This issue will be particularly important in the coming decade when prisons, already largely filled to capacity, can expect significant growth in sentenced populations. Projections of anticipated growth over at least 20 years are also needed to estimate if long-range prison population growth warrants additional prison capacity. The paper discusses methods of projecting prison populations, organized by increasing complexity of the projection model: using the current year's population as baseline; multivariate regression analysis which invokes other variables known to have a causal relationship with prison population; projections based on demographic-specific incarceration rates; disaggregated flow models; and microsimulation models. It concludes that the last two approaches are most appropriate for policy impact estimates and describes four steps for developing an impact estimate: characterizing the subset of court cases to which the policy applies, translating policies into corresponding values of policy variables, formulating the behavioral model characterizing the court's response to the sentencing policy, and calculating the projected change in prison population resulting from responses to the changed policy. To demonstrate this methodology, the paper summarizes an impact analysis of a mandatory minimum sentence bill first considered by the Pennsylvania legislature in 1976. Tables, equations, 13 footnotes, and 13 references are supplied.