A recent New York Times article, What Really Makes Us Fat, highlights how scientific research may sometimes lead us to wrong conlusions.
In terms of diet and nutrition, as the results of research roll in, science is beginning to agree that the food behaviours humans had for thousands of years (higher fat & protein, lower carbohydrates) were better than the typical high carbohydrate (and fructose!) diets of the last 50 years – contradicting some of the advice that had come from earlier research.
Science is necessarily slow to examine these things, because of the need to study one variable at a time while trying to control for confounding factors. Unfortunately, such attempts to reduce error can invite it. The biggest downfall to the scientific model being that it fails to see the bigger picture. It takes hundreds of years & thousands of studies built upon studies before research can piece together the real, broad picture.
For example, researchers are still trying to sort out whether carbs, fats, or protein are better – the same question they’ve been examining for the last 100 years. Never mind that western practices have introduced a gazillion other changes to our diet in that time – namely processing, pesticides, over-farming, early picking, increased shelf life, genetic modification, additives etc (please comment & add to this list!).
Our children’s children’s children will be long dead before the scientific community has done enough controlled studies to catch up to today’s situation – and then they’ll still be hundreds of years behind because of the continued changes in diet and food production! So what do we eat while we wait for science to be able to tell us what’s healthy? We can look back to the diet eaten over 100 years ago – when obesity, diabetes, metabolic syndrome, food allergies etc weren’t an issue – and copy it as best we can.
If you’re passionate about diet & healthy living, please comment. We would love to hear about any reading (books, blogs etc) you can recommend!
Yes! In fact, the standard methods of performing research hit a critical flaw when it comes to diet. Every body is different. Not all people metabolise fat (or proteins or carbohydrates) in the same way. So a standard, double-blind, randomised control study will always hit a road block when studying a group of people as though their bodies are all the same. This was highlighted in a very interesting program I saw ages back called “Why Aren’t Thin People Fat”. There were a wide variety of ways that these congenintally thin people didn’t get (very) fat even on a very high calorie diet with no exercise.
It’s an area where we simply can’t depend on science for all the answers, but have to consider our design and the Creator’s intent, and use a whole lot of common sense.
Within-group variability is often assumed to be random when it may be systematic and an effect of sub-groups – this is a big issue that applies to a lot of research. Stratified random sampling works if you know something in advance about differences in the population and their potential impact on your investigation. But what about factors we don’t yet know about? There are statistical methods to check whether sub-populations exist in your research data, but I think they tend to be poorly understood and often overlooked. These issues don’t invalidate research, but they highlight the importance of applying some common-sense, as you say.