Barnes, Deborah M. "Drugs: Running the Numbers," Science 240 (24 June 1988), pp. 1729-


Barnes, Deborah M. "Drugs: Running the Numbers," Science 240 (24 June 1988), pp. 1729-

There have been caveats throughout this bibliography concerning the weaknesses and the
limitations of estimates of just about everything: the size of the corn crop, the rate of
unemployment, the trade deficit, and so on. These numbers are all "soft" and based on inadequate
data. (And note, there is no one trying to hide the size of the corn crop, unemployment, or the
trade conditions of this country.) On the other hand, when it comes to estimating crime rates,
drug use, abused children, and that sort of thing, the "softness" of the data becomes almost

Here is an article concerning the estimates given Congress by federal bureaucrats who deal with
drugs. Congress would like to know how big the problem is and if the problem has been
changing though time and how successful the federal government has been in its efforts to deal
with the problem. Congress has listened to the bureaucrats who try to provide the best figures
they can, but how good are those figures? This article suggests that "some may not be valid and
others reflect an inappropriate use of raw data." (p. 1729) Congress was recently told that 10.6
million Americans are alcoholics and another 7.3 million abuse the drug. That seems
straightforward; those are numbers. But where did they come from? They derive from a 1979
national survey and they are extrapolations from the survey data. One problem is that not
everyone knows what an alcoholic is. The problems with a definition of "alcoholism" are
enormous. Then, too, "The validity of the survy data is at issue because people respond to
questions from their own memories and willingness to be candid." (p. 1729) This article notes
that of all the substances which are abused, alcohol is the most intensely studied and yet experts
refer to the 10.6 million estimate with such disclaimers as "I don't know where they get their
figures from." (p. 1079) Such extrapolations are risky numbers and can be used for political

A serious point is to be made concerning the standardized definition of alcohol abuse. The
Diagnostic and Statistical Manual compiled by the American Psychological Association has not
been consistent in its definition of alcohol abusers. How are researchers to project estimates of
their surveys onto the inconsistent categorizations of official diagnoses? If one compares the
official definition of the APA with the official definition of the bureaucrats charged with dealing
with alcoholism, further confusion immediately develops. And alcohol is the most intensively
studied drug and it is not illegal. When one gets around to the illegal drugs, the complications

The governmment does survey illegal drug users but all data are based on self-reports of use. And
there are other difficulties as well. For example, the survey did not include homeless and
institutionalized people, who probably have a higher incidence of drug use. Also, government
routinely analyzes high school students concerning drug use, but dropouts are not included in
such surveys as they are no longer in school! Still these data get used because they are the only
data available and it is better to use them than have nothing at all - or so say the statisticians.

Another statistic put frequently before a gullible public refers to the dollar costs of drug abuse.
The government uses the figure of $150 billion for 1988. Where did that come from? Note
should be made that almost half of the estimate derives from off-the-cuff estimates of "lost
productivity," which seems a fun and games statistic.

"‘If I had my preferences, I would avoid having the figures sound spuriously exact,'" is the way
the scientific director of the Alcohol Research Group put it. That is a reasonable statement. But
the final paragraph suggests the nub of the problem: "‘Another difficulty is the way the political
system handles the numbers,' says Room," (the previously quoted director). "‘Typically,
government reports and policy statements begin with a litany of numbers, usually the highest
credible numbers that can be extrapolated from research data.'"(p. 1731) That is just it: these
numbers are used for political purposes. Science is used for political purposes. It is, in our
society, reason and justification for what the bureaucrat wants to do.