Flawed Logic Produces Flawed Research

You should always be wary of any dietary studies that are reported in the news that don’t seem logical or go contrary to common sense. For example, a study may be published that states that eating some particular food is bad and provide convincing reasoning and statistics to back up its claim. The medical community and even government officials and the media may jump on the bandwagon and proclaim the item to be unhealthy and recommend that we avoid it at all costs. Governmental agencies may even make policy decisions or design educational programs based on this flawed research.

Later, new studies may demonstrate that the previous study was wrong and that instead of being unhealthy, the particular food is harmless or even a good nutritious food that should be included in the diet. Once someone has it in their mind that a particular food is bad, it is hard to convince him otherwise. Thus a controversy ensues. This is what has happened with eggs, avocados, and coconut oil, to mentions just a few.

Often times, researchers, influenced by political correctness and medical/nutritional dogma, plunge into studies seeking to find fault with a certain type of food. Their premise is flawed, which leads them into the wrong direction. Data can be manipulated in order to arrive at a conclusion that supports the flawed logic. Statistics, one of the most useful tools of the researcher, will often be incorporated into the calculations.

You should always be wary of statistics. Although very useful, they can be misused. Statistics can be used to prove just about anything. To show the absurdity at which statistics can be taken, yet seem logical, we can look at a recent study on obesity.

Apparently, in the future, everyone will be fat — or so warns a new study published online in the medical journal Obesity. About 66 percent of American adults are now overweight or obese, according to government estimates, and the report makes the dire prediction that, based on current trends, by 2048 the figure will reach 100 percent! That means everybody over the age of 18 will be overweight.

But that projection, which presumes a linear increase in the number of people who are overweight, is logically impossible. The reasons are well outlined in a column “The Numbers Guy” in the Wall Street Journal written by Carl Bialik.

Employing that same logic to telephones, “13 out of every 10 adult Americans by then won’t have landlines,” Mr. Bialik wrote. “The phone forecast is impossible, of course, but it’s arguably no less solidly grounded than the obesity forecast.” Mr. Bialik also interviewed top statistician Donald Berry, chairman of the department of biostatistics at the University of Texas M.D. Anderson Cancer Center.

“Extrapolations are dangerous,” Dr. Berry explained. “Especially dangerous is to assume that trends are linear. Otherwise we’d conclude that Olympic swimmers will one day have negative times, there will be more Internet users than people, and more people on Earth than molecules in the universe.”

To read Mr. Bialik’s column, click here.

Another way researchers add bias into studies is to combine several variables, but attribute all or most of the effects of the study to only one variable. This variable, however, in reality could have absolutely no influence on the results. As a hypothetical example, let’s assume ABC Aluminum Company wants to find a cheap way to dispose of fluoride, a toxic waste byproduct of aluminum manufacturing. If they could create a demand for fluoride, they would not have to pay the high cost of having it disposed. They could sell it instead and make a huge profit. So, they commission a study to show that fluoride can cure scurvy, and thus prove it is an important nutrient that should be added to our diet.

The method used to prove fluoride cures scurvy is simple. You give test subjects (animals or humans) a dietary supplement that contains fluoride, plus some vitamin C and perhaps a few other vitamins for good measure. Well, it comes as no surprise that the new dietary supplement is successful in curing scurvy. Therefore, the study has demonstrated the nutritional need for fluoride in order to have a well balanced diet. We all know that vitamin C by itself can cure scurvy, so the fluoride really had little or nothing at all to do with the positive results of the study. Believe it or not, his type of research is done all the time. We may not recognize it because we don’t always know how the variables used individually affect the outcome.

An excellent real life example of this is with soy. Soy foods have been credited with all sorts of health benefits, but perhaps none so appealing as the assertion that it can lower high cholesterol.

The notion was cemented in 1999, when the Food and Drug Administration allowed companies to claim that 25 grams of soy protein a day, in a diet low in saturated fat and cholesterol, “may reduce the risk of heart disease.” The agency evaluated studies — including an industry-financed analysis published in The New England Journal of Medicine in 1995 — concluding that soy protein could cut cholesterol.

But studies since have raised doubts. In 2006, an American Heart Association advisory panel reviewed a decade of studies and determined that soy products had no significant effects on HDL (“good” cholesterol) or triglycerides, and little or no ability to lower “bad” cholesterol, or LDL. Another study, published in The American Journal of Clinical Nutrition, found that consuming 24 grams of soy protein daily had no “significant effect on plasma LDL” even in people with mildly elevated cholesterol.

Soy protein by itself has shown to be all but useless in lowering cholesterol. Only when soy is combined with a high fiber diet and plant sterols (two other variables in the mix) is soy shown to be of any value in this respect. Both fiber and plant sterols have independently been shown to lower cholesterol. So obviously, if you combine soy with a high fiber diet and plant sterols the combination will lower cholesterol.

It is shenanigans in medical/nutritional research like this along with statistical manipulations and other “tricks of the trade” that make the results of many studies suspect, especially if a certain industry stands to profit from it. ■