Correlation or Cause

David Mendosa Health Guide July 22, 2011
  • We can divide everyone into two types of people -- those of us who have diabetes or those who don’t. Or we can divide us into men and women. Or into people who can read and write compared with those who can’t.

    “Wait a minute!” you might exclaim. “That’s already a lot more than two types.”

    Precisely. And that’s just the problem with all those studies that purport to show that two situations that occur today have a cause-and-effect relationship.

    A greater proportion of people in those countries with a high rate of literacy have diabetes than in those countries where people are less educated. That’s a correlation between diabetes and literacy. It doesn’t prove that education causes diabetes.

    Maybe the rate of diabetes has gone up in this country because more people live in homes that have air conditioning. Those two variables do correlate. But that doesn’t prove that one caused the other.

    We have so many types of people in fact that each of us is unique. And no study can possibly control for every type of difference among us.

    Yet scientific researchers continually bring forth new studies correlating one sort of change with another. These studies do have a real value.

    They can give us hints about what researchers need to investigate further. Correlation studies can suggest theories worth testing in a randomized double-blind, placebo-controlled study.

    In the past few weeks I have been thinking a lot about correlation or cause because of a study presented at last month’s convention of the American Diabetes Association in San Diego. HealthCentral sent me there to report on the latest findings about diabetes.

    One study that got lots of attention seemed to show that people who drink diet soft drinks get fatter. Researchers at the University of Texas Health Science Center in San Antonio showed that diet soft drink users experienced 70 percent great increases in waist circumference compared with non-users. You can read the abstract of their oral presentation, “Diet Soft Drink Consumption Is Associated with Increased Waist Circumference in the San Antonio Longitudinal Study of Aging,” online.

    I’m not questioning that they did find a correlation. But where is the mechanism? How can we get fatter by drinking something that has no calories?

    The researchers don’t say. Maybe people who drink diet sodas are the sort of people who eat more junk food. Maybe diet soda drinkers have a type of gene that makes them want to eat and drink more than others. Who knows.

    What we do know is that this is preliminary research that needs further testing. We also know that switching to soft drinks sweetened with high-fructose corn syrup isn’t likely to shrink our waistlines.