demonstrate their skills do their true strengths and weaknesses reveal themselves. (Of course if you're going to interview--and almost everyone does--make sure it's structured)
A couple strengths the authors leave out: work sample (sometimes called "performance") tests are easier to defend legally, since you're measuring an observable KSA rather than a construct like intelligence, and they give candidates a more realistic preview of the job. Heck, after doing a work sample a candidate may decide the job's not for him/her. Finally, they tend to be well received by candidates, more so than many other types of assessment.
This is my favorite quote:
"...figure out whether candidates can do the job. Research has consistently shown that one of the best predictors of job performance is a work sample. If you're hiring a graphic designer, get them to design something. If you're hiring a salesperson, ask them to sell you something. If you're hiring a chief executive, ask them to say nothing -- but reassuringly."