Statistics Stink

            “There are three kinds of lies: lies, damned lies, and statistics,” former British Prime Minister Benjamin Disraeli said in a quote that is mentioned, without fail, at the beginning of every statistics course. This is so because statistics, contrary to public opinion, can so easily be used as a form of opinion; it does not take much effort to find a statistic that supports a claim that you want to make. And in the world of cancer and medicine, that is a very dangerous proposition.

            Throughout The Cancer Chronicles, George Johnson mentions countless examples of statistical abuse, for all of which he promptly provides the necessary context to understand them. There was no better example of this than advertisements touting the anticarcinogenic effects of fruits and vegetables: “Throughout the 1990s, the news was filled with reports of miraculous anticarcinogenic effects from consuming nature’s bounty. The National Cancer Institute began pushing its 5 A Day program. Eat that many servings of fruits and vegetables and you would be a long way toward beating the odds against cancer” (Johnson 24). What these companies did not mention, however, was that these conclusions were reached based upon flawed experiments, known as voluntary retrospective studies, in which individuals with and without cancer were asked to report what they remembered of their diets. This undoubtedly could have led to what can be called circumstantial bias: that is, those without cancer would be more inclined to recall the healthy foods of their diet, whereas those with cancer could have been more likely to recount the negative aspects of their diet. And thus, we reach an abuse of statistics: a conclusion based off of hopelessly biased data, and one that clearly impacted American society for countless years thereafter.

           

            This is only reflective of half of the problem; what is also not known about fruits and vegetables is that the antioxidants that they possess, as well as fiber and certain types of vitamins, have anticarcinogenic effects only to a point. Only in moderation can these foods improve an individual’s chances of avoiding cancer; past that point, they actually worsen those very same odds. As if that was not enough, the changes in these odds often are made to sound far greater than they actually are: a 20 percent decrease in an individual’s chances of developing cancer sounds tremendous on paper, but when that twenty percent decrease is from a 1.50 percent chance to a 1.20 percent chance (fairly in line with actual decreases seen by many of these products), that decrease does not sound nearly as significant. This would be another abuse of statistics seen in the medical world.

            Statistical illiteracy, largely a byproduct of insufficient education, has therefore played a huge role in developing societal attitudes towards the sciences. Often times, science is used as the explanation to eye-popping statistics, but when the reality is exposed, the result is a distrust of the sciences. This reality is quite unfortunate, as both statistics and the sciences are given poor reputations in the media as a result. Statistical literacy and more specialized education would surely go a long ways towards solving these problems.