In 2008 (through March), Tiger Woods, the number 1 golfer in the
world, hit only 55.95% of the fairways on his Tee shots, ranking him at
170 among professional golfers. His putting average wasn't much better.
With 28.93 per round, he ranked 58th.
On October 14, 2007, Brett Favre set the all-time quarterback career
interception record in a game against the Washington Redskins, his
278th.
Reggie Jackson holds the major league baseball strikeout record with 2597.
These facts, considered in isolation make Tiger Woods a bad golfer,
Brett Favre a poor quarterback and Reggie Jackson a lousy hitter. Yet
each is on the short list of the greatest ever in their respective
sports. How can one reconcile these facts against total performance?
Why is one number, even a seemingly very important one, so misleading?
The answer is both apparent to even a non sports enthusiast and is the
key to explaining the complex statistical methodology of multivariate
analysis to a lay audience.
There are many more facets to each of these games than Tee shots,
errant passes and unsuccessful batting moments. In the cases of Tiger,
Brett and Reggie, those other factors more than compensated for poor
performance in a single area. Such multiple factors can be
statistically analyzed through a method known as multivariate analysis.
Each factor is tested for its influence on the whole, while other
factors are set constant, as though they did not exist.
| Component of Game (Variable) | World Rank |
| Driving Distance |
289 yds |
38 |
| Driving Accuracy |
55.95% |
170 |
| Greens in Regulation |
74.54% |
1 |
| Putting Average |
1.696 |
3 |
| Eagles (Holes per) |
108 |
6 |
| Birdies Average |
4.83 |
1 |
| Scoring Average |
67.3 |
1 |
| Sand save percentage |
60% |
32 |
| World Ranking |
1 |
Claims of Lead toxicity in a child producing (you name it) cognitive
disorders, behavioral problems, delinquency, ADHD, etc. all suffer from
this same misuse of data. They take studies of large populations which
have controlled for numerous variables — maternal IQ, home life,
socioeconomic status, birth-associated abnormalities and many others
and only then can lead effects by teased out. In fact, absent such
controls, lead effects cannot be found in most of those studies. At
most, lead may contribute 4% to IQ.
Extrapolating from these studies to an individual child is exactly
like choosing one sport statistic – Tiger Woods' driving accuracy,
Brett Favre's interception numbers or Reggie Jackson's strikeouts and,
from those, concluding that each is or was a poor competitor. That is
patently wrong and so is the method by which lead claims are judged.
Ronald E. Gots, M.D., Ph.D.
THE NMAS GROUP
2301 Research Boulevard
Suite 210
Rockville, MD 20850-3204
800-566-1669
info@nmas.com
