Separating generalizable from source-specific preference heterogeneity in the fusion of revealed and stated preferences

Preference heterogeneity is one of the central behavioral concepts in applied econometrics. Its centrality particularly evident choice modeling literature, notably its widespread application to environmental and health economics, marketing, transport. Despite conceptual empirical advances preference heterogeneity, generalizability different decision contexts data generation processes remains an open question. The basic premise this paper that latent sources can be decomposed into components general others specific them. We study structure with goal reliably identifying common (presumably generalizable) not heterogeneity. contribution both methodological, leading testing five rival model specifications which together elucidate present two same behavior. In application, we find multitrait-multimethod has best fit most sensible interpretations, indicating while each source contributes uniquely certain components, contribute also (generalizable) Recognition separability versus source-specific will lead more reliable accurate demand forecasts assessments welfare impacts.