Yet in this study the Wonderlic is keyed to position, not performance in that position.
We know that people largely stay within their positions, and we know that metrics like the Wonderlic and the drafting Combine strongly influence both draft decisions, and fielding decisions. We know that such metrics are designed to correspond with existing ideas: in this case, that quarterbacks were smart and tailbacks not so much.
It is plausible that this correlation represents bias in the decisions of the coaches, not real ability.
In fact, that's what suggested by Malcolm Gladwell. He points out that if you sort by Wonderlic scores, those who come out on top haven't really panned out as great quarterback. If we then look at some of the best quarterbacks of all time, you find that, on the Wonderlic, players are all over the map.
I didn't like Gladwell's talk because he ignored baseball: teams like Oakland and Boston have started to understand what metrics do work in assessing player performance (Michael Lewis wrote about this in detail in "Moneyball").
By ignoring something because it doesn't fit in his neat theory, Gladwell loses credibility on this topic.
You might like to know that Malcolm was asked in an interview elsewhere about Moneyball, and he mentions it as one of his favorite books, and indeed he calls it one of the most important works of nonfiction of the past decade. Apparently he's found a way to resolve these difficulties, but not in a way that found it into a 20 minute talk -- I guess it has to wait for the book.
Yeah I have noticed that. Reading his stuff reminds me of junk food, kind of nice at the time but not long before either I want more, or regret having read it. I think its the journalist style writing that I find unsatisfying (his books are one of the few I end up leaving lying around, unfinished and never seem to want to pick up again).
> It is plausible that this correlation represents bias in the decisions of the coaches, not real ability.
Gladwell's point is that there is a matching problem, i.e., that quarterbacks with high Wonderlic scores are drafted ahead of those with low Wonderlic scores. In fact, one of the players on his high scores list (Tony Romo) went undrafted, while some of those on his low scores listed (McNabb, Vince Young, and Marino) were drafted in the first round.
His examples don't support his matching problem thesis. Further, people that follow the NFL know that Wonderlics aren't a huge factor in selecting players.
> Yet in this study the Wonderlic is keyed to position, not performance in that position.
It's automatically keyed to performance in the position since it's based on NFL players, who are the people at the 99.99th or 99.999th percentile at that position.
I think he means that of the small proportion of people who are good enough to make the NFL the coaches will have picked those that conformed to their biases.
I.E. QB A and QB B are both on the edge of joining the NFL for team C. The coach cannot pick between them and so he goes for QB A who does better on the Wonderlic test.
Anecdotal, but the article does cite the Patriots offensive line as high scorers on the Wonderlic. They are also considered probably the best offensive line in football.
But, yeah, Gladwell raises a lot of surprising points in that talk.
Yet in this study the Wonderlic is keyed to position, not performance in that position.
We know that people largely stay within their positions, and we know that metrics like the Wonderlic and the drafting Combine strongly influence both draft decisions, and fielding decisions. We know that such metrics are designed to correspond with existing ideas: in this case, that quarterbacks were smart and tailbacks not so much.
It is plausible that this correlation represents bias in the decisions of the coaches, not real ability.
In fact, that's what suggested by Malcolm Gladwell. He points out that if you sort by Wonderlic scores, those who come out on top haven't really panned out as great quarterback. If we then look at some of the best quarterbacks of all time, you find that, on the Wonderlic, players are all over the map.
Malcolm Gladwell discusses this, on his talk about 'the mismatch problem'. I highly recommend it. http://www.newyorker.com/online/video/conference/2008/gladwe...