Sample and Response Bias Corrections of Transit Passengers’ Socioeconomic and Travel Characteristics Collected via Onboard Surveys: Sensitivity to the Boarding-to-Alighting Estimates used in Determining the Corrections
Corresponding Author: Rabi Mishalani, The Ohio State University
Presented By: Rabi Mishalani, The Ohio State University
Abstract
Transit agencies collect large-scale onboard surveys of passengers’ socioeconomic and travel (SE&T) characteristics on an infrequent but regular basis (approximately every five years). The surveys are subject to sample and response biases. Independently conducted passenger boarding-to-alighting (B2A) surveys can be used to correct for the biases, but such surveys are costly, intrusive, and time-consuming. However, the information in the B2A survey can be estimated using data already being collected from Automatic Passenger Count (APC) and Automatic Fare Collection (AFC) systems and conceivably used to correct the SE&T characteristics for sample and response biases.
A large-scale empirical study was conducted to investigate the sensitivity of the adjusted SE&T characteristics to different data sources used to estimate the B2A flows. Extensive survey data were collected on greater Columbus’s Central Ohio Transit Authority (COTA) bus routes, and available APC and AFC data were obtained and processed to estimate B2A passenger flows. The empirical results indicate that expensive and labor intensive B2A flow survey data collected at sample sizes typical of those seen in practice offer little, if any, systematic improvement in correcting SE&T characteristics, compared to using B2A flows estimated from APC and AFC data that are already being collected on an ongoing basis. That is, similar correction accuracy could be achieved at greatly reduced cost by using available automatically collected data.