Inferential reasoning in policy analysis: knowledge use under uncertainty and complexity
本文提出一种诊断性政策分析方法，结合多种知识类型，通过迭代推理过程促进快速政策学习，帮助分析师在不确定和资源受限条件下更快得出可行结论。
Contemporary policy problems often involve calls for action in contexts marked by high levels of uncertainty and value-laden conflict. This poses a challenge to traditional approaches to policy analysis based on expectations of the availability of relatively uncontested norms and evidence and, often, more or less unlimited time. This article proposes a diagnostic approach to policy problems which combines multiple knowledge types (scientific, craft, practical, theoretical and intuitive) in an iterative inferential process of problem framing, analysis, deliberation and adaptation. This kind of inferential reasoning promotes rapid policy learning through better inference, data collection and use. Based on ‘abduction’, this approach allows analysts to work productively under knowledge, time and resource constraints by making provisional but transparent inferences which are then iteratively updated as new evidence emerges. Similar but less constrained than Bayesian approaches, such an orientation helps policy analysts arrive at actionable conclusions and recommendations faster and more accurately than traditional inductive or deductive methods.