Document Type

Conference Paper

Publication Date


Subject Area

place - australasia, mode - bus rapid transit, operations - frequency, operations - service span, ridership - drivers


ridership drivers, Bus Rapid Transit (BRT), Australia, service level, public transport infrastructure


Bus Rapid Transit (BRT) systems are an increasingly popular public transport option in Australia and internationally. They provide rail-like quality for bus services for a fraction of the cost of fixed rail. Many claims of high and increasing ridership have resulted from BRT system development; however it is unclear exactly which aspects of BRT system design drive this.
This paper undertakes an empirical analysis of factors influencing ridership on 77 BRT and non-BRT bus routes in Melbourne, Sydney, Adelaide and Brisbane. Explanatory variables considered included service level, frequency, speed, stop spacing, separate right of way share, vehicle accessibility, employment and residential density, car ownership levels and BRT infrastructure quality.
The paper reviews previous research associated with transit ridership at a route level and then presents the methodology and results.
Two multiple regression analyses were undertaken to explore the influence of the explanatory variables on ridership. The first considers overall ridership (boardings per route km, BRK) and identified a statistically significant model (R2=.81). The largest influence on BRK was vehicle trips per annum (β = .82), consistent with past research, followed by vehicle accessibility (low floor buses, β = .16) and population density (β = .14). The second considered patronage per vehicle kms (PVK) which explores ridership drivers after accounting for service levels. Results for this were statistically significant but with a less powerful model, adjusted R2 = .44. There were four explanatory variables including average speed (β = -.42), weekday frequency (β = .41), BRT infrastructure ranking (β = .29) and vehicle accessibility (β = .25). An alternative form of BRT infrastructure quality was also tested but did not improve the explanatory power of the modelling.
The paper concludes with a discussion of the various influences on ridership and recommendations for existing policy and future research associated with this field.