Anthony Sibillin - BRW - April 2011
Hiring the wrong person for a job is a costly mistake. So employers will do anything to avoid making it – including, increasingly, subjecting candidates to psychological or psychometric tests.
As paper-and-pencil make way for screen-and-mouse, the cost of psychometric testing is falling and demand from employers is rising.
While reliable data on the percentage using it for selection is hard to find, experts reckon it may be as high as one-third of small and medium employers, rising to more than two out of three big employers.
But does psychometric testing work? Naturally those with a financial interest in the answer are keen to get across the message that it does.
Most commentators “now accept that the use of personality profiling and other similar measures in pre-employment screening significantly improve the predictivity of the selection process”, claims the website of one Queensland-based testing company, whose clients it says include Domino’s Pizza Enterprises, Village Roadshow and Westpac Banking Corp.
However, those that don’t have a financial interest in the answer say the jury is still out on pyschometric testing. “It is a big decision to take somebody on, so firms should invest in a high-quality selection process,” Nick Wailes, an associate professor at the Sydney University Business School, says.
“[Getting it wrong] has significant costs and can affect organisational performance.” Although psychometric testing may be part of that process, it should be only a small part, he says.
“It is not a silver bullet.”
Should a firm decide to use psychometric testing it should know that there are three different tests: those that test a candidate’s intelligence; others that assess their personality; and, finally, those designed to test the candidate’s behaviour in a variety of situations similar to those they are likely to confront on the job.
Of these, doing well on a general intelligence test is the clearest sign of a good hire, according to a survey of 85 years of research findings by American academics Frank Schmidt and John Hunter.
Even then, the survey concedes, the sign is only really clear for an entry-level job. For one needing experience, subjecting candidates to a hands-on simulation of part or all of the job will be more telling, according to their research.
Personality tests typically test for five traits: conscientiousness, emotional stability, extraversion, openness and agreeableness. Most studies find that only two of these traits – conscientiousness and emotional stability – are “positively correlated with job performance in virtually all jobs”, report Schmidt and Hunter.
So what these tests boil down to is the mantra “hire smart people who are aren’t lazy or crazy”: hardly a revelation and some might think thin justification for spending thousands of dollars on psychometric testing.
The justification for behavioural tests is even thinner, Wailes says. Unlike personality tests, which assume an individual’s traits are fixed, the newest piece of psychometric artillery makes the more realistic assumption these will vary from situation to situation.
For instance, a manager may be agreeable in formal meetings but a tyrant one on one. In theory, behavioural tests, when tailored to the situations candidates are likely to find in the job being filled, should help employers weed out unsuitable candidates.
In practice, Wailes notes, “these tests are quite new, so there is not a huge amount of evidence around them”.
If there is so little evidence for psychometric testing, why is it so popular? Blame the human weakness for appearance over substance, Chris Jackson, a business psychology professor at the University of NSW Australian School of Business, says. “It is no different from buying a car,” he suggests. “Like a typical car buyer who doesn’t know that much about cars, an HR [human resources] director will go for what is shiny and glossy.”
Even psychometric testing’s defenders admit many employers choose the wrong tests or misinterpret results. But they insist testing is a useful extension to recruiting by resume-followed-by-interview.
“It enables the organisation to get behind candidate ‘spin’ that can appear with well-rehearsed applicants,” says Chris St Clair, managing partner of PsychWorks, which has run psychometric tests on behalf of BHP Billiton, Sara Lee and Luxottica, among others.
Griffith Business School senior lecturer Ashlea Troth says employers should make sure “the capabilities assessed by psychological testing are actually relevant to the job. “This is more likely to occur if a thorough job analysis has been conducted beforehand – that is, information about the knowledge, skills and abilities needed to perform the job.” This should then drive what needs to be assessed and what type of test to use,” Troth says.
Filling graduate and other jobs with many applicants is made easier by psychometric testing, University of NSW psychology lecturer Joanne Earl notes. “If you advertise a vacancy and are inundated with applicants, your options are to screen these using an automated system, a team of people or to use psychometric screening tests,” she says. “Generally the cut-off scores for these screening measures are set quite low and then combined with one or two other pieces of information to decide who gets considered further for a telephone interview or other screening interview. “I think the more important question is: ‘Why wouldn’t you assess [using psychometric testing] if it provides you with additional information that cannot be obtained elsewhere?’”
If the additional information increases the probability of hiring the right people, the answer is obvious. But this is an assumption not a fact.
More fatal to that assumption than the lack of independent supporting evidence is the possibility that candidates are “gaming” tests – that is, answering questions to produce the psychological profile they know a potential employer is looking for.
“Vendors say you can’t get better at them,” Sydney University’s Wailes says. “But there is some evidence that the more you do them, you tend to perform better because you know what you are looking for.”
Another worry is that psychometric testing highlights the mediocre at the expense of the exceptional.
Like all statistics, it shoehorns test-takers into a limited number of categories. If a person gives a particular set of answers it means, on average, they are suited to a particular set of jobs.
Or does it? Presumably, there are introverted property agents and extroverted engineers who are nonetheless good at what they do.
Such outliers are likely to be culled by the increasing number of employers in thrall to a practice that has no clothes.
