Directors of Research and Evaluation in school districts were surveyed about their methods in making school enrollment projections. The questionnaire covered school district size, minority enrollment, staff responsibility for performing projections, acceptable levels of accuracy, methodology used, special considerations, and use of the data. Fifty responses were received, primarily from school districts of over 30,000 students. The official with primary responsibility for enrollment projections was commonly employed in the department of planning or management information services, research and evaluation, or student accounting. Principals and other administrators also had input into the process. Two-thirds of the respondents aimed for a district-wide enrollment projection error rate of one percent or less. Error rates for individuals grades and for time periods beyond one year were higher. The data were primarily used for budgeting, personnel planning, building utilization, transportation, and capital improvement planning. Two approaches were generally used: (1) projection, using historical data on birth and enrollment trends; and (2) prediction, which combines this data with other variables thought to influence enrollment, often using a multiple regression equation. (This paper presents a detailed model for enrollment projection, and the questionnaire, The Survey on Enrollment Projections). (GDC)