DEVELOPMENT OF TIME SERIES BASED TRANSIT PATRONAGE MODELS. EXECUTIVE SUMMARY

This document is a summary of research addressing the use of time series models to both evaluate service and fare policy changes by public transit operators, and forecast transit system ridership at the route, service, and system level. The data used in model development and testing is from Portland, Oregon and covers the period 1971 through 1982. Models are developed at the system, sector and route levels, and are used to assess the impacts of past changes in service level and fare, as well as to forecast future transit patronage. The statistical approach used was developed by Box and Jenkins for time-series data and is therefore more appropriate and powerful than the more traditional regression analysis. Of particular interest is the identification of the lag structures and functional forms that constitute the relationships between transit ridership, level of service, travel costs, and market size.