Every five years, the federal government’s Dietary Guidelines Advisory Committee comes up with new recommendations about what Americans should or should not be eating. I and a team of researchers have studied the data that these guidelines are based on and have come to the conclusion that the data are scientifically flawed.

That’s because most of the data on which dietary guidelines are based were gathered by asking people to recall what they had consumed in the recent past—something people are notoriously bad at remembering. Human memory and recall are too inaccurate and imprecise to be used as tools to collect scientific data. The data collected are literally uncorroborated anecdotes.

This reliance on bad data is why we often end up with conflicting dietary guidance—one study suggests that coffee causes cancer…another that it prevents cancer.

But if government dietary guidelines cannot be trusted, what do we need to do? We need to exercise more. Physical activity and exercise are major determinants of health and are largely ignored by federally funded nutrition ­researchers.

Many ­Americans now suffer from cardiovascular problems, diabetes and obesity. These problems are caused less by what we eat than by underexertion. An obese but active person has lower cardiovascular health risk than a thin but inactive person.

If your doctor has given you dietary guidance related to a specific health issue, do follow that advice—limit salt consumption if you have high blood pressure, for example. But beyond this, perhaps we should spend less time sitting around worrying about the details of our diet and more time doing anything other than sitting around.