2 ms·
High-quality studies don't just do a linear regression. For instance, it's possible to control for confounding factors by keeping a factor constant and varying
by YeGoblynQueenne 6y ago
High-quality studies don't just do a linear regression. For instance, it's possible to control for confounding factors by keeping a factor constant and varying others, in a regression model. I believe this was for example done in studies that showed the correlation between smoking and lung cancer, or high blood pressure and heart disease and so on.
In any case, when faced with a large body of literature with some conflicting results the thing to do is to look at reviews or meta-analyses. From my reading there is no argument that the majority of meta-analyses show a strong correlation between high levels of saturated fats in the diet and high incidence of cardiovascular disease- and if I remember correctly, even the so-called "French paradox" is not observed in high-quality reviews and meta-analyses, but only in smaller studies.
Note also (from the linked paper):
>> The French paradox concept was formulated by French epidemiologists3 in the 1980s.
In a similar manner, you can find many studies showing benefits from the consumption of olive oil, but the majority of them are from Spanish, Greek and Italian research teams (I'm Greek. Perhaps not everyone has the same context as me but the number of times I've heard Greeks say how healthy olive oil is... And how red wine is good for the heart. Greece, of course, produces both).