In addition to the author's observations on the obstacles to conducting meaningful nutrition experiments, I would like to cite two other issues that I have been thinking about a lot.
On the one hand is the seeming impossibility of experimental science's ever arriving at a comprehensive understanding of complex systems. Because the scientific method relies on a severe simplification of problems in order to support or refute a hypothesis about relatively few variables at a time, it just isn't set up to analyze the interactions between a myriad of factors. A telling illustration of this is one author's description of the many calculations that you would have to make to optimize your purchases in a typical supermarket trip. The author was critiquing the inability of neoclassical economics to realistically account for complex, non-rational human behavior, but I think it accurately describes the extent to which any academic study must simplify and make rigid assumptions, so much so that the study is responding to an invented reality that is very far from actual reality.
Homo economicus, the fictional actor envisioned by the neoclassicals, performing calculations instead of interacting with reality, could be diagnosed as “autistic” more easily than the economists who created him. More advanced and evolved than the average homo sapien consumer, this idealized construct is capable of analyzing an infinite string of data in an infinitesimally small period of time – all with seamless prescience and precision. Take as an example a trip to the supermarket, where actors are charged with calculating which basket of goods will maximize utility and minimize cost. With the number of combinations increasing exponentially with the number of options, the actor faces 100 combinations given 2 options when told to choose 0-10 units of each. But given just 30 goods, told once again to choose 0-10 units of each, the consumer faces 10exp30 combinations. Even if the consumer could rule out 99.9% of the combinations and calculate each remaining combination in one-billionth of a second, he would be faced with a task lasting 32 billion years, or a period longer than the age of the universe. Homo economicus does not even bat an eye.Anyway, if I am right in this agnostic view of certain aspects of reality ("agnostic" pertaining to the inability to truly arrive at knowledge of something), then the NYT author is not entirely on the mark in his explanation of the lack of concrete evidence about the causes of obesity and other health problems. It is not so much the fact that we haven't gotten our act together to perform a watertight, rigorous long-term nutrition experiment that would exhaustively explain obesity, but rather that there could never be such an all-explanatory experiment, because the human body and human society are simply too complex to describe through such means.
The second tie-in to an issue that I think a lot about is the nefarious influence of commerical interests on sound science. Much of science as practiced today is shaped and bent by commercial interests, either in the questions asked, the techniques used to respond to them, or even what is done with the results of an honest experiment. In the case of nutrition, I am not as aware of how the corporate agenda feeds into academic research, but I am very aware of how nutrition science is often channeled into popular culture through dishonest books, dangerous fad diets, uncritical news coverage, and perhaps worst of all, the bogus nutrition claims made on the packaging and advertising of all sorts of processed junk food. So again, beyond the legitimate problems the NYT author points out with conducting meaningful, telling nutrition research, I also see a commercial agenda distorting what we know or think we know about nutrition and obesity.