Not All Transfers Are Created Equal: Towards a Framework Relating Transfer Connectivity to Travel Behaviour

Abstract

Walking from origins to transit stops, transferring between transit lines and walking from transit stops to destinations—all add to the burden of transit travel, sometimes to a very large degree. Transfers in particular can be stressful and/or time‐consuming for travellers, discouraging transit use. As such, transit facilities that reduce the burdens of walking, waiting and transferring can substantially increase transit system efficacy and use. In this paper, we argue that transit planning research on transit stops and stations, and transit planning practice frequently lack a clear conceptual framework relating transit waits and transfers with what we know about travel behaviour. Therefore, we draw on the concepts of transfer penalties and value of time in the travel behaviour/economics literature to develop a framework that situates transfer penalties within the total travel generalized costs of a transit trip. For example, value of time is important in relating actual time of waiting and walking to the perceived time of travel. We also draw on research to classify factors most important to users’ perspectives and travel behaviour—transfer costs, time scheduling and five transfer facility attributes: (1) access, (2) connection and reliability, (3) information, (4) amenities, and (5) security and safety. Using this framework, we seek to explicitly relate improvements of transfer stops/stations with components of transfer penalties and changes in travel behaviour (through a reduction in transfer penalties). We conclude that the employment of such a framework can help practitioners better apply the most effective improvements to transit stops and transfer facilities.

Acknowledgements

This study has been supported by grants from the California Department of Transportation. Special thanks to Mark Miller and Adina Ringler for comments and editing assistance. We would also like to thank the anonymous reviewers of this manuscript for their valuable comments and suggestions. Any errors or omissions are the responsibility of the authors and not of the funding agencies.

Notes

1. Other attributes of transfers are: seamlessness, flexibility, safety, security, comfort, convenience of both transferring and taking care of errands (e.g. buying a cup of coffee, a magazine and/or a newspaper), ease of payment, ease of vehicle access/egress, in‐vehicle time, seat availability, staff friendliness/helpfulness, familiarity of service, ease of comprehension, ease of finding out information and image of public transport (California Department of Transportation and California Division of Mass Transportation, 1985; Horowitz and Thompson, 1995; Department for Transport Local Government and the Regions (DTLR), 2000; Kajima Institute Publishing, 2002).

2. Liu et al. (1997) also state that a typical transit user in New York–New Jersey area in their study would “walk to a transit station, board a bus or the subway system, make one or more transfers, and finally walk to the destination”. Among US transit operators, 60% of all public transit trips are made without a transfer, about 30% involve one transfer, and about 10% involve two or more transfers (American Public Transportation Association, 2007). Given such data, an example of a transit trip involving one transfer could hardly be characterized as atypical, while not in the majority.

3. For the sake of elucidation, each of these attributes in Table 1 is assigned time that one might reasonably experience on a given transit trip: (1) 8 minutes, (2) 4 minutes, (3) 20 minutes, (4) 6 minutes, (5) 30 minutes and (6) 6 minutes. In practice, of course, these trip‐component times vary substantially from system‐to‐system, person‐to‐person and trip‐to‐trip.

4. TP_n and TP_b are equivalent to Interchange I and Interchange II, respectively, in Wardman’s (2001b) study.

5. This example does not include mode‐specific constant in Equation (1).

6. Most of the British studies reviewed by Wardman (2001a) were conducted using stated preference techniques, two‐thirds were suburban studies, and 7% were studies of inter‐urban travel. Balcombe et al. (2004) provide detailed data on this issue from a variety of reports.

7. By choosing to pay 75 cents to avoid an average of 5.3 minutes of additional waiting, Hess et al. (2005) estimated the value of waiting time for this group of students to be US$8.50 per hour.

8. Rapid developments in the sophistication and diffusion of information technologies present significant potential to improve transit travellers’ travel experiences. For example, accurate, real‐time vehicle/train arrival information (at stops and stations, and/or on cell phones) may substantially reduce travellers’ perceived wait times by increasing the certainty of transit service (Dziekan and Vermeulen, 2006).

9. In this paper, we cite reports published from the Transport and Road Research Laboratory (TRRL) and the Transport Research Laboratory (TRL), which are the same agency in the UK. In 1992, TRRL became TRL (Transport Research Laboratory, 2008).

10. For low‐frequency, longer headway services, passengers are much more likely to time their arrivals at stops and transfer facilities in order to reduce their wait times.

11. When service is relatively frequent (i.e. headways are 12 minutes or less), travellers are far less likely to time their arrivals at stops and stations according to transit service schedules. In such environments, passenger arrivals at stops are more or less random, so that half of the headway can be used to estimate the waiting time for relatively high‐frequency service.

12. This defines transfer penalties in the narrow sense discussed earlier.

13. Given the varying data sources used in these studies, the results are not directly comparable.

14. It is not clear why Wardman et al. (2001) estimate smaller transfer penalties in their stated preference survey study of Edinburgh and Glasgow in Scotland: 4.5 minutes in‐vehicle time for bus‐to‐bus transfers, 8.3 minutes for auto‐to‐bus transfers, and 8 minutes for rail‐to‐rail transfers.

15. Wardman (2001a) found in his meta‐analysis of value‐of‐time research that the studies published in the 1960s and 1970s largely used revealed preference data, while those published after the 1980s were predominantly conducted using stated preference data. Wardman also found that the values of wait time and walk time estimated in the revealed preference data were 143% and 46% larger (respectively) than those conducted with stated preference data. These differences may be due to the fact that stated preference respondents tended to ignore unrealistic walk and wait times and other complex travel alternatives (Wardman, 2001a). Given these differences, we argue that Guo and Wilson’s (2004) study is exceptionally well‐designed to collect and analyse revealed preference data.

16. Even the most comprehensive report on public transport demand by Balcombe et al. (2004) provide only the valuation of transfer‐related attributes, not elasticities. In this study, ‘[i]ndicative’ elasticities for in‐vehicle time, wait time and walk time are inferred and reported from the attribute valuation of these components, but not for transfers.

17. In addition to this study by Wardman et al., a few studies in the UK and Australia estimated monetary values per trip, instead of using equivalent values in in‐vehicle time, for bus stop/station characteristics by stated preference data analysis. These studies include Steer Davies Gleave (1996), MVA Consultancy (2000), ATOC (2002) and Hensher and Prioni (2002).

18. Given the importance of safety and security perceptions cited in many other studies, it is perhaps surprising that security variables, represented by closed‐circuit television and staff presence, were found to have lower values than information variables. This may reflect self‐selection bias in the sample because all of the survey respondents are current bus users who make transfers as part of their journey. Those who feel unsafe making such journeys are unlikely to make the transit trip and respond to the survey.

19. In the USA, recent work by Guo and Wilson (2004) is a notable exception, though their analysis focuses primarily on the characteristics of transit systems, and much less on the characteristics of the users. To address this significant gap in the literature, we are embarking on a substantial data collection exercise to allow us to estimate simultaneously how the characteristics of both environments and users affect the perceived out‐of‐vehicle burdens of transit travel.

20. We do not apply the confidence intervals of values in Wardman’s review since they would give us negative values on the lower end.

21. This paper primarily focused on the valuation of walk, wait and transfers within the concept of generalized costs of travel that can be used in a utility function of discrete choice model comparing different travel modes. In addition, combining the generalized costs of transit trip with the price elasticity of demand and ridership, we can estimate the change in ridership.