Assorted Links

  • Correlation between fat intake and brain-test scores. “Those women who reported the highest saturated fat intake also had, on average, the worst scores on reasoning and memory tests.”
  • How many iPads does it take to change a textbook market? A perfectly good physics textbook is now available for free download (pdf). The author of the post, a physics professor at William and Mary named Marc Sher, does not understand what’s going on when he refers to “the textbook publishers’ price-gouging monopoly” and their “outrageous practices”.  Textbooks cost so much because students can be forced to pay that much. This has nothing to do with publishers, I submit, and everything to do with the power professors have over students. Sher would reply: All the textbooks are expensive. And I say: So what? If students could choose not to buy $200 textbooks, none would be sold. Zero. And future years would see no more $200 textbooks.

Thanks to Jonathan Graehl.

10 Replies to “Assorted Links”

  1. I imagine the people w/ high saturated fat intake weren’t eating a paleo or low carb diet but were instead people who ate a lot of Mcdonald’s and other fast foods. In other words, the group, having more of the obsese and dumb, probably had more people on the left side of the bell curve to begin with.

    Contrast this with the kind of people who use olive oil, eat avocado, etc., who are middle to upper class people who are somewhat informed and follow trendy health advice, so they would be more likely to have more on the right side of the bell curve to begin with.

    Seth: I agree. This is common in epidemiology: People who do the “good” (= approved) thing are in better health than people who do the “bad” (disapproved) thing. But not because the “good” thing (such as avoiding saturated fat) is actually good, but because of other ways the two groups differ.

  2. Seth –

    I am not a scientist and I am curious to know what you think of this study as it correlates to your consumption of butter.

    Thank you,

  3. The Kanazawa Hypothesis would predict the saturated fat result. Still, I second Patti, this is a good change to challenge my bias that quality animal fat is good. Personally, I feel much sharper when running on animal and coconut fat.

  4. The fat correlation must simply be due to on confounders. Those kinds of conclusions fail to take into account basic inherent properties of the different fats. Long term controlled interventions show less lipofuscin, less collagen crosslinking, higher energy expenditure, less peroxidation, lower liver trigs, less body fat, etc with *saturated* fats. Also, there are similar results when compared to glucose, but you have to search deep, as it is hard to find a useful study in that regard (ie, not comparing corn starch to sucrose+lard+corn oil). There is one study in rats that uses a “cereal-based” diet, a “high polyunsaturated fat” diet, and a “meat and milk based” diet. The last comes out best.

  5. In the article, the researcher appeared to make a claim of efficacy for an intervention based on the results observational study with likely confounders and no clear causation argument. The journalist went right along and published it. This is typical but doesn’t reflect well on either the journalist or the researcher.

    Seth: I disagree. The article explicitly stated the limitations (“this one can’t prove…”). And the researcher is correct when he says “Our analysis suggests if you substitute out 5 percent of your saturated fat calories with 5 percent monounsaturated fats, you could have a 50 percent lower risk.” It does “suggest” that conclusion in the sense of “make more plausible” that conclusion. I believe the correlation is due to confounders but I am not absolutely 100% sure of that.

  6. Is there a good review anywhere showing how often results from correlative studies are supported or disproved by randomised controlled trials?

    Seth: I haven’t seen such a review.

  7. Dearieme: I believe John Ioannidis published something about this. His conclusion was that non-interventional studies should be given more credit.

  8. dearieme,

    Why compare unrelated studies just because they are observational-turned-controlled? What makes you think their methods and/or results would provide any insight to a brand new observational result? But just for kicks, Harvard’s Walter Willet and his team have been wrong with many [all?] of their conclusions from observationals.


    A blanket statement like that is completely useless though, as different studies vary significantly.

    These correlation studies with unsaturated fats go against the basic properties of fats, saturated being theoretically superior. Except, we don’t even need to say “theoretically” because the work has been done already! We don’t have the clear results in humans because they live too long to do well-controlled interventions that look at aging or aging indicators.

    Short term mice/rat studies, which are inconsistent anyway, that use industrial lard and/or sucrose tell us near-nothing about diet. They are useful for studying the physiology of obesity, but any diet advice given based on them is a set-up, or the authors [of the paper] are stupid.

Comments are closed.