Predicting Transit Ridership in Response to Transit Service Changes

Abstract

This paper reports the development of a simplified form of the multinomial logit model and its application to the prediction of travel mode shares for a range of transit service changes. An earlier version of this model, the incremental logit model, can be applied to the prediction of transit ridership changes in response to improvement in an existing transit mode. This paper describes an extended version of the incremental logit model which can be used when a new transit service is introduced. Further, this extension is formulated so that it may be applied with the nested logit model as well as the more commonly used non‐nested model. The extended incremental logit model provides the capability to predict the ridership impact of transit introduction or service changes using only information on existing mode shares and changes in transit service. A simple example is presented to illustrate the application of this model.