Most Personality Tests are Shams
June 12th, 2007 . by Brad SmartMOST PERSONALITY TESTS ARE SHAMS!
We get lots of questions about testing, because clients would love to have some sort of inexpensive screening device that cuts the candidate pool down to just a few to call and then to interview. Unfortunately, the fact is that most personality tests are validated in a devious way that makes the tests look useful, when they are not. I’ll explain why:
I have on my desk the validation manual for one of the many personality tests used to screen candidates for sales positions. It’s a thick manual. When the test is “tested,” it’s administered to present employees, with the promise that their score will not be known to their employer. OK, so the current employees have no motivation to fake answers; they are honest, and so sales people might admit they are not highly motivated, that they call on present customers when they should call prospects, etc. Then the test purveyors match the results against performance and, VOILÁ, the people who sell the most are the ones with the highest test scores. The test appears valid! But it’s not.
The only honest way to validate a test is to administer it to candidates for selection; don’t even score the test for a year, then pull out the test, score it, and then see how the scores correlate with performance. When real selection candidates take the test, they fake it – sales candidates say they are highly motivated, that they of course call on prospects, etc.
Years ago I validated several personality tests that were being used by one of the largest retailers. We had 2,000 subjects. The company thought the tests were highly valid, but when I validated them properly (with hiring candidates, not present employees, taking the tests) the value of every one of the tests was ZERO! And unbelievably, the sham test purveyors are telling the truth when they say their methods are approved by the American Psychological Association.
Recommendation: get the validation manuals and don’t be intimidated by all the statistical mumbo jumbo. Just look for how the tests were validated. If the manual does not say that candidates for hire knew they were not hired yet when they took the test, you have a problem. Call the test purveyor and challenge the validation results. Or, administer the test to 200 candidates the way I described, and see for yourself it the test really works. Don’t hold your breath!
I am surprised at the over-generalization of these comments. I am a top-grading company, and we use profile testing very successfully to make sure people are in their area of strength. It is easy for a manager to see whether someone actually lines up with that score, but it is not a right/wrong test, it is a profile that actually tells you where that employee will work best with their skillsets. You still have to topgrade them based on their performance and belief in company vision, etc.
Perhaps we are talking apples/oranges, but it strikes me odd that you would state “most” of these are shams.
Thanks for your comment! I’ll grant you that many companies use personality profiles, such as Myers-Briggs Type Indicator, to get a feel for personality or decision making styles, and they don’t actually have cutoff scores — they just get some additional information that they incorporate in phone screen interviews and face-to-face interviews. Great! No problem. BUT … when companies eliminate candidates because scores were not high enough, THAT’s where companies are getting taken to the cleaners.
There isn’t one personality test I have seen that has honestly validated the test and shown improved hiring. Ie, the “predictive validity” of the tests, as described in the test manuals, is totally bogus, because the the “subjects” are present employees (who tell the truth) rather than selection candidates (who … umm .. “put their best foot forward”).
Hey, I’m receptive — if you use a test to screen out candidates and you think it really is valid as a predictor of job success, do what dozens of companies have done: send me the validation manual, and I’ll review it at no cost. In 30 years I’ve reviewed dozens of manuals, discovered the sham, the company contacted the test purveyor to ask if validation was on true selection candidates or present employees, and NOT ONE test purveyer has come back with a study done properly. And back in my youth, when I did a lot of test validation studies, I did not find even one personality test that was valid with true selection candidates.
Put more positively, if you or anyone can show a personality test valid in selection, topgrading professionals would love to have it, we’ll suggest that clients use it — hey, it would save all of us time!
… but I’m not holding my breath!