I’ve seen her critique, but with her being a blogger and me being neither a statistician, epidemiologist, nor a nutritionist, and after seeing the resulting fight about methodology in the comments by people who claim to be those things, and knowing that Campbell stands by his claims even after reading her critique and has written his response, and with me having a rather low base rate confidence in any given study being correct anyway, it didn’t influence my opinion that much in any direction.
I’m pretty much out of my depth insofar as methodology is concerned in this case so, I’m not really in a position to evaluate anything on those grounds, yet. It at least feels more productive to just skim tons of abstracts and get a big picture idea than to analyze whether one particular study is correct.
Also, there’s a big difference between the claims of The China Study book which cites a range of findings (some of which I believe fall into the “made sense at the time but are now outdated” category), and the findings of the China-Oxford-Cornell study in particular, and as I read her critique it seems she’s getting those two confounded.
PS—The “meat, especially red meat, is carcinogenic” claim is one of those for which Cambell provided fairly robust support from other studies, not relying on the China-Oxford-Cornell data alone. I think subsequent research found that preserved meat (deli, smoked, etc) and maybe various common methods of grilling explain the carcinogenic factor, but don’t quote or trust me. (Not that I have strong priors against meat being carcinogenic, the “ancestral environment” arguments might not hold for extremely late-stage diseases like cancer)
me having a rather low base rate confidence in any given study being correct anyway, it didn’t influence my opinion that much in any direction.
If your prior is that the study likely doesn’t provide much value then this might not change your opinion. On the other hand reaffirm your priors should mean not putting much weight in Campbell case.
On the other hand reaffirm your priors should mean not putting much weight in Campbell case.
Not sure I understand what is meant by this.
Was trying to say: The data brings value. I don’t trust conclusions drawn from data. I also don’t trust Minger’s belief that the data analysis was obviously flawed because I see many people arguing over that. Rather than investigating further about Minger vs. Cambell on an argument that has experts disagreeing, it’s more worthwhile as a reader to just provisionally assume the data analysis is passable and read more on the topic elsewhere, because the risk of being mislead by flawed data analysis or other methodological issues is ever present, and in a field like this one is better off reading widely and look for broad trends and conceptual replications than one is by reading extremely closely. (And hoping any wrong beliefs brought about by bad data analysis fall away because the other experiments don’t support them.)
That’s what I mean by “low base rate”. If I’m just reading to get a big picture of reality in a field, rather than dive into the difficult rabbit hole of “are the methods ok”, I just operate under the assumption that there’s always an x% risk of any given study being flat-out wrong about everything and keep reading more without worrying about it.
(Which is why “Minger disagrees with those methods” falls into the “well, methods are frequently complicated and controversial and you’ve already factored that in so don’t worry” box. If other experts unanimously chimed in agreement with Minger, or if what she wrote about the flaws seemed obviously true to me, it would be a different matter).
I’ve seen her critique, but with her being a blogger and me being neither a statistician, epidemiologist, nor a nutritionist, and after seeing the resulting fight about methodology in the comments by people who claim to be those things, and knowing that Campbell stands by his claims even after reading her critique and has written his response, and with me having a rather low base rate confidence in any given study being correct anyway, it didn’t influence my opinion that much in any direction.
I’m pretty much out of my depth insofar as methodology is concerned in this case so, I’m not really in a position to evaluate anything on those grounds, yet. It at least feels more productive to just skim tons of abstracts and get a big picture idea than to analyze whether one particular study is correct.
Also, there’s a big difference between the claims of The China Study book which cites a range of findings (some of which I believe fall into the “made sense at the time but are now outdated” category), and the findings of the China-Oxford-Cornell study in particular, and as I read her critique it seems she’s getting those two confounded.
PS—The “meat, especially red meat, is carcinogenic” claim is one of those for which Cambell provided fairly robust support from other studies, not relying on the China-Oxford-Cornell data alone. I think subsequent research found that preserved meat (deli, smoked, etc) and maybe various common methods of grilling explain the carcinogenic factor, but don’t quote or trust me. (Not that I have strong priors against meat being carcinogenic, the “ancestral environment” arguments might not hold for extremely late-stage diseases like cancer)
If your prior is that the study likely doesn’t provide much value then this might not change your opinion. On the other hand reaffirm your priors should mean not putting much weight in Campbell case.
Not sure I understand what is meant by this.
Was trying to say: The data brings value. I don’t trust conclusions drawn from data. I also don’t trust Minger’s belief that the data analysis was obviously flawed because I see many people arguing over that. Rather than investigating further about Minger vs. Cambell on an argument that has experts disagreeing, it’s more worthwhile as a reader to just provisionally assume the data analysis is passable and read more on the topic elsewhere, because the risk of being mislead by flawed data analysis or other methodological issues is ever present, and in a field like this one is better off reading widely and look for broad trends and conceptual replications than one is by reading extremely closely. (And hoping any wrong beliefs brought about by bad data analysis fall away because the other experiments don’t support them.)
That’s what I mean by “low base rate”. If I’m just reading to get a big picture of reality in a field, rather than dive into the difficult rabbit hole of “are the methods ok”, I just operate under the assumption that there’s always an x% risk of any given study being flat-out wrong about everything and keep reading more without worrying about it.
(Which is why “Minger disagrees with those methods” falls into the “well, methods are frequently complicated and controversial and you’ve already factored that in so don’t worry” box. If other experts unanimously chimed in agreement with Minger, or if what she wrote about the flaws seemed obviously true to me, it would be a different matter).