Is a personality test the right selection mechanism for your needs? In trying to answer that question, an important consideration is: Does the job allow for the expression of personality facets?
There's a concept in psychology called situation strength. It refers to how the "strength" of a situation impacts the display of personality via behavior, and it's something important to remember when using measures of personality to predict job performance.
A situation's "strength" refers to the environment under which personality aspects are displayed; think of it like a rulebook. If Job A is described perfectly in exacting detail with very little room for deviation ("Place container A over part B..."), does one's personality really matter when it comes to successfully performing the job? Or thought of another way, if the rulebook repeatedly emphasizes using an aspect of personality (e.g., extraversion), how will people without the ability to express that behavior consistently fare?
If you're a software programmer, it probably matters little in the grand scheme of things how extraverted you are; analytical ability is likely much more important. Similarly, if you work in a call center and follow a script, differences in openness to experience probably don't mean much; extraversion is likely more important.
But what about something like conscientiousness--could the strength of a situation impact the relationship between this aspect of personality and job behavior? That's the question that Meyer and colleagues set out to answer, and an answer is published in a recent issue of the Journal of Organizational Behavior.
The way they went about trying to solve this puzzle was to conduct a meta-analysis at the occupation (rather than job) level. What did they find? A few things:
1) Uncorrected correlations between conscientiousness and performance varied widely between .06 and .23.
2) Correlations appear slightly stronger when using overall performance as the criterion rather than task performance (not surprising given previous research).
3) Stronger correlations were found in "weak" occupations.
What does this mean? Well, for one it reinforces the fact that the answer to "Do personality tests work?" varies greatly depending on what the job is and how you measure performance. But perhaps more interestingly, it suggests that at an occupation level we can expect that for jobs that come with built in rules regarding behavior ("strong" occupations), measures of personality aspects such as conscientiousness may not predict performance as well as for jobs with more flexibility ("weak" occupations).
So the next time you're thinking about using a personality inventory for selection purposes, consider: To what extent will incumbents be allowed to express their personality?