In Evidence-Based Policy: A Realist Perspective, Ray Pawson argues that current methods for synthesizing research fail to account for the true complexity of social interventions. He critiques the meta-analytic approach for treating programs as simple medical dosages, which often results in systematic obfuscation by ignoring the human reasoning and contextual factors that drive change. Pawson asserts that social initiatives are actually theories incarnate that unfold through long, active implementation chains rather than through standardized, reproducible treatments. Consequently, he proposes a realist synthesis that moves beyond the traditional hierarchy of evidence to focus on "what works for whom in what circumstances." This methodology seeks to unearth the generative mechanisms and specific conditions that cause a policy to succeed or fail in different environments. Ultimately, the text provides a blueprint for an accumulating explanation of policy effectiveness by prioritizing theoretical understanding over simple statistical averages.