Well said! I spotted it in a conversation with Dennis Mangan. I guess it’s good to see that someone else has brought up these points.
Here’s the rest of the conversation:
Unfortunately, it seems, Dennis Mangan – like many in the medical establishment – appears permanently locked into a rather unscientific way of thinking. The problem of disentangling correlation and causation is ignored. Uncontrolled observational studies are taken as proof of whatever medical fad in vogue at the time. It is possible to “control” for various variables (since we can surely think of all the relevant ones) to see if the variable of interest is a causal factor.
James Thompson had a recent post on research conducted in this matter (Psychological comments: Is it healthier to eat 7 vegetables or 7 scientists?). Thompson references an article written by David Colquhoun, where Colquhoun references the work of John Ioannidis (here Colquohoun, quoting Ioannidis):
The gist is given by the memorable statement
“Almost every single nutrient imaginable has peer reviewed publications associating it with almost any outcome.”
and the subtitle
“Definitive solutions won’t come from another million observational papers or small randomized trials“.
Colquhoun lays bare the problem:
The problem, of course, is that humans are very good at attributing causality when it does not exist. That has led to confusion between correlation and cause on an industrial scale, not least in attempts to work out the effects of diet on health.
The likely truth:
We can probably say by now that no individual food carries a large risk, or affords very much protection. The fact that we are looking for quite small effects means that even when RCTs are possible huge samples will be needed to get clear answers. Most RCTs are too short, and too small (under-powered) and that leads to overestimation of the size of effects.
and what must be done:
The best thing that can be done in the short term is to stop doing large observational studies altogether. It’s now clear that inferences made from them are likely to be wrong.
At the very least, it would seem to me there’s one idea that would be a very helpful step in the right direction:
Correlational health studies would be more impressive (and useful) if they included IQ and personality data.
— JayMan (@JayMan471) April 6, 2014
As for this particular finding, the apparent link between depression and processed food consumption, assuming it’s real at all, what’s behind it? Well, what does processed food consumption correlate with (hint: negatively)? IQ. This is indeed the subject of James Thomspons’ latest post (IQ, Neuroticism, booze, and those damn vegetables again).
On that, one very large study (of Swedish conscripts) found that IQ was negatively correlated with severe depression (and other things like schizophrenia). So there’s that, anyway.
Previously: Gary Taubes on Obesity and Bad Science
For some reason, this seemed appropriate: