“The Plural of Anecdote is Not Data” (Part 3 of 4)
This is an excerpt from the intro of 12 Steps to Whole Foods.
Advances in the field of nutrition are taking place faster than ever in history. For example, just this decade, the “master hormone” leptin has been discovered, which governs the other hormones. New data calls into question the popular counsel of the past decade to eat 4-6 small meals daily: leptin research suggests that we should eat three meals daily and allow our bodies much rest from digestion. In just 2004, a class of glyconutrients (sugars) have been found to have powerful healing properties, which disputes a decade of anti-carb “experts.” Many people become frustrated by all the new information and competing voices telling us what to eat, what not to eat, and why. So that you don’t give up and “throw the baby out with the bathwater,” I have a bit of common-sense advice that super-simplifies the essence of a statistics class.
The main way to push through the inevitable cognitive dissonance is to read and learn all you can: 12 Steps to Whole Foods is a good start, and you may also consider the reading list on www.greensmoothiegirl.com (I am adding to it shortly). When you encounter contradictions, consider several things.
First, what is the funding behind the research? You don’t have to become paranoid to examine whether research was undertaken to objectively examine an issue, or to promote an agenda. It’s simply a part of being a savvy consumer of information in an age when we are all bombarded with thousands of voices.
For instance, if a study tells you that drinking wine daily prevents heart disease, use your critical thinking skills. Why did researchers study wine instead of grape juice–or better yet, grapes? Before you go out and stock up on a year’s supply of wine, ascertain if you can who paid for the study. Was it the wine growers of Sonoma Valley? Often studies in the modern age are funded, second-level, by an industry wanting to promote a product (often one that is under fire), even if the legitimate-sounding researchers named in the media, such as a university, are not directly linked to a motive. When that is the case, researchers know they are to publish whatever they can that is favorable to a product or industry, and publish nothing they find that is unfavorable.
Second, is the study valid? This is the highest standard in statistics and research, and it means does the study measure what it purports to measure? This seems simple enough, but it is in fact a difficult thing for researchers to achieve. If wine drinkers have much less cancer than beer drinkers, wine must be preventing cancer, right? Not necessarily. Maybe wine drinkers are a higher socioeconomic class than beer drinkers, in the aggregate, and beer drinkers also eat more fast food and smoke at higher rates.
Third, in tomorrow’s post.