Fear Merchants

Once upon a time, in the not too distant past, the 2004 election was coming to a close. The setting is the Vice Presidential debate between two very different personalities: Dick Cheney and Jonathon Edwards. Both of them were throwing out statistics like they were going out of style and both used different statistics to inform their stance on the same topics. Both were trying to make everyone believe that without their guidance America would fall apart. I'm sure I'd thought about this kind of thing before but it was the first time it really stuck.

About a month ago ESPN came out with a report about a "scholarly paper" by a University of Penn professor and a Cornell graduate student about racism in the NBA. The academics claimed that white referees are biased to calling fouls on black players. They garnered substantial slant looks because they dug their data out of box scores that had no racial designation on the referees or who called what foul on what player. Only the names of those in the crew were posted below the box score. You would have to go back and watch every game. This was the talk of all sports radio only a couple of weeks ago and cast the NBA officials as bigots. The players didn't buy it and neither did the talk show hosts. That's good, one for common sense and about a hundred for statistics.

Let me take you into a hypothetical situation where the statistical issue is self selection. A study comes out and claims, "Most People Taking Mamograms Find They Have Breast Cancer" and it asks if you've taken your mamogram recently. The knee jerk reaction if you're a woman is "Holy Cow, I might have breast cancer," but it may just be a problem of self selection.

(Also, I just want to point out here that in some situations people think "I don't want a mamogram because that will result in me having breast cancer." This is a problem of causation and causality, a pretty way of saying the two really don't have anything to do with each other)

Most of the women going out and getting mamograms are the people that have a family history of breast cancer or an accute interest outside of the standard normal interest, maybe their friend died or they hear about other people and believe they need to do it. This is self selection. Let me illustrate it another way.

Occasionally my email backs up, some are kept after their initial reading and some are instantly deleted but I found one the other day from the Templeton Foundation about spirituality among college age students. Likely, they'll use the survey to say something like "College Students Are More Spiritual Now Than They've Ever Been" or "The Death of Religion, The Rise of Spirituality". It was, however, a very long survey and unless you were interested in spirituality you wouldn't fill it out. It took way too long. This is the essence of self selection.

Maybe if Templeton is interested in getting a more accurate cross section of students to ask about their spirituality they would visit a mixture of private and public universities and take a solid sample size from universities in few different regions of the country. Offer money to the people taking the survey, five dollars for twenty minutes. They can't leave before twenty minutes and get the money. That will guard against Christmas tree-ing. That is something they could really use for data.

I'm sure most of us have heard that we only use 10% of our brains. It's not true if you talk to people in the field. It's really only 3%. Just kidding. Truly, our synapses are firing all over our brains. These sort of strange statistics are used all the time and they enter the vernacular and are never cited again until an academic or someone with curiosity decides to search it out. For all we know it could be as unreliable a report as the racist NBA officials. It can really get to be hazardous like the 2004 VP debate. It's just a load of statistical silliness.

We as people have got to start asking some questions and not be taken with fear at the hearing of some number. We have a responsibility to be careful about what we tell others. At the same time, it's our responsbility to verify what someone else says. Let me close with a real example.

Someone said, "The instances of cacerous deaths are going up because we inject our food with growth hormones and it isn't pure." Some notes:

Cancer wasn't as easily detected in the 1950's
People live longer. They don't die of illnesses I became familiar with playing Oregon Trail.
Is it listed in nominal terms? Have the statistics adjusted for a greater population?
What kind of cancer did the person refer to? Skin cancer has been increasingly observed and everytime you get a cancerous spot removed it counts.

There are too many questions with that statement. So don't run out to the store and buy organic tomatoes on account of fear. Buy them because you like the taste or you don't agree with pesticide use, but don't do it because of fear. By the way, does anyone know how Trident came up with the 4 out of 5 dentists thing?