A single glass of wine or beer at the age of 14 can push a young teenager along the path to binge drinking, say scientists.
So begins the Daily Telegraph's report on some research into the psychological factors in adolescents that might predict whether they become "binge drinkers" at sixteen. The thing is that this isn't what the researchers are saying. They aren't arguing that if you allow your fourteen-year-old daughter a half glass of champagne on New Year's Eve she will become, in short order, a raging alcoholic. Yet that is what is being implied here.
The research (or rather what is being reported) tells us that the researchers have a jolly model based on a series of 'personality' factors that has a 70% chance of predicting that a given young person will be 'binge drinking' at age sixteen. Now my gut instinct is that, like lots of psychological metrics, the model is deeply flawed. However, if we accept what it is saying, it is still a pretty blunt instrument that will both fail to identify young people 'at risk' and also identify young people who aren't 'at risk'.
Finally, we are having all population solutions - don't let your child drink - proposed for what is clearly not an all population problem. Indeed, although the size of the whole study cohort (2,000) is given we have no idea how many of the young people surveyed actually drink. We know that over 80% of adolescents don't drink at all and that most of those who do drink, don't so so dangerously. So even with the study's broad definition of 'binge drinking' (based on self-reported drunkenness) the numbers involved are likely to be very small. The certainty of the findings has to be questioned given these numbers (plus the likelihood of young people overstating consumption).
It is useful to study adolescent development and, clearly, risk-taking behaviour could be a predictor of many things (some good like being future innovators, creators and entrepreneurs) but to argue that giving a child a drink will "push" them into binge drinking is a complete misrepresentation of the research.