Estimating Ridership of Demand Response Transit Services – Scope

(Center Identification Number: 79060-02-A)

Principal Investigator
Jeremy Mattson

Project Scope
Few studies have been completed that provide a method for estimating ridership for demand-response transit services. Lack of data for demand-response service characteristics and geographic coverage has limited the estimation of such models. TCRP Report 161 {2013) provides methods for forecasting demand for rural passenger transportation, including general demand-response services. The demand model developed in this report includes some service characteristics, such as size of service area and service miles, but it lacks other service characteristics. Factors such as fares, span of service, reservation requirements, and other service characteristics will likely impact ridership. These data are not included in the rural National Transit Database (NTD), but recently collected data from demand-response transit providers in North Dakota and Florida provide detail about these service characteristics, as well as geographic coverage. These data, when combined with rural NTD data and population and demographic data from the U.S. Census and the American Community Survey (ACS) allow for the estimation of a more detailed demand model. Results could be used to forecast demand for new demand-response services; estimate the impact of service changes, such as changes in geographic coverage, span of service, fares, etc.; and project future ridership based on projected population and demographic changes.

Tasks/ Deliverables
1) Literature review
Previous research on demand for transit services in general and demand-response services specifically will be reviewed.

2) Develop Methodology
A method will be developed to estimate demand for demand-response transit. Ridership will be estimated as a function of service characteristics and service-area characteristics. Service characteristics include span of service, fares, service type, reservation requirements, eligibility, and potentially other factors. Service-area characteristics include population served and demographic and geographic characteristics of the area served. Demographic variables of interest include the population of older adults, people with disabilities, and households having no vehicle. Geographic characteristics of interest include population density, travel distances, and whether the service area is largely rural, urban, or suburban.

3) Collect Data
Service data will be collected from the rural NTD and previously conducted surveys of transit agencies in North Dakota and Florida. Population, demographic, and geographic data will be collected from the U.S. Census and the ACS.

4) Estimate Demand Model
A regression model will be estimated in SAS. Ridership will be estimated as a function of service and service-area characteristics. One. model will be estimated using the full set of data from the rural NTD, and a second model will be estimated using more detailed data collected from transit agencies in North Dakota and Florida.

5) Evaluate Model Results
The performance of the models in predicting ridership will be compared to that of other models published in the literature. The model will be used to predict ridership for a selection of demand­response transit agencies around the country, for which the necessary data will be collected. Ridership for these agencies will also be predicted using other published demand models. The accuracy of the models developed in this study will be compared to that of previous models to evaluate its performance.

6) Prepare Final Report
A full technical report detailing the methods and results will be published, as well as a short, non­technical summary. The result can be used as a tool for estimating ridership for demand-response services. Results will be presented at conferences and submitted for journal publication.

Jeremy Mattson
Ranjit Godavarthy
Jill Hough

Project Schedule
June 2015 and conclude by March 31, 2016

Project Budget
Total Project Cost     $38,366


Leave a Reply

Please complete the simple math problem prior to submitting your comment. This reduces spam. Thanks! * Time limit is exhausted. Please reload CAPTCHA.

This site uses Akismet to reduce spam. Learn how your comment data is processed.