Ist est nicht ein Wonderlic?

Wednesday, April 29, 2009

Test-taking is a bad way to measure if people are bad, but especially this bad test.

The Wonderlic Test is a 50-question test to be completed in a time of 12 minutes which is given to prospective candidates as a sort of occupational IQ test. The questions themselves span verbal, spatial and mathematical skills and increase in difficulty as the test progresses.

The scoring system is based entirely on how many correct answers you give, not how difficult the questions are, so doing them in order is advisable, and skipping over any that give you even the slightest pause increases your chances of finding one suitable for you. If you finish with this method before the 12 minutes are up, you can go back to the ones you've skipped.

A score of 21 correct answers is supposed to represent the average, and therefore an IQ of 100, but each position the test is used for is subject to a different baseline. For example, while a security guard might fall within his peers with a score of 17, a chemist can be expected to rope in a much more respectable 31 for herself.

The most notable and intriguing aspect of the Wonderlic Test is that is given as part of the NFL's drafting process. More...As a disclaimer, I know nothing about football, but evidently the scores required to be acceptable are, perhaps understandably, lower than average, which raises the question of why the test is administered in the first place. Examples: An Offensive Tackle is the brains of the operation, statistically hitting a 26 on the test. More in line with the Everyman, the Tight End typically scores a 22. No doubt harassed for this within the football community (though not by nerdy bloggers), the bar is set at 16 for a Halfback.

Keep in mind that a score of 10 is considered literate, though how they'd read any of the questions, all of which have some written aspect, I don't know. Now, the above scores for football players are potentially what you might expect, but here at Weaselsnake, we like to challenge our preconceived notions. Pat McInally, a graduate of Harvard University who punted and was a Wide Reciever (avg score: 17) for the Cincinnati Bengals, received a confirmed perfect score. Ryan Fitzpatrick, Quarterback (avg score: 24) for the Buffalo Bills and later the St. Louis Rams (and who, in a terrible case of research-related perspective, is younger than I am) was rumored to have gotten a perfect score on the Wonderlic, but he asserts he failed to fill in at least one question, and therefore could not have done so. Don't worry, Ryan. You get a perfect score for modesty.

Reducing someone's intelligence to a numerical value is intrinsically facile. Human beings have a capacity for understanding that exceeds the bounds any intelligence test has. Even using the test scores as a metric for choosing between two equally qualified candidates fails to take into account the personality and other unknowable factors that comprise a person. They could just be a really lousy test-taker. I've seen brilliant people - titans of understanding - bomb tests that were leagues beneath them for panicking during the test. Some people perform better under pressure than others. 12 minutes isn't much for 50 questions - that's 1 question every 15 seconds, doable for easy questions, not so much for the toughies at the end. The argument that you want someone who performs better under pressure could be made, but not without examining the differences in how the competing candidates perform under normal circumstances.

In addition, the Wonderlic gives no venue for discussion or potential problems with the questions themselves. An awkwardly-worded question could have several correct answers - leaving the candidate to choose which she thinks the test is going for. Once you've done that, the game is over - a tester is no longer thinking about the correct answers to questions as written but how to outmaneuver the test.

Consider this question: Which of the following months have the closest number of hours of daylight and nighttime as September? A) January B) March C)May D)November?" The wording allows two possible interpretations of the question. The answer could be November, as it has the closest number of daylight hours as September has daylight hours. The answer could be May, as it has the closest number of daylight hours as September has nighttime hours. However, on the test, there's only one correct answer. (In case you're wondering, the test wanted "May".)

Test-taking is by its very nature simply a model for getting at how a person thinks, or how much a person knows. It's a necessary evil in schools, in order to measure progression and comprehension (though not necessarily aptitude), but is also often used to judge the competence of a teacher, the sufficiency of a school's resources and the performance of a principal. And here's where I'd take umbrage. Scored tests - tests which have correct answers - are not a fair means of simplifying a candidate for a job. The stakes are too high for the candidate for it to be entrusted to a 12-minute task done alone, and the potential for abuse of using the test as a cover for hiring discrimination too great. A candidate doesn't get her answers back, has no idea what she scored, and can therefore be denied for any reason based on her test (The legality of this is dubious at best, but the point is sound; the test puts the candidate in an unfair position.)

Many companies give tests of another nature, those which focus not on answers but on reasoning - how the person arrived at their answer, or how long it took them to be confident their answer was correct. It's common in the tech industry, which uses brain teasers as a way to gauge your problem-solving skills and capacity to think outside the box.

A famous one involves a closed door and three light switches on the wall beside it. You are asked to determine which switch lights the lamp in the room beyond the door, and given some arbitrary limitation, like being able to open the door only once (and then close it immediately afterwards). The questioner being there allows for necessary information to be given upon asking ("Will the door stay open?" "Can I go in the room?"), and so is helpful to the test-taker, but is moreso valuable for the questioner who gets a chance to see how the wheels are spinning. Does the person draw a diagram, does she do mental arithmetic, what sorts of questions does she ask?

This way isn't perfect either, but at least you can determine who is cocky, who likes a challenge, and who might stride confidently over the line of "thinking outside the box", landing in "potential mental patient" territory ("Rip the switches from the wall and follow the wires!".) It might tell you who you might like to work with, based on how creative or flippant their answers are. It might tell you who is not afraid of admitting defeat, and who will doggedly attempt to figure it out.

What it won't tell you, though, is who can catch a 40-yard pass.


Post a Comment