Statistical Decision in Forecasting Planning Data

Application of probability theory provides a rational and systematic approach to the problem of reliability prediction of such basic planning input data as population and economic forecasts. Furthermore, Bayesian decision statistics permit the utilization of a quantitative method for determining the best possible planning decisions with available information and planners' past experience and judgement. In this study the concept of reliability in land use-transportation planning is examined. As an example, the problem of county-level population forecasting in the southeastern Wisconsin region is considered. The reliability values associated with the population projections generated by different methods and for various counties are established. Also examined are the optimal number of different sets of projections to be considered for a county. The procedure outlined here can assist planners to arrive at an optimum level of design year population based on economic considerations, past experience and the amount and nature of information available.