ISSN 1055- 1425
January 2009
This work was performed as part of the California PATH Program of the
University of California, in cooperation with the State of California Business,
Transportation, and Housing Agency, Department of Transportation, and the
United States Department of Transportation, Federal Highway Administration.
The contents of this report reflect the views of the authors who are responsible
for the facts and the accuracy of the data presented herein. The contents do not
necessarily reflect the official views or policies of the State of California. This
report does not constitute a standard, specification, or regulation.
Report for RTA- 65A0194- 15762
CALIFORNIA PATH PROGRAM
INSTITUTE OF TRANSPORTATION STUDIES
UNIVERSITY OF CALIFORNIA, BERKELEY
Thinking Outside the Bus: Understanding
User Perceptions of Waiting and Transferring
in Order to Increase Transit Use
UCB- ITS- PRR- 2009- 8
California PATH Research Report
Brian D. Taylor, Hiroyuki Iseki,
Mark A. Miller, Michael Smart
CALIFORNIA PARTNERS FOR ADVANCED TRANSIT AND HIGHWAYS
Thinking Outside the Bus: Understanding
User Perceptions of Waiting and
Transferring in Order to Increase Transit
Use
Brian D. Taylor
Hiroyuki Iseki
Mark A. Miller
Michael Smart
Final Report for RTA- 65A0194- 15762
January 16, 2009
i
Preface
This document is the Final Report for the Research Technical Agreement ( RTA) between the
California Department of Transportation ( Caltrans) and the University of California at Los
Angeles ( UCLA). The RTA is entitled “ Tool Development to Evaluate the Performance of
Intermodal Connectivity ( EPIC) to Improve Public Transportation”. Caltrans’ primary interest in
this research was interconnectivity among transportation modes in California and the
development of a methodology to evaluate connectivity performance, which could provide a new
and needed tool to improve passenger transit trips.
This project was a collaborative effort between UCLA and the University of California at
Berkeley ( UC Berkeley). The overall project Principal Investigator was Professor Brian Taylor at
UCLA, and Professor Samer Madanat served as the Principal Investigator for UC Berkeley. Mr.
Mark Miller was the Project Manager working with Dr. Hiroyuki Iseki of the University of
Toledo; at the start of the project Dr. Iseki was a Post- Doctoral Researcher at UCLA.
Additionally, two Graduate Research Students at UCLA, Mr. Michael Smart and Ms. Adina
Ringler, were members of the project team. Professor Taylor provided overall technical guidance
and support to the project team for all project tasks. In addition to managing the project, Mr.
Miller conducted research in the areas of reviewing the literature, designing and administering
both project surveys and the institutional interview guide, and documenting research findings.
Dr. Iseki developed the transfer penalties/ travel behavior conceptual framework as part of his
review of the literature; he also worked on designing the transit passenger survey and analyzing
its responses as well as documenting its findings. Mr. Smart worked on designing and
administering both project surveys and the institutional interview guide, analyzing responses to
the transit operators survey and the institutional interview guide, and documenting their findings.
Ms. Ringler worked on designing and administering the transit passenger survey, analyzing its
responses, and documenting its findings. Additional information about the four authors of this
report is provided in the “ About the Authors” section of this report.
The two key products of this research are:
- The transfer penalties/ travel behavior conceptual framework, which was based on our
review of the state- of- the- practice for evaluating intermodal and intramodal connectivity, and
- The preliminary transit connectivity assessment tool
The framework allowed us to consider various attributes of transit stops, stations, and transfer
facilities and guide us in our subsequent analysis of user perceptions of walking, waiting, and
transferring experiences. Our research findings, especially the preliminary Assessment Tool,
have taken substantive steps toward determining the connectivity of transit systems, its
influences on travelers’ satisfaction with transit services, and ways that public transit systems can
reduce the burdens of out- of- vehicle “ travel” times to help make public transit more attractive
resulting in ridership increases.
This Final Report has integrated each of our project’s components into a cohesive product
documenting the significance of transit connectivity’s contribution toward increasing transit
ii
usage.
In this report, we describe transit trips made with transfers, the types of transfer venues, and
transit connectivity. We follow this with a discussion of our transfer burdens/ travel behavior
conceptual framework. After this we discuss the three types of stakeholders we focused on in our
assessment of transit stops, stations, and transfer facilities; next we present the methodological
approach we employed in this assessment. We then discuss our findings together with
presentation of our Attribute Assessment Tool. Finally, we discuss next steps for this line of
research.
iii
Acknowledgements
This work was performed under the sponsorship of the State of California Business,
Transportation and Housing Agency, Department of Transportation ( Caltrans), Division of
Research and Innovation ( DR& I) ( Interagency Agreement # 65A0194). The contents of this
paper reflect the views of the authors, who are responsible for the facts and the accuracy of the
data presented herein. The contents do not necessarily reflect the official views or policies of the
State of California. The authors thank Mr. Bruce Chapman of DR& I who managed this project
for Caltrans. The authors also thank Ms. Adina Ringler, a former Graduate Student Researcher
for this project at the University of California, Los Angeles, for her significant contributions to
this research. The authors would also like to thank former UCLA student Mr. Ting Sit who
helped design and administer the survey, and collate and analyze the data; UCLA students Mr.
Syed- Abrar Ahmed, Ms. Lanka Ranasinghe, and Ms. Karla Vasquez who helped administer the
survey; UCLA student Ms. Vanessa Fernandez who translated the survey from English into
Spanish; Los Angeles County Metropolitan Transportation Authority ( Metro), Metrolink, Santa
Monica Big Blue Bus, and Culver CityBus for allowing us to survey their passengers; the nearly
750 Los Angeles County transit riders and nearly 200 transit managers representing agencies
around the U. S. who took time out of their busy schedules to give us their thoughts and views on
how to make waiting for transit a less onerous experience; Ms. Michelle Tse and Ms. Rowena
Barlow in the Business Office of the UCLA School of Public Affairs and Mr. Ken Castro- Oistad
and Ms. Virginia Anders of the UCLA Office of Contracts and Grants Administration for their
administrative support of this project.
iv
ABSTRACT
This report presents the results of its research of interconnectivity among transportation modes in
California and the development of a methodology to evaluate connectivity performance, which
could provide a new and needed tool to improve passenger transit trips. The two key products of
this research are the transfer penalties/ travel behavior conceptual framework, which was based
on our review of the state- of- the- practice for evaluating intermodal and intramodal connectivity,
and a preliminary transit connectivity assessment tool. The framework allowed us to consider
various attributes of transit stops, stations, and transfer facilities and guide us in our subsequent
analysis of user perceptions of walking, waiting, and transferring experiences. Our research
findings have taken substantive steps toward determining the connectivity of transit systems, its
influences on travelers’ satisfaction with transit services, and ways that public transit systems can
reduce the burdens of out- of- vehicle “ travel” times to help make public transit more attractive
resulting in ridership increases. In our research to learn more about how wait times at stations
and stops are perceived, and how they can be made better, we surveyed approximately 750
passengers at stops and stations in Los Angeles County, as well as 175 transit operators
nationwide. From our analysis of the passengers/ users perspective, one principal finding stands
out clearly:
The most important determinant of user satisfaction with a transit stop or station is frequent,
reliable service in an environment of personal safety, and only indirectly the physical
characteristics of that stop or station.
Our principal finding from our analysis of transit managers perspective precisely matches that of
the transit user investigation:
For operators, safety- and security- related factors far outweighed other
attribute factors at transit stops, stations, and transfer facilities.
This report further develops a Preliminary Assessment Tool that transit operators can use to
guide their efforts at improving existing transit stops and stations, or in developing plans for new
facilities. The Preliminary Assessment Tool, sketched briefly, guides the operator in:
1. Determining the priority of improvements to stops and stations
2. Devising a user perception survey for stations and stops of particular interest, and
3. Analyzing the survey results to produce a ratings matrix using Importance-
Satisfaction Analysis
Key Words: transit connectivity, intermodal, intramodal, stops, stations, transfer facilities, transit
managers, attributes, safety, security
v
EXECUTIVE SUMMARY
Transit travelers expend a great deal of their time, energy, and patience outside of buses and
trains – but the in- vehicle experience captures the lion’s share of attention from transit managers.
A typical door- to- door trip involves walking from one’s origin to a bus stop or train station,
waiting for a vehicle to arrive, boarding the vehicle, traveling in the vehicle, alighting from the
vehicle, and then walking to one’s final destination. In many cases, the trip also involves
transfers, contributing to both their actual and perceived burden of transit travel. This research
examines ways to increase the attractiveness and reduce the perceived burden of the time spent
outside of vehicles during transit trips.
In order to learn more about how wait times at stations and stops are perceived, and how they
can be made better, we surveyed approximately 750 transit passengers in metropolitan Los
Angeles, as well as 175 transit operators nationwide.
We surveyed passengers at stops and stations and asked them to assign a level of importance to
each of a list of attributes, and then to tell us how satisfied they were with each attribute. We
combined these two scores using Importance- Satisfaction Analysis to identify which attributes
passengers found most important and which needed the most improvement.
We surveyed transit operators, asking them to do two things: to rate by importance a series of
objectives for transit stops and stations, and also to guess how their operators would respond to a
user perception survey ( described above). We used the former to construct a rank- ordered list of
transit operators’ priorities for stops and stations, and the latter to see just how accurately
operators understand their riders’ priorities.
From our analysis of the passengers/ users perspective, one principal finding stands out clearly:
The most important determinant of user satisfaction with a transit stop or station is frequent,
reliable service in an environment of personal safety, and only indirectly the physical
characteristics of that stop or station.
From the sixteen attributes we examined, users ranked safety and on- time performance most
important, and amenities least important:
Most Important
1. I feel safe here at night
2. I feel safe here during the day
3. My bus/ train is usually on time
14. It is easy to get around this stop/ station
15. There are enough places to sit
16. There are places for me to buy food or drinks nearby
vi
Least Important
A companion part of our analysis compared how transit managers and neighboring communities
viewed transit stops and stations. Perhaps reassuringly, our principal finding precisely matches
that of the transit user investigation:
For operators, safety- and security- related factors far outweighed other
attribute factors at transit stops, stations, and transfer facilities.
Telephone interviews confirmed this finding, with most interviewees stressing the importance of
safety and security. One interviewee told us that “ safety trumps all” other concerns. Following
safety and security, operators rated the following attributes as most important:
2. Reducing pedestrian/ vehicle conflicts
3. Schedule coordination
4. Minimizing operating costs
We also compared transit managers’ views of what was important to their
riders with riders’ own views from our analysis of Los Angeles County
transit riders. While transit operators appear to have a fairly accurate
understanding of what attributes are important to their, there are several
points of disparity:
• The transit managers surveyed correctly assumed that safety and security were very
important to riders, but they tended to underestimate the importance of specific
safety- related factors, such as the presence of security guards and emergency
assistance.
• It also appears that, controlling for other factors, transit managers overestimate the
importance of station cleanliness and schedule information to their riders.
This report further develops a Preliminary Assessment Tool that transit operators can use to
guide their efforts at improving existing transit stops and stations, or in developing plans for new
facilities. The Preliminary Assessment Tool, sketched briefly, guides the operator in:
4. Determining the priority of improvements to stops and stations
5. Devising a user perception survey for stations and stops of particular interest, and
6. Analyzing the survey results to produce a ratings matrix using Importance-
Satisfaction Analysis
vii
Table of Contents
Preface i
Acknowledgements iii
Abstract iv
Executive Summary v
List of Tables viii
List of Figures ix
1.0 Thinking Outside the Bus: Waits and Transfers in Transit Travel 1
2.0 Transit Connectivity: Improvements at Stops, Stations, and
Transfer Facilities 6
3.0 Transfer Penalties/ Travel Behavior Conceptual Framework 7
4.0 Three Perspectives on Transit Stops and Stations – Users, Managers, and
Neighbors 12
5.0 Methods of Investigation 14
6.0 Primary Findings 17
7.0 Preliminary Assessment Tool: Putting Research into Practice 22
8.0 Next Steps / Future Research 25
9.0 About the Authors 27
10.0 Bibliography of Project Documentation 29
11.0 Appendices of Interim Deliverables 30
Appendix A: The Effects of Out- of- Vehicle Time on Travel Behavior:
Implications for Transit Transfers
Appendix B: Evaluating Connectivity Performance at Transit Transfer Facilities
Appendix C: Evaluating Transit Stops and Stations from the Perspective of
Transit Users
Appendix D: Evaluating Transit Stops and Stations from the Perspective of
Transit Managers
viii
LIST OF TABLES PAGE
TABLE 1 Classification of Transit Stops, Stations, and Transfer Facilities 3
TABLE 2 Survey Questions on User Importance and Satisfaction 15
ix
LIST OF FIGURES PAGE
FIGURE 1 A Transit Transfer Trip 1
FIGURE 2 Simple Bus Stop: Downtown Los Angeles 2
FIGURE 3 Los Angeles Union Station 2
FIGURE 4 Conceptual Wait/ Walk/ Transfer Impedance Framework for
Public Transit 8
FIGURE 5 Hierarchy of Traveler Wait/ Transfer Needs 18
FIGURE 6 I- S Ratings Graph Template 21
FIGURE 7 Stop/ Station Evaluation Flow Chart 23
1
1.0 Thinking Outside the Bus: Waits and
Transfers in Transit Travel
A typical door- to- door transit trip involves walking from one’s origin to a bus stop or train
station, waiting for the vehicle to arrive, boarding the vehicle, traveling in the vehicle, alighting
from the vehicle, and then walking to one’s final destination. In many cases, the trip involves
transfers; travelers alight from one transit vehicle, move to a new stop or platform, wait for
another transit vehicle, board that vehicle and continue this process until they reach their last stop
or station at which they walk to their final destination. Figure 1 shows a schematic diagram of
the major components involved in a transit trip involving a transfer.
FIGURE 1 A Transit Transfer Trip
Source: Metropolitan Transportation Commission, Transit Connectivity Study, March 2006.
Transit stops and transfer facilities are obviously not all the same and can differ relative to
numerous factors, for example with respect to:
• Physical size of the station or facility
• Travel modes serving the location
• Number of lines per transit operator
• Number of operators, and
• Amenities offered to travelers there.
At one extreme, we can have the bare minimum of attributes: An on- street bus stop that serves
two lines of the same transit agency with only posted time- point schedules, no real- time bus
arrival times, and not even a bench for waiting passengers to sit on ( Figure 2 Simple Bus Stop:
Downtown Los Angeles).
At the other end consider Union Station in downtown Los Angeles, which, as an off- street
facility, accommodates both intermodal and intra- modal ( bus, shuttles, light rail, heavy rail,
commuter rail, and inter- city rail) transfers among different transit agencies and different lines of
the same agency ( Figure 3 Los Angeles Union Station).
.
2
FIGURE 2 Simple Bus Stop: Downtown Los Angeles
FIGURE 3 Los Angeles Union Station
3
We describe in Table 1 how transit stops, stations, and transfer facilities may be grouped by the
following set of factors in which wait and transfer locations can differ:
• Volume of passengers and activities
• Number of interfacing routes
• Number of interfacing modes
• Physical configuration
• Investment in facilities
• Transit center type ( community, regional, or other), and
• Whether or not it is a joint development with commercial use of facility.
TABLE 1 Classification of Transit Stops, Stations, and Transfer Facilities
Source: Fruin, John J. 1985. Passenger Information Systems for Transit Transfer Facilities, In
Synthesis of Transit Practice, 7, edited by N. C. T. R. D. P. ( U. S.). Washington, D. C.: Transportation
Research Board National Research Council.
4
Thus, transit stops and transfer facilities vary greatly. For example, there are
• Bus stops
• Light rail stations
• Heavy rail stations
• Commuter rail stations
• Ferry docks, and
• Terminals
In general, the more transit users at stops and transfer facilities, the more complex a transfer
facility is. We highlight the following three types of transit stop/ transfer facilities:
A transit mall is a special street set aside for exclusive use of buses and/ or light rail
vehicles in a city center or other high activity center that focus on pedestrian movement
and activities, and include design components that are related to both transit and urban
design, such as waiting shelters, the use of landscaping, street furniture, shopping and
other civic activities. Transit malls are often combined with a development of adjacent
property, which consists of shopping and office activities as well as transit- related retail
and services.
A transfer center is a facility whose primary purpose is to facilitate easy transferring
between transit modes and routes and can be combined with transit- related developments,
concessions to accommodate users with convenience shopping, ( e. g. newsstands, snacks,
flowers, and teller machines) or coordinated with a full scale shopping center. Such
centers are usually located entirely or partially off- street. They also incorporate a more
elaborate and extensive shelter and more passenger amenities than ordinary bus stops.
These centers are typically located in suburban or edge- of- city locations in the
metropolitan area with sufficient area to allow access and circulation of multiple travel
modes as well as automobile parking.
An intermodal terminal is a facility that provides key transfers among transit modes,
which may include local bus, bus rapid transit, intercity bus, light rail, heavy rail,
intercity passenger rail, ferry, or automated guideway transit. Such facilities may also
have a variety of other services and connections, including parking, drop- off, ticket
vending, and information booths. These facilities are a fixed location where passengers
interchange from one route or vehicle to another that has infrastructure, normally only
shelters and/ or benches.
5
2.0 Transit Connectivity: The Key to the
Wait/ Transfer Experience
Public transit passengers typically must wait for and transfer between buses and trains during
their journeys, and this constitutes the connectivity between distinct parts of a passenger’s transit
trip from origin to destination. Thus, the travel time spent outside of transit vehicles while
waiting and transferring plays a significant role in the passenger’s overall transit trip experience.
But what exactly is transit connectivity? How does one define, measure, and evaluate
connectivity? Although the importance of transit transfer connectivity has been recognized for
several decades, surprisingly little of what researchers have learned about out- of- vehicle travel
behavior today explicitly informs transit planning practice. Efforts to improve connectivity at
stops and stations have proven to be less effective than expected for the following reasons:
• Both practitioners and researchers tend to pay more attention to quantity
and quality of in- vehicle travel for its more intuitively obvious effects on
ridership.
• Stops and stations vary in size, modes served, location, and amenities; they
are hard to analyze comprehensively using uniform criteria
• Most of the literature on stops and stations is descriptive in nature and lacks
a theoretical framework to explain how improvements of transfer facilities
affect people’s travel behavior and, in turn, overall transit ridership.
Most previous studies of transit stops, stations, and transfer facilities have compiled laundry lists
of out- of- vehicle trip attributes that contribute to or detract from travelers’ transfer experiences;
however, they have largely failed to consider the relative importance of each of these attributes ─
positive and negative ─ or whether and how these attributes influence ridership separately or in
concert with another. As a result, we know little about which attributes are most important, under
which circumstances, and in what combinations with other factors. In other words, we know
very little about the effects of stops, stations, and transfer facilities on transit ridership and
network performance. This state of knowledge based on past studies of the subject is incomplete
because it fails to guide transit agencies toward planning practices that effectively improve the
quality of transfers at transit centers that actually result in a ridership increase.
In our research on transit stops, stations, and transfer facilities we have addressed these
shortcomings by developing a theoretical framework for understanding the relationship between
transfer- facility attributes and travel behavior, which we discuss below.
6
3.0 Transfer Penalties/ Travel Behavior
Conceptual Framework
The concept of the transfer penalty represents generalized costs — including monetary costs,
time, labor, discomfort, inconvenience, etc. — involved in transferring from one vehicle to
another, between the same mode, or different transportation modes ( e. g. bus to train, walk to bus,
etc.). We use the term transfer penalties in two ways. Viewed broadly, transfer penalties are used
to represent all of the monetary, time, and labor expenditures involved in waiting and walking,
experiencing discomfort, worrying about safety, and any other inconvenience and emotional
stress involved in waiting and transferring, and thus can generally be viewed as an impedance to
travel.
Viewed more narrowly, transfer penalties are the impedance in transferring, excluding easily
quantified factors, such as waiting time, walking time, and transfer fares. In other words, a
narrow definition of transfer penalties considers costs beyond the monetary and time costs
associated with transferring.
For the more easily quantified transfer penalties, such as walking and waiting times, there are
differences between actual and perceived values for these times. People perceive time differently
depending on the circumstances. While actual waiting time is the difference between a
passenger’s and his/ her vehicle’s arrival at a boarding location, perceived waiting time can be
considerably longer depending on waiting conditions such as vehicle arrival time uncertainty,
comfort, security, and safety. Thus the generalized cost of waiting can greatly increase beyond
the cost of actual waiting time.
Traveler’s perceived walking distance and time can also be substantially greater than their actual
walking distance and time. Physical conditions and adequate information are both important in
determining both actual and perceived walking distance and time. The shortest walking time is
determined by the most direct path and a traveler’s walking speed. When a traveler is familiar
with a stop location or transfer facility, walking paths can be direct and walking times
minimized. However, unfamiliar stops or facilities and/ or poor information lead to wandering,
stress, and uncertainty about how and where to make the connection. Thus, the location, layout,
and information at transfer stops and stations can significantly influence the perceived transfer
experience as well as actual walking distance/ time and waiting time, and both affect the
likelihood of using transit in the future.
Differences in actual and perceived travel, waiting, and transfer times can be viewed as different
valuations of time for different activities, and such different valuations of time for different trip
attributes are weighted differently. In choosing a travel mode, travelers make decisions based on
their perceived total generalized cost of taking a trip by various modes, which can depend
substantially on their perceptions of travel ( including transfer) attributes, such as time, labor,
comfort, and safety.
The perceived burdens of waiting time, walking time, and transferring suggests the following
three broad categories of factors contribute to transfer penalties:
7
A common rule of thumb is that walking and waiting time are considered by transit users to be
two to three times as onerous as in- vehicle travel time.
• Operational factors, such as headways, reliability, on- time performance of
service and availability of adequate information.
• Physical environmental factors at facilities related to safety, security,
comfort, and convenience
• Passenger options, such as whether they are forced to wait or whether they
can be productive while waiting.
Given this, transit managers can take various measures to lower the burden ( or generalized cost)
of waiting, walking, and transferring by addressing both actual and perceived waiting time,
perceived walking time, transfer burdens, and fares paid. Figure 4 presents our conceptual
framework for determining the generalized cost of transferring in the overall context of transit
travel. Perceived waiting and walking time are determined by actual time plus the weights that
travelers assign to waiting and walking, which vary by the attributes, conditions, and
environments of stops, stations, and transfer facilities.
We group the factors listed above into four groups:
1. The monetary cost of a transfer ( fare);
2. Factors that affect the actual transfer time and distance;
3. Factors that influence people’s perception of waiting and
walking ( e. g. the weights users assign to waiting and
walking), and
4. Other factors that affect perceptions of transferring that are
not taken into account by the first three groups.
The matrix at the bottom of Figure 4 notes which aspects of three factors – transfer fare, time
schedule and operation, and transfer facilities – affect four aspects of traveler impedance: ( 1)
monetary cost, ( 2) actual travel time and distance, ( 3) perceived travel time and distance, and ( 4)
other penalties. This is discussed further below.
8
FIGURE 4 Conceptual Wait/ Walk/ Transfer Impedance Framework for Public Transit
9
Transfer fares
In the context of the total costs of a transit trip, the penalty of a transfer fare is typically relatively
small. It is often free or quite low for most intra- urban transit services, For short trips, however,
transfer fares can be relatively large on a per- mile- traveled basis, and may disproportionally
affect the burden of short trips with transfers.
Schedule and operation
Service frequency, schedule adherence, and schedule information ( both posted and real time)
affect both actual and perceived waiting time. Obviously, increasing service frequency reduces
average waiting and transferring times. Poor coordination between lines, modes, and systems,
and lack of schedule adherence can significantly increase transfer wait times; not surprisingly,
improved coordination has been shown to increase transfer rates.
As noted above, frequent service can substantially ( and nonlinearly) reduce the perceived burden
of waiting. And frequent, reliable service has been shown to substantially reduce transfer
burdens because travelers can count on short average wait times and can reliably time their
arrival at stops and stations to minimize waiting.
Transfer facilities
Physical attributes of transfer facilities likely affect walking time and effort, waiting time and
effort, convenience, comfort, safety, and indeed many other components of transfer burdens. In
general, “ passenger friendly” and “ user friendly” transfer facility attributes can be grouped into
the following five categories:
1. Facility design can affect access by defining the distance between alighting and
boarding locations, improving off- vehicle passenger flow, and providing clear and
comprehensible directions. Perimeter- oriented bus depots, for example, have been shown
to increase transfer walk distances and inhibit pedestrian flows. Further, confusing or
incomplete signage, or poorly located ticket machines and information kiosks can
significantly increase both the actual and perceived distances walked in stations and
transfer facilities.
2. Connection and reliability are determined by time schedules and schedule adherence,
and have been repeatedly shown to strongly influence transfer burdens and transit use.
3. Complete, concise, and easy- to- understand information has been shown to reduce the
actual ( by reducing wandering) and perceived burden of transferring, especially for new
or occasional transit users.
4. Amenities, such as benches, shades, water fountains, and rest rooms, affect comfort
and convenience while passengers are waiting and transferring. Through increased
comfort and convenience, these amenities can affect perception of waiting and walking
time as well as other burdens of transferring.
5. Security and safety also influence perception of waiting, walking, and transfer burdens.
Safety and security can be a “ deal breaker” for travelers if levels of perceived risk exceed
thresholds over which they will no longer consider traveling by transit, and will instead
travel by other modes or forgo the trip entirely.
Even though the passenger perspective regarding transit connectivity is of utmost importance, we
have found many references in the literature to transit connectivity from the transit operator and
10
the neighboring community’s perspective. Looking at these three aspects has provided us with a
complete picture of assessing transit connectivity. We begin to explore these additional
perspectives together with continuing our in depth examination of the passenger perspective in
the next section.
11
4.0 Three Perspectives on Transit Stops and
Stations – Users, Managers, and Neighbors
In assessing how effectively stops, stations, and transfer facilities operate, we identified three
primary stakeholder groups from whose perspectives such evaluations have been performed.
These are:
• Passengers/ users
• Transit Operators
• Neighboring Communities/ Businesses & Residents
Passengers/ Users
Passengers/ users are the clients who use stops, stations, and transit transfer facilities and who
have specific desires and expectations for such facilities. Previous travel behavior research
suggests that transit users’ principal concerns are with quickly and easily boarding their desired
vehicle. Toward that end, users desire:
• Minimum transfer time and distance,
• Convenience,
• Comfort, and
• Safety and security.
Which of these is most important under what circumstances, however, is less well known.
However, when transfer facilities are designed and/ or renovated to make transferring more safe
and secure, pleasant, faster, and less problematic, people accept facilities more favorably and are
more likely to accept the necessity of transferring in their transit trips.
Transit Operators
When a transit operator owns the property under which a stop or transfer facility sits, it can
largely control the design and operation of the stop or facility. In most cases, however, transit
operators do not own the land under their stops and stations and must therefore work and
negotiate with a wide variety of public and private stakeholders.
Neighboring Communities/ Businesses & Residents
Any transit stop or transfer facility ─ whether it is located in an urban or suburban environment,
or whether it hosts intra- modal or intermodal transfers ─ does not exist in a vacuum. It and its
users necessarily interact with adjacent neighborhoods and districts. As such, the people who
live and/ or work near the stop or facility, and the people who own and operate commercial
establishments in the vicinity of the stop or facility have a stake in the facility that may be largely
unrelated to its utility to transit users. These include:
• Community image and pride ─ architectural, cultural, and historic
12
preservation
• Joint development and partnerships
• Safety and security
• Environmental impacts on surrounding neighborhood
• Neighborhood economy / local employment
• Physical and social impacts on neighboring land uses
Accordingly, the research described below sought explicitly to examine perceptions of transit
stops and transfer facilities from the differing perspectives of these three groups. And it is to this
research we now turn.
13
5.0 Methods of Investigation
In our investigation of each of the three stakeholder perspectives, we employed a variety of
research methods:
Passengers/ Users
We designed and administered a user survey based on the five principal transit stop and station
attribute categories thought in the literature to affect transfer penalties:
• Access: Management of passenger flow control and directional information
• Connection and Reliability: Distance and time to make connections; on-time
performance/ frequency of bus/ train
• Information: What, where, and how passengers acquire information
• Amenities: Comfort, service, weather protection, and cleanliness of
station/ stop
• Security and Safety: Station/ stop equipment, infrastructure, or personnel
that provide passengers with a safe and secure environment
Our objective was to provide an accurate portrait of transit riders at the system- wide level, by
service- type, by time of day and day of week, and by location. This portrait included the
following information:
◦ Demographic characteristics of riders at every transit transfer facility in terms of:
• Age
• Gender
• Income
• Ethnicity
• Car availability
• Modal preference
◦ Trip characteristics, including
• Trip purpose
• Pre- and post- trip mode
• Transfer rate
• Time of day and day of week
• Service type;
◦ Frequency of use, and
◦ Evaluation of transit services and amenities
For each of the five attribute categories, the research team crafted a series of specific
questions. The resulting survey, which was made available in English and Spanish, consisted
of 29 questions and was self- administered to 749 transit users at 12 transit stops and stations
around metropolitan Los Angeles. In total we approached 1,023 transit users and 274 of them
refused to participate in the survey yielding a 73% response rate. Moreover, the 749 surveys
were not entirely completed as some users had to stop providing responses to catch their bus
or train. The survey was designed to assess the importance of and satisfaction with various
14
aspects of transit stops, stations, and transfer facilities from the transit rider’s perspective. The
dozen transit stop and transfer sites were selected to secure the widest possible variation in the
following:
• Transfer facility types ( See Table 1)
• Available modes ( bus, rail)
• Type of passenger loading ( on- or off- street)
• Time of day
• Weather
A significant component of the survey was soliciting respondents’ views on their satisfaction
with, and level of importance of, various stop/ station attributes ( listed in Table 2). A copy of the
User Survey Instrument may be found in Appendix C of this report, which consists of a copy of
our interim deliverable documenting our evaluation of transit stops and stations from the
perspective of transit users.
Transit Operators
We designed a transit system manager survey to collect the following information from
respondents:
• Operators’ estimation of how important various evaluation factors are to their
own passengers
• Operators’ views of what evaluation factors are important from their own
perspective
The survey was administered by means of a web- based online nationwide survey of transit
managers. The survey instrument ( which is available in Appendix C) was designed to both
mirror many of the questions in our user survey, and to ask about political and operational
concerns not directly related to passenger use of stops or stations. From the Federal Transit
Administration’s 2005 National Transit Database we selected all 400 transit operators with at
least one fixed- route/ fixed- schedule transit line in service in the United States. We sent the
general manager of each an electronic invitation to either respond to our survey or to designate a
member of his/ her staff to do so. We received a total of 175 completed responses, for a 43%
response rate.
Neighboring Communities/ Businesses & Residents
Finally, we developed a set of questions that were used during telephone interviews with a
representative sample of transit operators in the United States in order to gain further insight into
the transit operators’ perspective, as well as to gather illustrative anecdotes about transit stops
and stations. Twenty agencies were selected by a weighted sampling methodology, with the
probability of inclusion in our sample weighted by the agency’s annual ridership figures. Of
these, 8 agencies participated, for an effective response rate of 40%. During these interviews, we
also gathered data on the role of stop and station neighbors – both private and commercial – in
shaping the design, location, and operation of transit stops and stations. These interviews
focused in particular on community advocacy for and against the location, re- location, and/ or
expansion of transit stops and stations. Due to budget limitations, however, we did not survey or
interview stop- or station- adjacent stakeholders directly.
15
TABLE 2 Survey Questions on User Importance and Satisfaction
16
6.0 Primary Findings
From our analysis of the passengers/ users perspective, one principal finding stands out quite
clearly:
The most important determinant of user satisfaction with a transit stop or
station is frequent, reliable service in an environment of personal safety, and
only indirectly the physical characteristics of that stop or station.
In other words, most transit users would prefer short, predictable waits for buses and trains in a
safe, if simple or even dreary, environment, over long waits for late- running vehicles. This is true
even if such long waits occur in the most elaborate and attractive transit stations and especially
so if users fear for their safety. While this finding will come as no surprise to those familiar with
past research on the perceptions of transit users, it does present a contrast to much of the
descriptive and design- focused research on transit stops and stations.
In total, we examined sixteen stop and station attributes ( listed in Table 2), using a technique
known as the Importance- Satisfaction Analysis method, which seeks to identify those attributes
passengers find most important ( importance level) and those attributes in need of the most
improvement ( satisfaction level). Respondents’ level of satisfaction with each attribute under
current conditions at the 12 survey sites in the Los Angeles metropolitan area indicates that users
are least happy with factors related to access, followed by some factors related to security and
safety and connection and reliability. When we considered the level of satisfaction and
importance ratings in tandem, factors that require improvement pertain most to security and
safety and connection and reliability, and least to amenities. Of the sixteen attributes, users
ranked safety and service quality factors as most important ( the top six of the sixteen attributes)
as shown in the following list:
Most Important
1. I feel safe here at night ( 78%)
2. I feel safe here during the day ( 77%)
3. My bus/ train is usually on time ( 76%)
4. There is a way for me to get help in an emergency ( 74%)
5. This stop/ station is well lit at night ( 73%)
6. I usually have a short wait to catch my bus/ train ( 70%)
17
In contrast, stop and station- area amenities were ranked as least important by users:
Least Important
11. It is easy to get route and schedule information at this stop/ station
( 62%)
12. There is a public restroom nearby ( 59%)
13. This stop/ station is clean ( 58%)
14. It is easy to get around this stop/ station ( 57%)
15. There are enough places to sit ( 50%)
16. There are places for me to buy food or drinks nearby ( 34%).
However, when we statistically related users’ satisfaction with various stop/ station attributes
with their overall satisfaction with their wait/ transfer experiences, we got similar, though not
identical, results:
Most Important
1. It is easy to get around this stop/ station.
2. I feel safe here during the day.
3. Having security guards here makes me feel safer.
4. it’s easy to find my stop or platform.
5. The stop/ station is well lit at night.
6. My bus/ train is usually on time.
In contrast, the following stop and station- area attributes were ranked as least important ( bottom
six of the sixteen attributes):
Least Important
11. This stop/ station is clean.
12. There is shelter here to protect me from the sun or rain.
13. There is a way for me to get help in an emergency.
14. There are enough places to sit.
15. There are places to buy food or drinks nearby.
16. There is a public restroom nearby.
Informative, rank- ordered lists like these can be problematic if users “ split their votes” among
similar, yet important factors, such as “ I feel safe here at night” and “ This stop/ station is well- lit
at night.” To correct for this problem, we employed an ordered- logit regression model to
measure the independent influence of each of 16 wait/ transfer attributes on overall user
satisfaction. This analysis tends to eliminate all but one of closely related factors, while
18
elevating presumably less- important factors that independently influence users’ overall levels of
satisfaction. The results of this modeling exercise are telling:
Most Important
1. My bus/ train is usually on time.
2. Having a security guard here makes me feel safer.
3. This stop/ station is well lit at night.
4. I feel safe here during the day.
5. It is easy to get around this station/ stop.
6. The signs here are helpful.
Of the 16 stop and station attributes that we evaluated, transit users assigned the highest
importance to factors related to security and safety, and then to factors related to connection and
reliability. In contrast, stop and station- area amenities were ranked as least important by users.
We do not claim that amenities are not important to travelers; more than half ranked information,
the presence of public restrooms, cleanliness, and ease of navigation as important attributes.
However, travelers definitely prefer safe, frequent, and reliable service over these other factors.
Based on this analysis we have identified a simple hierarchy of transfer burdens perceived by
users, shown in Figure 5. This figure summarizes the findings from our transit user investigation
succinctly.
Figure 5 Hierarchy of Traveler Wait/ Transfer Needs
In addition to surveying transit users, we conducted an nationwide online survey of transit
operators, asking them about their objectives at transit stops, as well as about their perceptions of
users’ and neighboring communities’ priorities for stops and stations. From our analysis of the
survey results, we find that transit operators’ top priority is precisely the same as that of the users
19
of their systems:
Safety and security related factors far outweighed other attribute factors at
transit stops, stations, and transfer facilities.
Following safety and security (# 1), ten other factors cluster relatively closely as important factors
in the views of the transit managers surveyed. We list them in order of priority:
2. Pedestrian/ vehicle conflicts
3. Schedule coordination
4. Operating costs
5. Stop/ station equipment reliability
6. Comfortable environment
7. Adequate stop/ station space
8. Inter- agency coordination
9. Facilitate passenger flows
10. Accommodate vehicle movements
11. Protect passengers from weather.
The survey results further suggest that transit operators value user- oriented attributes such as
physical comfort and seamless transferring higher than other, non- user oriented, attributes. This
may be due to the immediacy and constancy of user- related factors such as the provision of clean
and comfortable transfer stops and stations, while non- user attributes such as joint development
typically occurs infrequently.
Our online survey results show that, while transit operators appear to have a fairly accurate
understanding of what attributes are important to their riders at transit stops and transfer stations,
there are several points of disparity. While operators correctly assumed that safety and security
were very important to riders, they tended to underestimate the importance of specific safety-related
amenities, such as the presence of security guards and emergency assistance. It also
appears that, controlling for other factors, operators overestimate the importance of station
cleanliness and schedule information to their riders. We note, however, that there was a
mismatch in geographical coverage for this comparison; our riders’ survey collected data from
Los Angeles County transit riders, while our operators’ survey collected data nationwide. It is
likely that this mismatch has overemphasized some disparities, while downplaying others. These
findings should be considered preliminary and further research should examine both subgroups
that cover the same general location. Next steps and follow- on research are discussed in a later
section of this report.
Our telephone interviews served to highlight these findings. Interviewees relayed to us many
anecdotes in which safety and security concerns “ trumped” all other concerns. For example,
20
comfort concerns ( ample and comfortable seating) often defer to security concerns ( benches that
are not conducive to sleeping). Another telephone interviewee told us of a station redesign that
resulted in a safer environment for pedestrians, but which was far less aesthetically pleasing. Yet
another interviewee from a city with a “ very high murder rate” told us that city police are present
at station design meetings, and that personal safety and security concerns always outweigh
aesthetic, design, and passenger comfort concerns. Less obvious and more nuanced tradeoffs are
made throughout the set of objectives; our ranking describes the propensity of transit operators to
value one attribute more highly than others, and assigns estimates of the magnitude of these
propensities.
Additionally, we talked to transit operators about the role of the community in planning,
operating, and maintaining transit stops and transfer facilities. We heard from many respondents
that the community often serves as opposition, and that its input usually comes indirectly through
politicians and community leaders. Furthermore, we heard that community concerns are typically
voiced in response to planned changes, rather than during initial planning stages.
We also determined that other stakeholders ( specifically local government entities) control the
design and location of most transit stops and stations. We also found that adjacent businesses and
residents exert significant influence over the location, design, and operation of stops and stations.
Often, transit agencies have surprisingly limited control over the siting and design of stations and
stops.
21
7.0 Preliminary Assessment Tool: Putting
Research into Practice
Based on the findings reported above we have developed a 3- step process ( synthesized in Figures
6 & 7) that transit operators can employ as a tool to guide them as they consider making
improvements to already existing transfer facilities or developing initial plans for new facilities.
Step 1: Use the Hierarchy of Traveler Wait/ Transfer Needs ( Figure 5 above) to determine
the priority of improvements to any stop or station. We endeavored in this
research to produce generalizable findings from our analysis by surveying a large
number of transit users at a wide variety of facilities. 2
Step 2: For transit stops and stations serving particular user populations ( children,
immigrants, the elderly, etc.) or for stops/ stations in unique environments
( adjacent to airports, amusement parks, hospitals, etc.), the user perception survey
instrument developed and tested in this study can be used to survey the
perceptions of passengers.
Step 3: Use the survey results to conduct an Importance- Satisfaction ( I- S) Analysis
( documented in detail in Appendix C) to produce an I- S Ratings matrix showing
Average Importance and Satisfaction ratings for the users and/ or stops surveyed
as shown schematically in Figure 6 below.
FIGURE 6 I- S Ratings Graph Template
Satisfaction Rating
Average
Average Importance Rating
Region 4
Exceeding Expectations
Region 3
Less Important
Region 1
Needs Improvement
Region 2
Continue Improvement
22
• Region 1 is an area where – for the surveyed users or stops – facility
attributes have above- average importance but a less than average level of
satisfaction, meaning that these attributes should be high priorities for
improvement.
• Region 2 is an area where attributes have above- average importance and
above- average level of satisfaction, meaning that priority should be given
to maintaining the quality of these attributes.
• Region 3 is an area where attributes have less than average satisfaction levels
but also less than average importance ratings; improvement to such
attributes are warranted only at low cost or if all of the attributes in
Regions 1 and 2 have been fully addressed.
• Region 4 is an area where attributes have above average levels of satisfaction
and importance ratings less than average; such attributes exceed
expectations and warrant no further attention.
We suggest that transit operators employ this 3- step process in successive stages using the flow
chart below ( Figure 7). This chart guides users in identifying the order – consistent with our
research findings – in which to improve a targeted transit stop or station. We have thus
structured the flow chart so users’ priorities in the Hierarchy of Traveler Wait/ Transfer Needs
( Figure 3 above) are addressed in order of importance: first with Safety and Security attributes,
second with Connections and Reliability attributes, third with Access and Information attributes,
and lastly with Amenities- related attributes.
23
Figure 7 Stop/ Station Evaluation Flow Chart
24
8.0 Next Steps / Future Research
The major milestone of this project was the development of a conceptual behavioral framework
of the passenger’s wait/ transfer experience based on our review of the state- of- the- research of
travel behavior. We used this framework to capture both transit user and manager perceptions of
transfer burdens, which allowed us to advance considerably the body of research on transit stops
and stations that to- date has been largely descriptive.
The findings of our research, together with the development of our preliminary assessment tool,
have taken substantive steps toward:
• Determining the connectivity of transit systems and how this connectivity ( as
well as other service attributes) influences travelers’ satisfaction with transit
services, and
• Examining how public transit systems can reduce the burdens of out- of-vehicle
“ travel” times in order to help make public transit more attractive
resulting in ridership increases.
There are, however, limitations to our research conducted to date. In this project, we surveyed
over 700 transit users to determine the factors affecting their perceptions of waiting, walking,
and transferring during a trip. Within each category of attributes, the users’ satisfaction level was
correlated with data from a detailed inventory of 12 stops and transfer facilities in Los Angeles
County to identify significant linkage between users’ perceptions of transit services and the built
environment at stops, stations, and transfer facilities. While we secured a large number of
surveys of users’ perceptions, the fact that these were collected at just a dozen, locations –
though diverse for Los Angeles County – did not give us sufficient variability in the facilities
data inventories to statistically link the physical and operational characteristics of transit stops
and stations with users’ perceptions of them. In other words, we were unable to evaluate the
relative importance of facility attributes in directly determining users’ overall satisfaction levels.
Nonetheless, our evaluation framework has provided us with a strong theoretical foundation to
expand our study of transit users and facilities beyond Los Angeles County. Accordingly, we are
working with Caltrans to develop a follow- on scope of work to the research reported here,
specifically, to:
• Evaluate user perceptions across a wider cross- section of users and a much
wider array of transit systems;
• Expand our stop/ station Assessment Tool to apply to a broader range of
transit user populations and operating environments;
• Embark on a field implementation phase; and
• Expand our stakeholder analysis to include the perceptions and motivations of
local governments that control the location of development of most transit
stops and stations.
We aim in our next phase of this research to expand our inventory of stops and stations from 12
to 50 across California, with a goal of surveying approximately 2,000 users. This expanded
approach will help make the findings of this effort considerably more generalizable to cities and
25
transit operators in large and small cities around California. Moreover, by field testing the
findings of our Phase I and II work at specific transit stops and stations, we can conduct before
and after testing to determine if, indeed, this research can help transit operators attract more
riders by cost- effectively addressing the specific aspects of waiting for and transferring among
transit vehicles that transit users find most burdensome.
26
9.0 About the Authors
Brian D. Taylor, University of California, Los Angeles
Brian Taylor is a Professor of Urban Planning, and Director of the Institute of Transportation
Studies at UCLA. His research centers on both transportation finance and travel demographics.
He has examined the politics of transportation finance, including the influence of finance on the
development of metropolitan freeway systems and the effect of public transit subsidy programs
on both system performance and social equity. His research on the demographics of travel
behavior has emphasized access- deprived populations, including women, racial- ethnic
minorities, the disabled, and the poor. His work in this area has also explored the relationships
between transportation and urban form, with a focus on commuting and employment access for
low- wage workers. Professor Taylor teaches courses in transportation policy and planning and
research design. Prior to coming to UCLA in 1994, he was an Assistant Professor in the
Department of City and Regional Planning at the University of North Carolina at Chapel Hill,
and before that a Transportation Analyst with the Metropolitan Transportation Commission in
Oakland, California.
Hiroyuki Iseki, University of Toledo
Hiroyuki ( Hiro) Iseki is an Assistant Professor in the Department of Geography and Planning at
the University of Toledo. He received his Ph. D. in Urban Planning at UCLA. He specializes in
transportation, land use, environmental planning and policy analysis, and the application of
quantitative analysis methods ( econometrics and GIS) to planning. His work focuses on
balancing efficiency, effectiveness, and equity in urban planning and policy. He has published
extensively on transit planning, management, and finance, and is an active member of the
Committee on Public Transportation Planning and Development of the Transportation Research
Board. His recent research projects have examined subjects such as: 1) privatization of public
transit service, 2) a policy framework design for an interoperable transit smart card system, 3) an
evaluation of the effects of transit transfers on travel behavior, and 4) a public- private
partnership as a new strategy for highway financing.
MARK A. MILLER, UNIVERSITY OF CALIFORNIA, BERKELEY AND LOS ANGELES
Mark Miller is a Research Specialist at the California PATH ( Partners for Advanced Transit and
Highways) Program at the University of California, Berkeley where he has worked for eighteen
years. During this time his research has focused on developing evaluation frameworks and
methodologies and performing impact assessments of intelligent transportation systems ( ITS)
technologies in the setting of field tests and case studies. His work experience has been well
balanced between quantitative and qualitative investigations covering both technical and non-technical
( deployment, societal, and institutional issues) aspects of ITS. Mr. Miller has
significant work experience in the areas of transit operations research and policy and behavioral
research, including bus rapid transit, and commercial vehicle operations.
Since 2003, Mr. Miller has also been a Visiting Scholar at the Institute of Transportation Studies
( ITS) within the School of Public Affairs at the University of California, Los Angeles. Since
arriving at UCLA, Mr. Miller has worked to expand PATH’s visibility and presence in southern
California and to forge and strengthen ties between the Los Angeles and Berkeley campuses of
the Institute of Transportation Studies, worked closely with graduate students on PATH projects,
and given lectures on Intelligent Transportation Systems.
27
Michael Smart, University of California, Los Angeles
Michael Smart is a second- year doctoral student at UCLA. His research interests include
transportation and social equity, access to the labor market, and travel behavior. He has recently
completed an analysis of the SAFETEA- LU earmarking process and is currently working on a
detailed look at carpooling among immigrant communities. He graduated with a Master’s degree
in City Planning from the University of Pennsylvania in Philadelphia and has previously worked
with Professor Myron Orfield at the University of Minnesota Law School’s Institute on Race and
Poverty.
28
10.0 Bibliography of Project Documentation
Interim Project Deliverables
1. Iseki, H., B. D. Taylor, and M. A. Miller, The Effects of Out- of- Vehicle Time on Travel
Behavior: Implications for Transit Transfers, January 2006.
2. Iseki, H., M. A. Miller, A. Ringler, M. Smart, and B. D. Taylor, Evaluating Connectivity
Performance at Transit Transfer Facilities, February 2007.
3. Iseki, H., A. Ringler, B. D. Taylor, M. A. Miller, and M. Smart, Evaluating Transit Stops
and Stations from the Perspective of Transit Users, August 2007.
4. Smart, M., M. A. Miller, and B. D. Taylor, Evaluating Transit Stops and Stations from the
Perspective of Transit Managers ( Draft), September 2007.
Conference Compendiums & Proceedings and Journal Articles
1. Iseki, H. and B. D. Taylor, “ Not All Transfers Are Created Equal: Toward a Framework
Relating Transfer Connectivity to Travel Behavior”, 86th Annual Meeting of the
Transportation Research Board CD- ROM Compendium of Papers, Transportation
Research Board, Washington D. C., January 2007.
2. Taylor, B. D., H. Iseki, A. Ringler, M. J. Smart, and M. A. Miller, “ Hate to Wait: Relating
the Wait/ Transfer Experience to Perceived Burdens of Transit Travel”, Accepted for
Presentation and Publication at 48th Annual Conference of the Association of Collegiate
Schools of Planning, Milwaukee, Wisconsin, October 2007.
3. Iseki, H. and B. D. Taylor, “ Style versus Service”, Submitted for Presentation and
Publication to 87th Annual Meeting of the Transportation Research Board,
Transportation Research Board, Washington D. C., January 2008.
4. Smart, M., M. A. Miller, and B. D. Taylor, “ Transit Stops, Stations, and Transfer
Facilities: Evaluating Performance from the Perspective of Transit System Managers”,
Submitted for Presentation and Publication to 87th Annual Meeting of the Transportation
Research Board, Transportation Research Board, Washington D. C., January 2008.
29
11.0 Appendices of Interim Deliverables
Appendix A
The Effects of Out- of- Vehicle Time on Travel
Behavior: Implications for Transit Transfers
Appendix B
Evaluating Connectivity Performance at Transit
Transfer Facilities
Appendix C
Evaluating Transit Stops and Stations from the
Perspective of Transit Users
Appendix D
Evaluating Transit Stops and Stations from the
Perspective of Transit Managers
APPENDIX A
Appendix A
The Effects of Out- of- Vehicle Time on Travel Behavior:
Implications for Transit Transfers
( Deliverable # 1)
Under Contract 65A0194 for Project
Tool Development to Evaluate the Performance of
Intermodal Connectivity ( EPIC) to Improve Public Transportation
Submitted to:
Bruce Chapman
California Department of Transportation
Division of Research and Innovation
1227 ‘ O’ Street, 5th Floor
Sacramento, CA 94273- 0001
Prepared by:
Hiroyuki Iseki, 1 Brian D. Taylor1, and Mark Miller1,2
1Institute of Transportation Studies
University of California, Los Angeles
School of Public Affairs
3250 Public Policy Building, Box 951656
Los Angeles, CA 90095- 1656
2California PATH Program
Institute of Transportation Studies
University of California, Berkeley
Richmond Field Station, Bldg. 452
1357 S. 46th Street
Richmond, CA 94804
January 18, 2006
i
EXECUTIVE SUMMARY
This report constitutes an interim deliverable for the Project “ Tool Development to Evaluate the
Performance of Intermodal Connectivity ( EPIC) to Improve Public Transportation” under
Contract 65A0194 with Caltrans. Our primary objective in this project is to develop an
evaluation tool that transit agencies can use to assess the quality of service at transit transfer
facilities and use the findings of such evaluations to improve travel connectivity. Such
improvements, can, in turn, help the overall transportation system operate more smoothly and
can make transit a more attractive travel option and thus can eventually contribute to increases in
ridership. This report focuses on a review of the literature in the area transit transfer facilities
with particular emphasis on studies of the perceived burdens of transferring by passengers and
their travel behavior as this is potentially a rich source of information to be used as input in the
design of the evaluation tool.
Many factors affect travel choices, including time, labor, cost, security, convenience, and
comfort of the entire trip. As such, privately- owned automobiles have many advantages over
traditional fixed- route public transit in providing higher levels of accessibility, flexibility,
convenience, comfort, and safety against crime. The relative burdens of public transit service
vis- à- vis private automobiles help to explain why the majority of personal travel in metropolitan
areas is in private vehicles, which poses a daunting challenge to transit managers. Given that
travelers tend to consider out- of- vehicle travel time ( walking, waiting, transferring, etc.) to be
substantially more burdensome than in- vehicle travel time, attracting travelers to public transit in
significant numbers requires transit agencies to focus increasingly on improving transit users’
experience outside of their vehicles – walking, waiting, and transferring.
As cities have grown more dispersed and auto- oriented, the relative burdens of out- of- vehicle
transit travel have increased. In an effort to accommodate increasingly dispersed patterns of trip-making,
transit systems in many U. S. metropolitan areas have adapted “ hub- and- spoke” route
systems, which require transit users to frequently make transfers among lines and systems. In
larger metropolitan areas with many transit operators, where the number of transferring
passengers can be very high, transfer centers to facilitate passenger transfers are central parts of
transit networks. Given the importance of out- of- vehicle times on travel choices, intermodal
connectivity at such transfer facilities is a critical part of overall transportation network
effectiveness. Transfer facilities that integrate various transportation modes in one location
encourage people to use transit service by reducing the burdens of transfers.
What aspects of walking, waiting, and transferring do travelers find to be more burdensome,
and what can transit managers do to cost- effectively increase the attractiveness of transit travel?
This report examines this question by carefully reviewing the literature on the perceived burdens
of transit travel.
We find that, despite its importance, efforts to increase connectivity at transfer facilities have
proven less effective than expected for the following reasons: 1) not enough attention has been
given to the effects of out- of vehicle travel on ridership; 2) it is difficult to comprehensively
analyze transfer facilities using uniform criteria due to a large variation in size, modes served,
location, and amenities of transfer facilities; and 3) there is a lack of a framework to theorize the
effects of transfer facility improvements on people’s travel behavior and transit ridership. In
particular, the lack of causal clarity in the research on transit transfer facilities is an enormous
drawback. Most previous studies of transit stops, stations, and transfer facilities have compiled
laundry lists of positive and negative attributes, but have largely failed to consider the relative
importance of each of these attributes, or whether they influence ridership differently alone or in
ii
concert with other factors. As a result, we know little about which attributes are most important,
under which circumstances, and in what combinations. Past studies on the subject have failed to
lead transit agencies to implement planning practices that can effectively improve the quality of
transfers at transit centers. Bridging this knowledge gap can lead to improvements of transfer
facilities that will result in a ridership increase.
In this literature review, we identify the gaps in the current literature on factors influencing
transit ridership, transfer penalties, and transfer facility improvements. We address the lack of a
theoretical basis for understanding the relationship between transfer facility attributes and travel
behavior and provide a brief review of determinants that affect transit ridership. This framework
situates transfer penalties within the total cost of a transit trip. Finally, we examine the attributes
of transfer facilities that influence transit transfers.
We situate the literature of travel behavior and valuation of time in the transfer penalties
framework. Transfer penalties is a concept that represents generalized costs— including
monetary costs, time, labor, discomfort, inconvenience, etc.— involved in transferring from one
vehicle to another between the same or different transportation modes, and is well- established
theory in the travel behavior literature. When a traveler finds the total generalized cost of her/ his
trip by transit lowest among different means of transportation, she or he chooses to travel by
transit. Value of time is another important concept in examining the relative importance of
factors that influence people’s travel behavior, particularly in mode choice. The transfer
penalties framework provides the theoretical backbone for the importance of improvements
pertinent to transit transfers.
According to previous studies on transfer facilities, we found that within a typical transit trip,
a transfer accounts for approximately one quarter of total generalized costs ( or time). The
shorter the trip is, the more significant the impact of the transfer. Among several factors
associated with a transit transfer, waiting time is generally the most important component to
determine total generalized costs ( and time) as long as safety and security are ensured. Time
schedule and certainty of arrival time are two important factors to determine actual waiting time.
In comparison to actual waiting time, perceived waiting time is very important in determining
whether or not a traveler uses transit service. Perceived waiting time is affected by factors, such
as safety, security, comfort, whether waiting is forced or not, and acquired knowledge about the
arrival of the next vehicle.
In the examination of various attributes of transfer facilities that are thought to particularly
influence transit transfers, we make a clear connection between improvements at transfer
facilities and changes in people’s travel behavior due to a reduction in transfer penalties. In
other words, we distinguish two categories of improvements that are related to transit transfers:
1) those that affect actual time and costs of making a transfer, and 2) those that affect people’s
perception of transfer penalties. From this perspective, we identify the connection of transfer
costs, time scheduling, and five evaluation criteria associated with transfer facility attributes that
affect transfer penalties: 1) access, 2) connection and reliability, 3) information, 4) amenities,
and 5) security and safety. The effectiveness of transit agencies’ efforts to improve attributes of
transfer facilities can be understood in terms of the effectiveness to improve travelers’
experience at these facilities, reduce transfer penalties, influence travelers’ behavior in mode
choice, and eventually contribute to an increase in transit ridership.
We find that in order to improve the quality of transit transfers, transit agencies can work on
the operational aspects that influence transfers ( such as time schedule, on- time arrival, and
transfer fare) and the physical aspects of transfer facilities ( such as distance to make a transfer,
ii i
lighting, seating, signage, streamlining, circulation lines, protections from weather, visibility). It
is also an option for facility management to provide amenities at transfer facilities, such as
commercial establishments including news stands, coffee shops, convenience stores, and dry
cleaning stores. Physical aspects of transfer facilities can also affect walking time to travel
between locations where people alight and board vehicles for transferring. Such aspects can also
influence travelers’ experiences at facilities, and therefore their perceptions of waiting time,
walking time, and transfer penalties.
Because few studies have examined how the effects of physical improvements on transfer
facilities affect travelers’ choices to use transit service, it is important to investigate this issue in
greater detail. At the same time, it is important to recognize that improvements of service
operation are likely to have more significant impacts than physical improvements in facilities
alone will have.
We conclude from this review that there are three ways to enhance the scope of study from
our proposal: 1) as most transit transfers are intra- modal, these should be examined in addition to
intermodal transfers, 2) operational and managerial attributes of transfer facilities should be
examined in addition to the physical attributes of such facilities, and 3) steps need to be taken to
begin to develop more systematic, quantitative tools for evaluating transit transfer facilities.
Finally, it is important to examine the relative effectiveness of improvements on physical
attributes of transfer facilities as well as service operation whenever possible.
Key words: transfer facilities, travel behavior, transfer penalties, generalized costs, value of time
iv
LIST OF TABLES PAGE
TABLE 1 Direct and Indirect Factors Influencing Transit Ridership 5
TABLE 2 Measures Available to Transit Agencies 7
TABLE 3 Typical Transit Trip and Its Associated Time and Costs 15
TABLE 4 Overall Time Valuations ( relative to in- vehicle time = 1.0) 17
TABLE 5 Valuation of Transfer Penalties 21
TABLE 6 Estimated Subway- to- Subway Transfer Penalties at the MBTA 23
TABLE 7 Factors Affecting Attributes of Transfer Penalties 26
v
LIST OF FIGURES PAGE
FIGURE 1 Conceptual Framework to Determine the Cost of Total
Transfer Penalties 25
vi
TABLE OF CONTENTS
SECTION PAGE
EXECUTIVE SUMMARY i
LIST OF TABLES iv
LIST OF FIGURES v
PREFACE 1
1.0 INTRODUCTION 1
2.0 FACTORS INFLUENCING TRANSIT RIDERSHIP 3
2.1 Relative Effects of Factors Internal to Transit Agencies on People’s
Travel Behavior 6
3.0 THE FRAMEWORK OF TRANSFER PENALTIES WITHIN TOTAL
TRAVEL COSTS OF TRANSIT TRIPS 13
3.1 Example of Transfer Penalties in a Typical Transit Trip 13
3.2 Valuation of Time Associated with Components of a Transit Trip 17
3.3 Weighting of Time Associated with Elements of a Transit Trip 18
4.0 FACTORS INFLUENCING TRANSFER PENALTIES 24
4.1 Transfer Fare 26
4.2 Time Schedule of Transit Service 27
4.3 Transfer Facilities 28
5.0 SUMMARY AND CONCLUSIONS 29
6.0 BIBLIOGRAPHY 31
1
PREFACE
While private automobiles provide door- to- door travel, public transit requires people to walk to
bus stops and rail stations, wait for services, and often make transfers from one vehicle or mode
to another. Good interconnectivity in the transit system is essential to reduce the burden of
walking, waiting, and transferring and to provide a high quality of service for transit trips.
However, the transit system in California lacks interconnectivity between transportation modes
and often fails to efficiently serve the public that travel by public transit. To improve
interconnectivity in the transportation system, it is important to develop a methodology to
evaluate the quality of transferring in order to improve transfer facilities. Such improvements at
transfer facilities lead to a provision of seamless travel for transit users.
The research project, Tool Development to Evaluate the Performance of Intermodal
Connectivity ( EPIC) to Improve Public Transportation, will assist the California Department of
Transportation ( Caltrans), regional and local transportation related entities, transit operators, and
other stakeholders in evaluating interconnectivity issues pertaining to travel and in identifying
opportunities and solutions for improving transportation systems. This project contributes to
Caltrans’ goals of Flexibility and Productivity by assisting it in providing the appropriate tools to
contribute to a transportation system ─ with both intermodal and intra- modal components ─ that
maximizes safety, security, reliability, mobility, and access.
The larger scope of our research addresses the following three questions: First, what factors
at transfer facilities are important from the transit users’ perspective relative to determining their
travel behavior? Second, what factors at transfer facilities are important from the operators’
perspective relative to improving efficiency in transit service operation? Third, what factors at
transfer facilities are important from the neighboring community perspective that allow the
community to benefit from the presence of and services provided by such facilities?
In this literature review, we address the first question and investigate factors at transfer
facilities from the users’ perspective in relation to their travel behavior. This is the first step to
develop a tool to evaluate the performance of connectivity to improve public transportation. We
found it essential to: 1) understand where to improve the quality of transfers positioned within a
group of factors that affect transit ridership, 2) establish a conceptual framework to relate
improvements at transfer facilities to people’s travel behavior, and 3) identify a systematic
classification of transfer facility attributes in relation to the developed conceptual framework.
By understanding these factors, we will be able to identify improvements at transfer facilities
that will effectively lead to a transit ridership increase. While this literature review is theoretical
in developing a conceptual framework to relate improvements at transfer facilities to travel
behavior, we are producing a second literature review that examines the current practice of
evaluating connectivity based on attributes of transit facilities from the traveler, operator, and
community perspectives.
1. INTRODUCTION
When people choose to travel by foot, bike, bus, rail, or private automobile, they consider many
factors, such as time, labor, cost, security, convenience, and comfort for the entire trip— from
door to door. Needless to say, private automobiles have significant advantages in most aspects,
which helps to explain why over 86 percent of all metropolitan person trips in 2001 were in
private vehicles ( Hu and Reuscher 2004). Private vehicles – cars, trucks, vans, and motorcycles
– once owned, provide many benefits over public transit, including greater mobility, accessibility,
flexibility, convenience, comfort, and safety against crime. This poses a daunting challenge to
2
public transit agencies aiming to improve their transit service to compete with private vehicles.
Given that travelers tend to consider out- of- vehicle time ( walking, waiting, transferring, etc.) to
be substantially more burdensome than in- vehicle time, attracting travelers away from private
vehicles in significant numbers will require transit agencies to focus increasingly on improving
transit users’ experience outside of vehicles – walking to and from stops, waiting for vehicles,
and transferring between vehicles.
The importance of intermodal connectivity has been recognized for a long time. The
Committee on Intermodal Transfer Facilities of the Transportation Research Board in 1974
emphasized the importance of identifying factors to measure and be used to optimize total
transportation network effectiveness:
“ The intermodal transfer facility determines total transportation network
effectiveness. As a connecting node, the facility integrates the various
transportation modes to maximize the number of users. A poor connector would
discourage potential users or cause them to be diverted to other modes. Poor
transportation system operating practices sometimes introduce crowding and
delay, which can be attributed wrongly to inadequacy of the transfer facility.
There is a need to establish factors that optimize total transportation network
effectiveness. More information is required on the effect of system operating
practices on modal transfer efficiency and space use, and procedures should be
developed to improve efficiency and reduce space requirements, passenger
inconvenience, and delay ( Committee on Intermodal Transfer Facilities 1974).”
Attention to improving the connectivity of transit – between lines and systems – has been
increasing for some time. The ongoing suburbanization of U. S. metropolitan areas puts
traditional fixed- route transit service at a growing competitive disadvantage with private vehicles,
and makes serving increasingly far- flung trip origins and destinations increasingly costly. In
response, many cities, such as Boise ( Idaho), Sacramento ( California), and Seattle ( Washington)
( Pratt and Evans 2004), have adapted so- called “ hub- and- spoke” route systems to serve growing
service areas, increasing transfers in the process. A hub- and- spoke model derives its name from
a bicycle wheel, which consists of a number of spokes jutting outward from a central hub. In the
abstract sense, a location is selected to be a hub, and the paths that lead from points of origin and
destination are considered spokes. This transit model requires that people be routed through a
transfer station and make transfers among lines and systems before reaching their destination. In
larger metropolitan areas with many transit operators, the number of transferring passengers can
be very high. In such places, transfer centers are used to facilitate passenger transfers from one
line to another, from one mode to another ( car to/ from bus, bus to/ from rail, etc.), or from one
system to another and are central parts of transit networks.
Despite long- time recognition of its importance, efforts to address connectivity at transfer
facilities have proven less effective than expected. First, although connectivity at transfer
facilities is very important, both practitioners and researchers generally pay more attention to
quantity and quality of transit vehicle services ( in- vehicle travel) for their more intuitively
obvious effects on ridership. Second, because transfer facilities vary in size, modes served,
location, and amenities, it is hard to comprehensively analyze transfer facilities using uniform
criteria ( ITE Technical Council Committee 5C- 1A 1992). Third, most of the literature on
transfer facilities lacks a theoretical framework for how improvements of transfer facilities affect
people’s travel behavior and, subsequently, the overall ridership of the transit system. This lack
of causal clarity in the research on transit transfer facilities is an enormous drawback that this
research seeks to overcome.
3
Most previous studies of transit stops, stations, and transfer facilities have compiled laundry
lists of attributes that contribute to or detract from travelers’ transfer experiences, but have
largely failed to consider the relative importance of each of these attributes, or whether and how
they influence ridership separately or in concert with other factors ( Rabinowitz et al. 1989; Fruin
1985; Kittelson & Associates 2003; Vuchic and Kikuchi 1974; Evans 2004). As a result, we
know little about which attributes are most important, under which circumstances, and in what
combinations with other factors. In other words, we know very little about the effects of transfer
facilities on transit ridership and network performance. This state of knowledge based on past
studies on the subject is incomplete because it fails to guide transit agencies toward planning
practices that effectively improve the quality of transfers at transit centers that actually result in a
ridership increase.
This literature review addresses the lack of a theoretical basis for understanding the
relationship between transfer- facility attributes and travel behavior. We do this by placing the
literature in a transfer penalties framework. The concept of transfer penalties refers to
generalized costs — including monetary costs, time, labor, discomfort, inconvenience, etc.—
that is, those costs involved in transferring from one vehicle to another and, between the same or
different transportation modes, and is well- established theory in travel behavior literature
( Ortuzar and Willumsen 2004).
The implications of intermodal transit systems and the factors that affect transit ridership are
discussed at three levels in this report. First, reviewing past studies on determinants of transit
ridership, we find that policies and programs that transit agencies use to increase ridership have
had only limited effectiveness. We have found that transit use is determined largely by factors
outside the control of transit agencies, such as patterns of urbanization, regional economy, and
demographic factors. Second, we introduce a framework that places transfer penalties within the
context of total travel costs of a transit trip. The concept of travel costs is drawn from travel
behavior modeling, and has been examined extensively in transportation economics, engineering,
and planning literature. Value of time is another important concept in examining the relative
importance of factors that influence people’s travel behavior, particularly in mode choice. This
section provides the theoretical backbone for the importance of improvements pertinent to transit
transfers. Third, we examine factors thought to particularly influence transit transfers. In doing
so, we make a clear connection between improvements at transfer facilities and changes in
people’s travel behavior through reduction in transfer penalties, so that we will have in turn a
clear connection between transit agencies’ efforts to reduce transfer penalties and increased
ridership. From this perspective, we identify the relationship among transfer costs, time
scheduling, and five evaluation criteria of transfer facilities which affect transfer penalties: 1)
access, 2) connection and reliability, 3) information, 4) amenities, and 5) security and safety.
The final section summarizes the gaps in the current literature by clearly defining the objective
of this study, establishing a foundation for research on transit transfer facilities, and proposing an
agenda for further research on transit transfers. 1
2. FACTORS INFLUENCING TRANSIT RIDERSHIP
According to economic theory, transit ridership is determined by the level of service supplied in
the system and travel demand in the service area. Transit systems operate in diverse urban
1 Our second deliverable — a continuation of the review of the literature — focuses on reviewing aspects of
transfer facility evaluation and directly addresses the project’s research questions and explains the next steps in our
research that leads to the project deliverables.
4
environments where a variety of factors have been shown to influence service operation and
travel demand. While aggregate travel demand is subject to people’s socio- economic status,
residential and work locations, and the state of the regional economy, transit agencies determine
the level of service supply by taking into account their operating and financial conditions. Thus,
actual consumption of transit services ( i. e. transit ridership) can be considered a function of a set
of factors that affect transit demand and a second set of factors that affect transit service supply.
Factors that affect transit ridership, according to criteria by Taylor and Haas ( 2002) and
Transport and Travel Research Limited and European Commission ( TTRL & EC) ( 1996), can be
grouped into three categories. 1) External factors, such as physical geography and population
demographics; 2) Indirect measures, which include policy factors external to public transit
agencies — such as land use freeway plans; and 3) Direct measures, which include policy factors
internal to public transit agencies— such as service frequencies and fare levels ( See Table 1).
External factors directly affect transit travel demand and are not easily influenced by local
governments or transit agencies. 2 External factors include factors such as population and
employment growth, the regional economy, salary scales, residential and workplace locations,
and migration of people.
Indirect policy measures can be influenced by regional governmental actors ( TTRL & EC
1996). Local governments may be able to implement indirect measures to increase the relative
attractiveness of public transit services and influence peoples’ decisions about whether to take a
trip and on which mode ( TTRL & EC 1996). These measures include regulation, taxation, and
pricing for automobile use, land use planning, measures to reduce travel demand, and
enhancement of non- motorized modes. While indirect policy measures can strongly influence
transit use, they are usually outside of the control or influence of transit systems from the
perspective of transit agency managers ( Taylor et al. 2002).
Direct measures are under the control of transit agencies, according to the framework of the
study by Taylor et al. ( 2002). These measures enhance the advantages of public transit in
absolute terms, and make public transit more attractive as a mode of transportation. These
measures are related to the level of service provided, fare structure, service frequency and
schedules, route design, and service information.
Although transit agencies have a variety of measures to take, their effectiveness is limited,
compared to the impact of external factors. Direct policy instruments ( or direct measures) have
little influence on changes in people’s choice over transportation modes for travel ( TTRL & EC
1996; Taylor et al. 2002). TTRL & EC ( 1996) recommends that the most effective strategy is to
“ combine direct and indirect measures through a combination of physical, flow control and
relative pricing measures.” Despite their relatively low effectiveness, continuous efforts to
incrementally improve service by transit agencies are important by helping to provide mobility
and accessibility to transit dependents, reduce traffic congestion, improve air quality, and other
issues related to automobile use.
2 Here we distinguish travel demand that arises to meet people’s needs to travel to conduct other activities and
consumption of service that reveals actual movement of people by driving and taking public transit. In other words,
travel demand exists even though it may not be met due to the insufficient level of supply, as treated in general
consumer theory.
5
TABLE 1 Direct and Indirect Factors Influencing Transit Ridership
INDIRECT MEASURES
Improving the competitive position of public transport
CAR OWNERSHIP
Taxation of car ownership
Restrictions on car ownership
Road pricing
CAR USE ( AREA SPECIFIC)
Traffic calming
Access restrictions
Car vehicle specification
CAR USE ( GENERAL)
Fuel tax
Restrictions on car use
OTHER
Information on traffic conditions
Land- use planning
Tele- communications / tele- shopping
Flexible working hours
Increase in road capacity
Improvements to non- motorized modes
DIRECT MEASURES
How to improve the offer of public transport
PRICING
Fare levels
Ticketing regimes/ fare structure
Ticketing technology
Subsidy regime
Fleet size
SERVICE PATTERN
Extensiveness of routes
Distance to/ from stops
Service frequency/ travel time
Operating hours
SERVICE QUALITY
Vehicle characteristics
Bus/ rail stop quality
Interchange quality
Quality/ Number of staff
PRIORITY MEASURES
Link priority/ right- of- way
Junction priority
Quality regulations
REGULATORY REGIME
Market regulation
Operational regulations
INFORMATION
Information provisions
Publicity/ promotion
OTHER
Park- and- ride
Integrated approach
Source: Taylor et al ( 2002) and TTRC & EC ( 1996)
6
2.1 Relative Effects of Factors Internal to Transit Agencies on People’s Travel Behavior
Transit agencies can use direct measures to increase the relative attractiveness of transit service
to encourage people to choose transit among various modal options. In this section, we review
the effects of these measures that transit agencies can control, and carefully examine what
aspects of a trip are influenced by these measures.
Many studies on the subject prior to 1990 examined the impacts of various measures on
transit ridership or modal shift to transit service at an aggregated level. Subsequently, the focus
shifted to a disaggregated analysis using discrete choice models, which can take into account
various characteristics of individual travelers and trips. Since the impacts of various measures
are likely to vary by socio- demographic characteristics of travelers ( e. g. age, income, auto
access) as well as by trip characteristics ( e. g. trip purpose, travel time of day, trip length), it is
necessary to examine the impacts of various direct measures on people’s choice of travel mode
by different market segments ( Cervero 1990; TTRL & EC 1996). Past studies have reported that
changes in service quality, such as frequency of service and schedule reliability, have more
significant impacts on ridership than fare changes. However, few studies have examined how
improving transit facilities affects ridership ( Cervero 1990; TTRL & EC 1996; Paine et al. 1967;
Wachs 1981).
Table 2 presents an array of approaches available to transit agencies to increase ridership,
some of which are drawn from a list of direct measures in the TTRL & EC study ( 1996). In this
table, italicized items are related to transferring. The concept of elasticity is often used to
examine the effect of some measure on transit ridership. In this case, elasticity is defined as the
ratio of a percent change in ridership to a percent change for that measure. For example, when
transit ridership decreases by 10 percent with a fare increase of 20 percent, fare elasticity is - 0.5
(=- 0.1/ 0.2). 3 Since it is an algebraic calculation, it requires numerically quantifying a change in
some measure. For this reason past studies have primarily focused on measurements that can be
easily quantified, such as fare, service output, and headway, and less on other measures that can
be only qualitatively evaluated.
Fare and subsidy
Of all measures, fare elasticity has been examined the most in past studies. Cervero ( 1990)
reviewed studies up to 1988 with a focus of transit pricing and found that fare changes have
relatively small effects compared to changes in service quality, such as average headway and
speed. Most studies Cervero reviewed reported estimated fare elasticities between - 0.1 and - 0.5.
Similarly, the review of TTRL & EC ( 1996) reports fare elasticities in the range of - 0.2 to - 0.5.
In general, fare elasticities are approximately half of elasticities of changes in service quality.
Gaudry ( 1974) has found similar conclusions in his regression study that compares relative
effects of factors on transit ridership.
Studies on the effect of transit subsidies report a range of elasticities from + 0.2 to + 0.4 based
on a review of 11 international cases ( Bly, Webster, and Pounds 1980; TTRL & EC 1996).
However, the mechanism of the effect of transit subsidies on ridership is complex. While transit
3 When elasticity is between negative infinity and negative one, demand is elastic, which means the percentage
change in quantity is greater than that in price. When elasticity is between negative one and zero, demand is
inelastic, which means the percentage change in quantity is smaller than that in price. The negative sign indicates
that an increase in price leads to a decrease in demand, and vise versa.
7
subsidies certainly help to keep fares lower and increase service supply more than without
subsidies, it is not clear which of these two factors is a main cause for an increase in ridership
increase. Since part of the subsidies is often used to increase labor compensation, subsidies do
not increase service output in the same proportion ( Lave 1985), which, in turn, reduces the
effects of subsidies on ridership.
A fare structure is likely to significantly influence ridership especially when it varies by time
of day and trip distance, since it influences people’s mode choice of travel differently for
different socio- demographic groups and for different trip purposes. However, there has not been
much study done in this field. Smartcard technology is also related to fare structure, but is still
very new with little, if any, evidence of its impacts on ridership ( TTRL & EC 1996).
8
TABLE 2 Measures Available to Transit Agencies4
Group Factor Elasticity
Fare and subsidy fare level - 0.5~- 0.1( half of that of service quality)
subsidy regime + 0.2~ 0.4 ( its effect is not clear)
ticketing regime/ fare structures, -
ticketing technology ( smart card) -
Service supply: ( vehicle- km of bus service) + 0.2~ 0.7
route, stops -
station distance - 0.57~- 0.49
operating hours -
Service quality twice as much effects on
ridership as fare changes)
service frequency/ scheduled
journey time
-
waiting time - 0.54
Reliability -
vehicle speed ( in- vehicle travel
time)
- 1.16~- 0.59
vehicle speed ( in- transit time) - 0.54
link priority/ segregated right of
way
-
junction priority -
vehicle characteristics -
fleet size -
Transit facilities bus/ rail stop quality -
station facilities -
bus stop quality, station facilities -
terminal/ interchange quality -
park and ride -
information provision -
Others safety/ security -
publicity/ promotion -
market regulation -
number and quality of staff -
operational regulations/ quality
regulations
-
Service supply: route, stops, and operating hours
Since ridership is determined by service supply and travel demand, the level of service
supply certainly influences ridership. Elasticities of ridership to service supply measured by
vehicle- kilometers of bus services are in the range of + 0.2 and + 0.7 ( TTRL & EC 1996).
4 Vehicle speed is the only factor in this table that is estimated by a discrete choice model study. Other factors are
estimated by aggregate models or not specified at all in the literature.
9
Routing and the degree of route extension influence a transit system’s coverage area, and
therefore potentially influence ridership. The effect of these factors on ridership significantly
varies by area. Some scholars critique the expansion of transit service into suburban areas as
having the effects of lowering productivity, efficiency, and therefore effectiveness of transit
service ( Lave 1985; Garrett and Taylor 1999).
In contrast, the number or density of stops shows a relatively large impact on ridership,
because it affects access distance and walking time for transit users. Transit service demand with
respect to walk time is very sensitive ( Cervero 1990). TTRL & EC ( 1996) cites a study by
Gordon and Wilson ( 1985) to report that demand for light and heavy rail have elasticities of -
0.568 and - 0.485 respectively with respect to walking distance. Station distance also determines
distance that rail users may have to walk to access and therefore affects walking time, which is
perceived to be very onerous by travelers.
While some users may have a preference for longer operating hours, there has been no
careful study to separate the effects of longer operating hours from the effects of an increased
total service supply due to longer operating hours. In other words, is it the earlier and later hours
that attract riders, or simply the greater number of vehicle runs?
Service quality: service frequency/ scheduled journey time, vehicle speed, link priority/ segregated
right of way, junction priority, vehicle characteristics, fleet size
From TTRL & EC ( 1996), service frequency “ refers to average frequency, length of operating
day/ week, and reliability.” The most important objectives of scheduling and frequency
adjustment in service quality are to reduce overall travel time and improve convenience for
passengers ( Evans 2004). 5 Scheduling changes can be made to improve the reliability of service
that results in both actual and perceived waiting time for passengers and less anxiety ( Evans
2004). While frequency of service, headway, and reliability influence opportunities for waiting
time at stops/ stations, vehicle speed is a main factor to determine travel time ( or in- vehicle time).
In general, changes in service quality, such as average headways and speeds, have twice as much
effect on ridership as fare changes ( Cervero 1990).
It is very difficult to reliably measure service elasticities in response to multiple service
changes that often occur simultaneously – such as schedule changes that accompany a fare
increase. Further, most transit ridership data are in terms of unlinked trips, while travelers make
linked trips ( walk – wait – ride – walk, or walk – wait – ride – walk – wait – ride – walk in the
case of a trip with a transfer), where the out- of- vehicle aspects of the links have the largest
influence of perceived travel burdens. Such methodological challenges notwithstanding, Evans
( 2004) reports an elasticity of 0.5 in response to service frequency changes. When changes in
service hours and frequency were accompanied by aggressive marketing, such as direct mail
campaigns, free ride coupons, and image building by new bus paint designs in Santa Clarita and
Santa Monica, California, each transit system experienced significant ridership increases with
elasticities of + 1.14 and + 0.82 respectively ( Evans 2004; Mass Transportation Commission
1964).
5 In the TCRP report 95, Evans ( 2004) list the following types of scheduling and frequency changes for
discussion: 1) frequency changes, 2) service hours changes, 3) frequency changes with fare changes, 4) combined
service frequencies, 5) regularized schedules, and 6) reliability changes. Combined service frequencies is the
approach to offer a combination of different transit services on the same corridors to accommodate diverse trips
taken by different groups of transit users.
10
In general, higher values of elasticity are achievable when frequency changes are made to
transit lines with previous service schedules with 60 minute or 30 minute headways and when
riders are mainly in middle and upper income groups ( Evans 2004). On the other hand, elasticity
tends to be relatively low when previous service already has short headways and the majority of
patrons are from lower income groups ( Evans 2004). In addition, different groups of transit
users have different responses to frequency changes. Off- peak riders are often more sensitive to
frequency changes than peak period riders ( Evans 2004). Since transit dependents are likely to
use transit service even though service quality may not be satisfactory, an increase in ridership
due to frequency changes is often attributed to an increase in new discretionary ( choice) riders
who are likely to be in middle and upper income groups ( Holland 1974).
Scheduled journey time and vehicle speeds affect in- vehicle travel time. Cervero ( 1990)
reports in- vehicle travel time elasticities in the range of - 0.59 and - 1.16 from two mode choice
studies ( McGillivrary 1969; Domencich, Kraft, and Valette 1968), in which the high end
represents an elasticity in the peak period. Gaudry ( 1974) reports elasticity of 0.27 for in- transit
time, compared to fare elasticity of - 0.15.
Service frequency and reliability determine travelers’ waiting time at transit facilities. Transit
riders are found to be very sensitive to out- of- vehicle time, and among various types of out- of-vehicle
time, waiting time is the most onerous factor to transit users ( Cervero 1990). Gaudry
( 1974) reports elasticity of - 0.54 for waiting time.
Reliability is one of the most important factors to attract transit ridership. Commuters in
attitudinal studies conducted in Baltimore and Philadelphia considered “ arrival at intended time”
as the second most important for work trips, following “ arrival without accident ( Evans 2004).”
Similar results were shown in a survey in Boston and Chicago; “ arrival at intended time” is more
important than travel time, waiting time, and cost measures ( Evans 2004). Improvement in
reliability and speed in urban bus services in Britain in the 1970s significantly increased
ridership ( TTRL & EC 1996). In the study conducted by Horowitz and Thompson ( Horowitz
and Thompson 1995), time- scheduling and reliability are the second most important attribute at
transfer facilities following safety and security. Douglas ( 1991) found in a study in New Zealand
that the value of expected delay was 8 times as much as that of walk time for rail users ( TTRL &
EC 1996). Waiting time with uncertainty of arrival of the next vehicle increases the value of
waiting time by a factor of two ( Webster 1977).
Link priority, segregated right of way, and junction priority generally influence ridership
through their impacts on variability of travel time and in- vehicle travel time. The effect of bus
lanes has been found to be less than expected in the studies reviewed by ( TTRL & EC 1996).
While one study reports that the reduction of travel time by increased speed of a light rail line
using junction priority from 33 minutes to 22 minutes increased ridership by 10 percent, the
measure of junction priority is not developed enough and it is still difficult to evaluate its effect
( TTRL & EC 1996).
It is also difficult to quantify vehicle characteristics, and there is no hard evidence to support
particular vehicle characteristics, although people generally prefer comfortable rides by rail
vehicles to those by buses.
Transit facilities: Bus/ rail stop quality, station facilities, terminal/ interchange quality, park and
ride, information provision
The quality of transit facilities can have significant impacts on attracting ridership to transit
systems in several different ways. Since one of the main functions of transit facilities is to
11
accommodate users’ waiting time, factors such as comfort, security, safety, and convenience,
influence people’s experience in taking public transit service, and therefore increase their
likelihood of choosing transit service over other modes. However, past studies provide little
evidence that clearly indicates a direct connection between qualities of transit facilities and
ridership. As we discuss in later sections of this report, qualities of transit facilities can
indirectly affect transit demand and ridership by improving travelers’ experience at facilities. 6
Some studies examined the values transit users placed on components of terminals ( e. g.
including waiting facilities, lifts/ escalators, catering facilities, and information displays),
terminal/ interchange quality, and park- and- ride facilities. Survey respondents in the study by
Douglas ( 1991) value improvements on stations as much as those on trains ( TTRL & EC 1996). 7
However, the effects on transit demand are unknown ( TTRL & EC 1996). 8 The only study that
took into account a component of transit facilities in a discrete choice model is the study by Guo
and Wilson ( 2004), which showed that the presence of escalators to assist level changes for
transferring at subway stations could reduce transfer penalties.
At the same time, it is not difficult to think that a small change in transit facilities will not
dramatically change people’s travel behavior. A study in Lima, Peru, showed that bicycle
storage and easier access for the handicapped by replacing stairways did not have a statistically
significant impact on people’s choice of travel mode in the stated preference survey, while
increase in feeder service to rails and in bus rapid transit service to downtown were found
important ( Martinez 2003).
Travel time interconnectivity at transfer facilities is very important. This is determined
mainly by vehicle scheduling: “ Specific benefits from adjusting frequencies so that services
interconnect efficiently. Values of waiting time on transferring ( or interchange) and delays are
high ( TTRL & EC 1996).” Several studies in recent years developed models to minimize the
uneasiness, inconvenience, and other costs associated with transit transfers. 9 These studies used
a modeling approach to optimize time- related functions such as time tables and vehicle
dispatching to reduce waiting time ( Shayer 2004).
In the survey study by Douglas ( 1991), respondents placed a value of seven New Zealand
cents on at- stop ( rail) information in addition to having leaflets, and also placed a similar value
on a telephone inquiry system, and real- time information ( TTRL & EC 1996). However, no
6 One of the main problems in past studies that evaluated the qualities of transit facilities is a lack of a conceptual
framework that explains how facility improvements can affect transit demand and ridership and how cost effective
those improvements are. For example, although almost all transit users would like to have shelters and benches at
bus stops, the presence of shelters and benches does not necessarily increase ridership, as the presence of
refrigerators and laundry machines at bus stops, for an extreme example, does not necessarily increase ridership.
This lack of causal clarity in the research on transit transfer facilities is an enormous drawback when transit agencies
implement transit facility improvements in order to increase the overall ridership in the transit system.
7 Network Southeast have values for station appearance, station facilities ( including catering) and information,
although there is some debate about the plausibility of some of these values ( See Cuthbertson et al., 1993).
8 London Underground and British Rail have determined the values passengers place on terminals ( Case study
2.5). A look- up table of interchange ( or transfer) penalties has been developed based on distance and connection
time, to take into account that certain interchanges are more onerous than others. Evidence from Thameslink
suggests that this method may have underestimated the penalty of cross London interchanges, which has
implications for other rail schemes.
9 These studies include Bookhinder and Desilets ( 1992), Chowdhury and Chien ( 2001), Chowdhury ( 2001), and
Boile ( 2002).
12
study has been found that provides evidence of a significant effect of route- specific service
information on an increase in ridership.
There are other measures listed by TTRL & EC ( 1996). These include publicity/ promotion,
market regulation, number and quality of staff, and operational regulations/ quality regulations.
These measures, however, lack hard evidence of their effects on transit demand.
Safety and Security
While it may not necessarily attract new ridership, improving the built environment to reduce
overall crime may have a significant impact on regaining transit users’ confidence. Transit
security is a serious concern in most metropolitan areas of the United States. Studies that
examined the relationship between transit facilities and crime show certain built environment
attributes contribute to higher and lower crime rates. Crime rates were higher for bus stops near
alleys, multi- family housing, liquor stores and check- cashing establishments, vacant buildings,
and graffiti and litter ( Loukaitou- Sideris et al. 2001; Liggett, Loukaitou- Sideris, and Iseki 2001).
In contrast, good visibility of the bus stop from its surroundings, large numbers of pedestrians,
and the existence of bus shelters contributed to lower crime rates ( Loukaitou- Sideris et al. 2001;
Liggett, Loukaitou- Sideris, and Iseki 2001).
While the studies found that the most important predictor of crime is the location of bus stops,
appropriate design and layout of the physical characteristics around transit facilities at the micro
level can affect opportunities for and likelihood of criminal activity ( Liggett, Loukaitou- Sideris,
and Iseki 2001). In the case of the light rail system that runs through the median of the Century
Freeway ( I- 105) in Los Angeles, the study found a high crime rate at park- and- ride facilities
adjacent to stations. While these parking lots are partially fenced and adequately lit, a lack of
pedestrian activity reduces the level of ambient surveillance and may facilitate criminal activities
( Loukaitou- Sideris, Liggett, and Iseki 2002). Platforms of five stations with high crime rates are
located in the middle of the freeway median and isolated from surrounding neighborhoods
( Loukaitou- Sideris, Liggett, and Iseki 2002). These stations are likely to suffer from little
visibility and natural surveillance as well as several hiding places ( under stairs and behind
pillars), and result in higher crime rates. There is certainly correlation between the built
environment at and around transit facilities and the incidence of crime. The sense of security is
so important in people’s choice of travel mode, time of travel, and route that it may completely
deter taking transit. Therefore, transit agencies should maintain a certain minimum level of
security, taking measures of policing and improving the built environment.
Overall, measures available to transit agencies have only limited effects to increase ridership
in comparison to the effects of external factors and indirect measures in policy options that are
outside the control of transit agencies. Past studies provide more information on the effects of
factors that are easily quantified, such as fare, service output, and headway, on ridership, and
have resulted in an understudy of other measures that can be only qualitatively evaluated. There
is no clear theoretical framework to relate qualities of transit facilities to transit demand,
ridership, and travel mode choice. The majority of past studies that examined the effects of
various factors used aggregated analyses that are not capable of examining the effects of qualities
of individual transit facilities. Although disaggregated analyses using discrete choice models are
capable of such examinations, only few studies actually took into account qualities and
components of transit facilities. All of these contribute to a lack of evidence to evaluate the
effects of qualities of transit facilities on transit ridership.
13
In addition, it is also important to take into account cost effectiveness as well as political
feasibility of adopting various policies and programs, including improvements of transit facilities,
so that policy makers and planners can choose the best strategies to increase transit ridership.
3. THE FRAMEWORK OF TRANSFER PENALTIES WITHIN TOTAL TRAVEL
COSTS OF TRANSIT TRIPS
“ Understanding what affects the transfer penalty can have significant
implications for a transit authority. It can help identify which types of
improvement to the system can most cost- effectively reduce this penalty, thus
attracting new customers, and helping determine the value of improvements to key
transfer facilities ( Guo and Wilson 2004).”
The concept of transfer penalty represents generalized costs — including monetary costs, time,
labor, discomfort, inconvenience, etc. — involved in transferring from one vehicle to another
between the same or different transportation modes, and is well- established theory in the travel
behavior literature ( Ortuzar and Willumsen 2004). The concept of travel costs is drawn from
travel behavior modeling, and has been examined extensively in transportation economics,
engineering, and planning literature. In the transportation literature, the term “ transfer
penalties” is used in two different definitions. In a broader definition, transfer penalties is a
general term to represent all of the monetary costs, time, labor, inconvenience, and emotional
distress pertinent to making a transfer, and generally work as an impedance factor for travel. In
this broader definition, transfer penalties consist of factors, such as transfer fare, walking time
and labor, waiting time and labor, comfort, safety, and convenience ( Liu, Pendyala, and Polzin
1997). 10 In contrast, in a narrower definition, transfer penalties are an impedance factor in
transferring after excluding factors that we can easily quantify, such as waiting time, walking
time, and transfer fare. In other words, transfer penalties in the more narrow definition are the
penalties beyond the monetary and time costs associated with making transfers ( Liu, Pendyala,
and Polzin 1997).
3.1 Example of Transfer Penalties in a Typical Transit Trip
In the following example, we will use a description from Currie’s article ( Currie 2005). A
typical one- way transit trip consists of the following attributes ( minutes in parentheses are
numbers that we chose for this example): 11
1) access by walking from a trip origin to a bus stop ( 8 minutes),
2) wait at a bus stop ( 4 minutes),
3) travel in vehicle from a bus stop to a rail station ( 20 minutes),
4) transfer from a bus stop to a rail station, involving walking ( 6 minutes), waiting ( 10
minutes), and other transfer penalties,
10 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, magazine, and 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.
11 Liu, Pendyala, and Polzin ( 1997) also states 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.”
14
5) travel in vehicle from a rail station to another ( 30 minutes), and
6) egress from a rail station to a trip destination ( 6 minutes).
Assuming we can convert all of time, fare, and qualities of travel into generalized cost, a formula
to compute the total generalized cost ( TGC) for this trip looks like:
TGC = {( Walkt * Walkw) + ( Waitt * Waitw) + ( IVTt * IVTw)
+ ( NT * TPb) + MSCm} * VOT + Fare ----- Eq. ( 1)
Where:
Walkt: time in minutes walking to and from the transit service
Walkw: passenger valuation of walk time to and from transit stops
Waitt: time waiting for transit vehicle to arrive at the transit stop
Waitw: passenger valuation of wait time at transit stops
IVTt: travel time in transit vehicles
IVTw: passenger valuation of in- vehicle travel time
NT: number of transfers
TPb: transfer penalty, including transfer walking and waiting in a broader sense12
MSCm: mode specific constant for transit mode m
VOT: value of travel time
Fare: average fare per trip
Following the definition of transfer penalties in both the broad and narrow senses, we can further
decompose TPb:
TPb = ( Walktt * Walkw) + ( Waittt * Waitw) + TPn ----- Eq.( 2)
Where:
Walktt: time in minutes walking to make a transfer
Waittt: time waiting for transit vehicle to make a transfer
TPn: transfer penalty, including transfer walking and waiting in a narrow sense
In Eq. ( 2), weights represent different valuations of time for different attributes. Weights, in
this context, can be interpreted as the differences between actual travel time and the time
perceived by a traveler. In a mode choice, travelers make their travel decisions based on the total
generalized cost of the trip in their calculation, which partly depends on their perception of
transfer attributes, such as time and other burdens associated with different segments in transit
trips.
Table 3 shows time and costs associated with components of a typical transit trip. Walking
in Eq. ( 1) is further divided into different segments of a trip: 1) ingress, 2) transfer, and 3) egress.
12 TPn and TPb are equivalent to Interchange I and Interchange II respectively in Wardman’s study ( 2001), which
will be reviewed in a later section.
15
This example includes two kinds of waiting time: 1) waiting at a bus stop for the initial segment
of trip and 2) waiting for making a transfer. It also has two types of in- vehicle time and two
types of fare for bus and train. This example does not include mode specific constant in Eq. ( 1).
We assume the monetary value of in- vehicle time is $ 7.50 per hour— half of an assumed
wage rate of $ 15 per hour. We use average valuation of walking time, waiting time, and other
transfer penalties according to a study by Wardman ( 2001). Monetary value of walking time,
waiting time, and other transfer penalties are computed to be $ 12.45 per hour, $ 11.03 per hour,
and $ 1.32 per transfer respectively based on our assumptions. We have intentionally made costs
associated with other transfer penalties comparable to other costs in this example— and $ 1.32 for
“ Other transfer penalties” in Table 3.
In this example, transfer penalties, including transfer walking and waiting time, account for
26 percent of the total generalized cost of the trip. In the fourth column which assumes that
people can make a transfer without waiting, the total travel cost decreases by 11 percent. In the
fifth column, which assumes no waiting time for transferring, the total travel cost decreases by 7
percent. In the sixth column which assumes no waiting and walking time ( for example, a timed-transfer
across a platform), the total travel costs significantly decreases by 18 percent. The
proportion of costs associated with transfer penalties in total costs can be reduced from 26
percent to 9 percent in the case that transit users have to spend for neither waiting nor walking.
Thus, the significant portion of the total generalized cost of a trip can be attributed to transfer
penalties, and can be reduced by providing timed- transfers which do not require transit users to
wait or walk long distance to transfer. We will extensively review these transfer penalties in a
later section.
TABLE 3 Typical Transit Trip and Its Associated Time and Costs
Typical
No transfer
waiting
No transfer
walking
No transfer
walking &
waitin
Time ( min.) Cost Cost Cost Cost
Access by walk from trip origin to bus stop 8 $ 1.66 $ 1.66 $ 1.66 $ 1.66
Wait at a bus stop 4 $ 0.74 $ 0.74 $ 0.74 $ 0.74
Bus fare ($ 1.35) - $ 1.35 $ 1.35 $ 1.35 $ 1.35
Travel in vehicle from a bus stop to a rail station 20 $ 2.50 $ 2.50 $ 2.50 $ 2.50
Transfer Penalities
Transfer from a bus stop to a rail station: walking 6 $ 1.25 $ 1.25 $ 0.00 $ 0.00
waiting 10 $ 1.84 $ 0.00 $ 1.84 $ 0.00
Other transfer penalties* - $ 1.32 $ 1.32 $ 1.32 $ 1.32
Travel in vehicle from rail station to another 30 $ 3.75 $ 3.75 $ 3.75 $ 3.75
Train fare ($ 1.35) - $ 1.35 $ 1.35 $ 1.35 $ 1.35
Egress from a rail station to a trip destination 6 $ 1.25 $ 1.25 $ 1.25 $ 1.25
Total 84 $ 16.99 $ 15.16 $ 15.75 $ 13.91
Reduction in total costs - 11% 7% 18%
% of transfer penalties in TOC 26% 17% 20% 9%
Weight Hour Minute
Wage 2.00 $ 15 $ 0.25
In- vehicle travel 1.00 $ 7.50 $ 0.13
Walking** 1.66 $ 12.45 $ 0.21
Waiting** 1.47 $ 11.03 $ 0.18
Other transfer penalties** 17.61 $ 132.08 -
*: Other transfer penalties is further weighted by 0.01 to make its cost comparable to other costs.
**: The ratio relative to in- vehicle time is taken from Wardman ( 2001).
16
In the above example, we assumed that weights ( or valuation of time) for different attributes
are constant. However, weights for different attributes vary by differences between perceived
time and actual time.
People perceive tim