Putting a price tag on social ills is a common way to garner media attention, and if the numbers are scary enough, they can have a real impact on public opinion — and by extension, on public policy.

For example, a recent British study (summarized by Join Together Nov. 2) deemed alcohol “the most harmful drug” in terms of societal impact: higher even than heroin or crack cocaine. The researchers were diligent, using numerous criteria and complex statistical models to come up with their numbers.

But can we really calculate the financial and social impact of human behavior, with so many potential confounders — age, economic status, geography, and plain old biology, to name just a few?

Maybe, said Carl Bialik, ?Numbers Guy? columnist at the Wall Street Journal, in his Nov. 6 column. “Sticker shock is the point of these studies, which attempt to classify societal ills in terms anyone can understand,” said Bialik.

“That doesn’t mean the studies lack merit,” he continued. “Crime and drugs that shorten life spans or diminish quality of life impose real costs in productivity and medical spending.”

But the numbers can be misleading. In the British study, for example, alcohol was found to be more dangerous than heroin or crack based on its wide use and association with drunk driving and crime.

However, tobacco takes first prize for killing more users than any other drug, including alcohol, said Bialik.

Bias in criteria selection can’t be ruled out in the British study, either. David Nutt, professor at Imperial College and organizer of the panel that assessed the harm data, is an outspoken critic of Great Britain’s tolerance for alcohol compared with other drugs.

In the end, there’s no arguing the power of numbers, even when what they represent is unclear. What is clear? When offered a price estimate on human behavior, let the buyer beware.