Explanations, Predictions, and Weight of Evidence: Rigor with a Qualitative Approach?

How should we weigh complex sets of evidence?  In this talk, it will be argued that an explanatory approach to weight of evidence analysis provides a clear way to analyze and assess complex sets of evidence, a way which is both rigorous and accessible.  An explanatory approach utilizes explanations to account for sets of evidence, including divergent and/or multi-disciplinary evidence.  Explanations serve to conceptually structure evidence, and can thus provide testable accounts for why evidence looks the way it does, for why divergent evidence is present, and for why evidence is or is not relevant.  Divergent expert opinion can be explicated in terms of divergent explanatory accounts.  To prevent ad hoc explanatory moves and add rigor, explanatory accounts should be both checked for internal consistency and used to generate testable predictions.  The more central the explanation to the account, the more important it is to utilize it to generate a testable prediction.  Competing explanatory accounts, once developed and clarified, can then be compared with respect to the available evidence and their predictive success.  Thus, the explanatory approach can be made rigorous.  In addition, the explanatory approach may be preferrable in a policy context because it can be presented to a non-expert audience.  Finally, the talk will compare the explanatory approach to more quantitative approaches.  It will be argued that developing and clarifying the competing explanations is needed before a more formal method such a causal net (captured by an acyclic graph) can be applied.  The question then becomes what the quantitative approaches add to the explanatory approach in practice.