On 'Flaws And Fallacies In Statistical Thinking'

While unpacking the Nth hernia of books in our new house, I found an old friend that I had totally forgotten about. Flaws and Fallacies in Statistical Thinking by Stephen K. Campbell. It was published in 1974, but as I skimmed through some of the examples, particularly the ones involving misuse of statistics in advertising and politics, the material seemed as fresh as today.

The same bogus claims are nightly made on network TV that were made back then. Politicians still make the same no-content affirmations that they did then. Is there a quantitative difference? I don’t know, but the introductory quote to Chapter 1 caught my eye:

Statistical Thinking will one day be as necessary for efficient citizenship as the ability to read and write.

H.G. Wells

This quote is particularly poignant when we realize that H.G. Wells was truly prescient in this instance, just as he was in so many other predictions. However, what he overlooked is that a culture can lumber along with inefficient citizenship for a long time. I maintain that with the information explosion, the need for education in statistical analysis and decision theory has increased greatly, but that need is not being satisfied. In a real sense, citizens are probably less efficient in the way Mr. Wells meant now than they were in his day because the possibilities to disseminate faulty claims is greater, and most people do not have the tools to fight back.

Lest you think I am over-reacting, consider another quote. This one is the introduction to Chapter 4.

It ain’t so much the things we don’t know that gets us in trouble. It’s the things we know that ain’t so.

Artemus Ward

Have you ever met anyone who knows that Darwin was wrong and can cite scientific studies that disprove evolution? Have you ever met anyone who knows the essential characteristics of the various races of humanity and therefore can rank order them? Have you ever met anyone who had talked to a god about you and received advice on how you should change your life?

Most good scientists relish finding something they don’t know because that is the grist of a good investigation and might even lead to new discoveries. Oftentimes making progress starts with unlearning something we knew that “ain’t so.” That’s a difficult thing to do. We all value the things we think we know, and admitting that something valued is just plain wrong is not easy.

At least that’s what nine out ten psychologists say in a recent survey published on a Web site somewhere.

More to the point, the author gives several examples of eccentric methods used to make economic predictions. These are all attempts to predict the market performance based on data thought to be relevant.

  1. Jupiter-Saturn cycle. This is pure astrology. For the uninitiated, “Adverse psychological conditions prevail when two planets are in conjunction on the same straight line from the sun or in squares at a 90-degree angle, with the sun at the center. Favorable conditions exist when the planets form a 60-degree sextile with the sun or a 120 degree trine.” Please don’t address questions about this to me.
  2. Levels of Lakes Michigan and Huron. When the level drops, the load capacity of ore boats drops, shipping expenses rise, and that propagates through the economy.
  3. Average size of tabs and tips at Sardi’s restaurant. Presumably this famous restaurant samples the optimism of the shakers and movers.
  4. Purchases of gourmet foods. Advocates of this claim gourmets switch to plain food immediately before a recession starts. They don’t offer a reason. It’s as mysterious as woolly caterpillars.
  5. Women’s hemlines. Higher hemlines mean good times ahead – economically good times. Since there is an absolute limit to the height of a hem, this is actually a pessimistic hypothesis since it puts an upper limit on the growth rate.

It’s easy to completely dismiss these eccentricities, but I think that is dangerous. It’s prudent to ask why such things would be posed and why some of them continue to be used. That is different from assigning any credibility to them. A recent book entitled “Bigfoot Exposed” makes the argument that even though we have no evidence of a large primate roaming northwest America, the study of the propagation and longevity of this myth is important. Why do hoaxers ply their trade when neither money nor fame will result? In a similar way, I think that we should seriously consider the logical eccentricities of others as a study by themselves. For if we understand why they attract adherents, then maybe we can begin to understand how we ourselves are fooled by similar eccentricities without knowing it.

For those who wish to delve further into decision theory without wading through a lot of equations, I have posted a tutorial on elementary decision theory. It shows examples of faulty physicians’ diagnoses (important for those considering surgery) and how to evaluate anti-terrorist activities (important for everyone). That tutorial can be found here.