Suspicious stats

One of the latest in the seemingly endless rounds of alarming statistics is that one out of 12 American children has some form of disability. With all the things that are supposedly getting worse, you have to wonder how our life expectancy keeps increasing. A cynic might even wonder if the increasing availability of money from the government has anything to do with the increasing number of "problems" that need to be "solved" by government programs.

One way of telling whether a given statistic is a fact or an artifact is to ask whether the definition used fits the thing that is being defined. Buried in the news story about the children with disabilities is the fact that the definition of "disability" has been expanding over the years.

A child who is likely to be diagnosed as autistic today might not have been some years ago. Yet that is seldom mentioned in alarming statistics about the escalating number of cases of autism. As the author of a couple of books about late-talking children, I hear regularly from parents who tell me that they are being asked to allow their children to be labeled "autistic," in order to get either the government or their insurance company to pay for speech therapy.

It is amazing that, with something as serious -- indeed, catastrophic -- as autism, statistics are thrown around without mentioning the variation in what is being diagnosed as autism. In something much less serious, such as sales receipts at Wal-Mart, a comparison of how much money was taken in this year, compared to last year, will almost certainly make a distinction between sales receipts at the same stores as last year versus sales receipts that include new stores opened since last year.

In other words, they notify you of changing definitions behind the numbers. Otherwise, the statistics could mean almost anything. If it is important enough to do this for Wal-Mart sales, it certainly ought to be important enough to do it for autism.

Regardless of whether the old or the new criterion for autism is better, they are different criteria. Statistics should tell us whether or by how much autism has risen by any consistent standard. Moreover, those who diagnose autism range from highly trained specialists to people who never set foot in a medical school.

Another set of statistics whose definition is at least questionable are statistics about the incomes of high school dropouts versus those who have more education. Since most high school dropouts resume their education at some later time, are these statistics really counting all -- or even most -- dropouts? Or just the minority of dropouts who never enter a classroom again?