Prison population is the product of two factors: prison admissions and length of stay. Overall the current projection model has thus far produced accurate short-term (up to two years) estimates, but inaccurate projections thereafter. Projections beyond the 3-year period have error rates in the 8 to 13 percent range and in no consistent direction. In addition, there are several factors associated with the current base projection which need to be modified, including exclusion of jail credits factor, assumption of constant rate of new admissions, and sensitivity of length of stay equations to data errors. Collectively these factors, unless modified, will cause the current SCDC base projection beyond three years to be inaccurate. Recommendations include adjusting the new admissions assumption using a desegregated demographic method, verifying the accuracy of cases classified as first and second degree burglaries, and verifying the accuracy of cases classified as ineligible for parole and/or receiving five-year concurrent sentencing enhancements. 2 references.