This study analyzes the policy process through which two states determined education beneficiary programs of newly adopted state lotteries. Tennessee, in 2003, followed the regional pattern of allocating all lottery proceeds to merit-based college scholarships. North Carolina, in 2005, bucked this trend by allotting no lottery revenue for merit aid. This comparative case study draws on extensive interview and archival data to test the explanatory power of a revised multiple streams model. Our findings indicate that both stable characteristics, such as government structure and intra-state policy trends, and anarchical influences, such as the timing and strategies of policy entrepreneurs, matter most. (Contains 1 figure and 3 footnotes.)