No matter our growing sophistication about the manifold ways in which social and political influences penetrate and shape what is seen as scientific inquiry, the ideas of the policy science establishment seem fixed in an earlier era of unbridled confidence that science is science, even social science is science, and politics is politics. As for how science should be used in the formulation of social policy, the view is easy to summarize, based as it is on a simple model of rational decisionmaking. Presumably the objective is to alter some condition deemed a social problem, whether poverty, or insecurity, or dependency, or social disorganization. That objective is determined not by policy science, but by the political process. Policy science rather has to do with identifying the key variables or chains of variables that bear on this politically determined objective, and with identifying and measuring the effects of manipulating those variables. The effort is of course susceptible to great complication, as researchers try to cope with multiple intervening and interactive variables and with difficult problems of measurement. But complexities aside, the social scientist uses multivariate analysis in the effort to identify and measure ostensible cause-and-effect relations which are then the basis for scientifically informed interventions by government to alleviate social problems.

This dominant model in turn generated a dominant critique. For several decades, thoughtful social thinkers have agreed that this view of the uses of social science rested on an impossibly simplified view of real-world policy decisionmaking.