Public debate and decision-making in our society suffer from the glorification of statistical data. After a particular social phenomenon, such as marihuana use, has been defined as a problem, armies of social scientific researchers set out to analyze and describe the problem. A sophisticated computer technology instantly translates millions of bits of data into correlations, probabilities and trends. The most striking findings are then fed to a data-hungry public. The result is data overload.

Descriptive information about the nature and scope of marihuana use as a behavior is an essential component of the policy-maker's knowledge-base. However, such information does not in itself have social policy implications. The policy-maker must define goals and evaluate means; only after he asks the right questions will statistical data suggest useful answers. Unfortunately, a tendency exists in the marihuana debate to assign prescriptive meanings to descriptive data without testing the underlying assumptions. Further, the data have often been accumulating in a fragmented way. No overall plan was devised beforehand; the result has been an ad hoc use of available data triggered by individual research interests rather than by long-term policy needs.

What does it mean that 24 million people have tried marihuana? Some have suggested that it means marihuana ought to be legalized. But does it mean the same thing if 15 million tried the drug once and have decided not to use it again? And does it mean the same thing if popular interest in the drug turns out to be a passing fancy, which wanes as suddenly as it waxed?

On the other side of the controversy, what does it mean that a substantial percentage of the public would favor increased penalties for marihuana use? The prescriptive implications of a democratic impulse may be offset by a preference for individual freedom of choice. Also, this segment of public opinion may have been influenced by incorrect information, such as unwarranted belief in marihuana's lethality or addiction potential. So, although the policy-maker must be aware of political realities, he must not allow his function to be supplanted by public opinion polls. This is an area which requires both awareness of public attitudes and willingness to assert leadership based on the best information available.