Reductions in the funding of compensatory education along with observed discrepancies in per-pupil cost among equally effective schools created a compelling case for the study of the determinants of cost-effective Chapter 1 programs in the state of Minnesota. Cost-effectiveness (CE) indexes for Chapter 1 reading and math services reflected a ratio between a district's weighted mean normal curve equivalent gain and the average per-pupil cost. A random sample of 141 Chapter 1 programs participated in the study with a database consisting of a merged file of fiscal and classroom variables. Multiple regression analysis was used in a cross-validation design to develop CE models predictive of Chapter 1 reading and math programs. The final set of predictors was conceived as two composites: the first set largely under the control of state education agency administrators and the second under local education agency program managers. The first composite is a district size effect: Chapter 1 projects in larger districts are more cost-effective than those in smaller districts. The primary implication of this finding points to the need to consolidate smaller Chapter 1 projects into larger districts. The second composite shows cost-effective practices and policies found in larger Chapter 1 projects. These include teacher aide programs, shorter time blocks for math instruction, and an emphasis on small group, inclass instruction. Five technical appendixes are included: expanded regression models, regression summary tables, correlation matrix, T-test tables, and cross-validation tables. (Author/TE)