As part of an effort to reassess its enrollment forecasting methodology and techniques, the Florida Division of Community Colleges conducted a national survey of state directors of community/junior colleges to identify successful forecasting practices. A conceptual framework for analyzing responses was developed from a review of the literature on enrollment projections and used to summarize the state of the art of enrollment forecasting. Four categories of approaches were identified: (1) trend extrapolation approaches, i.e., curve-fitting techniques, such as simple and moving averages, which utilize historical enrollment data as the basis for projections, and causal models, such as cohort-survival techniques, which give consideration to independent factors in addition to historic enrollment data; (2) subjective judgments, which involve the use of expert opinion or futuristic approaches; (3) combination approaches, which use both extrapolation and subjective methods; and (4) no state-level forecasting. The survey revealed that nine of the 46 responding states conducted no systematic state-level forecasting; 12 states used curve-fitting techniques; 18 states used causal models; one state based its enrollment projections solely on subjective judgment; and six states reported using combination approaches. The survey report details methodology and findings by state and concludes with observations on responsibility for, success of, and desirability of enrollment forecasting. (AYC)