T. Dehue

By this author

Testing policies: On the history and assumptions of «social experimentation» (translated from English by Irina Sirotkina)

Issue 4 from 01.12.2001

According to this article, the contemporary definition of a “true” experiment in social sciences as a random group experiment is not derived from an alleged transcendental logic of science. Neither does it stem from any research lab. The author argues that social experimentation, and particularly the random group experiment, epitomizes the values of 20th-century Western bureaucracies and takes them to their extreme. Originally the word “experiment” was introduced into social thought as a natural science metaphor, but the central meaning of the word in contemporary social science exemplifies the 20th-century administrative aspiration of ruling by technique rather than tradition, of replacing the individuality of both the governors and the governed by impersonality.

This point is illustarted by a recent Dutch social experiment with heroin addicts. In spring 1997, the Dutch minister of Health wanted to establish the effects of handing out free heroin to incurable abusers. More specifically, she wanted to find out whether the bad condition and social misbehavior of severe heroin addicts is caused primarily by the heroin itself or by its high price on the illegal market. She argued before parliament that, for statistical reasons, at least 750 severely addicted participants would be needed. Moreover, she explained that the project should be conducted according to the exemplary experimental design in which experimental groups are compared to control groups. In order to ensure comparability and valid statistical inference the 750 participants should be assigned randomly to either the experimental group that would receive free heroin or the control group that would only get the unpopular substitute drug methadone.” During half a year both the experimental group and the control group would be extensively screened as to differences in their health and behavior.The minister finally succeeded. But the currently running project has already taught something quite unexpected. Rather than letting themselves be willingly monitored in exchange of free heroin (as the minister expected), rather than besieging the heroin post (as the conservatives anticipated), quite some junkies quit at an early stage, for, as they explained, “this heroin does not taste good.” The failure of the experiment shows that understanding addiction demands more than measuring physical and psychological characteristics. And, most importantly, it reveals the arbitrariness of the core assumption of social experimentation which is the taken for granted supposition that regulated society is everybody's best of all possible worlds.

82 0