Historiography taking issue: Analyzing an experiment with heroin abusers

This article discusses the predicament of historians becoming part of the history they are investigating and illustrates the issue in a particular case. The case is that of the randomized controlled trial (RCT)—more specifically, its use for testing the effects of providing heroin to severe heroin abusers. I counter the established view of the RCT as a matter of timeless logic and argue that this research design was developed in the context of administrative knowledge making under twentieth‐century economic liberalism of which it epitomizes some central values. I also argue that the applicability of the RCT depends on the degree to which its advocates can define the issue to be studied according to its inherent values. Next, I demonstrate how advocates of an RCT with heroin provision in the Netherlands steered the political discussion on heroin provision and how the values of economic liberalism also shaped the results of the Dutch maintenance experiment. In addition, I relate how my analysis of this experiment became part of political debates in the Netherlands. Contrary to my intentions, adversaries of heroin maintenance used my critique on the heroin RCT as an argument against heroin maintenance. Such risks are inherent to historiography and sociology of science aiming at practical relevance while challenging treasured scientific beliefs. I conclude that it still seems better to expose arguments on unjustified certainties than to suppress them for strategic reasons. © 2004 Wiley Periodicals, Inc.