BUSINESS INTELLIGENCE AND THE SEARCH FOR TRUTH
by Inderpal Bhandari, executive editor at large

Facts or Imaginative Exercises?

As the technologies of business intelligence take hold in corporations, the question is being asked: How can one discern which findings of a knowledge discovery program are true and which are spurious?

If the history of scientific discovery is any guide, this fear is well founded. A recent article in the New York Times alerted me to an organization called the American Council on Science and Health. Based in New York, the council releases a popular report, Facts Versus Fears, that promises to be interesting reading. The report reviews the greatest unfounded health scares of the last five decades. The article in the New York Times draws on examples of those scares to make the point that it is wise to look closely at any new study before espousing its recommendations.

What caught my eye in those examples were the amounts of money that were lost on account of some of these studies. First, Alar, a chemical used to ripen apples, was accused of being a cancer-causing agent on the basis of a study in 1973 that showed that one of its byproducts caused tumors in mice. However, subsequent tests by the National Cancer Institute and the EPA failed to show that Alar caused cancer. But the damage was already done. The apple industry and the Department of Agriculture lost $390 million dollars as parents replaced apple juice with less nutritious drinks in the diets of their children.

Second, asbestos. Asbestos is harmless in solid form. But when it becomes airborne and inhaled, it can sometimes cause cancer. Asbestos also does not burn, does not conduct electricity and is often resistant to chemicals, which makes it suitable for fireproofing. Consequently, its installation was required in schools until 1973, when the EPA banned its use in schools. Soon after that, Congress swung into action. Schools were required to clean up the asbestos hazard, resulting in 6 billion dollars being spent by 1990. Expert opinion was divided on this issue. Many believed that the asbestos removal efforts, by increasing airborne asbestos, created a far greater hazard to the children than simply leaving the solid asbestos in place would have been.

Why do such mistakes occur? Gore Vidal once opined that it was the spirit of the age to believe that any fact, no matter how suspect, was superior to any imaginative exercise, no matter how true. Scientific studies do indeed lead to the discovery of facts. But the studies themselves are fundamentally imaginative exercises undertaken by scientists. In that light, there really is no difference between facts and imaginative exercises. They are but two sides of the same coin. And, imaginative exercises are prone to error.

The bottom line is that to evaluate a fact, one must evaluate the underlying imaginative exercise, namely, the methodology that was used to discover the fact. When we make use of a computer program to assist in the discovery of knowledge, the situation does not change. On the face of it, one may be disposed to think that the troublesome imaginative exercise has been replaced by an unfailing mechanical mind that will not err. There is some truth to that, in that the use of a computer will generally reduce operational errors. However, with regard to methodological errors, there are now two imaginative exercises to consider, namely, that undertaken by the programmer and that undertaken by the user of the program. Either one or both could lead to errors, hence, the popular saying: To err is human, but to really foul things up requires a computer.
---
Inderpal Bhandari can be reached via http://www.virtualgold.com