Resolving empirical controversies with mechanistic evidence

Abstract The results of econometric modeling are fragile in the sense that minor changes estimation techniques or sample can lead to statistical models support inconsistent causal hypotheses. fragility undermines making conclusive inferences from empirical literature. I argue program evidential pluralism, which originated context medicine and encapsulates normative reading Russo-Williamson Thesis claims need both difference-making mechanistic evidence, offers a ground for resolving disagreements. analyze recent controversy regarding tax elasticity cigarette consumption smoking intensity. Both studies apply plausible but report results. show evidence allows discriminating representing genuine relations accidental dependencies data. Furthermore, discuss differences between biological social mechanisms across disciplines. economists mainly rely on mathematical represent possible (i.e., could produce phenomenon interest). Still, claiming actuality represented requires establishing crucial assumptions these descriptively adequate. exemplify my approach assessing quality economics with an analysis two rational addiction.