“Green tea lowers cancer rates.”

“Yoga helps breast cancer rates.”

“Coffee may cut alcohol liver damage.”

These and other sensational health-related headlines permeate the news. Each story touts some remarkable research result, promises some fabulous resolution, or warns of some impending doom.

And you should not believe any of them. You should be a critically thinking consumer.

First, the authors are not researchers, nor adequately trained in the health sciences. They are controlled by their news agency's gatekeeping policies. They do their best to interpret the research results in a prescribed number of words, time limit, and (in some cases) slant. (Sensationalism sells, especially in television.)

Second, “correlation is not causation.” Correlation is a relationship between two things that is expressed as either a ratio or as a percentage. Positive correlation means that when the scores of one factor go up, so do the scores of the other factor, to the degree of the correlation. An example: the more deadline pressure you feel when you begin a project, the greater your risk for making mistakes. But the pressure did not cause the mistakes.

Negative correlation is an increase in one factor with a corresponding decline in another. Increased body temperature is correlated with declining health. The fever didn't cause the decline, nor did the decline cause the fever. A third factor is the causal agent.

So let us look at those headlines again. Was it the ingestion of green tea/yoga exercise that caused the cancer rates to go down? Or was it the fact that the person who drinks green tea/does yoga is more health conscious and careful with diet, exercise, and other cancer-prevention behaviors, than others?

Third, any result needs to be replicated by multiple researchers, under similar and differing conditions, in order to substantiate and give strength to our belief in the initial results. Notice the “hedge words” in the third headline above. “May” is a far cry from “does.”

Fourth, no research project is free of confounds, which are conflicting explanations or influential factors that could explain or alter the results. It is simply impossible to so tightly control an experiment as to eliminate all confounds. At the very least, a tightly controlled, laboratory experiment produces results that are only valid under those tightly controlled laboratory conditions. The more tightly controlled the experiment, the less applicable it is in the outside world.

Fifth, a statistically significantly different result may not be clinically significant. For example, a statistical difference on a renowned depression index is only a few points, but unless the second score lies within a lesser-severe category of depression severity, there has not been much progress made.

Sixth, researchers are human. They and their assistants can make mistakes in data collection and evaluation. They can interpret data incorrectly. They can get so excited about the results that they come to extreme conclusions. Even worse, their financial sponsors may demand a positive spin on the outcome (this is especially true in drug research), or they may be frantic for publishable results in order to keep their jobs.

Does this mean that we ignore or hold in disdain all research? Absolutely not. Much good has come from research efforts across time. It simply means that we should be cautious about accepting reports of research results on face value only.

Grace MacDowell, M.A., of Auburn, is completing her doctorate in

clinical psychology. She can be reached at drmac2be@yahoo.com