"Rather than using traditional question-and-answer interviews to evaluate job candidates, employers should observe candidates as they perform some of the job's actual tasks."

It is often argued that question-and-answer interviews are ineffective for candidates evaluation; instead, it is offered to observe how candidates perform some actual professional tasks. I can only partially agree with this proposal, because "task observation" approach has both pros and cons. Let me briefly analyze the key points.

What is definitely positive about testing the candidate with actual tasks is that this allows to identify the candidate's actual professional level, not the one he claims during the interview. It is hard to detect whether his statements about past experience are true; a candidate can easily exaggerate the complexity of tasks he had apparently resolved. When placed against real-life challenging task, the candidate will either handle it or fail - there is no option to cheat or boast.

Another pro-factor of allowing candidate to perform actual tasks is ability to observe his approach to the problem he never met before. A skilled professional not only has extensive background in his area, - he is also aware of how to handle a challenge he never met before. This ability is extremely hard to gauge during a question-and-answer session. However, this "task-testing" approach has some serious limitations that prevent me from considering it best practice for any interview. Firstly, the interview by a recruiter is an excellent opportunity to discover personal traits of the candidate, his morale and attitude to his future job position.

Being a decent psychologist would allow recruiter to obtain information that was not even expressly spoken by the candidate; this is of course impossible during task-performing session, and therefore one should not hurry to abandon the practice of interviews. Another problem with observing job simulations is that such observation is time-consuming and requires attention of a skilled professional in a similar area. A human resource specialist may be unable to properly evaluate how well the candidate handled the task, thus rendering the test totally useless. Involving an expert for every session would alleviate this problem, but experts' time is too costly for the company to waste it during multiple interviews.

Overall, I would advise to combine the two practices to get the most information about the candidate. The "basic level" of recruiting process could be an interview with HR specialist; once the candidate proves to be adequate and worth further consideration, the company can invest its experts' time into observation of candidate's test assignment session.