Computational social science has seen a shift from theoretical models to empirical agent-based models (ABMs) grounded in data-driven correlations between behavioral factors defining agents’ decisions. The community’s interest in theoretical ABMs with behavioral theories setting stylized rules that guide agents’ actions is fading, especially when it concerns policy-related simulations. However, it remains unclear to what extent theory-driven ABMs mislead, if at all, possible policy decisions when compared to the outcomes of models with empirical micro-foundations. This is especially relevant for pro-environemntal policies that increasingly rely on quantifying cumulative effects of individual behavioral changes, where ABMs are so useful. To address this methodological dilemma, we propose a comparison framework to quantitatively explore the differences between theory- and data-driven ABMs. Inspired by the existing theory-based model, ORVin-T, which studies the individual choice between organic and conventional products, we designed a survey to collect data on individual preferences and purchasing decisions. We then used this extensive empirical microdata to build an empirical twin, ORVin-E, replacing the theoretical assumptions and secondary aggregated data used to parametrize agents’ decision strategies with empirical survey data. We compare the models in terms of key outputs, perform sensitivity analysis and three policy scenario.