Scientists and engineers in small firms are far more likely than their large firm counterparts to enter entrepreneurship. We label this phenomenon the small firm effect and explore its origins. In particular, we identify four classes of explanations for the small firm effect--preference sorting, ability sorting, opportunity cost, and the possibility that workers in small firms develop entrepreneurial human capital--and examine the empirical evidence for each. We find that preference sorting does play a role in generating the small firm effect: small firms attract those with prior preferences for autonomy who are similarly drawn into entrepreneurship. Similarly, ability sorting plays a role: those who ultimately become entrepreneurs may be d...