I meant to get back to this but took a break from the forum. I was reminded when a guy at work brought it up the other day.
Going back through my copy of
The China Study, I find no experimental evidence for his generalization from casein in rat chow to total animal protein in rat chow, let alone in human diets. Furthermore, I see no basis for extrapolating the monster doses of casein in the rat chow to the realistic range of casein in a human diet. (NB: The 5% level that was completely safe for the rats equals a quart of milk or 1/4 lb of cheddar cheese every day on a 2000-calorie diet.) Since he doesn't cite any evidence and nobody here has been able to offer any either, I infer that it doesn't exist.
Which leaves his argument standing on epidemiology. Campbell devotes only one chapter out of his book to the actual China-Cornell-Oxford Project; the rest is stats (mostly between-countries) hand-picked from all over the place to make his point. Admittedly, he has a lot of them. Like everybody, I've never seen the actual 900-page China Project paper, but I have read reviews of it that call into question his interpretation of the data, a lot of which seem to be first-order correlations. The tabulated data I've seen on the web don't support him at all; in fact, they showed a statistically insignificant
negative correlation with meat and a significant positive one with wheat. Meanwhile, the massive WCRF/AICR meta-analysis found milk consumption to be
protective against colorectal and (weakly) bladder cancer (and weakly risky for prostate cancer).
In other words, the epidemiological data on this are mixed, period. Here's an example from Campbell's very own Cornell:
http://envirocancer.cornell.edu/factshe ... .dairy.cfm
Now, the problem I have is that anytime a contrary study comes up, such as one showing milk consumption reducing the risk of breast cancer, the response from the vegan crowd is invariably, "Well, of course there's the odd study here and there, but they were probably methodologically shaky [not that I'm going to look into it], and we know that the
consensus and
preponderance of evidence [by which they mean
The China Study, i.e., the lay book] is
crystal clear overall." This amounts to a circular argument.
I'm of course not arguing that eating meat/dairy ad libitum is safe. But unless I've missed something, we need to stop running around citing this book as
proof that any amount of animal protein consumption causes cancer and that a pure vegan diet is safer than one with a small amount of animal products. (Geez, I've even seen this book taken to mean that
all cancers are caused
only by eating animals. E.g., "I read about cancer in wild herbivores; how's that possible when they weren't eating meat?")