Research published in the British Medical Journal earlier this week said that moderate drinking was only associated with reduced mortality for women over the age of 65, with "little to no protection present in other age-sex groups". This finding was reported (accurately) by the Daily Mail and the Times.
But statistician David Spiegelhalter has raised questions about the research, saying that the number of non-drinkers in the study was too small to make this conclusion.
In this case deaths of moderate drinkers were compared to those of people who said they'd never been drinkers. For example, there were 36 deaths of never-drinkers under 65, a comparatively low number. That means that differences between them and the moderate drinkers would have to be very large in order for us to be confident that drinking had a protective effect.
Spiegelhalter graphed some of the figures and found that, "All the observed data are compatible with the kind of 10 to 20% protection that has been previously suggested."
While the results were compatible with a small beneficial effect, the research can't give us a definitive answer over whether or not drinking moderately is beneficial.