You would like to test the experience curve (i.e. learning-by-doing) hypothesis: as physicians gain experience, their quality improves (where quality is measured by risk-adjusted outcomes.) You have cross-sectional data showing that physicians with higher volumes have better risk-adjusted outcomes. While the data are consistent with the experience curve hypothesis, they allow another interpretation that is quite different. Briefly offer a competing hypothesis that is consistent with the data.1 (Hint: consider the possibility of reverse causality.)