Skip to main content

Posts

Showing posts from November, 2011

Why the new study on alcohol and breast cancer got it wrong - again

Big news! The latest study on the association between alcohol and breast cancer found what all the numerous prior studies using the same methods found: even small amounts of consumption increase the risk, regardless of the type of alcoholic beverage, even red wine. But as I point out in my book Age Gets Better with Wine, they are simply repeating the same mistakes and failing to see the big picture. Here’s why:
Self-reporting bias. Studies such as this, which seem to derive power from their large numbers, only magnify the errors if the data isn’t reliable. The nurses in this study were asked to fill out questionnaires on their drinking habits and other lifestyle factors every 6 months. It is widely acknowledged that this retrospective self-reporting is highly unreliable. So having a hundred thousand or even a million participants doesn’t yield stronger data, it just magnifies the error. Statisticians are of course aware of this and attempt to make adjustments according to known behavio…