The common reporting of effect sizes has been an important advance in education research in recent years. However, the benchmarks used to interpret the size of these effects--as small, medium, and large--do little to inform educational administration and policy making because they do not account for program costs. The author proposes an approach to establishing cost-effectiveness benchmarks rooted in an explicit economics-based decision-making framework and assumptions about the decision-making context. To be considered large, the ratio of effects to costs must be at least as large as the ratios for substitute interventions. Evidence related to class size, prekindergarten, and other interventions is discussed to illustrate the calculation of the cost-effectiveness ratios, how the evidence can be used to develop benchmarks, and how the benchmarks can be useful for researchers and policy makers. The development of benchmarks is intended to encourage cost-effectiveness analysis as a standard part of policy analysis, thereby providing more evidence to increase the validity of the benchmarks and, ultimately, improving policy decisions. Recent cost-effectiveness research in health care policy illustrates the potential value of cost-effectiveness benchmarks in education. (Contains 26 notes and 5 tables.)