Simplified linear programming approach to the estimation of enrollment transition rates : estimating rates with minimal data availability

Unfortunately, the traditional methods of estimating these rates are fraught with serious problems. This paper suggests an alternative method, one which has minimal data requirements and can be quickly computed using a microcomputer. The proposed method requires neither exact equality in the mathematical relations, as in the simultanenous equation option, nor a long time series of data, as in the least-squares option, nor the shaky assumption that transition rates do not change over the longer period. It does not require data on enrollment by age and grade, which are not available in many countries, and assumptions can be made as restrictive or relaxed as one feels comfortable with. Only 3 years of consecutive enrollment data and 2 years of data on population of school entrance age are required. There are two disadvantages, however. The first is methodological: rates that seem unstable from grade to grade may occasionally be produced. The second is that the method is impossible to apply by hand; it requires access to a microcomputer and linear programming software.