[gabrielgoh]

Sure, counterfactual causation is the gold standard in science, but it is extremely difficult to pin down experimentally (in this case it requires an experiment, or a natural experiment, where people are forced to drink 2 drinks daily over a span of 30 years. This is not only impossible but unethical.). The judgement of causation from correlation is usually an art not a science, with the use of such things as instrumental variables or other such "soft" techniques, which boil down to opinion and expert knowledge.

The article itself states that the authors believe there is an element of causation in the study

"Alcohol might represent a modifiable risk factor for cognitive impairment, and primary prevention interventions targeted to later life could be too late."

The "may" is, of course a kind of scientific hedging, but I do not think the journalists made an unfair reading of the paper. Bear in mind many of these articles are made in consultation with the authors themselves. If you have any problems with the conclusions of the article, you should take it to the authors of the paper itself, not the journalists

[epistasis]

I think the request here isn't to change this paper, or the article, but for the headline to be "Even moderate drinking correlated with atrophy in brain areas..."

This is on the editor's heads.

[gabrielgoh]

i'm saying that headline says too little. The authors of the paper strongly suggest that there is a causative effect in the paper, and I think its very reasonable this is reflected in the title of the headline.

[edit] Even the title of the original paper is "Moderate alcohol consumption as risk factor for adverse brain outcomes and cognitive decline" which suggest causation. why should the headline be weaker than the title of the paper itself?

[epistasis]

In this setting, "risk factor" means "associated." It does not mean causes, and it really misinforms the public to jump to that. Even if the authors strongly suggest causation, but stop short, the most that should go in the headline is "Study strong suggests that moderate drinking causes.."

Jumping to conclusions in the popular press hurts the reputation of scientists. Journalists and editors have been doing this for years, especially on topics that sell papers, specifically what to eat and what to not eat, how to gain or lose weight, or how to live longer.

[gabrielgoh]

Fair point about the risk factor. I take the edit back!

But I think our disagreement is philosophical. Ultimately I feel much of medical research is prescriptive - and thus there is no misrepresentation, except perhaps amount of uncertainty that is associated with these studies.

I think hese uncertainties are poorly communicated by scientists themselves, and just appending "maybe" to every headline adds very little to the conversation.

[epistasis]

Yes, I agree that we probably disagree on philosophical grounds. I would definitely favor a more bland article of maybes than one that was more bold. Perhaps I'm one of those scientists that's bad at communicating, but the uncertainty is where there are discoveries to be made, and I tend to focus on it :)