In this paper we posit that the lack of consensus about empirical tests of resource based view (RBV) could be the result of endogenous resource picking on the part of firms. If resources are endogenously selected, regression based methods that examine their connection to firm performance will be mis-estimated. We show that traditional remedies for endogeneity do not resolve this problem when returns to resources are heterogeneous (as theorized under RBV) and when managers act with at least partial knowledge of the expected, idiosyncratic return (as theorized under the strategic factor market hypotheses). As such, we develop a Bayesian approach that solves this endogeneity problem by directly incorporating resource picking into the modeling framework. We illustrate the validity of our approach through the use of a comprehensive simulation study and show that our proposed approach outperforms traditional linear models (including traditional cures of endogeneity and unobserved heterogeneity) under a variety of conditions. Our findings suggest that: (1) research in strategy requires a more careful and deeper understanding of potential sources of endogeneity and (2) the use of Bayesian methods in management can help develop more theoretically motivated empirical approaches to hypothesis testing.