The 1987-88 Schools and Staffing Survey (SASS) looked at a national sample of elementary and secondary public and private schools. Components of the SASS, a data collection effort of the National Center for Education Statistics, were surveys of teacher demand and shortage, schools, administrators, and teachers. Approximately 13,000 schools and administrators, 65,000 teachers, and 5,600 local education agencies composed the SASS sample. This study was conducted to develop generalized variance functions that could be used to approximate the sampling error associated with an estimate of interest from the SASS. These generalized variances were designed for the user who does not have half-sample replication software available, but requires an approximation to the sampling error associated with his or her estimates of interest. A generalized variance function is a mathematical model describing the relationship between the variance or relative variance of a survey estimator and its expectation. The method for determining the generalized variance functions was tested in a pilot test, and generalized variance functions were then developed. The first appendix presents summary pilot test results, and the second contains tables of the generalized variance functions for the different survey components. (Contains 9 figures and 15 tables.) (SLD)