Lots of people think it’s obvious: vegetarians are healthier than omnivores in Study X. Therefore vegetarianism is healthy. This is such a common line of argument that I draw your attention to Denise Minger’s slides for her talk at the Ancestral Health Symposium, which I have already blogged about. The slides make clearer than I did what she said. It’s all excellent but the best part is at the beginning where she points out the many confounds in the studies of Dean Ornish, Neal Barnard, Caldwell Esselstyn, and John McDougall. They change many things but — as Denise put it — What caused the benefits? It must be the vegetarianism! Paleo humor.
There are lots of other interesting things in Denise’s talk, as you will see, such as the health of religious vegetarians.
Wait, what? Correlation is not causation?
One of the slides is misleading. The exercise comparison was 70% compared to 80%. Proper charting says you always start your bar chart at 0, the person that did the chart started at 50 or so. Otherwise, interesting, not a lot of info, just the last slide that gives a website to look at. Will make me look at it a little closer when I do comparisons in the future.
BTW, bought your book and have been losing weight. It’ll be nice when I get down to my desired “ideal”. Maybe then I’ll be able to run again without getting injured all the time.
Shane, the data Denise was talking about (Ornish, etc.) is not correlational — that is, from a survey. It’s from experiments. Moreover, one of Denise’s slides is about “religious vegetarians” — what we can learn from them. The data in that slide is correlational. I believe that data is helpful. To claim, as Denise did (and I agree), that you can learn something helpful from survey data is roughly the opposite of “correlation is not causation”.