Transportation 19: 201-220, t992 1992 KluwerAcademic Publishers. Printed in the Netherlands. Development of a route level patronage forecasting method PETER R. STOPHER Louisiana Tran~ortation Research ~nte~ 4101 Gourrier Avenue, Bamn Rouge, ~ 708~, USA Accepted 20 January 1992 Key words: transit, operations, quick response, service, demand Abstract. The purpose of this paper is to describe the development of a model designed to predict changes to ridership resulting from small changes in the service provided by a given bus line. The paper describes the methodology used and rationale for the methodology; the manipula- tion of data required to produce the model and for use with the model; and the models themselves. A brief description is provided of some results from hypothetical service changes. Because actual changes and their consequent ridership changes were not available to the author, only the hypo- thetical resutts can be reported. The model was intended to be able to respond to the following types of service changes at a route level, by service period and route type; including each of changes in headways by service period, elimination of service in a time period, route extensions and new routes, route shortenings and short-lines, changes in the service period (i.e., addition or elimination of a trip within a service period to extend or shorten it), and the combined effect of any two or more such actions. The primary output from the model was designed to be the change in boardings on the route, and possibly on other routes that would be impacted through transfer patterns. Candidate alternatives Based on the literature review of previous efforts to construct route-level patronage forecasting models (Stopher & Mulhall 1992) and the requirements to develop a model that is readily used by service planners and other staff of a transit property, the potential methodology for this model was narrowed down to a choice between two alternative methods. The two alternatives that were considered are: An elasticity-based model * An econometric model based on socioeconomic and service data Each of these methods is described briefly in this section and the advan- tages, disadvantages, and data requirements are described.