CHOICE-MODEL PREDICTIONS OF CAR-POOL DEMAND: METHODS AND RESULTS

The results of a number of car-pool strategies were predicted by using disaggregate choice models. Car pooling is explicitly considered as an alternative mode only for work trips. However, the effects of carpooling incentives on interdependent travel choices and vice versa are also predicted. Forecasts are made by applying the models to each household individually, using revised values of the appropriate independent variables to simulate the particular transportation alternative being analyzed. These household predictions are then summed to represent predicted areawide changes in travel behavior. Before and after data from the implementation of car-pooling incentives and transit-service improvements were used to test the validity of the model's forecasts. Three such tests are reported. The results indicate that the work-trip modal-choice model successfully captures the effects of changes in level of service on modal choice. The predicted effects of several significant car-pooling strategies are presented. In general, traveler response to many car-pooling incentives is small. The most significant changes in travel behavior are predicted for those parking-related policies that combine disincentives for driving alone with incentives for car pooling. /Author/