A study was conducted to (1) assess the relation of different CETA management decisions at the prime sponsor level to program goal achievement and (2) assess the conditions under which specific management decisions seem most likely to maximize goal achievement. Members of a research team visited 15 prime sponsorships throughout the United States to collect data on seven different external local conditions and systematically investigate eight areas of local management decisions. Seventeen Ohio prime sponsorships for which there was detailed data available were used as a comparison group. The conclusions were that (1) CETA at the local level is not a highly constrained system; (2) local programmatic choices are not determined by factors such as unemployment, the history of pre-CETA manpower programs, demographic composition of the community, the nature of persons served, or the activities of the Department of Labor; and (3) program performance is significantly affected by factors over which local managers have considerable control; program mix; programmatic priorities; and management decisions in the areas of staff quality, data collection and use, monitoring and evaluation of programs, degree of subcontracting for service delivery, use of requests-for-proposal for subcontracting, open decision making centered in the advisory council, and conflict management strategies. (EM)