How Evidence Alters Understanding and Decisions

have been gradually developed to portray the idealized versions of each kind of use. One of these is the instrumental model. Under this model, consists of making a decision, and social science evidence is assumed to be instructive to the decision. Though there are variations on this theme, the model usually assumes that once the evidence is available, the decision is relatively straightforward. The instrumental model has recently become something of a straw man in discussions of social problem solving, as observers have pointed out deviations from the model on such things as the sequence of events in social problem solving (Rein & White, 1977), the ability of social science to clarify issues (Cohen & Weiss, 1977), the need for findings from social science to be confirmed by ordinary knowledge (Lindblom & Cohen, 1979), the procedural ambiguities inherent in group decisionmaking (March & Olsen, 1979), and the tendency of policies to change incrementally rather than abruptly (Weiss, 1980). The other model of use could be called the conceptual model. Although it does not account for decisions, it does account for a good deal of social science use (Rich, 1977; Weiss, 1977). Under this model, evidence is not specifically instructive, but it is nevertheless relevant, and consists of thinking about the evidence. The most important difference between the conceptual model and the instrumental model is this: Whereas the central feature of the instrumental model is the decision, the central feature of the conceptual model is the human information processor, and that fact leads to several other important distinctions between the models. For instance, human information processors are assumed to already have a considerable body of knowledge before receiving the evidence. They approach the evidence with a well developed, internalized model of the substantive issues at hand. If participants work primarily in education, for instance, they may approach new evidence with knowledge about child development and the learning processes; about how teachers teach and how to tell a good one from a bad one; about the different kinds of effects that different kinds of policies or programs are likely to have and why; and about how well the administrators of existing educational programs are doing their jobs. The presence of this knowledge is essential to the conceptual model. Because the evidence is no longer instructive to a decision, as it was in the instrumental model, it must have a bearing on something else. In the conceptual model, it influences the user's working knowledge of the issues at hand. A second important difference between these two models has to do with the authority that the evidence may have: Whereas a decison may follow almost automatically from the instructions contained in the evidence, information pro-