Why do experiments report different critical exponents for the Anderson metal-insulator transition in doped semiconductors? Simulations of sulfur-doped silicon point to a key factor: hybridization between impurity and conduction bands near the Fermi energy.