Technology, Technophobia, and
In the early fall of 1998, there was a story on the news services about a statistics teacher who had his students do an assignment of either flipping a coin one hundred times and recording the results or making up the entire assignment. He could always tell which was which by the fact that those who made up the assignment failed to have the runs of heads or tails that would be expected as the random outcome of flipping the coin. Simply stated, random does not mean uniform. This is a point that needs to be repeated over and over again.
There is a growing concern in the United States over “clusters” of cancer and/or birth defects (Gawande 1999). As would be expected, the immediate suspicion is that there are hidden chemicals (meaning man-made chemicals) in the environment that are the cause of the cluster (Sachs 1998). In epidemiology, clusters of a disease often become the basis for investigation into a possible local cause. Obviously, a cluster that is several hundred or even several thousand times what would be expected in a random distribution affords more than sufficient reason to initiate a thorough epidemiological investigation. Over the years, some of these inquiries have discovered disease vectors and environmental causes both manmade and natural (Gawande 1999). Current concerns about clusters and “cancer mapping,” however, are, at best, well within a range that would be expected by random variation. In fact, given the number of different types of cancers, it has been estimated (for the state of California) that any particular census tract would have a “better than even” chance of having “a statistically significant but perfectly random” elevated level of a particular type of cancer (Gawande 1999, 37).