-
Title
-
Eng
A Case Study of Bicycle Facilities Network Implementation: The Seattle Bicycle Master Plan
-
Date
-
2011
-
Creator
-
Eng
Bombara, Justin
-
Subject
-
Eng
Environmental Studies
-
extracted text
-
A CASE STUDY OF BICYCLE FACILITIES NETWORK IMPLEMENTATION:
THE SEATTLE BICYCLE MASTER PLAN
by
Justin Bombara
A Thesis: Essay of Distinction
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
June 2011
©2011 by Justin Bombara. All rights reserved.
This Thesis for the Master of Environmental Study Degree
by
Justin Bombara
has been approved for
The Evergreen State College
by
________________________
Martha L. Henderson, PhD
Member of the Faculty
________________________
Date
ABSTRACT
A Case Study of Bicycle Facilities Network Implementation:
The Seattle Bicycle Master Plan
Justin Bombara
Cities across the United States have increasingly adopted Bicycle Master Plans (BMPs)
to promote the bicycle as an environmentally and economically sustainable form of urban
transportation. BMPs embody a community’s vision for integrating bicycles into existing
transportation infrastructure, while outlining the policies for adoption necessary to
support cycling. Despite widespread adoption of BMPs by state and local governments,
minimal attention has been given to factors that influence successful plan
implementation. Through a review of the literature from the planning and public health
fields, this study assesses the empirical support for the proposition that creating bicycle
facilities increases ridership. Interviews with key participants involved in
implementation of the Seattle Bicycle Master Plan from 2007-2009 are analyzed to
identify factors influencing implementation of projects as part of the Bicycle Facilities
Network. An “ideal” policy implementation framework is used to structure interviews
and is subsequently assessed for its applicability in the Seattle case. A positive
correlational relationship is found between bicycle facilities and bicycle use, although
self-selection cannot be ruled out due to limitations of the research designs currently
employed in the literature. Several factors are identified that are critical to BMP
implementation, including the presence of dedicated funding, a Complete Streets policy,
the political will of elected officials, and public support of constituency groups.
Problematic for implementation are a lack of streets space, public opposition, and
expenses associated with capital projects. The implementation framework used to
structure interviews was successful in identifying major influences on implementation
and its use is recommended for future case study research on bicycle facilities
implementation.
Table of Contents………………………………………………………………………………………………..
iv
List of Tables………………………………………………………………………………………………………………… vi
Acknowledgements……………………………………………………………………………………………………… vii
1. Introduction……………………………………………………………………………………………………………. 1
2. Literature Review and Methodology………………………………………………………………………
2.1 Review of Bicycle Policy Implementation Literature…………………………………………..
2.2 Study Question and Methodology: Bicycle Master Plan Implementation..………….
2.2.1 Policy Implementation Evaluation Framework….………………………………………….
2.2.2 Interview Instrument..………………………………………………………………………………….
2.3 Study Question and Methodology- Theory Behind Policy Implementation…………..
6
7
11
14
19
20
3. Data………………………………………………………………………………………………………………..........
3.1 Factors Influencing Implementation of the Seattle Bicycle Master Plan………………
3.1.1 Context and Leadership: Policy Framework and External Constraints……………
3.1.1.a Complete Streets Ordinance…………………………………………………………………..
3.1.1.b Seattle Climate Action Plan…………………………………………………………………….
3.1.1.c City of Seattle Comprehensive Plan…………………………………………………………
3.1.1.d Street Space..............................................................................................
3.1.2 Resourcing-Timing…………………………………………………………………………………………..
3.1.2.a Environmental Awareness and Concern………………………………………………….
3.1.2.b Increase in Gasoline Prices…………………………………………………………………….
3.1.3 Political Stability…………………………...................................................................
3.1.4 Resourcing: Staff, Time, Funding and Expertise……………………………………………..
3.1.4.a Staff…………………………………………………………………………………………………………
3.1.4.b Funding: Bridging the Gap Levy (BtG)………………………………………………………
3.1.4.c Funding: Capital Project Cost……………………………………………………………………
3.1.4.d Expertise………………………………………………………………………………………………….
3.1.5 Monitoring………………………………………………………………………………………………………
3.1.6 Leadership: Single Implementing Agency……………………………………………………….
3.1.7 Ability to Enforce Compliance…………………………………………………………………………
3.1.8 Agreement on Objectives to be Achieved: Mix of Facilities…………………………….
3.1.9 Leadership: Policy Champion………………………………………………………………………….
3.1.10 Support: Flexible Attitude Toward Public Reaction………………………………………
3.1.11 Support: Public Trust, Support, and Opposition……………………………………………
3.2 Factors Influencing the Decision to Bicycle……………………………………………………………
3.2.1 Influence of Bicycle Lanes and Pathways on Bicycle Ridership………………………
3.2.2 Influence of Socio-demographic Characteristics on Bicycle Ridership……………
3.2.2.a Gender……………………………………………………………………………………………………………….
3.2.2.b Race…………………………………………………………………………………………………………
3.2.2.c Age………………………………………………………………………………………………………….
3.2.2.d Education Level……………………………………………………………………………………….
3.2.2.e Income, Employment and Workforce Characteristics………………………………
3.2.2.f Housing Characteristics……………………………………………………………………………
3.2.3 Influence of Individual and Attitudinal Factors on Bicycle Ridership………………..
22
23
27
28
30
33
34
36
37
40
41
43
44
46
48
51
53
57
61
65
69
74
77
82
86
91
92
92
93
93
94
95
96
iv
3.2.3.a Bicycle Ownership………………………………………………………………………………….
3.2.3.b Car Ownership and Attitudes Towards Motor Vehicles………………………….
3.2.3.d External Support…………………………………………………………………………………….
3.2.3.e Attitudes Towards Cycling, Transit and Walking……………………………………
3.2.3.f Individuals’ Health/Attitudes Towards Exercise…………………………………….
3.2.3.g Environmental/Economic Awareness……………………………………………………
3.2.4 Influence of Environmental Factors on Bicycle Ridership……………………………..
3.2.4.a High Traffic Densities/Presence of Automobiles……………………………………
3.2.4.b Weather Related Variables……………………………………………………………………
3.2.4.c Geographic Factors: City Population and Density………………………………….
3.2.4.d Neighborhood Level Characteristics: Trip Origins………………………………….
3.2.4.e Trip-specific Variables: Distance and Destinations…………………………………
3.2.4.f Slope………………………………………………………………………………………………………
3.2.4.g Road Quality………………………………………………………………………………………….
3.2.5 Influence of Policy Level Variables on Bicycle Ridership.................................
3.3 Assessment of Models in Relation to Implementation Framework…………………….
96
97
98
99
100
101
102
102
103
104
105
106
107
108
108
109
4. Analysis and Conclusions………………………………………………………………………………………….. 114
References…………………………………………………………………………………………………………………… 128
Appendices………………………………………………………………………………………………………………….. 133
Appendix A. Background: Goals and Objectives of the Seattle Bicycle Master Plan….
133
Appendix B. Seattle Bicycle Master Plan Thesis Interview Questions……………………….
136
Appendix C. Predictive Models in the Literature: Attitudinal,
Socio-demographic, Environmental and Policy Level Factors
Associated with Bicycle Use………………………………………
138
v
List of Tables
Table 1: Theoretical Policy Implementation Framework……………………………………………
16
Table 2: Content Review of Seattle Bicycle Advisory Board Notes:
Potential factors Influencing Implementation of the Seattle Bicycle
Master Plan, 2007-2009………………………………………………………
18
Table 3: Factors identified by study participants that influenced implementation
of the SBMP 2007-2009……………………………………………………………………
25
Table 4: Seattle bicycle Master Plan Higher Costs Projects in
Areas of High Bicycle Demand………………………………………………………….
50
Table 5: Mile of Recommended Facilities……………………………………………………………………
135
vi
Acknowledgements
My deepest thanks to the advocates, planners and implementation staff at the City of
Seattle who were of great assistance in developing this research topic and providing the
data necessary to make a meaningful contribution to non-motorized transportation
planning. Their tireless efforts to provide the simple option of bicycling safely on public
streets are immensely appreciated.
I would like to thank my advisor, MES Director Martha Henderson, for her countless
edits, her flexibility, and her patience, all of which focused my thinking and contributed
in no small way to the successful completion of this thesis.
Gail Johnson, in the roles of both mentor and friend, offered needed support and
encouragement through the long days. Her wisdom and guidance were invaluable
throughout the process.
And finally, to the person who put up with it all, Janelle Drogo’s unending love, support
and patience throughout my time with the MES program was more than I could ever ask
for from a partner. Yes, you finally get me back.
vii
1. Introduction
The U.S. Energy Information Administration’s assessment of greenhouse gas
emissions for 2008 holds the transportation sector’s motor gasoline usage as
responsible for 1134.9 million metric tons of carbon dioxide, an amount corresponding
to 19.4% of the total emitted that year. As the predominant contributor to humaninduced global climate change, curbing emissions of CO2 in the transportation sector will
be a critical piece in addressing one of the defining issues of our time. Buried within
these figures is a dependence on the automobile for short trips. The 2009 National
Household Travel Survey found that when all trips are considered for 2009, 50% of trips
were 3 miles or less, with the automobile being responsible for 72% of those trips.
Bicycling currently only accounts for 1.8% of all trips less than 3 miles, while 85% of all
biking trips are 3 miles or less. Distances covered by a cyclist of average physical ability,
then, have the potential of supplanting many of the trips currently covered by the
automobile. Yet until recently, the bicycle has been considered by the Federal Highway
Administration as “one of the forgotten modes.” (FHA, 2010)
Efforts aimed at promoting the bicycle as a viable means of sustainable
transportation in the United States have stemmed from a growing interest at the
federal, state and local levels in addressing urban pollution, greenhouse gas emissions,
traffic congestion, and public health. Public policy has paralleled this interest to some
extent, with 33 of the largest 51 United States cities reporting public commitments to
the goal of increasing cycling, up from 2008 when only 25 cities reported having these
goals (Alliance for Biking & Walking, 2010). Federal spending on pedestrian and bicycle
1
improvements has followed suit, increasing at an exponential rate from only $6 million
in 1990 to $783 million in 2009 (FHA, 2010). While these funds account for only a small
percentage of the federal transportation budget, it is clear that the bicycle, previously
considered a fringe mode of transportation, has benefitted as a result of this remarkable
shift in priorities.
In service of these commitments, the adoption of Bicycle Master Plans (BMPs)
has been viewed by local governments as an essential component in making
communities more conducive to cycling (Litman et al., 2009). A BMP is a planning
document that embodies a community’s vision for integrating bicycles into existing
transportation infrastructure, while outlining the policies for adoption to support
cycling. BMPs identify goals, objectives and evaluation criteria for bicycle planning,
design, education, enforcement and encouragement, in turn outlining specific actions
for municipal agencies to implement these objectives (Litman et al., 2009). As of 2010,
25 states and 40 cities have adopted BMPs. Charting the longitudinal progress of states
and cities in adopting these plans on a nationwide level has not been possible, however,
as baseline data was not collected in previous years (Alliance for Biking and Walking,
2010).
In September of 2006, Seattle’s former Mayor Greg Nickels released Seattle’s
Climate Action Plan a strategy document outlining specific action items necessary for
the city to meet or exceed the international Kyoto Protocol’s climate pollution reduction
goals. Citing that nearly a quarter of Seattle’s greenhouse gas emissions come from cars
and that most trips are within five miles from home, the CAP directed the Seattle
2
Department of Transportation (SDOT) to complete the Seattle Bicycle Master Plan
(referred to in this paper as SBMP, or simply “the Plan”) in 2007 (City of Seattle, 2006).
The support of elected officials embodied in the CAP, coupled with the support of the
public and the concurrent passage of a dedicated funding source through the voter
approved “Bridging the Gap” transportation package, provided the necessary conditions
for SMP adoption in 2007. After three years of discussions, City Council members
passed the SBMP unanimously on November 5th, 2007 (Lindblom, 2007).
Central to achieving the Plan’s goals of increasing bicycle mode share and safety
in Seattle is the creation of a 450-mile network of bicycle facilities connecting the entire
city, referred to as the Bicycle Facility Network (BFN). Bicycle facilities, for the purposes
of this paper, refer to on and off-street bike lanes, including multi-use trails. The
creation of the BFN is considered “fundamental” to achieving the goal of the plan
(SDOT, 2007), representing the bulk of public investment in the Plan that will in turn
make lasting and extensive changes to Seattle’s transportation infrastructure. The
Network itself is defined as including all current bicycle facilities as well as proposed
ones that are detailed as specific action items in the Plan. The completed BFN is
expected to connect all parts of the city and, providing a bicycle facility within onequarter mile of 95% of Seattle residents.
The creation of bicycle facilities guided by a BMP represents an enormous
investment by localities in terms of staff, funding, and changes to existing street space.
Seattle, a medium size city of just over 600,000 residents (U.S. Census Bureau, 2010),
estimates the cost to implement the SBMP at $240 million (based on 2007 dollars)
3
(SDOT, 2007). By comparison, Portland’s 2030 BMP envisions 681 miles of facilities at a
projected cost of $613 million. Such large public investments beg the question of
whether there is empirical support for the creation of bicycle facilities as an effective
long term strategy for increasing the number of bicycle riders. Without such knowledge,
localities contemplating the creation of a BMP cannot be confident that bike lanes and
off-road trails are likely to have the intended benefit when implemented.
The purpose of this study is to identify factors that influenced implementation of
the SBMP during the period of 2007-2009. A policy implementation framework created
by Hogwood and Gunn (1984) is used as the basis for interview questions posed to key
respondents who had first-hand involvement with implementation of the SBMP during
the study period. Interview responses are analyzed by corroborate significant factors
and themes and are discussed in Section 3, Data. In light of Hogwood and Gunn’s
requirement that a policy or program be based on a sound theory of cause and effect, a
second question posed by the study involves whether a relationship can be established
between bicycle facilities and increased ridership. An assessment of the theory
underlying the policy, in terms whether a direct and causal relationship exists, is
necessary to determine the extent to which the policy is likely to achieve the stated end
of increasing ridership. To this end, an analysis of models in the peer-reviewed
literature that address bicycle ridership will provide a basis for evaluating the extent to
which it meets the criteria for causality with minimal linkages.
The study concludes with a brief assessment of the policy implementation
framework chosen in its relation to identifying factors through the interview instrument
4
employed by the author. The policy implementation framework created by Hogwood
and Gunn was successful in organizing the interview instrument to elicit responses from
study participants, with several critical factors identified. Refinements will be suggested
for future researchers to test against other U.S. cities in the process of implementing
bicycle plans based on the information provided by key respondents in the Seattle case.
Models detailing influences on bicycle use show that a positive relationship exists
between bicycle facilities and ridership, but a significant number of variables may serve
to mitigate the effect, depending on local variables. The direction of the relationship is
uncertain, and meeting Hogwood and Gunn’s requirement for causality is not possible
given the current state of literature.
This study contributes to the fields of transportation planning and public health
by assessing whether empirical evidence exists for the creation of bicycle infrastructure
in order to increase ridership. Understanding what factors influenced implementation
in the Seattle case will provide planners and advocates contemplating the creation of a
plan a “lessons learned” from a medium-sized, Pacific Northwest city.
The following section establishes the need for the current study by reviewing the
current literature available on bicycle policy and program implementation. Specific
study questions are then defined along with appropriates methodologies and data to
determine factors influencing implementation of the SBMP and bicycle ridership.
5
2. Literature Review and Methodology
The creation of a Bicycle Facilities Network outlined in SBMP Objective #1
(Appendix A) is a common approach to increasing bicycle ridership and safety.
Widespread enthusiasm for this approach exists across the U.S. in part based off the
successes of Northern European cities where bicycling represents a larger percentage of
the total transportation modal share and more extensive networks of bicycle
infrastructure exist. Pucher and Dijkstra (2003), in a widely cited article, assert that
policies aimed at non-motorized transport in the Netherlands and Germany are
responsible for the large share of total trips made by bicycles in these countries. Bicycle
modal shares of 28% and 12% for the Netherlands and Germany respectively, compared
to 1% of all trips in the US, combined with higher levels of safety in terms of injuries and
deaths, lead many to the commonsense conclusion that these countries’ policies are
responsible for successes. While the authors succeed in showing that both Dutch and
German cities have made substantial bicycle infrastructure investments, implemented
innovative traffic calming devices, and incorporated cyclists’ rights into both their traffic
education and enforcement regulations, such correlations are not evidence of causality
or a quantitative assessment of the effect on cycling. Such observations lack an
empirical basis for understanding the influence of any of these policy or design
prescriptions on subsequent increases in bicycling.
As discussed in the Introduction, a focus on infrastructure investments has
characterized the American approach to increasing bicycle use and safety, if to a lesser
extent than many of the Northern European countries with high bicycle mode shares.
6
Several key transportation acts during the 1990s provided devoted to the construction
of on-street bike lanes and off street trails. Notable is the $286.4 billion federal
transportation bill, SAFETEA-LU, which provides substantial funds devoted exclusively to
construct bicycle facilities (Krizek et al., 2009). The emphasis on creating both on and off
street bicycle lanes in order to make the existing streetscape a more viable option for
bicyclists is characteristic of the approach of many Bicycle Master Plans. Despite the
observation by some advocates and non-motorized planning professionals that many
BMPs do not get implemented, there is a dearth of empirical research that provides an
analysis of factors thought to influence the process of getting projects prioritized in
BMPs on the ground. In the following section, a review of the one academic study and
one professional guidance document on bicycle program and plan implementation is
presented in order to establish the need for the current study.
2.1 Review of Bicycle Policy Implementation Literature
Extensive searches of academic databases yielded only one academic study on
the implementation of bicycle policies. Stating that abstract common cycling goals set
through the National Cycling Strategy often translate into very different local results,
Gaffron (2003) surveys 92 British local authorities to explore the mechanisms that
contribute to implementation outcomes. The study is of interest in that local authorities
are required to prepare five-year local plans and encouraged, though not obligated, to
include policies for pedestrians and cyclists in these plans. While the inclusion of bicycle
facilities can’t be assumed, the study does identify both hindering and helping factors
7
for local bicycle policy implementation. Factors for cycling that were most obstructive
were lack of funding, lack of staff, and lack of staff time. Factors that were found to aid
policy implementation include having a national policy framework for cycling, other
national transport strategies and policies, and having a committed/motivated officer(s).
Overall, lobbying activities of local pressure groups and the political composition of the
city council were not found by respondents as having great influence over later
outcomes. Other important findings included the role of a policy champion is often
sufficient to influence implementation results, but this may also indicate that other
highly motivated actors who oppose the policy can have a restrictive influence. The
local authorities’ commitment to the policy through political will and resources were
likewise found to be important.
Lagerwey’s (2009) “Creating a Roadmap for Producing & Implementing a Bicycle
Master Plan”, provides a guide to BMP design and execution intended to be applicable
to a wide variety of local needs. The document identifies two factors external to the
implementing agency that are thought to assist with plan execution. First, the routine
accommodation of bicycle infrastructure during maintenance and capital projects is
listed as the “single best way” to implement physical improvements recommended in a
plan. This can be achieved through the adoption of a “Complete Streets” policy which
requires that the needs of all transportation modes be taken into account when repaving or building new streets (Lagerwey, 2009). These policies ensure that provisions
such as sidewalks, curb cuts, bike lanes, traffic calming and inviting crossings are
included in all road projects and not as an optional add-on (Alliance for Bicycling and
8
Walking, 2010). Second, the availability of a dedicated funding source is likewise seen
as beneficial, and considered a viable option to pursue even in increasingly turbulent
economic times. Dedicated funding sources for bicycle projects and programs have
included general funds and tax assessment measures approved by voters (such as levies,
the issuance of bonds, and millage, etc.) and through grants and partnerships from
public or private entities.
Other general strategies are aimed specifically at implementing staff focus on
the goal of increasing transparency, accountability, and information provision after plan
adoption. One such strategy details securing a place on the agenda of the group
responsible for monitoring implementation of the plan in order to start making monthly
reports. The “Roadmap” similarly suggests presenting an annual work plan to the
designated group that consists of measurable tasks for their review and approval, and
documenting all successes. Finally, the document highlights the need for ongoing public
outreach even after the adoption of the plan due to the potential for public backlash
that serves to slow or stop the plan.
The above study and policy guide identify that there is a clear need for
information related to factors influencing non-motorized transportation policy
implementation. These studies represent first steps for assisting policy makers in
making more informed decisions by allowing them to assess their own context in
relation to general strategies or factors that influence implementation outcomes. Yet
the scope of both studies is limited in important ways for the purposes of this study.
Gaffron (2003) does not discuss bicycle lanes specifically, in favor of a more general
9
“cycling policies”. This is necessary due to the author studying a large number of local
authorities that have likely adopted widely divergent policies under an overarching
national framework. The low response rate of the study, 45%, was identified by the
author as normal for such surveys, but may conceal a response bias on behalf of
localities that are not interested in cycling or walking policies. In such a case, important
factors that inhibit policy adoption and/or subsequent implementation may be
underrepresented.
Lagerwey’s (2009) “Roadmap” document, while a valuable summary of the
insights and experiences learned in Seattle, is limited in many respects. Only two pages
of the document discuss implementation specifically and in a general way. For example,
three of the six implementation steps suggested included (such as “Get the Plan
adopted”, “Document Your Success”, and “Seize the Day”) are more suggestive of
general elements to include in plan adoption and implementation and do not represent
an effort to systematically gather and corroborate data in relation to the process of
implementation.
The lack of studies dedicated to Bicycle Master Plans should be considered
alarming given that their adoption represents large public commitments requiring
agreement among elected officials, stakeholders, transportation planners, activists, and
consultants, with funding commitments of several hundred million dollars being
commonplace for medium sized cities. Currently, over 40 U.S. cities have developed
and adopted BMPs, a number that will likely continue to rise in coming years. The
literature review establishes that research has not kept pace with the adoption of these
10
plans and that there is little in the way of empirical guidance on the factors that will lead
to a higher probability of a successful implementation outcome. The state of the
literature warrants a closer look at what factors have proven to be important for
influencing the construction of bicycle lanes and trails, as envisioned in the Seattle
Bicycle Master Plan. To assist with filling the current gap in knowledge, the following
section outlines the questions to be addressed by the present study and method chosen
to analyze factors influencing Seattle’s implementation of a BMP.
2.2 Study Question and Methodology: BMP Implementation
An implicit assumption of this study is that the state of the literature
necessitates first describing what factors are thought to influence a plan before
explanatory or quantitative studies are conducted. Prematurely proceeding with
explanatory studies that assess the relative impact of various factors would neglect the
fact that the researcher may not currently know what those factors are. To assist in
this endeavor, the current study employs a case study methodology incorporating semistructured interviews to answer the primary study question:
Question #1: What are the facilitating and obstructing factors identified by
implementing staff that influenced the development of the Bicycle Facilities Network
during the 2007-2009 period?
The case study presents itself as the most appropriate methodology in that it is
unique in allowing the researcher to retain the holistic and meaningful characteristics or
real-life events (Yin, 2003). This was deemed a necessary attribute given the inherent
nature of political and organizational processes. Because the lines are not clearly
11
delineated between the phenomenon (implementation of bicycle infrastructure
projects) and context (community support, municipal policy constraints, budget
constraints, etc.), a methodology that will allow for collection of both was determined
to be of paramount importance. The case study methodology is also well suited to this
endeavor in that it does not require behavioral control of contemporary events,
allowing the author to address both exploratory and explanatory elements that are
inherent to the research question.
As opportunities for direct observation were not possible, and documentation
and records do not directly address the factors related to influences on SBMP
implementation, the use of interviews was deemed most appropriate to answer the
research question regarding influences on implementation. The use of interviews as a
data source is especially pertinent to the methodology chosen as they are considered
one of the most important and essential sources of case study information (Yin, 2003).
Participant selection consisted of identifying key informants with direct
involvement in implementation during the Short-Term Implementation Period (20072009) of the Seattle Bicycle Master Plan. Potential study participants identified in SBAB
meeting minutes as being involved directly with SBMP implementation were contacted
between May and June of 2010, yielding five study participants. Participants consisted
of two agency staff directly involved with implementation, a bicycle policy advocate, a
consultant involved with the plan, and a member of the Seattle Bicycle Advisory Board
that advises the implementing agency on program decisions. Three of the five
participants were directly involved with the development of the SBMP, either through
12
the Mayor’s Citizens Advisory Board or as City of Seattle planning staff. Four of the five
participants were employed as transportation professional in various capacities, with
the final participant being a professional in another field tasked with advising the City of
Seattle on decisions related to bicycle policy.
Due to the possibility of jeopardizing a study participant’s employment or
standing within the community should someone take offense to an interview response,
a Human Subjects Review was submitted to, and approved by, the Evergreen State
College. Interview participants were informed upon first contact that their participation
was to be completely voluntary and that there would be no compensation for their
time. Study participants were given the option of retracting statements and dropping
out of the interview at any time with no consequence. In order to allow participants to
provide information freely without fear of professional risk, all interviews are cited as
anonymous.
Interviews were conducted between August and November, 2010, or, less than a
year from the end of the study period. Semi-structured interviews with individuals (one
interview was agreed to with two participants concurrently, due to their limited
availability) were conducted over a period of one hour to ninety minutes either in the
offices of the study participant or a location deemed more suitable to them. Study
participants were asked a series of structured questions (Appendix B) developed by the
author and based on the synthesis of several theories of policy implementation
presented by Charles (2005), discussed in detail below. Participants were reminded
that the scope of questions refers to activities in the SBMP that took place after the
13
initial adoption of the plan in 2007 through 2009, and to limit speculation about events
that have not happened. Responses were audio recorded for later transcription and
analysis.
2.2.1 Policy Implementation Evaluation Framework
Designing an interview instrument able to address factors influencing SBMP
implementation required meeting two competing demands. First, interview questions
need to allow for open-ended responses that would reasonably ensure the major
factors influencing implementation would be covered. Second, questions needed to be
confined to areas most likely to affect plan implementation. The need for an
established policy implementation framework was deemed necessary to guide question
development by providing an existing structure that would increase the chances of
identifying factors while mitigating the possibility of researcher bias having an undue
effect on responses.
Literature specific to bicycle planning and policy implementation yielded little
precedent for adopting one approach over another. As discussed in the review of
literature, the author identified only one study that focused specifically on factors
influencing bicycle policy implementation. In conducting a literature review to provide a
basis for choosing a conceptual model to structure her study, Gaffron (2003) notes that
not only were there no widely accepted theories of the nature of policy implementation
process, there was no standard methodology for studying it. Gaffron concluded that
researchers, in analyzing program and policy implementation, have to rely on a
14
synthesis of the theoretical framework of implementation analysis, judgment of the
applicability of different approaches in their particular area and the conclusions they
draw from their own data.
Given the lack of precedent, academic databases were queried for an
implementation analysis framework that met several criteria, based off the author’s
best judgment of the applicability of the framework to non-motorized plan
implementation. It was determined that the policy implementation framework
synthesized by Charles (2005) from previous work in the field of implementation
analysis (Hogwood and Gunn, 1984; Ison and Rye, 2003 ; and Sabatier and Mazmanian,
1979) could reasonably provide a basis for developing a survey instrument to
comprehensively address a number of factors likely to affect SBMP implementation.
Charles’ framework (Table 1) synthesized work on good practice in implementation from
previous theory and studies in order to identify the key success factors for effective
implementation of regional traffic incident management program.
The use of Charles’ iteration of the “ideal” policy implementation framework is
justified due to its meeting several criteria established by the author. First, it was
determined that the framework covered the types of issues that may influence the
development of the SBMP. Monthly meetings minutes of the Seattle Bicycle Advisory
Board (SBAB, 2007-2009) provided an appropriate source of data to assist in this
assessment. A significant strength of this data source can be found in the wide range or
representation available at any meeting. Neighborhood bicycle activists, commuters,
bicycle policy advocates, various city agency staff, and members of the Board bring a
15
Table 1: Theoretical Policy Implementation Framework
Circumstances external to the implementing agency do not
Context: external
impose crippling constraints (HG); Relative priority of
circumstances
objectives not undermined over time (SM); Political stability
(IR)
Resourcing: time, skills, Adequate time and sufficient resources are made available;
funds
& required combination of resources is actually available
(HG); Leaders of implementing agencies possess significant
managerial and political skills (SM); Rationalize financing and
investment streams: allocate funding in a balanced way
(ECMT); program timing. (IR)
Theory: cause and effect Policy based upon a valid theory of cause and effect; &
relationship between cause and effect is direct (HG);
Program based on sound theory (SM).
Leadership: governance, Single implementing agency (HG); Provide a supportive legal
institutions
and regulatory framework: ensure the rules and regulations
clearly specify roles (ECMT); Policy champion clearly
dedicated to the task of implementation (IR)
Clarity: clear policy and Complete understanding of, and agreement upon, the
strategy
objectives to be achieved, and that these conditions persists
throughout the implementation process; & tasks are fully
specified in the correct sequence (HG); Policy contains
unambiguous directives and structure the implementation
process to maximize success (SM); Establish a supporting
policy framework (ECMT)
Coordination: good
Perfect communication and coordination , between the
communication and
various elements or agencies involved (HG); Improve
coordination
institutional coordination and cooperation: with
responsibilities commensurate with resources for
implementation to occur (ECMT); flexible and open attitude
toward public reaction (IR)
Compliance: require and Those in authority can demand and obtain perfect
obtain compliance
compliance (HG)
Support: stakeholder
Program is actively supported by constituency groups (SM);
support
encourage effective participation, partnerships and
communication (ECMT); Public trust and support (IR)
Monitoring: data
Improve data collection, monitoring and research: carry out
collection and
consistent monitoring (ECMT); Monitoring outcomes (IR)
monitoring
Note: HG refers to Hogwood and Gunn; IR refers to Ison & Rye; ECMT refers to European Council of
Ministers of Transport; SM refers to Sabatier and Mazmanian. Table from Charles, P (2005), pg. 611.
16
diverse array of professional backgrounds and experience to assist in advising the
Seattle Department of Transportation specifically on SBMP implementation issues. As
the intention of these meetings was not to assess and corroborate factors influencing
the Plan in a systematic way, data gleaned from this source is most beneficial in
providing a first look at influences on BFN development during the Short-Term
Implementation Period. Factors influencing implementation identified by the author
during the course of SBAB meetings were able to be categorized in five of the nine
categories presented by Charles (2005), the results of which are provided in Table 2
below.
The framework presents another advantage in that is has been used in previous
transportation related studies. Both Charles (2005) and Ison and Rye (2003) successfully
employed the framework to analyze implementation of a regional transportation
strategy, and travel planning and road user charging programs, respectively. Charles’
formulation of Hogwood and Gunn’s (1984) “perfect” implementation framework was
also determined by the author to meet the need of using concepts that are readily
understandable to practitioners in the field who may not be familiar with
implementation analysis, and would be confused by obscure jargon. A final
consideration was the adaptability of the framework to the current study’s purpose,
particularly the limited amount of time available to conduct interviews. Charles’
formulation of the framework does not present itself so much as an integrated system
but is modular in that questions can be developed around particular areas of the
framework and not others. For these reasons, and with little precedent to go off of,
17
Charles’ iteration of the original Hogwood and Gunn framework was chosen as a basis
for creating an interview instrument.
Table 2: Content Review of Seattle Bicycle Advisory Board Notes- Potential factors
Influencing Implementation of the Seattle Bicycle Master Plan, 2007-2009
Potential implementation issue identified in
SBAB Meeting Notes
Implementation
Framework
Category
Context: External
South Lake Union Streetcar danger to cyclists circumstances
Role of SBAB and SDOT (board and
implementing agency) in facilitating project
implementation
Bridging the Gap availability
Prioritization scheme for choosing
projects/opportunity cost
Coordination: good
communication and
coordination
Resources: Time,
Skills, Funds
Clarity: clear policy
and strategy
Producing list of projects and prioritizing
based on scoring system
Clarity: clear policy
and strategy
Businesses angry at bikers after
implementation
New ferry waiting lane precludes inclusion of
bicycle lanes
Support:
stakeholder support
Context: External
circumstances
Context: External
circumstances
Peak parking and bike lanes
Opposition to Plan from neighborhood
residents
Community Council opposition
Constituency support and issue creation
Complete Streets
SDOT potential budget cuts
Timing of funding
Support:
stakeholder support
Support:
stakeholder support
Support:
stakeholder support
Context: External
circumstances
Resources: Time,
Skills, Funds
Resources: Time,
Skills, Funds
# of times
mentioned by
participants
SBAB meeting
Nov 07. Mar 08. Apr
10 08. Mar 09. Sept 09.
Nov. 07, Mar. 08,
4 Aug. 08, Nov. 09
4 Oct. 08, Nov. 08
3 Sept. 09, Mar. 08
2 Oct. 07, Feb. 08
2 Mar. 08
1 Mar. 09
1 Oct. 09
1 Mar. 09
1 Apr. 08
1 Apr. 08
1 May 08
1 Aug. 08
1 Mar. 08
18
2.2.2 Interview Instrument
A comprehensive assessment of all factors influencing “perfect” implementation
included in Charles’ framework was determined to be inappropriate for the scope of this
study. For practical reasons, it was deemed necessary to keep interviews at
approximately one hour’s length in order to increase the response rate of participants
and decrease a respondent’s likelihood of dropping out of the study. Twelve core
questions were developed with several sub-questions for probing and follow-up that
would cover aspects of each of eight of the core categories the framework (Table 1). As
the framework only suggests factors that would support the ideal implementation of a
program, the absence or contradiction of such factors is assumed to obstruct
implementation. This assumption was used as a basis for developing questions whose
responses will allow for the identification of both positive and negative factors
influencing implementation of the Seattle Bicycle Master Plan.
The use of interview data was deemed by the author to be an inappropriate data
source to answer the question of whether Hogwood and Gunn’s requirement for a valid
theory underlying a policy or program was fulfilled (see “Theory: cause and effect”
category, in Table 1 above). The next section outlines the second study question, as
well as the need to employ a methodology on separate data that models the
relationship between bicycle facilities and bicycle ridership.
19
2.3 Study Question and Methodology: Theory Behind Policy Implementation
Hogwood and Gunn’s (1984) implementation framework states that an
important aspect of implementation is that the policy incorporates a valid theory of
cause and effect with minimal linkages, that “if X is done at time t(1) then Y will result at
t(2).” Several authors echo this sentiment, with one noting that successful plans and
programs that intend to increase bicycling need to be based on empirical knowledge of
who cycles, where, and why (Moudon, et al. 2005). As a condition for effective
implementation, Sabatier and Mazmanian (1979) similarly contend that a policy should
consider all major factors directly contributing to the problem within the scope of the
program and correctly relate each of these factors to the desired outcomes. Using
Hogwood and Gunn’s formula, a theory underlying the Seattle Bicycle Master Plan could
be phrased as follows: “If 450 miles of bicycle facilities are provided to be within ¼ mile
of 95% of Seattle resident over the 10 year (2007-2017) implementation period of the
Seattle Bicycle Master Plan, then bicycling will triple at the end of that time.”
With limited resources and no agreed upon model of determinants of bicycle
use, it is outside the scope of this study to venture predictions specific to the city of
Seattle in terms of tripling its share of cycling commuters from its 2007 rate of 2.27% to
6.81% in 2017. Assessing the theory behind the policy through the use of interview
data was also deemed problematic in that respondents may reflect unwarranted
assumptions about the positive effect of bicycle facilities on ridership. Indeed, Krizek et
al. (2009) state that such assertions “are often bantered about by planning agencies and
20
advocacy groups”, despite the fact that evidence that can reliably support such
statements has not been forthcoming.
A more appropriate data source can be found in through the use of predictive
models of bicycle use found in the peer reviewed literature of both the public health
and transportation planning fields. These models are less prone to be influenced by
professional assumptions due to the need to demonstrate statistical significance, and
can be critiqued for the strengths and weaknesses of their underlying study design.
Analysis of models of bicycle use is conducted through an assessment of their
conclusions in terms of several of Hogwood and Gunn’s criteria held up as ideals for
successful policy implementation. First, the ideal condition the implementation
framework assumes is that the relationship between a policy and outcome must be
positive and causal. Second, the relationship must be direct and have few if any
intervening links in the chain of causality. A third criterion assumed by the author
concerns the strength of the evidence: measures and constructs should be well agreed
upon and results should be repeatable across conditions, while adequately explaining
variations. An assessment of models in meeting these criteria will provide the basis for
determining how the evidence compares to the ideal theory aspect of the
implementation framework, and to what extent the conditions are satisfied. The second
study question is thus framed as:
Question #2: Does empirical evidence exist for the proposition that creating bicycle
facilities (in the form of on and off road lanes) has a positive, direct, and causal
relationship with bicycle ridership?
21
In the section that follows, the results of qualitative data obtained from
interviews will be used to explore the first research question to identify factors found to
be significant in influencing SBMP implementation outcomes. The results of models of
bicycle use will then be analyzed to answer the second study question concerning the
influence of bicycle infrastructure provision on ridership outcomes. As will be seen, this
assessment would be incomplete without also considering the often significant effects
that individual, attitudinal, socio-demographic, environmental, and geographical factors
have on bicycle ridership.
3. Data
The following section (3.1) will present the results of the qualitative data gathered
through the use of interviews with bicycle implementation staff, along with coding
techniques and the criteria for determining how responses are significant. Significant
factors affecting implementation are then discussed in detail according to categories in
Charles (2005) framework. Section 3.2 will review the results of models of bicycle
ridership and the use of ecological models to categorize results in terms of
individual/attitudinal, socio-demographic, environmental and policy level factors. A
discussion of the results in terms of Hogwood and Gunn’s (1984) criteria that the theory
underlying the policy demonstrates a cause and effect relationship with minimal
linkages will finish the chapter, in preparation for Section 4, which will provide a final
analysis and conclusions of the major study findings.
22
3.1 Factors Influencing Implementation of the Seattle Bicycle Master Plan
Qualitative data from interview responses with implementing staff provides the
basis for answering the first study question concerning the facilitating and obstructing
factors that influenced the development of the Bicycle Facilities Network during the
2007-2009 Short-Term Implementation Period. Interview responses were digitally
recorded and later fully transcribed to reduce errors due to recall. Interview transcripts
were analyzed in terms of the factors identified by respondents and their corresponding
influence on implementation. Interview responses were coded by the factor they
addressed and subsequently rated by the author for their effects on implementation
using a simple scale: Critical for implementation, supportive of implementation, neutral
or no effect on implementation, problematic for implementation, and barrier to
implementation. Categories were broadly defined according to the following scheme:
•
Critical for implementation: The factor was defined by the study participant as
being of paramount importance in regard to implementation outcomes. The
absence of the factor would have created a significant barrier to implementing
projects in the Bicycle Facilities Network.
•
Supportive of implementation: The factor was defined by the study participant
as contributing positively to implementation outcomes, but was not identified as
being a critical piece.
•
Neutral or no effect on implementation: The factor identified by the study
participant was not associated with implementation outcomes, or had no
discernible effect.
•
Problematic for implementation: The factor identified by the study participant
served to significantly impede projects, but not stop them entirely.
23
•
Barrier to implementation: The factor was determined by the study participant
to halt the construction of projects, and only in its absence could the plan move
forward.
A separate category, “not categorized”, was reserved for survey responses that were
either not well developed, conflicted on the influence of implementation, or too
contextual to readily assign a category to. Factors that were identified by three or more
respondents as having a particular effect on implementation were labeled “convergent”,
meaning that the effect of the factor on implementation was assessed similarly by the
majority of respondents and will be considered a significant study finding. If a factor
contained three or more responses that differed in the magnitude, but not the
association of the effect (i.e., two responses for critical for implementation, one for
supportive of implementation, but three positive overall), responses were likewise
labeled convergent and considered significant. Due to the study design generating a
large number of responses and a need for confidentiality, convergent factors identified
by participants are summarized below (Table 3) based on the effect reported. Factors
that were only mentioned by one interview respondent or that were not sufficiently
developed were not considered.
Overall, study participants identified the presence of a dedicated transportation
project funding source (“Bridging the Gap” levy), political will, an ordinance that
requires the consideration of bicycle facilities in all City transportation projects
(“Complete Streets”) and the support of constituency groups as being critical factors
influencing a positive implementation outcome.
24
XIb. Support- Consituency groups/public
X. Leadership: Policy champion
II. Resourcing-Timing
XIa. Support-public trust
V. Resourcing: Staff, Time, Funding,
Expertise
I. Context and Leadership:supportive policy
framework and absence of external
constraints
II. Resourcing-Timing
I. Context and Leadership:supportive policy
framework and absence of external
constraints
V. Resourcing: Staff, Time, Funding,
Expertise
Bridging the Gap
5
Complete Streets
Political will
Support of constituency
groups
Policy Champion- Cascade
Bicycle Club
Gas Prices
Public Trust
Adequate staff to
accomplish objectives
2
3
2
1
2
1
1
4
3
3
1
3
Seattle Climate Action Plan
Focus on the environment
3
3
1
Limited street space to
accommodate all uses
4
Funding Capital Projects
3
Opposition to projects/
bicyclists and infrastructure
XIb. Support- Consituency groups/public
the "unusual other"
Dependencies on other
VII. Leadership-single implementing agency agencies/publi process
1
3
XI. Support-flexible attitude toward public
reaction
Public process
III. Political Stability
Political stability
2
2
3
Mix of facilities- Sharrows
IX. Agreement on objectives to be achieved vs. bike lanes
VI. Monitoring-Improve Data Collection;
carry out consistent monitoring
I. Context and Leadership:supportive policy
framework and absence of external
constraints
V. Resourcing: Staff, Time, Funding,
Expertise
I. Context and Leadership:supportive policy
framework and absence of external
constraints
II. Resourcing-Timing
3
2
City Comprehensive Plan
Requisite expertise at
implementing agency
1
Focus on health/well-being
1
3
5
Data
Seattle Transporation
Strategic Plan
Not categorized*
Mixed Influence on
implementation
Barrier to
implementation
Problematic for
implementation
Neutral or no effect
on implementation
Factor
Supportive of
implementation
Implementation framework category
V. Resourcing: Staff, Time, Funding,
Expertise
I. Context and Leadership:supportive policy
framework and absence of external
constraints
VIII. Enforce Compiance
Critical for
implementation
Table 3: Factors identified by study participants that influenced implementation of the
SBMP 2007-2009 (by # of times factor was mentioned in interviews, categorized by
associated effect)
2
1
2
1
1
1
2
2
*response varied in many instances and did not converge
on any one factor, or were difficult to categorized for
contextual reasons.
25
Supportive of implementation was the presence of the Cascade Bicycle Club in
the role of policy champion, an increase in support due to high gas prices during a
portion of the study period, public trust, and having adequate staff to accomplish the
objectives of the SBMP. Seattle’s Climate Action Plan and environmental concern were
not characterized by study participants as having an influence on the implementation of
the Bicycle Facilities Network.
Problematic for implementation was the fact of limited street space to
accommodate all uses, difficulty funding capital projects such as bridges and multipurpose trails, and public opposition to projects. No significant barriers to
implementation were identified during the Short-Term implementation period. This
finding is not surprising given that the SBMP is in the early phases of implementation
and has flexibility in the projects it chooses, coupled with the fact that SDOT met its
projects goals over the three year study period.
Participants discussed the need to interact with other agencies and the
requirement for public process as having mixed, but ultimately positive, effect on
implementation. Left uncategorized were the roles of political stability and that of data
and performance measures in implementation outcomes. Responses in this case were
not sufficiently developed, speculative of events that have yet to happen, or could not
be adequately assessed during the study period for the Seattle case.
Overall, the interview instrument was successful in identifying factors of varying
influence on Bicycle Facilities Network implementation efforts. All interview questions
based off of aspects of implementation identified in Charles’ iteration of the framework
26
generated responses, with five framework categories (Context, Resourcing, Compliance,
Leadership, and Support, see Table 1 above) of eight having converged on responses for
one or more factors.
In what follows, a more in depth treatment is provided of the facilitating and
obstructing factors identified by key respondents during the course of interviews.
Factors identified by participants as influencing implementation are organized according
to categories in the implementation framework presented by Charles (2005) in Table 1
above. It is only through a consideration of the major themes present across
convergent responses (three or more respondents reporting a similar effect) that the
contextual influence of individual factors can be assessed in terms of the effect on
bicycle facilities implementation in the Seattle case.
3.1.1 Context and Leadership: Policy Framework and External Constraints
Hogwood and Gunn (1984) note that some obstacles to implementation are
outside the control of program administrators because they are external to the policy
and the implementing agency. They may be physical, such as a lack of space to develop
bicycle infrastructure projects, or political, when projects are unacceptable to certain
groups or interests (neighborhood associations, commercial trucking groups, etc.) that
can successfully lobby effectively against them.
The interview instrument had participants consider the effect of a supportive
policy framework, as outlined in Charles’ theory in the section on Clarity: Clear Policy
and Strategy (Table 1). While the external environment may serve to constrain a policy,
27
it may provide an environment that reinforces actions that lead to the successful
implementation of Bicycle Facilities Network projects. Several policies outlined in the
SBMP as supportive of the Plan include the Complete Streets policy adopted by City
Council in 2007, Seattle’s Climate Action Plan, and the “Bridging the Gap” initiative
passed by Seattle residents in 2006 that will make approximately $3 million available
annually to implement the Plan (SBMP, 2007). Alternatively, the possibility that existing
policies may conflict with Plan implementation in important ways needs to be explored
and assessed for its effect on the process.
Respondents considered the Complete Streets Ordinance critical to successful
implementation, while the Seattle Climate Action Plan was thought to have a largely
neutral effect as a framing document. The lack of street space to implement projects
was considered to be problematic, although responses did not converge on whether this
effect was more of an issue during Plan development or Plan implementation. The City
of Seattle Comprehensive Plan was not found to have convergent responses, but is
briefly discussed due to the potential constraining effect it could have on project
implementation. The following four subsections will discuss each factor related to the
policy framework and external constraints in more detail.
3.1.1.a Complete Streets Ordinance
The Complete Streets ordinance, passed by Seattle City Council on April 9, 2007,
was mentioned in the SBMP as one of several policies that “will play an important role
in building support for its full implementation”. (SBMP, 2007) The Complete Streets
ordinance requires SDOT to “plan for, design and construct all new City transportation
28
improvement projects to provide appropriate accommodation for pedestrians,
bicyclists, transit riders and person of all abilities, while promoting safe operation for all
users, as provided below.” (City of Seattle, 2007) The policy objectives could be
achieved through single projects, or incrementally through a series of smaller
improvements or maintenance activities over time. An example of the application of
the Complete Streets legislation would be that SDOT, when conducting a repavement or
capital project, would be required to consult the Bicycle Master Plan and consider
bicycle improvements to be included before the project is executed.
Four study participants out of five viewed the Complete Streets ordinance as
either a very supportive, or critical, policy in regard to implementation of the SBMP. As
a statute enacted by the City of Seattle, Complete Streets is a legal requirement that
institutionalizes the consideration and accommodation of projects outlined in the SBMP
that may have been overlooked in the absence of such a policy. One respondent with
experience throughout the U.S. working on BMPs noted that not only was it an
important piece in adoption and implementation in Seattle, but it may be the catalyst
for getting BMPs on the ground:
“[The Complete Streets ordinance] was a very, very important policy in terms of
getting the Bike Plan adopted because it said that we are going to be inclusive and
accommodate all modes in projects and programs... There’s a lot of plans that get
developed that don’t get implemented, but when it is combined, I think, with a
Complete Streets policy then it really motivates everyone to do that.”
According to one participant, the benefit of routine accommodation of
historically underserved transportation modes did not have any apparent influence on
the timing or complexity of projects. Its value, again, lay mainly in what advocates see
29
as an improvement to the planning process by considering the SBMP whenever
transportation projects overlap an area in the Bicycle Facilities Network:
“So, have projects been easier or harder to implement? I don’t know, but I think they
have been better implemented… the city is very aware that when they are doing
projects that they have to consider the BMP. I think it has improved the process,
whether it has sped it up or slowed it down, I don’t know.”
The importance of the ordinance is underscored by one advocate in noting that
the passage of a Complete Streets policy is no easy feat:
“You know, the Complete Streets ordinance took four years to craft and get passed, I
think it is the most important piece of policy language.”
Responses readily converge on the routine accommodation of bicycle facilities in
all city transportation projects as a critical factor in ensuring that they are not
overlooked. The legal requirement to do so has the effect of institutionalizing the
process and gives advocates standing in cases where the Complete Streets ordinance is
not adhered to. For these reasons, the existence of a Complete Streets policy is
considered to be a critical factor in the Seattle case.
3.1.1.b. Seattle Climate Action Plan
The Seattle Climate Action Plan (CAP), released in September of 2006, represents
former Mayor Greg Nickel’s overarching strategy for reducing global warming as
recommended by his appointed Green Ribbon Commission. The impetus behind the
CAP was Mayor Nickels’ concern that climate disruption would have detrimental
impacts on drinking water and electricity generation for the city, that reducing
30
emissions would improve public and environmental health while creating green jobs,
and that there was a need to address “the lack of meaningful federal action on an issue
so critical to the health of our city, country and our planet.” (City of Seattle, 2006) In
light of these needs, the document provides a series of action items the city, businesses
and individuals can take to reduce the city’s greenhouse gas footprint
The Climate Action Plan is specifically called out in the SBMP as providing a
supportive policy context, yet study participants expressed doubt that the CAP had any
direct effects on the implementation of projects. While not diminishing the importance
of mitigating greenhouse gases or the role that non-motorized transportation will play
in that endeavor, one study participant noted that:
“[The CAP] was certainly used to get grants, it was certainly mentioned in everything
we did, but it was sort of one of those umbrella efforts where you just took everything
you were doing and say. “Ok, we were doing all these things to improve climate
change.” There was a lot of verbiage about it, but I don’t think there was anything we
did….that would not have happened if we didn’t have that climate change initiative. It
didn’t give us a new policy that we didn’t have in terms of saying “yes” to bicycling. It
didn’t provide us with more funding that we didn’t have because that already came
through [the] Bridging the Gap [levy]. “
Another interviewee supported this point of view in that the CAP did not provide
any new framework that was either supportive or limiting of the SBMP, assisting
implementation only,
“…insofar as it said “Complete the Bicycle Master Plan.” In the implementation
strategies flowing from that, we are integrated into the updates to our Comprehensive
Plan and development, and we have an upcoming change to the Transportation
Strategic Plan (TSP) that is coming up. But the last TSP offered that policy
framework.”
Another respondent concurred with this same point, but noted that the policy
could be used to other ends:
31
“My understanding…. is that [the CAP] was sort of a large overarching policy. But I
don’t know if there were any methods of trying to implement that other than being
consciously aware of the policy when the city does projects. There was a recent
project called the re-channelization of Nickerson Street, which was and is very
controversial because the proposal was based on perceived safety of pedestrians and
also to a lesser degree cyclists, but there was to be an impact on freight traffic. So the
City Council had a hearing which a number of people participated in… Councilmember
O’Brien, one of the things he talked about specifically in that hearing was the
importance of implementing or complying with the policy of reducing adverse effects
on the climate with respect to this project.”
The outcome of invoking the need to comply with the city’s climate change
strategy appears limited, however. When the same respondent was asked if mentioning
the CAP affected the position of opponents to project implementation, he remained
doubtful:
“Not to a large degree. I think people understood it, heard it, but it wasn’t certainly a
central topic among all the presenters or Council members…I think most of the
commenters were more practical, “How is this going to affect my business?” “How is
this going to affect how I walk, how I cross, how I use the road?” And I think that is
more what people were personally concerned about that came and testified.”
As can be seen, participants were in agreement in regard to lack of direct effects
on implementation stemming from language in the CAP. The CAP’s usefulness lies
mainly in its ability to reframe the implementation of the Bicycle Master Plan in terms of
larger environmental goals. To the extent that it allows for the City to apply for grants
as a part of a larger environmental strategy, the possibility of a positive benefit exists.
However, it was not found to be particularly effective in addressing the immediate
concerns of an opposition that may be motivated primarily by what are perceived as
inconveniences stemming from a loss of existing entitlements. A strong convergence in
respondent opinion exists that the CAP, as a supportive document to the SBMP, did not
have any discernible effects on project implementation.
32
3.1.1.c City of Seattle Comprehensive Plan
An additional framework document that was thought to affect the plan was the
City of Seattle Comprehensive Plan. The Washington State Growth Management Act
(GMA) (36.70A RCW) requires local communities most affected by growth to engage in
comprehensive planning for the development of the community over a twenty year
period. The GMA requires these cities and counties to adopt regulations to protect
resource lands and critical areas, designate urban growth areas, and include mandatory
planning elements that must be addressed. Most relevant to the present study is the
Transportation Element, which requires jurisdictions to: identify expansions needed to
meet present and future demand; provide land use assumptions used in estimating
travel, facilities, and service needs; and provide level of service (LOS) standards for all
locally-owned arterials and transit routes, among other required tasks.
The Seattle Bicycle Master Plan mentions the Comprehensive Plan as being a
framework document, and it was seen by one interview participant as having “general
policy language there just saying that it’s the cities priority to promote bicycling and to
construct streets for all users.” No specific mechanism was mentioned by interview
participants through which the Comprehensive Plan positively impacted actual facilities,
separate and aside from the SBMP itself. One study participant brought up a potential
restrictive influence due to the methodology used to calculate the LOS standards
required by the Growth Management Act:
“The Comprehensive Plan is the force of law under RCW Chapter 36 under the GMA, so
whatever we say in there, whether its level of service, how we comply with our need
33
to provide adequate amount of infrastructure for the development that is going on in
the city… if our level of service is intersection delay for vehicles, that going to
undermine the SBMP because intersection delay essentially means taking a pipe
cleaner to every intersection and for lack of a better phrase, screw everybody else. I
still think our level of service metric is not completely adequate for allowing us the
flexibility with our public roadway network that we would want. The city uses
something called screen line grid, which is an acceptable approach in an urban area
but it is not fine grained enough in its analysis. We would like to see a multi-modal
quality of service model employed. When we get into the compromise between transit
buses and bike accommodation, if we have a diminishing level of transit service, that
can preclude some of the improvements that we had pushed for in the plan.”
The participant went on to mention specific projects that were modified so that
the existing LOS could be maintained and accommodate the heavy load of Metro Transit
buses on these routes.
While there was not a convergence of responses on the
subject of diminished LOS affecting projects, this example was notable in demonstrating
how a framework policy document could provide generally supportive policy language,
but still have a negative effect on projects as other modes are prioritized. This response
is suggestive of the need to assess the impact of metrics used to gauge an arterial’s level
of service to ensure that the bicycle is not precluded in favor of other modes.
3.1.1.d Street Space
“One of the big difficulties…of implementing the BMP in Seattle is kind of a
geographical problem. A lot of the streets are fairly narrow and there are a lot of
really tight spaces…it is really hard to add bike lanes. That means you have to take
away a lane somewhere, either parking or a roadway lane and a lot of lanes are
narrow. Physical room is a tough reality…People are used to parking here and having
a center lane for turning and what not. Those things are hard to get out.”
The issue of a lack of urban street space was brought up by three interview
participants as problematic, but not necessarily a barrier to implementation.
Participants defined this resource constraint in terms of narrow streets and the built-up
34
nature of the city. Similarly, existing entitlements that add to the convenience of motor
vehicle users represents a further constraint on space available in the form of
opposition to projects. One participant suggested that the type and quality of
infrastructure provided may be the downgraded from a designated bicycle lane to a
sharrows:
“We’re living in an urban environment and we are dealing with a scarce urban
resource which is street space and a lot of competition for that space. Everything from
the planting strips, to the sidewalks, the street trees, the parking, the vehicle lanes
and the streetcars and everything else. Anytime you see a shared lane marking it
means there probably wasn’t space for bike lanes. Not in all cases, but in many cases
that is why you have shared lane markings because you have run out of space.”
While agreeing that the constraint exists, another participant contradicted the
notion that street space was a problem at the time of implementation:
“There are constraints on just about every project that way. But that constraint was
not something that emerged when you do implementation, that emerged already
when we did the plan, that’s why the plan says “put a shared lane marking here”…
they already identified that constraint in the planning process, they didn’t have to
wait until the implementation phase to figure that out.”
This statement supports the notion that a lack of physical space is primarily an
issue at plan development. Another respondent contradicts this statement, however,
noting that in many cases the data was not available at plan development or competing
uses made planned facilities impossible to implement. There was a lack of knowledge
regarding,
“…on the ground realities that we couldn’t have known at the level of detail necessary
at plan development. We have had projects that were slated for bike lanes that have
wound up with shared use lane arrows for a number of reasons, some of them are that
Metro prefers not to weave in and out of bike lanes with buses.”
35
Taken together, the interview responses converged in characterizing a lack of
street space as diminishing the quality of project implementation by affecting the types
of facilities that were ultimately recommended. Regardless of whether the data is
available during the Plan development or only discovered at project implementation,
narrow street spaces, existing entitlements for competing modes, and related factors
can be shown in the Seattle case to negatively affect implementation outcomes by
constraining the options available to provide for a mix of facility types appropriate to
accommodate a variety of users.
3.1.2 Resourcing- Timing
The issue of the timing of implementation was called out specifically by Ison and
Rye (2003) as important to the process, while noting that it may be considered indirectly
in the question of external circumstances. The authors’ do not specifically define the
term “timing”, only differentiating it from “time” as a resource. For the current studies’
purposes, it is inferred from previous work that timing is best defined as contextual
events or trends pivotal in affecting implementation outcomes during the 2007-2009
period. Because policy issues tend to be highly interrelated and agencies are immersed
in economic and political systems characterized by change over which they have little
control, any particular policy decision may face erosion of political support over time as
other issues are deemed more important (Sabatier and Mazmanian, 1979). Just as
likely, a policy decision may receive more support if events make the issue more salient
in the eyes of the public. Events such as recent high profile bicycle/car accidents that
36
increase the sense of urgency around Plan implementation, increases in gas prices or
environmental awareness, national spotlighting of the Plan that builds support, or the
economic crisis (2007-present), are all context specific factors that may help or hinder
implementation efforts.
Participants were asked to discuss the impact of larger contextual events
occurring between 2007-2009, not intentionally willed by policy actors (to differentiate
between the policy framework, addressed above in section 3.1.1. Respondents were
given examples such as the current financial crisis, increasing environmental
consciousness and concern about global climate change, and increasing gas prices.
These examples appeared to have influenced responses to some extent, but
respondents were free to report whether a given trend influenced the Plan or not. The
following sections will discuss the impact of increased environmental awareness, which
was largely seen as having a neutral effect on implementation, and an increase in
gasoline prices, considered important for increasing ridership, and possibly, public
support for the Plan.
3.1.2.a Environmental Awareness and Concern
In addition to former Mayor Greg Nickels’ Seattle Climate Action Plan, several
indicators of increased awareness around climate change can be found during the study
period in the areas of advocacy, media, and research. As one indicator of the rapid
growth of interest in climate change, a study published by the Center for Public Integrity
reports that the number of interest groups lobbying on climate change in the U.S.
37
jumped by more than 400 percent (to 770) during the period of 2003 to 2008 (Lavelle,
2009) Also during the study period, Gallup conducted the first comprehensive survey of
global awareness and attitudes about climate change (Pugliese and Ray, 2009) again, a
possible indicator of the rising importance of this issue. Similarly in 2009, UNESCO, in
partnership with the United Nations Environment Programme, organized the
International Conference on Broadcast Media and Climate Change, bringing together
national broadcasters, scientific organizations and climate related agencies, resulting in
a resolution to give increased media exposure to the issue of climate change.
The growth in awareness of climate change and the environment during the
study period could be expected to lead to increasing support among Seattle residents.
Three study participants discussed the role of environmental awareness as it relates to
the Seattle Bicycle Master Plan, noting its limited to non-existent role in each
circumstance:
“…we see that there is only a certain segment of the public that operates from the
altruistic perspective in terms of determinants of personal choice. Most of the public
operates from an either enlightened or unenlightened self-interest model. So gas
prices can affect you directly, we don’t see much of the “I am doing my part to prevent
climate change.” The people riding will recognize and support their behavior
externally with a lot of rationale, but I don’t think it is a driver on this. There was a
significant series of focus groups that followed a US EPA conservation PSA series
around the country. Most of the participants in the focus groups were actually quite
offended by the notion that they had to go out of their way and change something
about their lives in order to make someone else’s life better. I think some of the
responses were “I recycle, I am doing my part.” So driving less, taking transit, all of
that….was found to not only be ineffective, but it had a very negative reaction from
the focus group audiences. Whether it would have the same reaction here is different,
plus we just saw a social marketing study that came out of Victoria recently, where
people were very aware of the negative consequences of their car use and a number of
them indicated that they chose to moderate their car use, switch modes, switch
directions pick their home based on their understanding of the negative implications
of car use.”
38
Another study participant concurs with this notion that environmental
awareness and concern about climate change hasn’t been seen as a motivating factor,
citing the current Mayor’s position as support:
“And even Mayor McGinn mentioned this at the Commute Seattle meeting that
environmental reasons aren’t primary, it’s more like the convenience of biking, the
cost savings of biking, you don’t get stuck in traffic- that could be a factor. The cost of
parking is a factors and the availability of bike parking is a huge factor- to be able to
get around by bike downtown is so efficient.”
Another expert on SBMP implementation echoes this sentiment:
“It is certainly nice it was there it was helpful, but I think environmental reasons get
overly rated. For example, if you get back to citizens support…you ask any person on
their bike why are you riding your bike, you are lucky to get one out of twenty
mentioning the environment. Now that may be a background reason. But today, it is
raining, and I have to make a decision about whether I am getting out on my bike and
it is probably a whole bunch of other personal factors that are much closer, right here,
that define that”
Notable in each of the examples above is that the topic was not the larger trends
that influenced support or opposition for SBMP project implementation but the
personal determinants of the decision to ride a bicycle. The possibility exists that
respondents misunderstood the question, which would invalidate the responses given
to some degree. The final quote above, however, suggests another dynamic at play- the
responses may be indicative of an assumption among non-motorized transportation
professionals that support for implementation of the SBMP is related to the decision to
ride a bicycle- the reasons why one chooses to do so are therefore secondary and
environmental awareness ranks low as a motivator. Regardless, while the data does not
lead to a definitive answer, increased awareness and interest in climate change and the
39
environment is considered by study participants to have a neutral effect on
implementation outcomes.
3.1.2.b Increase in Gasoline Prices
The timing of the SBMP short term implementation period coincided with a
larger trend of unprecedented increases in U.S. gasoline prices. Prices rose to an
average high nationally at $4.11 per gallon in August 2008, with a West Coast high of
$4.44 per gallon in early July 2008 (EIA, 2011). Articles ran during this time reported,
anecdotally, that gas prices were responsible for increasing number of bike sales and
usage, both locally in Washington State (Raderstrong, 2008) and nationally (Hurdle,
2008). An online survey sent out to bicycle retailers by the Bikes Belong Coalition (2008)
provides support to these notions, with 95% of bicycle shops reporting new customers
who cite gas prices as a reason for their bike transportation related purchases, and 80%
of retailers attributing increased bicycle sales to higher gas prices.
When asked about this trend, three respondents believed there to be a
correlation between gas prices and increased bicycle ridership, with the following being
a typical response:
“We were able to draw a correlation between increases in gas prices and increased
bicycle use, but at the same time we were making so many changes on the ground in
Seattle that it is really hard to isolate the variables. Nationally they saw a huge uptick
as well where the changes weren’t necessarily being made so we might be able to
draw a stronger connection.”
As with increased interest in climate change (discussed above), it was again
notable that while the original question specifically asked about contextual events that
40
impacted Plan implementation, respondents replied to question in terms of how
increased gas prices affected ridership. Similarly, it may be the assumption of
respondents that external factors are influential on ridership, which may in turn affect
the relevancy for individuals and the likelihood of support for the plan now that they
have a stake in its completion. Responding to another question regarding public
support, one informant’s statement provides support for this notion, asserting that,
“anytime you have a greater constituency it leads to more political support for
implementing the plan.” Making this determination is outside the scope of the study,
and while the respondents reported a supportive influence on cycling, it is not possible
to determine from the responses given whether there was a positive influence on
implementation and what that mechanism might be.
3.1.3 Political stability
Ison and Rye (2003) note the importance of political stability in the case of road
charging schemes, where many may have failed to advance beyond the initial stages
because of local authority elections and a subsequent change in the political
environment. Interview participants discussed their views on various political actors in
the process, with some conflicting opinions presented about the role of the State and
County in aiding or obstructing Plan implementation. Three interview respondents
affiliated with the implementing agency noted that overall, the city is largely
independent of the County and State in regard to implementation:
“The county is a presence [at the regional planning level]… as well as local
municipalities who have all been supportive and are actually interested, and I think we
as a larger municipality really have been influential and keep that momentum going
with other municipalities. They actually have been interested in implementing the
41
kinds of facilities that we have, but that doesn’t really speak to anything that really
influences our ability to implement our plan.”
A focus on the city is therefore most appropriate, with two respondents
characterizing the short-term implementation period as being one with relatively few
changing players, and consistent political support and stability throughout:
“…we had a favorable climate in King County. Then we had…Mayor Nickels, and then
as now, a City Council that seems to be completely in sync with the Mayor’s objectives
and policies for making changes for the benefit of the city as well as the world for
climate. To me, that has been a marvelous benefit as we don’t have agencies and
executives working at cross purposes.”
Another respondent, while in agreement with the foregoing comment, notes
that the same stable political environment makes it difficult to assess the role of stability
as a factor influencing the Seattle Bicycle Master Plan, as there is no basis of
comparison:
“Well, we only had one mayor during that period, so that didn’t change. And there
was some change in the City Council, but overall if you look at the votes and the
nature of the City Council, the vote split didn’t change much during that period so that
was pretty stable. One [City Transportation] Director that whole period so that was
pretty stable. I don’t think that this is a good case study to answer that question
[regarding the role of political stability]. There was no change so I would be
speculating. I have seen change make things a lot better and change can make things
a lot worse. I think the message is that people are important rather than whether
there is change or not and people really do make a difference, whose Governor or who
is Mayor, it really does make a difference day to day, that leadership.”
Respondents overall reported a supportive political environment in Seattle as
being an important factor in assisting implementation. The last respondent is correct in
that it is impossible to tell what the magnitude of the effect of stability on SBMP effort
is, as political conditions were consistently reported as conducive to plan
implementation throughout the process. It is only by identifying and comparing specific
42
variations in the political environment, with their concomitant effects on individual
projects, that a researcher would be able to get a sense of both the actual mechanisms
involved and the magnitude of their effect. The environment of consistent support can
only lead to speculative responses regarding the importance of stability. Implicit in the
responses, however, is that it was beneficial to have a consistent political environment
that is as favorable as or more favorable than the one present at policy adoption.
3.1.4 Resourcing: Staff, Time, Funding and Expertise
Even when a policy is physically or politically feasible, it may fail to achieve its
stated intentions if the means of implementing the policy are not adequately provided
for. A common reason for implementation failure is that too much is expected too
soon, especially when attitudes or behavior are involved (Hogwood and Gunn, 1984).
As the theory behind the Seattle Bicycle Master Plan is predicated on a change in
behavior among city residents in response to a more conducive cycling environment, it
is particularly critical that projects on the ground are not hampered by resource
constraints. Gaps in the implementation of BFN projects could create a barrier to
greater cycling if city residents see important destinations as inaccessible and
potentially dangerous. Gaffron (2003) echoes the importance of resources in her study
of British authorities implementing pedestrian and cycling policies, citing staff, time and
funding as providing the greatest potential opportunities, or barriers, to implementation
depending on the circumstances.
43
In addition to the consideration of staff, time and funding, the availability of
expertise among agency staff and others charged with implementation is a critical
element of this question. Charles included this criteria in his implementation framework
based on Sabatier and Mazmanian’s (1979) observation that policy support among
implementing staff is essentially useless if not accompanied by political and managerial
skill in utilizing resources. Among these skills are the ability to develop working
relationships in the agency’s subsystems, convince opponents and constituents that
they are being treated fairly, the ability to mobilize support among latent supportive
constituencies, and to make sure the program is fiscally responsible, among others.
Given that the scope of this interview question could easily occupy an entire
study on its own, the purpose is only to identify and explore the extent to which given
resources are seen by participants as affecting implementation outcomes. The influence
of appropriate staffing, expertise and the presence of a dedicated transportation
funding measure are discussed below. Discussions around funding turned to the issue
of capital projects, such as bridges and multi-user trails, which were seen by participants
as problematic in terms of implementation due to their complexity and associated costs.
3.1.4.a Staff
In anticipation of the increased volume of work required of SDOT’s Bike
Program, the SBMP recommended that an additional three full-time staff would be
needed to implement the Plan during the ten-year time frame. Three respondents
44
remarked on the importance of staffing, generally finding it to be adequate at current
levels for the completion of projects prioritized during the timeframe of the study:
“I think that staffing is a bigger concern than actually the financial aspect right now.
Because each year [project implementation] gets more complicated, the first few years
of the Plan you are doing the easier projects and each year the projects get to be a
little more challenging.”
“You really need the staff support, it takes a tremendous amount of work to put in
infrastructure; it is not an easy thing….there was additional staff but there certainly
was a struggle to get everything done. The good news is you have a whole bucketful of
money, the bad news is you have to spend that whole bucketful quickly and it’s hard.
They did it and they did a good job, but it was stressful and hard.”
Two more responses lend support for the importance of appropriate staffing
levels while noting that the declining economy is responsible for not only a staff freeze
but a potential loss of staff and its associated expertise. It should be noted that a
potential loss of staff was not thought to be a barrier to project implementation during
the scope of this study, but as a concern weighing on the minds of those closest
involved with the process, it may be assumed to represent the importance of
appropriate staffing levels.
“We are also in a no hiring mode for the entire city. Because of the economic forecast
and the revenue forecast I think that we are pretty much pressed staff-wise to do as
much as we are doing and do it well.
“[SDOT] staffed up post Bridging the Gap and the non-motorized section has been
refined a couple of times since then and now has the staff, the time, the funding and
the expertise. That could all change, layoff notices went out in SDOT this year and
people…there were people who were in the bike-ped program who were low people
on the totem pole who could be flushed out. The staffing and resources issue could
become a significant problem.”
The Seattle Bicycle Master Plan’s recommendation to increase the number of
staff, as well as concerns from respondents that current staffing levels may become
45
pressed due to the increasing complexity of projects and potential for cuts, testify to the
importance of providing for additional positions to respond to the heavy workload
associated with a BMP. Two participants made the connection that although dedicated
funding for the plan was in place during the study period, a lack of staff to study and
implement projects could prove a limiting factor. What qualifies as “enough” staff will
depend on a multitude of factors that affect workload and cannot therefore, be
assessed. For this studies purpose, responses testify to the supportive effect
appropriate staffing levels have on implementation.
3.1.4.b Funding: Bridging the Gap Levy (BtG)
In 2006, a year before the SBMP was adopted by City Council, Seattle voters
passed a nine-year, $365 million transportation levy for maintenance and improvements
know as Bridging the Gap (BtG). Complementing the levy are a commercial parking tax
($127.5 million) and an employee hours tax ($51.5 million). Over the life of the levy the
total expected revenue from the three sources is $544 million (SDOT, 2010) The main
purpose of the program is to address the maintenance backlog for paving, sidewalk
development and repairs, bridge repair, rehabilitation and seismic upgrades, transit
enhancements, and other maintenance work. Per the authorizing ordinance, 18% of the
levy is to be used for bicycle safety and pedestrian projects. More specifically, it
provided funding to support the 93 miles of bike lanes and sharrows that were installed
during the 2007-2009 period of SBMP implementation.
46
All five interview respondents emphasized the critical role that the BtG played as
a dedicated and consistent funding source that was available from the outset of Plan
implementation.
“There are characteristics of effective non-motorized plans that we go around the
state and train people and teach people… it has to have an implementation timeline, it
has to have dedicated funding. [These] are critical components because the plan
sitting on the shelf is the all too common scenario. You develop the plan and then you
can never fund it and you have no prioritization scheme involved, you have no
implementation schedule involved so it never goes anywhere. Eventually it becomes
outdated and outmoded and so even when you refer back to the plan or use it, it isn’t
a worthwhile exercise.”
Two more respondents agreed that the levy’s concurrency with plan
implementation was a critical piece, with one citing that it was likely to be the most
important factor influencing successful implementation:
“[It was] probably the strongest. I mean having the plan but then really
simultaneously having the funding package come through [at the same time].”
Two respondents, when given the hypothetical scenario that the BtG was never
passed by voters, were emphatic that projects would likely have not been possible.
One respondent discussed the likelihood that bicycle facilities would have been created
along a route he rode into work:
“They wouldn’t have been, there is no doubt about that. I’ve been cycling…since 2005
on a daily basis, it never changed in all those years and it certainly wouldn’t have if
the money hadn’t been approved. That certainly is a tangible benefit that I have seen
myself, I don’t doubt that the BTG levy has in fact allowed the BMP to be
implemented.”
Similarly, when asked to follow up on what the potential fallout would have
been for the SBMP if the levy didn’t received voter approval, another respondent
thought that:
47
“… we would have really had to rethink… about how to construct the [Seattle Bicycle
Master] Plan before we sent the final draft to Council for adoption that fall if we had
not had that initial revenue.”
Hogwood and Gunn (1984) state that politicians sometimes will the policy “end”
but not the “means”, so that expenditure restrictions end up starving a program of
adequate resources. It could be argued that Bridging the Gap represents the opposite
end of the spectrum in this regard. As a foundational funding source throughout the
majority of the ten year implementation period, a strong argument can be made based
off of interview responses that Bridging the Gap is the critical element in the Seattle
case for being able to implement the Seattle Bicycle Master Plan.
3.1.4.c Funding: Capital Project Cost
The presence of both Bridging the Gap transportation levy and the Complete
Streets ordinance allowed for a both a consistent funding source and for greater
efficiency through the routine accommodation of bicycle facilities. As such, funding was
generally seen by participants as being adequate for the 2007-2009 period, with some
exceptions:
“Projects that will be a problem: bridges, all the major capital projects. So for
instance, building a bridge across 47th and I-5, connecting the U-District to
Wallingford…even the most conservative estimate I can come up with…is $35M. $35M
is more than we have spent on bike capital projects in Seattle since the plan was
developed…these aren’t cheap. The Ballard Bridge, even the cheap way to improve
the Ballard Bridge, is a $4M project. And that’s cheap. It would be a substantial
improvement for very small dollars compared to the original parallel bridge proposal
that was in the SBMP which could have cost as much as $50M.”
48
The response above was shared by two other interview participants who
reported the construction of bridge and trail projects as problematic in terms of
implementation. Responses indicate that it was not the type of project in itself that was
obstructive, but rather the prohibitively high cost associated with the construction
coupled with the fact that alternatives for connecting bicycle facilities (particularly with
bridges) are not readily available. Two participants noted that the construction of capital
projects and trails presents a substantial opportunity cost to developing on street
bicycle facilities, as one project can drain the budget or several years and miles worth of
on-street bike facility construction. One participant believed the difference in costs
between on-street and capital facilities justifies the need for separate financial plans:
"There was not a shortage of funding [during the short-term implementation period].
Some of the trails projects were really expensive, sort of a separate deal, those are the
outliers and the ones that cost millions of dollars, some of those don’t have enough
money. In terms of on street system, I don’t think that funding was an issue… One
thing I would do differently in organizing a plan is that if you look at the plan, there is
sort of [on-street facilities] and then there are these outliers that cost millions of
dollars and then tend to be these bridges like the facility over Ballard Bridge and
possibly something over I-5. And those are so expensive that if you look at the total
cost of the plan and then look at the money spent it looks like the plan is underfunded.
Which is not quite fair. What I would do now is [include] everything but those outliers
and I would financially set those aside. We need to mention them because they are
important to happen, but neither do you want to take twenty years of bike funding
and build one bridge. That would be dumb too. On those really expensive ones you
have to find an opportunity to do it with something else. For example, on Ballard, it
looked like there was an opportunity to get some of the work done as part of the
monorail that was going to go in there, we worked closely with the monorail people
and then of course that didn’t happen. So there is not funding for those mega
projects, but setting those aside, there wasn’t a lack of funding for the other stuff.”
This current lack of funding for “mega projects” could prove to obstruct full
implementation of the plan, however, study participants responses converged in
characterizing the issue as problematic during the Short-term Implementation Period, as
49
advocates and city staff worked on long-term strategies and funding solutions capable
of moving these critical projects forward. The high costs of bridge and trail projects may
prove to obstruct full implementation of the Plan should adequate funding not be
provided for.
Table 4: Seattle bicycle Master Plan Higher Costs Projects in Areas of High Bicycle
Demand
Seattle Bicycle Master Plan Higher Cost Projects
•
Provide a bicycle facility connection between Downtown Seattle and the UW Campus
via Eastlake Avenue N.
•
Complete the Ship Canal Trail, including connection to the Fremont Bridge and Ballard
Bridge
•
Construct the Chief Sealth Trail Crossing of I-5 between S Spokane Street and S Lucile
Street (and provide a trail on the east side of I-5 between the Chief Sealth Trail and
the I-90 Trail).
•
Construct the Burke-Gilman Trail section between 11th Avenue NW and 17th Avenue
NW
•
Construct a new bicycle and pedestrian bridge across I-5 between Wallingford and the
University District
•
Provide a bicycle facility connection between the I-90 Trail and Downtown Seattle
•
Construct multi-purpose trail connections from the SR-20 Bridge to the UW Campus
and to Downtown Seattle as a part of the bridge reconstruction project
•
Improve the bicycle lanes on Alaskan Way S/E Marginal Way S between S Spokane
Street and Downtown, and complete the E-3 Busway Trail between S Spokane Street
and Downtown
•
Either rehabilitate the existing Ballard Bridge or add a new bicycle and pedestrian
bridge adjacent to the Ballard Bridge.
Source: Seattle Master Bicycle Plan, pg. 13 (SDOT, 2007)
All of the projects that appear on the SBMP’s list of higher cost projects in areas
of high bicycling demand (Table 4, below) consist of bridges and trail connections
considered to be critical in achieving the Plan’s goal of tripling ridership and increasing
the safety of bicycling in Seattle. Should funding be unavailable for these projects,
outcomes over the 10 year implementation period are expected to fall short of plan
50
goals. For these reasons, the majority of participants concern about these projects
appears well justified and the high costs of capital projects should be considered a long
term issue in terms of SBMP implementation.
3.1.4.d Expertise
Issues of expertise tended to focus on SDOT staff, with all three of the interview
respondents who engaged the question being current or former SDOT employees.
Interviews diverged somewhat from considering whether implementing agency staff
expertise affected SBMP during the 2007-2009 period, to respondents providing their
views on the SDOT’s reorganization of its bicycle program to a matrix model. A loss of
expertise due to an overreliance on one individual left agency staff in the position of
needing to pick up the pieces:
“With Peter Lagerwey (former Senior Transportation Planner with SDOT) we had a
pretty senior staff person with 25 years of historical knowledge who retired, so
regardless of funding, staffing, that was a huge loss to implementing a bike program.
So we felt it, because it takes us longer to do something that he would know,
especially about a trail. We are also in a no hiring mode for the entire city, so because
of the economic forecast and the revenue forecast I think that we are pretty much
pressed staff-wise to do as much as we are doing and do it well…staffing is a bigger
concern than the financial aspect right now.”
The loss of expertise and institutional knowledge discussed above was shown to
stress other resources, particularly staff, which had to quickly get up to speed in order
to perform the same duties that were lost due to the vacancy of this key position. All
three respondents discussed the need to disperse the knowledge and build resilience
across the organization:
51
“In theory you shouldn’t have a special bike section in a transportation department,
just like you don’t have a special car section. In theory the system should be
integrated and institutionalized so you don’t need that. I have never seen a city do
that successfully 100%. One of the problems is that on paper it looks good but the
people you hire don’t know anything about bicycling or walking, they didn’t learn it in
school and they are not likely to have learned it anywhere else they have worked. For
many years …we had one and a half/two people in the whole program and there was
a real lack of understanding or expertise throughout the larger department and then
slowly that changed by the time we started implementing the plan… in general I
wouldn’t say that would be a major barrier.”
This emphasis on the matrix model, where responsibilities and knowledge are
dispersed across the agency may ameliorate the effects of the loss of what may be
deemed a “policy expert”. Similar to Gaffron’s (2003) statement that the presence of a
policy champion may been indicative of the fact that bicycle policies are not well
institutionalized, expertise concentrated in one staff position presents a problem for
organizational resilience. Organizations with a separate pedestrian and/or bicycle
department may be particularly susceptible to a loss of staff that hold key expertise and
a lack of accountability to the Plan across programs. In the view of one participant,
however, the matrix model:
“…has its pros and its cons. The biggest pro... is institutionalizing implementation of
bicycle infrastructure in the city. So that the more [agency staff] who are aware of
[the Seattle Bicycle Master Plan] and the more people that do it, the more eyes there
are to make sure it gets done. Not everyone from what would be our Pedestrian
Bicycle Neighborhood Streets program and project development group could be at
every meeting so when you have people from our Traffic Operations implementation
group, they can be on the lookout for things. So, theoretically, it’s dispersing the
knowledge across the organization so that it will help with implementation.”
The cost of dispersing knowledge and responsibilities across the agency is that
staff must adapt to the new model and learn about other programs they are not
52
currently acquainted with. Another respondent notes this effect has been seen
throughout SDOT, without saying explicitly that implementation was impacted as a
result:
“It also has to do with a lot of education throughout the department to really
institutionalize. We have the financial resources, but the human resources… it is an
ongoing learning curve for folks. [Peter Lagerwey] kind of held a lot of the knowledge
for many, many years and we’ve kind of exploded the bicycle program to be
institutionalized throughout the department but there is a learning curve to that, so
that is what we are experiencing now with people coming up to speed, not only in our
group, but at other levels throughout the department.”
A definitive conclusion on the effects on implementation cannot be reached,
however, as participants overall did not comment specifically on the impact the staff
loss or organizational model change had on projects during this time. Agency staff did
note that there was a learning curve associated with dispersing information and
responsibilities throughout the organization in the wake of a major loss of expertise, but
it is assumed that they were able to cope successfully- the initial quote above supports
that “it wouldn’t be a major barrier”. The context provided is useful, however, in
identifying the important role of organizational structure in making expertise a more
resilient asset within an agency.
3.1.5 Monitoring
The SBMP recommends periodic monitoring through performance measures
(Action 4.13, SBMP 2007) that are to be evaluated on a bi-annual basis to ensure that
they are the most appropriate and cost effective measures for assessing progress
towards the Plan goals of increased safety and ridership. Monitoring the outcome of
53
implementation efforts is seen by Ison and Rye (2003) as closely related to the question
of whether the policy demonstrates a cause and effect linkage, assisting implementing
staff in refining the plan to make it as effective as possible within the available
resources. It is this second aspect of monitoring that we are most interested in here.
This interview question therefore looks to establish whether the performance
measure data gathered simply allows for the program to provide evidence of its overall
progress towards established Plan goals, or whether a feedback loop has been
established that allows the outcomes to inform future implementation efforts. Study
participants did not mention the SBMP’s Strategic Performance Measure related to
facilities, the percentage of bicycle network completed, as being particularly useful in
assisting plan implementation. More important was the use of a GIS tool that is allowed
for prioritization of projects based off several parameters:
“ I guess the way I see [data] influencing our projects is how we select projects, we
have a whole prioritization process, a GIS base, we have five categories and we assign
points to different roadways or projects and we include collisions in that analysis. We
also have some demand, not necessarily based on our bicycle counts, but we look at
neighborhood centers and major employers and tie that in. So I think that is probably
the most influential where we are trying to address high collision locations through
the projects we are trying to select every year.”
New data, then, can be used to queue up projects that are most likely to prevent
the likelihood of accidents and fatalities, assisting in the earlier accomplishment of Plan
goals. Another respondent echoes the important role of using geographically based data
to increase ridership and reduce fatalities, while addressing broader social and
environmental justice issues in historically underserved areas of the city.
“There is feedback. We had a soul searching moment last year… we had four fatalities
in the city, a number of very serious life threatening injuries, and none this year. The
54
year is not over but we are really asking ourselves, “Did we do anything wrong here, is
there a false sense of security?” So the metrics, the counts, crashes, fatalities, they all
feed back into the (GIS) system and are instructive of how we accomplish the goals of
the plan…what we asked for this Spring and got was for the city to incorporate the
health inequity information that was used in the GIS modeling for the Pedestrian
Master Plan into the Bicycle Master Plan so that in terms of our geographic focus,
whether its low income or disadvantaged communities…communities with high crash
and fatality rates… that we are accomplishing a broader set of goals. And if sedentary
lifestyles are influencing health equity/health outcomes in South Seattle…because
their street network is not what it is in North, Northeast, Northwest Seattle there
aren’t the diversity of travel options…they act as barriers they keep people from
walking and bicycling, they contribute to poor health outcomes and air quality
problems, they contribute to disengagement socially which we know contributes to
poor health outcomes. So in incorporating that other metric we are going to better
address the safety and mobility needs. By factoring those in to the GIS tool we can
help prioritize the planning.”
Data played a clear role in prioritization of projects that may assist in plan goals,
but it says nothing of the quality of the outcomes. The same respondent notes that the
dogged pursuance of more facilities, however prioritized, can lead to a shortsightedness regarding the appropriateness of facility types for a given project.
“…the downside to rigidly adhering to metrics, how many lane miles were dropping in
is it can lead you to a quantity over quality approach that can be detrimental. The
quality comes down to how much time you can put into really looking at configuration
of interchange. If you are just trying to hammer [bicycle facilities] out …that’s not
helpful if their just slapping down stuff that we come back and find that there were
problems with. I can’t point to any really significant examples, but the early
implementation of sharrows was sharrows were not located as per city instructions.
On 34th at Stone Way, sharrows in the middle of the right turn lane where we thought
about configuration issues and we said, “Wait a second, why aren’t we moving bikes
across the turn lane and queuing them up with the thru-traffic over there.” There will
be facilities around this town that need to be reworked.”
Again, a respondent found that an exclusive focus on metrics, specifically the
number of miles of bike lanes installed and crash counts, creates a problem by not
addressing more complex projects that represent barriers to cycling. The resulting
discontinuities in the Bicycle Facilities Network that result are seen as stemming from
55
strong political support the Plan currently has, leading to a desire of public officials to
act while the opportunity exists:
“Well, certainly they do the counts that are going to…tell you how well you are doing
and not what you are doing right and what you are doing wrong. What the city has
been doing is focusing on just getting miles of bike lanes and shared lane markings in,
but what they haven’t done is gone back and addressed spot locations that are known
barriers. And if you don’t address those locations, you aren’t going to get a big jump
in cycling…The question comes in “Do you go for those spot locations or do what they
have been doing [focusing on getting lanes installed]. I would argue they are doing
the right thing. And so there is a little bit of a delay in terms of realizing the benefits
of what you have been doing because right now there is a window open…those
windows don’t stay open forever. Portland went through a period where they hardly
put in any new bike lanes for a number of years. When that point comes, then you will
have an opportunity to circle back and do those spot locations and remove those
barriers. But you can always do that, that’s not cyclical. But the massive ability to put
in miles and miles of new facilities probably is. So short term, I don’t think the city is
getting all the benefit in terms of increasing bicycling that it could, long term. I think
there is going to be an opportunity to come back and fix those spots and then you’ll
see a huge change. So that’s the strategy and it is one I agree with but it is also not
real satisfying because I think you’ll find that there was an initial surge in numbers,
but I suspect that it is going to plateau at this point for a bit…”.
Responses fell short of converging on a singular influence for the role of data in
affecting future project implementation. Responses did, however suggest that data
collection could lead to positive implementation outcomes through the prioritization of
projects that will accrue safety and health benefits for a greater amount of time. An
overreliance on metrics in terms of getting bike lanes on the ground, while support and
funding is present, may result in early successes, but respondents suggested that the
long-term appropriateness of a given facility, as well as the lack of critical connections,
may serve to limit the effects of the Bicycle Facilities Network overall until key barriers
have been removed.
56
3.1.6 Leadership- Single Implementing Agency
Hogwood and Gunn (1984) specify that a necessary condition of implementation
is that there is a single implementing agency which depends minimally (or ideally not at
all) on other agencies for success. This view is predicated on the idea that
implementation requires a complex set of conditions and linkages to ensure a successful
outcome- the additional need for agreement among a wide array of participants in the
process will only serve to further diminish the potential of a successful outcome. The
theory holds that the number of decision points and need for consent among multiple
groups may serve to obstruct full program/plan implementation. The author framed the
interview question in a neutral way to avoiding assuming that the involvement of other
agencies is a liability to the process- it is just as likely that the need to consult other
actors for their expertise, criticism, and agreement with the implementing agency’s
approach may prove to be critical for predicting and preventing obstacles that may
serve to derail the process down the line.
All respondents acknowledged that the construction of bicycle facilities requires,
by necessity, interactions with a large number of agencies, stakeholders and the public.
Participants relayed that the involvement of other groups depended on the attributes of
the project itself; whether the project was a segregated bicycle lane or a sharrows,
whether it affects parking or businesses, and whether the number of existing vehicle
lanes would be restricted, among many others.
“There may not be a complete awareness of problems on a decision. From an agency
standpoint, they are trying to accomplish a goal which is to get a road repaved and in
doing so, how do they accomplish the Complete Streets [requirement]? Well that
57
means reducing parking. So what is the effect on business? The business owner says I
need those three spaces or otherwise I won’t have a business…I think that overall the
process was helped, not hampered, and improved. It took more time but I think the
result was better because they actually got some very good feedback and that
feedback is being seriously considered as part of the implementation.”
An agency insider notes that while the necessity of dealing with stakeholders is
essential for the projects longer term, it is the citizen input in particular that can be
problematic in terms of getting the requisite feedback and ensuring awareness about
the changes to be made on the ground level:
“Seattle is a very inclusive city, politically…we are process oriented, it’s just the nature
of Seattle. We have to do a lot of outreach, we have to get a lot of feedback… the
process that involves all the other stakeholders that we talk about, that is essential for
project success. Getting citizen input is really essential for project success, as well, but
that process is probably the one that you could say would maybe slow us down the
most.”
As established above, the conflicting opinions of affected stakeholders, in of
itself, is not necessarily problematic in that previously unidentified issues can be
brought to light. Indeed, disagreement on the best way to proceed may modify a
project in ways that can more comprehensively benefit all user groups. Another SBMP
advocate’s insight into the stakeholder process may serve to refine this sentiment in
noting that it is the evidentiary nature of the stakeholder input that can determine
whether the influence on implementation is problematic:
“[The SBAB’s] contribution insofar that they support bicycling is helpful and useful, but
anecdotes in my mind don’t contribute to the improvement [of projects]…But the
citizen involvement is good when it eliminates things you don’t know. And so citizens
advisory groups are good when they can bring a perspective that the department may
not have, but their challenges to technical assumptions can be counterproductive.
Generally, the interaction between agency staff and an advisory group doesn’t say
“Well here are my three peer reviewed reports, where is your evidence?” That’s not a
way to facilitate a dialog. And because they only meet monthly, having to turn back
around and refute assertions and assumptions can add delay and complication to the
58
project, and the inclination of the agency staff may be to placate the opposition…by
modifying the projects in a way that isn’t necessarily supported by the data. So we
value citizen input, citizen advisory groups and stakeholder groups to the extent that
they don’t undermine sound projects that are composed around generally accepted
standards and practices.”
As the participant noted, balancing the ideal of data-based decision making is
the need to facilitate a dialogue with outside users. Inherent in involving other affected
groups in the construction of SBMP projects is the notion of accountability, which
according to one participant, has the secondary effect of building long-term support for
the Plan itself:
“I think it is real important when you do any plan that you build in some
accountability [in terms of who approves projects], who do you report to? Late in the
fall, November or December, [SDOT] went to the SBAB with a work plan and said this
is what we are going to do next year and it includes a list of all the streets that are
going to get bike lanes or shared lane markings and they approve it. And then at the
end of the year you come back to them with a report card and say this is what we
promised and this is how we did. That is an important notification...Certainly
sometimes during the public process we will slow up a given project, on the other hand
the only reason you get to do all the projects is because there is so much public
support. So in the bigger picture, it makes things go faster. So the reason that Seattle
could put in so many facilities so quickly in a given year is because I think they got the
public process thing right. In the macro picture, it speeds thing up. In the micro
picture, it may slow down a given project.”
Another long-term effect identified by one study participant involved making city
processes more efficient and streamlined through interagency consolidation. The
coordination of needs, in this case, between multiple public agencies is seen as crucial
to avoid affecting the quality of recently installed facilities.
“…[SDOT tries] to get everybody together in one room and in the Right of Way
Improvement Manual the need to have this interagency consolidation on the project
planning was seen as imperative for a lot of reasons. I mean SDOT would go out and
put down brand new asphalt and will have asked City Light and Public Utilities if they
needed to do anything while they were in there, and City Light and Public Utilities
couldn’t get the work and the analysis done in time. And then two weeks after the
59
fresh asphalt is down you would see the spray paint markings where SPU is going to
come in and cut it back up. And then there is a big fight generally between City Light,
SPU and SDOT over the quality/standard of the repair so that it has the same project
life as the repaving did… So there is a turf war that goes on there, but they try to have
one consolidated agency working group.
Overall, interview responses related to the public process and the necessity of
dealing with multiple agencies and groups were categorized as having a mixed influence
on implementation. Projects were slowed down initially, but this was considered a
short-term effect. Responses ultimately contradicted Hogwood and Gunn’s
implementation framework by acknowledging the long-term benefits yielded due the
routine involvement of diverse and relevant groups in decision making. Three of the
respondents quickly qualified their initial assessments by noting that the long-term
effect of this involvement is better implemented projects that more comprehensively
address other public needs through the incorporation of insights that are not available
to staff. Greater accountability was though to engender greater public trust and in turn
support for projects, and involving groups with a stake in individual projects through an
interagency working group was thought to lead to efficiencies overall. Although the
responses differed in the aspects of implementing agency interactions that they
touched on, the results are suggestive that coordination of groups includes an initial
transaction costs that pays off in the long term by increasing public trust, reducing the
risk of unknowns, and minimizing redundancies between agencies that could serve to
degrade facilities.
60
3.1.7 Ability to Enforce Compliance
Noting that it is perhaps the least attainable condition of “perfect”
implementation, in addition to being a system that most of us would not want to live or
work under, Hogwood and Gunn (1984) posit that the potential for resistance to a policy
is limited if there is authority to secure total and immediate compliance from others
involved in the implementation process whose consent is required for the success of the
program. In such extreme cases, consent would have to be understood as symbolic, and
one would assume an inverse relationship between the amount of authority to demand
compliance and an institutionalized flexibility to public reaction. The question used for
the purposes of the survey opens up the possibility that other government entities may
be able to pressure for compliance with the Plan, as well as consider whether there are
consequences to overriding other actors involved with Plan whose short-term
opposition may in fact facilitate better long-term outcomes.
The implementation of the W. Nickerson Street rechannelization was brought up
in several instances by respondents as an example of implementation in the face of
opposition. The project, recommended as a part of the SBMP, would change the lane
lines on W. Nickerson Street, effectively reducing it from two travel lanes in each
direction to one travel lane in each direction along with a center-turn lane. Bicycle
facilities would be added in the space made available, in the form of a bike lane in one
direction and a sharrows in the other. National studies show that this level of traffic
could be accommodated within the proposed 3-lane configuration, it was expected to
61
slow traffic and also reduce rear-end collisions, side-swipe and angle collision, and make
pedestrian crossings more safe. (SDOT, 2009).
Yet many neighborhood residents were upset that the project would create
needless congestion despite the presence of the Ship Canal Trail, a recreational trail,
running directly parallel with the proposed rechannelization. In face of opposition,
Mayor McGinn remained resolute in directing SDOT to complete the project. Peter
Hahn, the Director of SDOT, said he had the authority to carry out the project without
further Council or Mayor actions.
Tom Rasmussen, chairman of the City Council’s transportation committee
responded to an outcry from neighborhoods and Democrat groups, saying that the
project should be delayed pending the completion of two other corridors are
completed. He considered several options to stop the project a) pass a budget proviso
withholding road-diet money, b) pass a recommendation for or against the plan, or c)
watch what happens, perhaps adding language repealing the road diet if things went
bad. Bike activists, in response, were able to successfully influence Councilmember
Rasmussen and he agreed to give the project a try. (Lindblom, 2010)
Clearly, the authority of SDOT and ultimately the Mayor was not absolute, but it
was substantial. It would take a veto-proof six council votes in order to stop the Mayor.
One respondent points to the unlikelihood of this happening given the legitimacy of
having a project in adopted plan:
“SDOT has the authority, as does the Mayor’s office, and without extraordinary
intervention the City Council does not. The City Council isn’t the executive so they
can’t direct the agency. Certainly the Mayor as the executive administrator of the city
had the authority to direct his department to move forward on a project that is in an
62
adopted plan. SDOT may also move forward so long as it is consistent with state law
and other standards and [the project] is in the plan…they are the actor that can just
drive over the top and say, “it is in the plan, we adopted the plan we went out we had
4,000 comments on the plan. The plan is the plan. You had an opportunity to
comment on the plan and didn’t.”
When asked whether such a response would generate fallout, another
respondent noted that for controversial projects, the city is more inclined to go through
its public process:
“On controversial project, yes, I think that that is right. But on ones that are not
controversial I don’t doubt the city makes a decision even if somebody complains, it’s
not seen as a big problem… I guess on Nickerson Street the city council was involved. I
think SDOT realized the Council didn’t need to be involved, the city could decide on its
own to implement a project yet it wanted the council to…well I think it was more for
support, do we have the support, and do we need to have the Council’s blessing? No,
because it has been budgeted. SDOT has the ability, probably legally to make the
decision without the Council saying yay or nay on it. I think it made sense almost as a
way to…process, say, what are the concerns of individual, and use the City Council
forum as a way to provide that feedback to the city and make the best decision. So I
think they went out and sought that feedback, to me that was fairly extraordinary and
it was smart too.”
Yet these two comment plays down the fact that an “extraordinary intervention”
of a City Council is still possible under the right conditions, such as overwhelming
opposition from affected stakeholders that makes the agency and Mayor’s position
untenable. Respondents tended to support the view that what allows project
implementation to move forward was not so much the legal structure, but the political
will and backing of elected officials to allow SDOT to move forward in the face of
opposition:
“I think it is important to remember that plans are not laws and there is no obligation
to do everything in a plan… It’s not a legal issue, it is a political will issue. In most big
projects there are one or two people who really don’t want it. What I have seen is
towns where they allow one percent of the population to stop something, are towns
that don’t get much done. So when I said figuring out that public process, part of that
63
is figuring out how to deal with the 1% of the population who will always be
negative…. [Political will] gives you permission to take the political risks and upset one
or two people because you also know there is a very well organized group out there
that represents thousands of people that have another opinion. That just gets down
to things like the Cascade Bike Club and let everybody know that you are well
connected.”
When two of the respondents were asked what would happen if opposition
continued to pursue using anecdotal evidence not supported by the data in order to
uphold a project, they similarly pointed to political will as being a key element for
obtaining compliance with the plan:
“That’s where political support is the key issue when we as the worker bees say, “Do
the management and elected officials have our back.”. We are used to it. It’s how
much we get the political backing with our projects.”
Both respondents replied that it was absolutely critical to have political backing,
which operates through support and trust in the implementing agency on the part of
elected officials who do have the power to obstruct or delay projects, if need be:
“We have had incredibly great management as well as political support for the
projects so if someone writes to the Mayor and say “I’m opposed to this project.”, the
Mayor’s office, instead of responding, will say “SDOT, I want you to send your
standard response, thank you for your input, your input is very valuable.” Maybe
some past administrations would have wanted to go into every detail and answer
every question but this administration has been very supportive of saying we have a
process, we understand SDOT’s process and we are supportive of that process.”
Convergence of opinion in the Seattle case demonstrates that political will is the
crucial element influencing implementation outcomes in the face of an opposition that
can serve to delay, halt, or substantially change recommended projects in the Plan.
Responses from implementation staff suggest an additional condition to Hogwood and
Gunn’s original formulation of the framework be taken into account- that those with the
formal authority to demand co-operation may only be able to use that power in
64
instances where public support and the will of elected officials is sufficient to overcome
opposition.
3.1.8 Agreement on Objectives to be Achieved: Mix of Facilities
Hogwood and Gunn (1984) compare examples of policy objectives in the theory
of planning, which we are told should be clearly defined, specific, mutually compatible
and supportive, against research studies that show that “real life” policy objectives are
often difficult to identify, may not be compatible with each other, or are vague and
evasive. Identifying whether aspects of the SBMP have been contentious due to a lack
of agreement on what the stated objectives are is therefore crucial; even if there is
legitimate discourse over the meaning of objectives in the Plan, it may serve to delay
projects or diminish the original intentions of the Plan in important ways.
The Seattle Master Bicycle Plan states the importance of providing a mix of
bicycle facilities. This view is predicated on the idea that different types of facilities are
appropriate depending on surrounding land use characteristics, available right-of-way
space, traffic volume and speed, and other roadway characteristics. Equal consideration
is given to individual bicyclists’ level of experience, in which some bikeways are more
preferred than others. Newer bicyclists, the Plan holds, often prefer off-road multipurpose trails and quiet neighborhood streets, while more experienced bicyclists prefer
bike lanes, wide curb lanes, paved shoulders, etc. (SBMP 2007) The issue of providing
for a mix of facility types came up consistently in interviews in response to the
implementation frameworks requirement for agreement on objectives.
65
The goals of tripling ridership and decreasing the number of bicycle accidents by
one-third are explicit in the SBMP, and may meet Hogwood and Gunn’s criteria of
having consensus and understanding of the overarching policy objectives to be
achieved. Yet it is the authors contention that implicit in these goals are assumptions
about what user groups should be served and what design treatments are effective.
This is in line with Charles’ inclusion in Hogwood and Gunn’s implementation framework
of the need for policies to contain unambiguous directives that structure the
implementation process to maximize success, as a lack of agreement on how to achieve
broadly conceived objectives may serve to unnecessarily hold up plan implementation.
Responses varied and no consistent effect on implementation was discernable. The
issue of determining the mix of facilities deserves development due to the presence of a
convergent theme across all interviews.
One respondent who was a key player in the initial creation of the Seattle Bicycle
Master Plan notes that even with consensus around the two Plan goals of increasing
ridership and decreasing accidents, methods for achieving those goals was a source of
debate from the inception of the plan all the way through current efforts:
“There was disagreement during the development of the plan on the composition and
mix of facilities. There were people who were particularly traffic adverse who were
offering up some pretty loony ideas about dropping jersey barriers down to segregate
the bicycle facilities…it was a small number, but it was a vocal minority. And then on
the other side you had a particularly vocal group of vehicular cyclists… they claim it’s
all education, that we don’t need any [separation of bicycle lanes]. And they say that
it’s just a form of segregation and second class citizenship for bicyclists, that all you
need to do is take the lane, in a 40 mph, 59,000 car [environment]… So we had the
complete segregation folks and the no facilities folks.”
66
Another respondent that provided guidance for the initial planning stages of the
SBMP similarly noted that underlying the consensus on objectives, broader
philosophical disagreements existed as to how to increase ridership and decrease
accidents.
“First of all, when we did the plan we had a bike advisory group that worked with us
and they unanimously supported those two goals which were really important.
Obviously in terms of how you reach those goals, there are a lot of different ways to
do that and that is why every year the work plan is taken to the SBAB and they
approve it by unanimous consent. Does that mean that everybody agrees with every
little thing in it? No. But it means there is an overall consensus. The Cascade Bicycle
Club always participates, they always have one or two people there at the Seattle
Bicycle Advisory Board meetings and they have been very helpful and supportive of
the whole thing. Having said that, can you find disgruntled people out there? Of
course you can. There has always been a small group of bicyclists who don’t want any
bicycle facilities...some of them like shared lane markings and some don’t.”
Debate and disagreement on how to reach objectives may simply represent the
demand of bicycle riders at various levels of proficiency advocating for a plan tailored to
their unique preferences and level of comfort while riding adjacent to motor vehicle
traffic. Another respondent discusses his experience on the Seattle Bicycle Advisory
Board,
“The SBAB did hear public comment. People come in and talk about their personal
experience, what they like and don’t like, the Board hears that and any members of
SDOT hear that. We have also heard feedback from the city that they would like to
deemphasize sharrows and emphasize other facilities in the coming years. I think that
certainly the Board agrees with that, again I don’t speak for the Board, but my point of
view is that that would be consistent with what we would like. So that has been a
change.”
While there is no convergence in responses, the above two quotes suggest that
by requiring consensus through the creation of a yearly work plan that is in turn
dependent on the approval by the Bicycle Advisory Board, Plan implementation can
67
successfully move forward by incorporating public comment while eliminating the
disproportionate influence of what may be termed “vocal minorities”. The potential
exists for implementation to be held up by disagreement over how to meet objectives,
yet interview responses are suggestive of the importance of having an accepted public
process that takes debate into consideration over how to meet the needs of diverse
users, ultimately being responsible for making a final decision.
Participants identified two other areas where changing understanding on how to
meet plan objectives may be necessary, and ultimately benefit implementation overall.
Again, while there was no convergence in responses, interviews suggested that
innovations that have taken place since Plan adoption may change the individual
facilities treatments that streets receive:
“And there has been a little bit of evolution as well since we have adopted the plan…
really the goals were to get facilities on major streets and arterials, as well as some
residential streets but its very arterial focused. Both in Seattle and other cities
throughout the U.S have been doing different kinds of facility types such as separated
facilities and cycle tracks and that is something that we are kind of adding to the plan.
And so just the backup a little bit, we take recommendations in the plan which we
take as recommendations in our design process we look at more of the details, and
that is maybe where, if there is any friction with those groups that sat in our planning
process, it’s in the details, its making those decisions as to “This is recommended” but
really when we get out there and measure the roadway and see what the options are,
our final design can vary from that initial recommendation. And occasionally that
means we can put in a facility that is not even in the plan.”
Finally, one respondent identified the need to change facility types from what
was originally outlined in the Plan in order to adjust to costs, when projects were
determined to be more expensive than originally thought.
While responses were not convergent on the influence on implementation, a
common thread that ran through all responses consisted of the need to adjust the mix
68
of facilities implemented due to changing conditions external to the Plan in order to
maximize the intended outcomes. Despite consensus on overall plan goals,
implementation actions will necessarily need to respond to user preferences as
bicyclists become acquainted with individual facilities treatments and attempt to
address their preferences through the public process. As cities increasingly experiment
with new road treatments such as bike boxes and cycle tracks, current BMPs will likely
adjust their recommended facilities to incorporate successes. The implementation
framework suggests that the lack of agreement on how to reach objectives may serve to
hinder implementation short-term until a consensus can be achieved, but it is just as
likely that the adjusting to new realities and changing user preferences will lead to
better outcomes overall.
3.1.9 Leadership: Policy Champion
Ison and Rye (2003) argue that an important influence on successful
implementation is the availability of what they call a “policy champion”. The role of the
policy champion is to provide leadership and direction during the implementation
process when a diverse range of stakeholders are involved and where alliances can
often be fragile. Indeed, Gaffron’s study (2003) of implementation of walking and
cycling policy in British authorities cites the presence of a local champion as often
sufficient to influence policy implementation. The study is quick to note that while this
may be a benefit to pedestrians and cyclists, it may cut both ways- the importance of
individuals means that a highly motivated actors or opposition groups may have an
69
undue restrictive influence on implementation efforts. The importance of a single actor
may also be an indicator that cycling policy implementation is not well embedded in
existing administrative and political structures and cultures. (Gaffron, 2003) The
interview question, unlike Ison and Rye’s brief treatment of the topic, allows for the
possibility that there may be consequences to a structure that grants one person
disproportionate influence over policy or plan outcomes.
Overall, the implementation framework proved useful in that the interview
respondents were readily able to identify champions whom they felt were pivotal in
affecting the implementation process. All respondents mentioned Cascade Bicycle Club,
or their advocacy and executive director, as policy champions that played a critical role
in successful implementation of the Plan:
“David Hiller (Advocacy Director at CBC) is probably the single most visible advocate
for cycling in the city and he is very knowledgeable and very active and very energetic
and very impressive. Now he doesn’t always create a single minded approach, he
sometimes has to take an approach that he feels is different than others are taking so
he has to raise objections when he feels it’s necessary. It’s championing bicycling, but
it’s not necessarily getting everybody on board with him in accomplishing that result.
But he is very good, very knowledgeable and I am very impressed by his experience
and commitment.”
Obtaining dedicated funding for the Plan, already identified as one of the major
pieces of policy implementation, was also attributed to the leadership of CBC:
“I think in the beginning Cascade Bicycle Club was a big [policy champion] that
supported and really got the funding to do the Plan…”
Part of CBC’s ability to position itself as a champion of the Plan lie in its political
ties, large constituency and increasing resources. One interview respondent, testifying
70
to Cascade influence and effectiveness in creating a political climate friendly to
bicyclists, remarked:
“… I know Cascade Bicycle Club was huge and lobbied to get certain people elected on
the advocacy level. I know people who ask “How is CBC voting” and that is how they
vote.”
This is more than simple anecdote- CBC is widely credited as having helped elect
Councilmember Mike O’Brien in his first run for elected office in 2009 and current
Mayor Mike McGinn. The organization’s influence has led candidates to seek its
endorsement and Club members have given thousands of dollars to its political-action
committee to elect pro-bicycle candidates. (Hefter, 2010b)
As the Club has transitioned from riding club to a professional advocacy
organization, its membership has tripled in under a decade from 3,700 members in 2001
to a current membership of over 13,000 in the Puget Sound area, and its influence
continues to grow. Boasting such a large and devoted constituency has earned the
organization monthly access to the Seattle City Council “bike caucus”, consisting of four
Councilmembers who hope to advance bicycle-friendly policies and prioritize funding for
bicycling initiatives. (Heffter, 2010) One staff member testified to this support as
crucial for its policy champion role in advancing implementation of the SBMP:
“…I think that this organization has been the most critical component on the
implementation side and it is because it is a major focus of ours and we have the
resources to do it…and we have the constituency we can leverage. Where they lead
we follow but we educate them on where to go.”
Responses readily lend themselves to placing Cascade Bicycle Club in the
category of being a critical influence to overall Plan implementation. While respondents
could clearly identify policy champions that were supportive of positive implementation
71
outcomes, what emerged from the interview responses, taken together, were strong
advocates and leaders at all levels of government, advocacy and research, affecting
change either as professionals or as citizens at the neighborhood level. While
neighborhood activists may not be defined as champions of the overall SBMP policy
implementation, their role in assisting in the implementation of individual projects was
seen as crucial:
“Every project has unique players involved… and you have some key citizens involved
with really promoting the projects. With implementation overall, it is generally “We
are doing this.” There is a lot of buy-in and I think when you get to every specific
project you have had someone who is very helpful and influential.”
An important factor emerged, however, in the need to not rely too heavily on
champions and what may be its corollary, the importance of knowledge and expertise
being institutionalized into processes. This requires at minimum, buy-in by
organizations (and therefore individuals), and a commitment to building resilience
within organizations to buffer the loss of any one champion. Noting the importance of
institutionalizing buy-in and ensuring a commitment throughout an organization, one
SDOT staff member remarked that:
“The SBMP is an accepted city document with high levels of accountability- so at all
levels of management there are huge levels of support. I am sure that every person
who works at SDOT has read the SBMP- it is pretty amazing. I think if you said
Transportation Strategic Plan or Comprehensive Plan there would be a lot of people
who have never heard of that.”
A board member echoed this sentiment in regard to SDOT:
“What I have observed in general is SDOT as an agency is very committed so it is not a
person but an agency to me that is even better than a single champion because we
have an agency that wants to accomplish this.”
72
One respondent who has assisted bicycle planning efforts in several capacities
discussed the need to build resilience into the implementing organization to buffer
inevitable changes that could unduly affect implementation efforts.
“I would say Seattle is kind of in the middle. Individuals still make a big difference, but
it is not institutionalized to the point where they don’t make a difference. I would also
say that there is a lot of depth here so that we can recover from one or two people
leaving pretty good usually. We’ve recovered with a new mayor, we have recovered
with a new director and things are going forward as much as ever….you have to build
depth in your organization. Then you will survive these ups and downs of being overly
dependent on one champion… you need to develop depth in your organization so that
it isn’t reliant on just a few people, because they move, they get married, they have
kids, life changes. And that will happen to everyone guaranteed.”
The results above suggest that, in the Seattle case, both Ison and Rye (2003) and
Gaffron (2003) are correct in identifying the role of policy champion as a critical
component of implementation success. Respondents were able to readily identify
champions at both the individual and organizational level and consistently characterized
the involvement as a supportive influence on implementation. Yet the respondents
were clear that institutionalization leads to greater resilience and is a critical component
to avoiding the risks inherent in relying on the leadership, motivation, and expertise of
the well-positioned few. The results do not lend themselves readily either way in regard
to Gaffron’s (2003) sentiment that the importance of a policy champion may indicate
that cycling policy implementation is not well embedded in existing structures and
cultures.
73
3.1.10 Support- Flexible attitude toward public reaction
“Seattle is known for its public process so that is the closest I can come to in terms of
ensuring that we had some flexibility…insofar that we can meet those needs and
address local concerns, the city through its public input process and through its
planning process, they go out and door hanger. When a new facility is coming to a
neighborhood, they’re going to every house, they hold an open house, they hold
meetings, they’ve got comment forms, and they have an online approach. They don’t
hide these things from folks.”
The above statement was typical of interview responses received regarding
flexibility in regard to public reaction. All five study participants agreed that flexibility to
public reaction was an integral element to the SBMP plan implementation process, in
terms of openness to new information brought about by local knowledge and the ability
to adjust projects accordingly. This was without exception related to Seattle’s public
process, which incorporates a high degree of public input through public meetings,
outreach and survey research. Part of the flexibility is being able to adapt the outreach
process commensurate with the complexity and potential impact of a given project:
“I would say there is a lot of flexibility, for each project there is a slight difference in
public outreach, there are standard thresholds for public projects, but if it is known to
be a challenging project in the neighborhood then you do a lot more public process, a
lot more meetings.”
This flexibility extended not just to the amount of public process the City goes
through for a given project, but into decision making that affects actual implementation
outcomes. One respondent outside the City notes that the flexibility is,
“…more in how it is carried out as opposed to some sort of a planned flexibility. What
I see in projects being implemented, especially ones that are major or controversial, is
that there is that effort being done to get that valuable information from people and
actually consider it, not just as a process. If it was just a process, then I would say here
is our checklist yes, we called a meeting, yes we invited reaction, yes we are going to
make the same decision, I haven’t seen that...A good example of that would be the
Capitol Hill Community Council saying “we really want this street car”…. the city was
74
very keen on that and very much aware of that and took it very seriously and actually
made a better decision. It seems to me that the flexibility is there even if it is not a
checkbox. They are actually doing it.”
City staff echoed the sentiment that the value of the adapting a project to local
knowledge may play a large part in of the quality of the project,
“When we go out to the community and we get feedback, it often does make a project
better…You get so much good feedback from people who live out there actually know
the road and it will make you think about something you didn’t consider and you can
tweak the design. That outreach and input is actually very valuable.”
Yet the flexibility toward public reaction was not without its setbacks in terms of
implementation. Participants relayed that, while leading to more successful outcomes,
flexibility towards public reaction can be burdensome and have a negative impact on
project implementation due to the workload involved and administrative needs. It is
not the value of the process that was problematic, but the degree:
“It would be fair to say that that process takes up tons of staff time where we could be
working on projects. Our own process hampers us from implementing. You can’t
double the number of lanes we do next year, but you can double the number we have
studied. We couldn’t do that next year, [the public process] is so time intensive.”
Delving further into the topic, the same respondent relayed that after a certain
threshold was reached, additional input from the public was unlikely to bring new
information to light that had yet to be considered.
“I just think that sometimes it is overboard. Could we have sufficed with 50
comments, or 100 comments, is 400 that much better? Most of those have very
similar trends. The people that are thoughtful and provide really good input about a
project, when you get into the 400 range you are not getting 400 of those comments.”
Separate from the burden on staff time, the process can also serve to hold up
individual projects as citizens contested the objective studies with local knowledge of
varying merit:
75
“Stone Way was a really good example. We did more study and public process on that
one then many other comparable projects just because it was a big deal….that was a
pretty straightforward project and we knew ahead of time that it would work, it had
standard numbers, there is things that clearly fit within a normal range, and there was
“were not quite sure if it will work or not”. This one was a slam dunker. But the
community wasn’t convinced that it would work, so we collected the numbers very
impartially and they showed what we knew they would … you don’t need to overly
test everything if you know what the outcome is.”
Yet the same respondent noted that even in cases where the outcome was good,
the incorporation of public comment has the ability to greatly hold up the project.
“Another one I worked on that really had a good outcome… a lot of neighborhood
issues, and other neighborhood cut through issues, and we just had to meet and meet
and meet. And even though it wasn’t that complicated of a project, it took us over 2
years to get it in because it was such an important connection.”
Given the views of study participants outlined above, Ison and Rye’s inclusion of
the need for a flexible and open attitude toward public reaction appears to be justified.
Yet there is a mixed influence on implementation. Overall, study participants believed
that local knowledge is taken into consideration throughout the implementation
process and has been shown to directly influence the final outcome of several projects.
While administrative burden associated with the public process was believed to hamper
outcomes, it is possible that this aspect is an indicator of a lack of systems capable of
handling what has become a routine part of the workload. One staff member discussed
how another city streamlined the public comment process through the use of a simple
online survey tool:
“We have to track every single comment and it is a lot of cutting and pasting into
spreadsheets. The City of Vancouver for example made it very clear as to how people
can comment and they had to actually fill out a survey on Survey Monkey- it does it for
you, it spits out graphs and in some ways analyzes the answers. You have people
entering their information and really getting something digestible out of it without
having the time commitment [of staff] to reenter it.”
76
It would seem appropriate for the implementation framework to consider an
agencies data needs, and corresponding systems in its analysis of whether adequate
resources exist. As the above example show, the need for timely data and routine
analysis that could just as easily be handled by databases and survey software can be a
drain on staff and expertise better spent working on other aspects of project
implementation.
Yet the need for public meetings, as discussed, still has the potential to
significantly slow down the more contentious or complex bicycle facilities projects.
Higher quality outcomes may be a product, as well as better coordination with
community and other agency needs while contributing to the public trust. The fact
remains, however, that the public process is inherently time-intensive and this
unfortunate by-product needs to be taken into account in the implementation theory.
For these reasons, flexible process for public opinion reaction should be considered as
having a mixed effect in terms of implementation: constricting in terms of resources
(staff and time) and beneficial in terms of data and creating public trust.
3.1.11 Support- Public Trust, Support and Opposition
Ison and Rye (2003) supplement Hogwood and Gunn’s original implementation
framework with the inclusion of public trust, which is considered to be a critical element
in the introduction of any new policy. Fundamental to this idea is that with the whole
process must be considered above reproach if it is to avoid widespread opposition.
77
Distrust in the implementing agency, it may be assumed, will hamper the
implementation process through the medium of public opposition if the program is seen
as benefitting special interests at the public’s expense, or if the public process is not
inclusive and transparent. The idea of public trust in regards to the SBMP incorporates a
sense of confidence in the ability of SDOT and the city of Seattle to take into account
public opinion and act faithfully and with full disclosure in executing their duties to the
benefit of the public overall.
Participants confirmed the support of constituency groups as either a critical or
very supportive factor influencing implementation of the plan and ultimately, affecting
SBMP goal outcomes:
“Anytime you have a greater constituency it leads to more political support for
implementing the plan, so we have seen ridership numbers go up quite a bit.”
One advocate well acquainted with the Plan deemed the role of public support
for projects as “key”, supporting the statement of a consultant familiar with bicycle
plans across the country that that, “if it wasn’t for public support, the plan wouldn’t
happen, it is as simple as that.” Responses were notable for their self-evident quality
characterized by little elaboration. Public support was simply seen as the backbone of a
strong implementation effort and the convergence of views lend credit to its
significance.
Four participants responded specifically to the term “public trust”, with three
agreeing that Seattle’s flexible and responsive public process played a supportive role in
engendering that trust. One respondent details his own experience as a Seattle Bicycle
Advisory Board member:
78
“When citizens come in [to SBAB meetings] and say there is a problem…we try to
include that information as best we can. The SBAB can only advise. We can’t make
decisions so we try to act as a conduit so the city can actually consider that feedback.
So from our view, really, we do as much as we can to make people feel that when they
come and talk- it is not just that they talk and it is over with. It is not just a venting
session, we actually try to do what we can to facilitate input from citizens. And just
from the people I have met from SDOT who have come to talk to us, I think they are
very sincere in taking information from citizens as best they can. I have never felt that
there was an unwillingness to seriously listen to what the public would like to be
considered”.
Following up on this sentiment, the same participant notes that a key element of
the process was its flexibility in responding to new data and public opinion, which may
be suggestive of a linkage between the creation of public trust and having a flexible
response to public opinion- another of Ison and Rye’s criteria covered above. Two
current city employees agreed with the board members opinion that the city doesn’t
just collect comments, but is responsive to public the public by incorporation comments
into actual decisions made:
“We do actually respond to everyone who comments and we take it really seriously.
Correspondence is huge and we do so much outreach and I think over time…at the
local level you have a lot of trust…people have worked with us and been through some
of these issues with SDOT and they have been involved in projects. It is just our whole
entire outreach.”
The public process not only has a receptive element through the incorporation of
public comment, but allows the City to engage the public proactively with information
that leads to greater transparency and trust, while dispelling myths that may be
perpetuated by opponents:
“It is really based on our structure, it is neighborhood based, it is just inherent in so
many of our practices. I would say that we approach each project differently in some
respects. I think using people who are well connected in the community, using
community councils, blogs and those sorts of outreach tools are really helpful, at least
for getting a constituency developed before the plan [implementation] really hits. And
79
also in messaging the project so neighbors are talking to neighbors and the word is
getting out about what your actual intentions are, that is really helpful.”
One conflicting view was offered by an advocate stating that public trust may be
a more critical component in the policy adoption, rather than implementation, phase:
“I don’t know that public trust component was as important or critical in plan
implementation. The public, or the engaged activist public, tends to be oppositional
on almost anything, change is bad…The public, insofar that they are voting to support
[the Plan and related measures] is conveying a certain amount of public trust, but I
think it is being driven largely by the stakeholders and the advocates in terms of
keeping implementation on track. I think there was public trust on the part of
allowing the city to do the plan…and the public trust in voting for the Bridging the Gap
streets levy, that we would meet the commitments set out in the financing package
the public trust.”
Responses by interview participants overall lends support in the Seattle case to
the notion that the creation of public trust is a supportive influence on plan
implementation.
The inverse of public support, public opposition, played a part as well, with three
participants responses centering around their view that the perception of bicyclists as
routinely breaking laws or obstructing traffic leads to complaints about bicycle facilities.
One respondent reported that the perception of reckless cyclists is,
“… one of the primary complaints, that bicyclists are an unusual other. Virtually
everywhere. And then the anecdotes, “Bicyclists don’t follow the law.” When you
pivot on that and say, “Well I was just at a red light the other day and, “Bam!”, the
light changed and three cars kept going through it. Should we ban cars from the
roads?” We need better compliance in terms of this public ambassador standpoint of
the small percentage of total trips, the 4% of total trips in the city of Seattle that are
done by bike.”
This response suggests that the real or perceived transgressions of city traffic
ordinances by the users of bicycle infrastructure can provide fodder for the opposition
to frame the user group as operating from a sense of entitlement. As the “unusual
80
other”, law abiding cyclists are in the sensitive position of being readily perceived as
infringing on long dominant modes of transportation (automobiles and buses), rather
than being incorporated into roadway users’ schemas about what is an acceptable and
safe form of transportation. Two respondents agreed that opposition to projects and
perceptions of disobeying traffic laws may be due to a lack of education regarding new
facilities types and cyclists rights to the road:
“Probably the only other factor that we didn’t talk a whole lot about, probably
because we are not involved, is education… I would say that one of the number one
letters we get has to do with rude or reckless cyclists… people have to understand why
a bicyclist would ride in a center lane. That is a whole different level of education. It is
trying to not get people so polarized one way or another from the motorist perspective
or the bicyclist perspective. You have these fringes that create huge amounts of angst.
Now if we have someone write to the Mayor about rude cycling behavior, we have this
standard response. There are so many people who think that the bicycle community
has a sense of entitlement.”
Echoing off this comment, an agency staff member laments the lack of education
that should have been provided concurrently with new projects unveiled during the
Short-Term Implementation Period:
“You really can’t expect to put a sharrow down on a street and expect people to know
what that means. Although now that they have been adopted at the national level,
you are seeing more and more of them. But I think that that is something we could
have done a better job on, but in the transportation department we don’t do
advertising all that well.”
The three participants who identified opposition due to the perception of cyclists
and infrastructure as impinging on existing roadway space were the most likely to have
to contend with opposition on a daily basis. As advocates and agency staff, they
identified these views as the number one complaint received from constituents.
Educating roadway users on how to interact in relation to new facility types and
81
increasing numbers of cyclists on the road was viewed by the study participants as
necessary to avoid confusion that may create backlash against future projects outlined
in the SBMP.
3.2 Factors Influencing the Decision to Bicycle
The previous section identified a broad range of contextual factors thought by
interview participants to be important for successful implementation of recommended
bicycle facilities in the SBMP.
An assessment of implementation only, however, would
fail to address whether the Plan is likely to have the intended effect of increasing bicycle
ridership. As the Bicycle Facilities Network is a core aspect of the Plan, it is reasonable
to expect that the costs and resources dedicated to providing infrastructure would be
justified by empirical evidence showing that increased access to bikeways is sufficient to
induce some people to ride their bicycle for utilitarian trips. Evidence that fails to
support this theory may ultimately prove to be problematic in terms of implementation,
as a policy based upon an inadequate understanding of a problem to be solved may
result in failure. In such a case, the underlying theory may be at fault rather than
execution of the policy (Hogwood and Gunn, 1984) and investing in improving bicycling
infrastructure may present an opportunity cost for more successful interventions,
ignoring the more complex interplay of factors that influence an individuals’ choice to
engage in non-motorized travel in favor of a policy prescription that is popular with
advocates, governments and current cyclists.
82
To answer the second study question regarding bicycle facilities effects on
ridership, a review of models in the literature that incorporate bicycle facilities as an
independent variable as an influence on ridership are presented. Analysis will be
conducted in terms of the policy implementation framework’s requirement for a valid
supporting theory with a direct causal relationship and few intervening links. It is also
necessary to critique the studies themselves, as the results may not be generalizable
outside the population of the study, measures may be problematic, or the results may
reflect response bias due to a low response rate or limitations of the study design.
As identified in the Introduction, a dramatic rise in facility provision and bicycle
transportation planning has taken place at all levels of government, with increases in
both the adoption of both BMPs and funding dedicated to facilities provision. Both
supporting and possibly stemming from the growth of these large public commitments
to cycling is a growing body of literature in both the public health and transportation
planning fields that focuses on the relative influences on an individual’s choice to use
the bicycle for both utilitarian and recreational purposes. An analysis of the literature is
provided to establish the strength of the available evidence to support the claim that
infrastructure investments, in the form of lanes and multi-purpose trails, will influence
on bicycle use.
The theory that creating bicycle facilities induces a change in behavior in noncyclists and cyclists alike must be supported by a valid theory of implementation if
outcomes are to be successful. In order to organize analysis of influences on human
behavior, researchers in the public health and physical activity fields have suggested the
83
use of ecological models of behavior to understand and identify targets for physical
activity programs and policies. Ecological models have increasingly been utilized by
researchers to gain a greater understanding of the relative influence of the social and
physical environment, and policies on physical activity (Pikora, et al. 2003). Proponents
of ecological models of behavior hold that environments restrict the range of a given
behavior by promoting and sometimes demanding certain actions and by discouraging
or prohibiting other behaviors; the implication is that environmental and policy
variables can add explanatory value above that provided by intrapersonal and
interpersonal factors (Sallis et al., 1998). This is not to say that environmental factors
are the only important variable to be considered- ecological models emphasize that
behaviors have multiple levels of influence and that a combination of these variables is
necessary to influence physical activity. (Saelens et al., 2003)
Ecological models have been used specifically in the literature regarding
influences on bicycle use, either stated explicitly or implicitly through frameworks that
are in many cases functionally identical. For example, Xing et al. (2008) employ an
ecological model in their cross-sectional analysis of the relative influence of individual
factors, social-environmental factors and physical-environmental factors on both bicycle
ownership and use in six small U.S cities in California. Explanatory variables are grouped
into these three categories, with individual factors being subdivided into both sociodemographic and attitudinal factors. T. Pikora et al. (2003), interview experts using a
Delphi panel study to inform an ecological model for understanding the potential
relative importance of environmental influences on walking and cycling. Claiming that
84
the evidence to date for the influence of the physical environment on physical activity is
sparse, the authors focus on developing four key themes that fall under the umbrella of
physical environmental factors: functional, safety, aesthetic, and destination. It should
be noted that the Delphi study did not examine the statistical associations of these
factors with increases in bicycling or walking. The purpose of the study was to create a
consensus of expert opinion regarding the relative potential importance of
environmental factors to aid in the development of an ecological framework. Rietveld
and Daniel (2004) likewise provide a general framework of factors that have a potential
impact on bicycle use as a basis for an econometric analysis of bicycle use in Dutch
municipalities. While not explicitly an ecological model, the framework shares nearly
identical categories such as individual features, socio-cultural factors, and local authority
initiative and policy variables that encompass what other models identify as
environmental factors.
The ecological model confirms that considering a wide array of factors is
necessary to determine if the provision of a network of bicycle facilities is likely to
significantly impact an individual’s choice to start cycling, and what other factors may
constrict or support preferred behavioral outcomes. Should another factor prove to
have a stronger influence, supporting a rival or alternative theory (and therefore policy
approach) may be necessary. For example, individual preference and demographic
factors, such as age, environmental concern, income and educational level, are in some
cases be found to be more important predictors in explaining bicycle use. Findings such
as these would suggest that increases in cycling may be attributed more to self85
selection, such as an influx of young, educated, and middle-class residents who already
bicycle and are attracted to a city’s amenities and pro-bike policies or marketing, rather
than access to infrastructure.
Using the ecological model as a general framework, the following sections will
use empirical research to provide evidence of the impact of bicycle infrastructure,
environmental, socio-demographic, and individual or attitudinal factors, and their
relationship to various measures of bicycle ridership.
3.2.1 Influence of Bicycle Lanes and Pathways on Bicycle Ridership
In what is widely considered the first study to test the relationship between facilities
provision and bicycle use, Nelson and Allen (1997) analyzed cross-sectional data of 18
U.S. cities and established that there is a significant positive correlation between the
number of bicycle pathway miles and the percentage of commuters using bicycles in
their journey-to-work. Each additional mile of bikeway per 100,000 residents was found
to be associated with a 0.075% increase in commuters using bicycles.
Building on the research of Nelson and Allen and lending strength to their
findings, Dill and Carr (2003) employ a regression model using new data to identify
factors associated with the percentage of workers commuting by bicycle in 33 U.S. cities
with populations over 250,000. The number of miles of Type II lanes per square mile
was found both significant and positive, with each additional mile of bike lane per
square mile associated with a roughly one percent increase in the share of workers
commuting by bicycle. As one potential explanation, also put forth by Nelson and
86
Allen, policy makers may be providing bicycle lanes in response to a large number of
cyclists in these cities.
Observing that from a policy perspective it is not clear what policy makers can do
to promote more bicycling, Rietveld and Daniel (2004) provide a quantitative analysis in
order to explain variations between municipal bicycle use for trips shorter than 7.5km.
An important aspect of this study is that it confines itself to only one country, the
Netherlands, where variations in bicycle use are less likely to be confounded by cultural
differences and more likely to result from variations in municipal policy. Route-related
environmental factors that had both significant and positive effect on bicycle use
included the speed of the trip (compared to the car), less stops per kilometer for a given
trip, fewer hindrances on a trip per kilometer. Safety, objectively measured through
victims of serious accidents, is an element that is also found to be important for
increasing bicycle use. Despite a negative correlation, it should be pointed out that this
study cannot attribute particular environmental interventions to promote cycling as the
cause of less serious accidents, only that accidents are negatively associated with
increased bicycle usage. Based on these findings, however, Rietveld and Daniel provide
support for the claim that both bicycle speed and convenience are essential elements in
promoting use. The authors conclude that the spatial design of networks that provide
direct routes with minimal stops for cyclists represents an effective policy that should be
used to increase bicycle use.
Parkin, et al. (2008) used aggregate data from the UK 2001 census to provide
evidence for the determinants of the choice to use a bicycle for work related trips.
87
Positive effects that were found to be significant in the model included the proportion
of the bicycle route that is off road, however, the elasticity was small, suggesting that
the creation of a large quantity of off-road facilities would only succeed in creating a
small increase in bicycle commuting. The proportion of route that has bicycle and bus
lanes did not have significant coefficient and was eliminated from the final model.
One study, often cited as evidence for the positive role infrastructure (bicycle
lanes, paths, and bicycle boulevards) may play in encouraging bicycling provides little
support for the conclusion that “infrastructure appears necessary to encourage bicycling
for everyday travel.” (Dill, 2009) Using GPS (global position system) trackers to monitor
a convenience sample of interested cyclists, the study provides evidence that the
distribution of bicycle travel differs significantly from that of the transportation
network. Utilitarian cycling travel during the study was 13% on bicycle/multi-use paths,
15% on secondary roads with bicycle lanes, and 10% on boulevards, compared to the
fact that these infrastructure categories represented 2%, 2% and 1% of the total
network, respectively. Overall, participants used bicycle infrastructure for about half
their travel, indicating that bicyclists are probably traveling out of their way to use these
facilities.
Several studies contradicted the findings that environmental influences were
positively associated with an individual’s choice to bicycle. Moudon, et al. (2005), in a
disaggregate cross-sectional study of urbanized King County, looked at the role of both
individual and environmental factors (both perceived and objectively measured,
through surveys and GIS respectively) thought to influence bicycling. The authors were
88
able to conclude that cycling mostly takes place irrespective of environmental prompts
or barriers, having more significant associations with personal factors. Objectively
measured route-related variables, such as the percentage of streets lined with bicycle
lanes, traffic speed and volume, number of vehicle lanes, topographical conditions and
street block size all were found to be insignificant in binary logit models. Only the
distance of an individual to the closest trail, measured both objectively and subjectively,
had a significant positive correlation with cycling.
de Geus et al. (2008) similarly found that the perceived environment was not a
significant predictor of cycling for transportation in areas with adequate cycling
infrastructure. Based on a self-reported survey of both cyclists and non-cyclists in
Flanders, Belgium (a cyclist being defined as cycling at least once a week to work in the
last 6 months), the results suggest that individual and social factors play a larger role
than environmental factors. The authors tested subjectively reported cycle lanes that
were present and in good condition, which were found to be insignificant.
Another study (Xing, et al. 2008) used cross-section survey data to identify the
relative importance of factors influencing both bicycle ownership and use while
controlling for the possibility of self-selection. Self-selection is an issue in correlational
studies where residents of a city choose to live in a particular area because it is
perceived as a supportive bicycling environment. This effect may erroneously attribute
increased bicycle use to physical or environmental interventions when in fact it should
be attributed to demographic changes in an area- people who already bicycle simply
moved to the area.
89
The influence of perceived infrastructure in the built environment showed no
significant effect on bicycle use or bicycling in the final model, including such facilities as
major streets with bike lanes, streets without bike lanes being wide enough to bike on,
well lit bike paths, networks of off-street bike paths, and bike lanes free of obstacles.
While there were significant associations of bicycle use and frequency with the
perception of the safety of bicycling to select destinations, this does not provide a basis
for the author’s suggestions that this may be an indirect role of infrastructure. No
evidence is provided as a basis for this assertion and other factors, such as lower traffic
intensity, may be responsible
Citing the field’s overreliance on cross-sectional studies as problematic for
establishing causality, Krizek et al. (2009) provide what may be the sole longitudinal
study to date. In investigating the impact of bicycle lanes established during the 1990’s
on subsequent bicycle use in Minneapolis and St. Paul, Minnesota, the authors are able
to conclude that bicycle facilities significantly impacted levels of bicycle commuting.
While a more nuanced treatment of the results is warranted, in general, the study found
that both for traffic analysis zones within defined buffers and outside of buffers
associated with bicycle facilities, increase in bicycle use took place during the period,
with areas inside the buffers showing a larger increase. The share of residents
commuting by bicycle living within defined buffers around individual facilities was
likewise found to be significant in almost all cases, except for near the University area,
where the initial rate was already very high. In some cases bicycle modal shares
doubled in areas with low bicycling, suggesting that there may be a diminishing marginal
90
return effect. Bridge improvements likewise were found to have significantly affected
commuters’ willingness to use bicycles to cross the Mississippi River, even though they
already had a relatively high bicycle mode share. Finally, the study found mixed results
in its analysis of the impact of facilities on final trip destination. As mentioned before,
model share for residents around the University of Minnesota did not change
significantly over the 10 year period, however, the share of trip did significantly increase
at this trip endpoint, suggesting that the facilities may have provided more benefit to
commuters coming from outside the area. Contrasted with this is the very slight
increase and decrease in final trip destinations in Minneapolis and St. Paul, respectively,
where as discussed earlier, residents in these areas were found to have increased their
share of cycling overall.
3.2.2 Influence of Socio-demographic Characteristics on Bicycle Ridership
A comprehensive analysis of factors influencing an individual’s choice to use a
bicycle for utilitarian trips would be inadequate if it only considered the influences
bicycle lanes and pathways. In keeping with the ecological model, socio-demographic
variables, such as age, sex, income, and related household characteristics have proven
to be significant and sometimes stronger correlates than environmental factors in
determining a person’s willingness to engage in non-motorized transport. Studies that
examine socio-demographic factors will be discussed in turn for their association with
increasing various measures of bicycle ridership.
91
3.2.2.a Gender
Studies reviewed consistently show that gender is significant predictor of cycling
rates, with males being much more likely to cycle then females in every study
considered. Through the use of data gleaned from the 2001 American Housing Survey,
Plaut (2005) was able to determine that for renters and homeowners identifying as
female, a negative correlation exists with the log of the probability of using a bicycle for
commuting divided by the relative to the probability of using a car. Gender was found
to be highly significant in two models that predict the number of times a week a person
bicycled for any reason, with males having odds ratios just over three times that of
women (Moudon 2006) . Cervero and Duncan (2003) similarly found a correlation
between male gender and the probability of choosing a bicycle for traveling. Troped et
al. (2001), found a correlation between the use of a community rail trail and respondent
identifying as male.
3.2.2.b Race
Race was found to be a strong predictor of cycling in several studies. Baltes’
(1996) aggregate study of 284 Metropolitan Statistical Areas found that the percentage
of the total population in an MSA that is Asian is a strong predictor of the percentage of
work trips taken by bicycle in 1990 in several regions. The percentage of the MSA that is
non-white was found to be negatively correlated with cycling trips, supporting one
study’s findings that non-white workers who both rent and own their homes were less
likely to use a bicycle for commuting rather than a car (Plaut, 2005) and another study
92
that showed that the percentage of non-whites were negatively associated with the
proportion of individuals in a ward cycling to work (Parkin, et al. 2008). A study of King
County, Washington, found that respondents identifying as White have odds ratios
between 3.5 to 5 times higher than non-whites in relation to the # of times a week the
individual chooses to bicycle for any reason (Moudon et al., 2005).
3.2.2.c Age
Age was found to be negatively correlated (Troped et al., 2001; Xing et al., 2008)
or a not significant predictor (Parkin, et al., 2008; Plaut, 2005) of various dependent
variables related to cycling across several studies. Rietveld and Daniel (2004) found a
correlation at the aggregate level between the proportion of young adults (ages 15-19)
in Dutch cities and the share of bicycle use. Moudon found a curvilinear relationship
showing that respondents in the age category of 25-45 as more likely to bicycle than the
youngest age category, 18-21.
3.2.2.d Education Level
Education level was strongly associated with the likelihood of bicycling across the
majority of academic studies considered. A study of five comparison communities in
California found a highly significant correlation between education level and whether
one chose to bike in the last seven days or not (Xing, et al., 2008) Both renters and
homeowners who were college graduates or have postgraduate schooling were found
to be more likely to choose a bicycle. (Plaut, 2005) An aggregate level study of
93
Metropolitan Statistical Areas (MSAs) found that when considering all MSAs in one
model, or stratifying MSA’s by region, the percentage of the population age 18-24
enrolled in school was significantly correlated with the percentage of work trips by
bicycle. Across several studies, education level and levels of enrollment can be shown
to have a high degree to association with a variety of cycling related variables.
3.2.2.e Income, Employment and Workforce Characteristics
The effect of income was found to be insignificant in one study looking at
whether one chose to bike or not bike in a given week, and how frequently (Xing et al.,
2008) This association appear again in an aggregate study comparing median family
income and its relation to work trips in metropolitan areas (Baltes, 1996). Baltes (1996),
in a model of Western U.S. MSA’s, was able to demonstrate that the percentage of
families living below the poverty level is negatively associated with the percentage of
work trips completed by bicycle. Parkin et al. (2008) similarly found that the index of
deprivation income score (a proxy of income) of both Welsh and English wards was
found to be negatively correlated to the proportion of individuals cycling to work-this
can be interpreted as the higher the rate of poverty, the less likely the individual was to
cycle. Plaut (2005) contradicted this effect in her study of commuting trends from the
American Housing Survey, finding that the salary of a worker was negatively associated
with the probability of using a bicycle over a car. While these results may suggest a
non-linear relationship in which those under poverty level and those with higher
94
incomes are less likely to bicycle, study results did not provide enough evidence to
support this conclusion.
Employment characteristics of individuals were also identified as a possible
influence on bicycle ridership, with two studies including employment related variables
in their final models. Moudon et al. (2005) found similar results in two models that
having less weekly work hours was positively associated with the number of times a
week a person bicycled for any reason. An aggregate study (Baltes, 1996) of all U.S.
MSA’s, stratified later into separate models by region, consistently found that the
percentage of the population that was unemployed was a strong and positive predictor
of the percentage of work trips by bicycle.
The percentage of various occupational categories represented in MSA’s were
found to be significantly associated with bicycling in several cases (Appendix C), with no
clear trend to explain why some are more likely to be associated with cycling outcomes
over others. The abundance of factors tested and the inconsistency of effects of factors
tested across studies do not readily lend support to the influences of income or
employment for the purposes of this study.
3.2.2.f Housing Characteristics
Plaut’s study (2005) of non-motorized community analyzed data from the 2001
American Housing Survey in order to see how commuting decisions are affected by, or
made jointly by, housing choices. The value of the owned units was associated with a
decrease in the probability of using a bicycle and the value of the rental payment was
95
associated with an increase in the chance of using a bicycle. Plaut found that the
number of persons in the household was insignificant in determining cycling outcomes,
while Xing et al. (2008) tested the same variable across several models, finding a
significant negative correlation affecting those who biked frequently as opposed to
moderately. In this study, housing size was found insignificant in determining who
bicycled and who didn’t. Finally, Plaut was able to predict a negative effect on choosing
a bicycle for transport with newer housing and rental properties. Many other aspects of
housing were found not to be significant predictors of choosing a bicycle over a car for
transport, including whether a parking space was available, if commercials properties
were nearby (i.e, a mixed-use neighborhood), the presence of green areas, and the
square foot/space of the unit.
3.2.3 Influence of Individual and Attitudinal Factors on Bicycle Ridership
The following section considers the influence of Individual and attitudinal factors on
bicycle use. Variables are represented that were included in models of bicycle use that
are associated with individual preference and are not due, primarily, to an individual’s
position in society.
3.2.3.a Bicycle Ownership
Bicycle ownership, as a prerequisite of bicycling and an indicator of interest in
cycling was positively and significantly associated with bicycling in the three studies that
included it as a factor. A study of 608 randomly selected adults in King County found
96
that ownership of a bicycle was significantly associated with the number of times
bicycling in a week, with two models resulting in odds ratios of 180:1 and 163:1
compared to those who don’t own a bicycle. (Moudon et al., 2005). Cevero and
Duncan’s (2003) study of two day travel activity for 15,066 randomly selected
households also found a positive correlation between the number of bicycles per
household and the probability of a person choosing bicycling for travel. Xing, et al.
Buehler’s (2008) study provides models on factors influencing who owns a bicycle based
on both individual and socio-demographic factors. Among those found to be significant,
those reporting that they are in good health and enjoy bicycling were more likely to own
a bicycle as opposed to those who reported that they need a car, like taking public
transit, and surprisingly, those that are pro-exercise. The results of these models,
however, should be cautiously interpreted as surveys received only had a 12.6%
response rate.
3.2.3.b Car Ownership and Attitudes Towards Motor Vehicles
The majority of studies reviewed found a highly significant negative relationship
between ownership of one or more motor vehicles and the likelihood of cycling.
Another recent study (Xing et al., 2008) found car ownership to be an insignificant
variable in the determining who bicycled and who didn’t bicycle, but found that it was
significant in determining who biked frequently vs. moderately. An aggregate study of
Dutch municipalities found that the number of cars per capita had a negative effect on
the share of bicycle use, mirroring a similar look in the U.S. by Baltes (1996) that found a
97
negative correlation when all MSA’s were considered on the percentage of bicycle trips
to work. Similar results were found in a study to determine the factors that determine
the probability of choosing a bicycle of using a motor vehicle, with both renters and
homeowners being more likely to do so if no car is available at the household (Plaut
2005). Plaut similarly found that owning homeowners with two or more cars had a
negative effect on bicycle choice, but the effect was not found to be significant for those
who were renters. Owning exactly one car per adult in King County Washington was
associated with a negative effect on the number of times a cyclist rode each week for
any reason (Moudon et al., 2005).
Attitudes towards motor vehicle ownership were of mixed importance for
determining the probability of cycling. Those that reported needing a car for
transportation were found in two models predicting the likelihood of whether one
bicycled or not in five California comparison cities were not found to be significant (Xing
et al., 2008). This variable did have a negative impact, however, on the likelihood of one
biking frequently as opposed to moderately, defined as five to seven days versus one to
four days a week.
3.2.3.d External Support
Only one study incorporated external support and reinforcement from friends
and family in its modeling of factors that influence cycling for transport to work at least
once a week (de Geus et al., 2008). While no other studies incorporated psychosocial
factors into their study, and thus they cannot be corroborated, the results of this study
98
of 343 Flemish adults discovered that those who reported high external self-efficacy,
having a partner or someone who cycles with them, and have someone who provides a
positive model of cycling behavior, were significantly more likely to have bike at least
once a week over the six months prior to the study. The results, while not conclusive,
show that a social element may be at play in influencing cycling rates, and that
increased cycling may in turn generate more cycling through positive modeling and
reinforcement.
3.2.3.e Attitudes Cycling, Transit and Walking
Perhaps unsurprisingly, a lack of interest in cycling significantly decreases the
probability of a respondent cycling at least once a week (de Geus et al., 2008). The
inverse was explored in another study (Xing et al., 2008) that integrated attitudinal
factors into its final models, finding that respondents who agreed strongly with the
statement “I like biking” were more likely to bicycle and to bicycle more frequently. The
same study found that survey recipients who favorably responded to the statement “I
like walking” were less likely to bicycle, but this was not a significant factor in regard to
the frequency with which the individuals cycled. Similarly, those who agreed strongly
with the statement “I like taking transit”, were more likely to bike moderately rather
than frequently, with no significant difference between those who bike and those who
do not. This could be seen as suggesting that walking and transit are competitive modes
with cycling, however, interpretation is difficult given the generality of the statements,
the lack of other studies that corroborate the relationship, and the low response rate of
99
the specific study in question. Complicating the analysis further, a study of King County,
Washington (Moudon et al., 2005) that incorporated transit users in two of its models
found them to be both insignificant (GIS Network Model) and positive (GIS Airline
Model) influences on the number of times a week the person cycled for any reason. As
can be seen, the mixed results of the few studies related to attitudes and behavior in
regard to alternatives to the automobile do not allow for a definitive answer on their
effect on bicycling rates.
Another attitudinal factor that was briefly explored in the Xing et al., (2008)
concerned the view that “Most bicyclists look like they are too poor to own a car”.
Responses in the affirmative were significantly correlated with the respondent not
bicycling. This influence on bicycling suggests that the unwarranted image of cycling as
a “fringe mode” may inhibit greater ridership numbers, no other study looked at this
variable and it cannot, therefore, be corroborated.
3.2.3.f Individuals’ Health/Attitudes Towards Exercise
An individual’s health, as a self-reported assessment, and attitudes about
exercise and the benefits of physical activity were found to be significantly related
across a variety of studies on cycling. Temporary illnesses/injuries were not found to
impact the use of a community rail trail, yet long term illnesses/injuries were found to
decrease the odds of using the trail to 0.43 compared to those that did not report an
illness or injury (Troped et al., 2001). Adult respondents in King County, Washington
(Moudon et al., 2005) who exercise at home were found more likely to bicycle more
100
during the week, mirroring an association in a study of five communities in California
that show that those who report being pro-exercise are more likely to have bike in the
previous seven days. (Xing et al., 2008). Contrary to these findings, de Gues, et al.
(2008) found no significant connection between self-reported physical well-being and
only a moderately significant influence of body image on the choice to cycle for
transportation at least once a week during the six months preceding the study.
Moudon’s study further contributes to an understanding of the influence of attitudes in
that those who strongly disagree with the benefit of physical activity, or reported no
vigorous physical activity, were less likely to bicycle during the week for either
commuting or recreation. Overall, the results of the four studies above support a
positive role for attitudes about health in regard to a variety of bicycle related
outcomes.
3.2.3.g Environmental/Economic Awareness
Two studies considered the impact of attitudes towards the environment in the
decision to bicycle at least once during the week. Environmental concern did not factor
into the final models of one study that considered bicycle ownership, bicycle riders, and
the frequency of which bicycles are ridden. (Xing et al., 2008) Ecological and economic
awareness was found by de Geus et al. (2008) as a significant predictor of biking to work
during the week, with an odds ratio of 1.71 compared to those who didn’t cycle.
Xing et al. (2008) also found a significant correlation between those who try to limit
driving as much as possible and those who cycle. Across models, those who self101
reported constraining their driving were found to be strongly associated with those who
biked versus those who didn’t, and those who bicycled frequently versus those who did
so only moderately.
3.2.4 Influence of Environmental Factors on Bicycle Ridership
The results of models that tested the influence of environmental factors are
explored below, with several significant effects on bicycle use present. Environmental
factors include variables related to geography, climate and urban design, including
population density, traffic density, neighborhood characteristics, and weather, among
others.
3.2.4.a High Traffic Densities/Presence of Automobiles
High traffic densities in neighborhoods and the presence of automobiles were
factors considered in five models reviewed for this study, both objectively and
subjectively measured. Troped et al. (2001) found that survey respondents were more
likely to use a community bikeway when they self-reported that no busy street barrier
existed en route to the bikeway. An objective assessment showing a busy street barrier
did not prove to be significant however, suggesting that what constitutes a busy street
and risky conditions are more likely to be a matter of perception. Another predictive
model looking at factors that influence the proportion of individuals in an area that cycle
to work found that the transport demand intensity, defined as the number of
102
employees for the area divided by road length, negatively influenced the outcome.
Other studies that looked at motorized traffic noise (Rietveld and Daniel, 2004), traffic
danger and traffic safety in neighborhoods (de Geus et al., 2008) found no significant
association.
One King County, Washington study (Moudon et al., 2005) relating
bicycle rates to the presence of motor vehicles and motor vehicle facilities found not a
negative influence, but a curvilinear one. The non-linear association suggested to the
authors that having a moderate level of traffic and auto-oriented facilities is more
desirable for cycling, compared to having too few or too much of them. The explanation
offered was that these conditions may offer a diversity of activities of interest to cyclists
in the form of sensory or visual stimuli.
3.2.4.b Weather Related Variables
Rainfall measured as number of days during the year rain exceeding 1/10th of an
inch for 18 U.S. cities and rainfall per year in millimeters in British wards were found to
have significantly negative effect on cycling when measured at the aggregate level in
two models (Nelson and Allen, 1997; Parkin et al., 2008). This same effect was only
found to be significant in one model that measured days of rain for 33 U.S. cities, but
only when D.C. and NYC were removed as outliers (Dill and Carr, 2003). A study of the
effect of rainfall as measured in millimeters on the cycling rates of Dutch cities was
found to be insignificant. The authors do note that there is less spatial variation for
precipitation across the Netherlands and it tends to rain evenly over the whole country.
103
Wind speed was a variable in two studies reviewed (Rietveld and Daniel, 2004;
Parkin et al., Wardman, and Page 2008) thought to negatively influence cycling by
increasing the effort the cyclist must make to ride against it. Both studies did not find
the variable to be significant in explaining cycling levels in the areas of study and it did
not make it into final models.
While studies considered other factors such as mean temperature and the total
annual hours of sunshine for the year in order to detect an outcome on cycling, weather
variables have not been explored to any great extent in the literature. From the above
review, we can conclude with reasonable confidence that rainfall will have an inhibiting
effect on cycling outcomes.
3.2.4.c Geographical Factors: City Population and Density
Plaut (2005), in her study of U.S. Metropolitan Statistical Areas (MSA), was able
to show a strong association between those that live in central cities of the MSA on the
West Coast of the U.S., and the log of the probability of using a bicycle over the
probability of using a motor vehicle. In line with this, a negative correlation was
detected for those who lived in rural areas and those who lived in secondary urban
areas in a given MSA. Baltes (1996) study of U.S. MSA’s found that the percentage of
population living in central cities was negatively associated with the percentage of work
trips made by bicycle.
Population density was found to significantly influence the proportion of
individuals cycling to work in a study of British wards (Parkin et al., 2008), but was not
104
significant in a model using 284 MSAs in the U.S. An increase in the population (in
thousands) was found to be associated with a decrease in the share of bicycle use in
Dutch cities, but the impact of total population was not measured in any other study
reviewed.
3.2.4.d Neighborhood Level Characteristics: Trip Origins
The influence of urban design, land-use diversity and density patterns on the
choice to use a bicycle was explored by Cervero and Duncan (2003) using a combination
of data on the built-environment in the nine county San Francisco Bay Area and travel
survey data detailing two days of household trip information for over 7000 households.
To deal with the issue of multicollinearity that will cause interrelated built environment
variables to contaminate the final predictive model, the authors use factor analysis to
extract two variables: a pedestrian/bike friendly factor (small blocks, four way
intersections, and five or more way intersections) for both trip origins and trip
destinations and a land use diversity factor, again for trip origins and trip destinations.
The land-use diversity factor, a measure of mixed land uses, employed residents to
retail/services balance, residential balance, and other related factors, was found to have
a moderately significant (probability 0.088) effect on the choice to cycle when
considered as a destination. Overall, the study found that demographic characteristics
of trip makers were far stronger predictors of bicycling choice than built-environment
factors. Moudon et al. (2005) found that the presence of convenience stores had a
negative impact on cycling in both models tested when captured as total parcel areas
105
rather than count. The authors suggest that larger convenience stores tend to be those
combined with gas stations, which may be explain their role as a possible detractor of
cycling.
Access to transit within a five minute walk from home was found to be an
important determinant in two models that predicted who bicycled in the previous seven
days, but it was not a relevant factor in a model determining the frequency of which
they biked (Xing et al., 2008). de Geus et al. (2008) found that the time predicted to go
to a bus, tram or metro stop was not associated with cycling for transport.
3.2.4.e Trip-specific Variables: Distance and Destinations
A variety of variables specific to individual trips was presented in the literature
with two consistent themes: trip distance and trip purpose.
Trip distance was found to be negatively correlated with the proportion of
individuals cycling to work in British wards, with an increasing coefficient distance to
work increased in the variable ranges of 2km-5km to 5km-20km (Parkin et al., 2008).
Another study (Cervero and Duncan, 2003), using disaggregate travel data on individual
trip characteristics in its modeling, found a similar negative result when comparing trip
distance in miles to the probability of a person choosing a bicycle for traveling in the San
Francisco Bay Area.
Trip purpose was also considered by Cervero and Duncan (2003) in their study of
how land-use diversity dimensions of built environments affects walking and cycling.
Recreational/ entertainment purposes and social purposes were found to have a
106
positive effect on cycling, in addition to the presence of food shops at the destination.
Weekend trips and those that involved shopping were found to be positive, but not
significant predictors, of traveling by bicycle. Xing et al. (2008) found that the average
self-reported rating of safety to various destinations such as grocery stores, post offices,
restaurants and elementary schools was found to significantly and positively correlated
in models that tested the factor against who biked and how frequently.
3.2.4.f Slope
Slope (rise/run) was found to be a significant factor influencing the use of a
bikeway in Massachussett’s, with survey respondents nearly twice as likely (Odds ratio
1.90) to use a trail when no steep hill barrier was objectively present (Troped et al.,
2001). Subjective reporting of a steep hill on the route to the bikeway, however, was not
found to be significant. The proportion of 1km squares with slopes 3% or steeper was
found be associated with a lower proportion of individuals cycling to work in British
wards (Parkin, Wardman, and Page, 2008), and the presence of slopes was found to be a
negative influence on the share of bicycles use in relatively flat Dutch cities (Rietveld
and Daniel, 2004). A study of the hilly San Francisco Bay Area contradicted these results
by finding that objectively measured average slope, calculated off of trip origins in
destinations, was not significant. Overall, studies reviewed support an expected
negative or insignificant effect of hilly areas on cycling.
107
3.2.4.g Road quality
Only two of the studies reviewed considered the effect of road conditions on
cycling, with both being aggregate measures that are not specific to the experiences of
individual cyclists and non-cyclists. One study of Dutch municipalities found that the
effect of pavement vibrations was either insignificant and had issues of multicollinearity
with other factors, and did not make it into the final study model (Rietveld and Daniel,
2004). The proportion of principal and non-principal roads with negative residual life
were both found to negatively affect the proportion of individuals in British wards who
cycled to work (Parkin et al., 2008). Xing et al. (2008) found that respondents’
assessments of the average comfort of biking on various street classifications was not a
significant predictor of who did or did not cycle, but these same perceptions do have a
significant positive effect on who does and does not own a bicycle. With only two
studies showing significant relationship for a subjective and objective measure of road
quality, the effect of this factor is presently indeterminable.
3.2.5 Influence of Policy level Variables on Bicycle Ridership
A broad category that could be termed “policy level” variables, was the focus of
the (2004) Rietveld and Daniel study, “Determinants of bicycle use: do municipal policies
matter?” The authors were able to determine that parking costs, measured in
eurocents per hour, had a positive effect on the aggregate share of bicycle use in Dutch
cities. The speed of cycling when compared with the car was found to be a positive
influence, with stop frequency for cyclists being negative. These results suggest that
108
one element that needs to be considered in making cycling a more attractive mode is
that it is able to compete favorably with other modes- in this example high parking costs
and a low stop frequency can be expected to make motor vehicles less appealing and
bicycles more so. As Rietveld and Daniel state, “this combination of push and pull
policies is a rather general result found in transportation research, and it also appears to
apply to bicycle use.” The only other finding of significance was related to an aggregate
measure of the degree of satisfaction in the study areas with bicycle policies, provisions,
etc. This had a significant positive influence on cycling as a larger portion of the modal
share, and may be somewhat corroborated with Dill and Carr’s (2003) model showing
that state spending per capita on bicycle and pedestrian improvements was significantly
related to the bicycle rates in 33 U.S. cities (when D.C. and NYC, are excluded).
The limited studies available that consider policies, coupled with the wide variety
of proxy measures that are used as indicators of municipal incentives and supportive
policies for cycling, proves problematic in terms of assessing the importance of variables
at this level. While there may not be agreement on what is relevant to measure, the
significant relationship between cycling outcomes and the variables presented should
be noted.
3.3 Assessment of Models in Relation to Implementation Framework
The ability to prove a causal relationship between bicycle facilities and increased
levels of bicycling has remained elusive in much of the existing research. Four of the
recent studies reviewed, (Nelson and Allen (1997); Dill and Carr (2003); Rietveld and
109
Daniel (2004); Parkin et al.,(2008)) often held up as providing evidence for the role of
lanes in inducing bike ridership, use cross-sectional or aggregate models that test factor
associations with the dependent variable at a single point in time. Three of the four
studies found statistically significant relationships between various measures of bicycle
facilities, with Rietveld and Daniel (2004) finding no significant relationship to a bike
network, and Parkin et al. (2008) only finding a relationship for off-road facilities (such
as bike trails). What may seem like favorable results overall is limited when taking into
account the inherent deficiencies of aggregate modeling- foremost is the inability to
prove causality in favor of statistically significant correlations that may be spurious or
misleading. One major issue identified in the literature is that of self-selection, in which
a correlation masks a causal relationship that move opposite of the intended direction.
In such a case, residents may have based their locational decisions in response to a
supportive bicycling environment and in turn create more demand for bicycle lanes.
This is separate from true causality, where bicycle lanes induce more to ride by making
utilitarian cycling more accessible to those who previously wouldn’t have considered it.
Disaggregate studies that rely on surveys and the subjective measures of the
presence of bicycle lanes are also present in the literature. These studies may prove
problematic in that there is no way of assessing the extent to which actual bicycle lanes
are present. Xing et al. (2008) found no significant relationship when respondents were
asked their perception on the presence of bicycle lanes, but were found to have a
marginally statistically significant (p=0.10) relationship when aggregate objective miles
bicycle lanes per square mile were modeled. These results do not pass the causality test
110
for the same reason elucidated above regarding aggregate studies. de Geus et al. (2008)
similarly used the presence of subjectively reported or perceived bicycle lanes and
found no significance.
Two studies attempted to remedy the problematic nature of self-reporting in
surveys supplemented respondents answers with objectively reported disaggregate
variables matched through GIS to the respondent’s location. Moudon et al. (2005)
found that the subjectively reported presence of bicycle lanes showed a highly
significant positive relationship with the number of times cycling a week for any reason,
while an objective measure of the percentage of streets lined with bicycle lanes, which
were not significant. Objective distance to the closest trail, however, did show a
positive relationship with bicycling. Troped et al. (2001) used a similar hybrid design
using a mixture of self-reported and objectively measured variables gathered through a
GIS tool, finding that distance to the nearest trail was negatively associated with the use
of a community rail trail- the closer the distance the person lived, the more likely they
were to make use of it.
Only one longitudinal study was identified that analyzed the effect of bicycle
facilities on mode share over time, with the explicit purpose of attempting to address
causality given the predominance of cross-sectional/correlational studies (Krizek et al.,
2009). Using aggregate U.S. Census data, the study identifies specific bicycle facilities
that were installed in Minneapolis and St. Paul during the study period of 1990-2000,
and test for effects using two different buffering techniques to define proximity. While
self-selection remains an issue, and the authors caution against any interpretation
111
inferring causality, the results are suggestive of the significant positive impact of bicycle
facilities on bicycle commuting. While these results may seem particularly promising
given the longitudinal study design, it is still appropriate to be cautious when inferring
causality. Yet the authors are able to reasonably conclude that self-selection is not the
case- a prior study conducted by Barnes and Krizek (2005) allows them to conclude that
even if all the “right type” of persons (affluent, male, between the ages of 18 and 40)
lived in one area and were absent from another, the difference in rates should only
differ by a factor of two. The differing rates of bicycle use between 1990 and 2000
were, however, many times that amount. While still falling short of showing a causal
relationship, these results are promising nonetheless for the positive role bicycle
facilities may plan in increasing ridership.
The foregoing discussion establishes a predominately positive correlation
between bicycle lanes and bicycle ridership across several models, with a few
exceptions showing no significance. Providing a study design that demonstrates a
causal relationship has been elusive given the reliance on cross-sectional study designs
and aggregate data. From the evidence provided, the overall relationship between
bicycle facilities and increased ridership can best be defined as a statistically significant
positive correlational relationship.
Hogwood and Gunn’s implementation framework requires that the relationship
between cause and effect be direct, as too many linkages will provide additional
opportunities for a break in the chain. The use of ecological models of behavior was
determined in the above review of literature to be well justified given the strong
112
associations many socio-demographic and attitudinal factors have on bicycle ridership.
The significance of these factors, however, may be problematic if they serve to mitigate
the effects of bicycle lanes. Overall, models found a significant correlation between the
presence of a high percentage of white, young, college educated males living in urban
areas that own a bicycle and are inclined to limit driving, and a higher proportion of
bicycle trips. Rainfall, road quality, and hills, on the other hand, may cancel out the
effects of bicycle lanes. Parkin, et al. (2008), Moudon, et al. (2005) , and de Geus et al.
(2008) all conclude that socio-demographic or individual factors played a stronger role
in predicting bicycle mode shares or the number of times a bicycle was ridden per week
than physical factors, and Xing et al. (2008) and Rietveld and Daniel (2004) do not show
significance for the effect of bicycle infrastructure, in comparison to many attitudinal,
socio-demographic and other environmental factors with significant and at time larger
coefficients. While it could be said that a direct causal linkage is to be taken as an ideal,
the factors that influence increases in bicycle modal share are numerous, correlational,
and at times contradictory. The evidence based off the literature presented clearly
shows that a policy intervention aimed at increasing bicycling modal share will be
impacted by numerous factors outside of its control.
The inability to show causality coupled with a large number of factors that may
serve to mitigate the potential impact of bicycle facilities may prove problematic for
meeting the SBMP goal of tripling ridership. Absent the provision of a well-accepted
model of bicycle use, predicting the outcome of the provision of bicycle infrastructure
will continue to prove difficult. Seattle precipitous climate and hilly landscape may
113
serve to mitigate the effects of bicycle lanes, while demographic factors may prove
favorable, adding an element of self-selection that increases ridership. While further
studies are required, the theory underlying the SBMP shows limited strength in
comparison to the framework’s ideal requirements for direct causality, and should at
best be interpreted as a significant positive correlational relationship.
4. Analysis and Conclusions
The study concludes by considering the implications of the study findings for
other cities engaged in Bicycle Master Plan implementation based off of the experiences
of implementation staff in the Seattle case. Refinements to the implementation
framework are offered based on evidence provided by interview participants that
contradicted the frameworks premises of elements required for ideal policy execution.
The limitations of the study findings are then addressed in light of the study design, and
suggestions are made for future research to assist in the development of an empirically
based understanding of the actual potential to increase ridership through the use of
bicycle facilities.
The use of the implementation framework devised by Hogwood and Gunn was
successful in assisting interview respondents in the identification of major factors
influencing the construction of bicycle facilities in Seattle, Washington over the 20072009 study period. Overall, the interview instrument employed was successful in
identifying factors of varying influence on Bicycle Facilities Network implementation
efforts. All interview questions based on Charles’ iteration of the framework generated
114
responses, with five framework categories (Context, Resourcing, Compliance,
Leadership, and Support, see Table 1 above) of eight having converged on responses for
one or more factors. These findings justify the use of the framework in both the current
and future case studies of BMP implementation.
Influences on Plan execution were found to be critical, supportive, neutral, or
problematic for implementation outcomes, yet no factors were determined to be
barriers to implementation of bicycle facilities in the Seattle case. This finding is not
surprising as respondents identified that enough funding was provided for during the
2007-2009 period and all projects prioritized were completed during this time. It is
reasonable to believe that factors exist that may serve to hold projects up, or the
absence of factors that were deemed critical to implementation in the current study
may serve to stall projects indefinitely. While these factors were not identified in the
Seattle case, future case studies of cities that have not been successful in meeting
recommended facilities goals may assist in a more comprehensive understanding of
factors that can hamper projects and lead to implementation failure.
The Bridging the Gap transportation package, a voter approved levy, was
determined by participants to be the most crucial factor in implementation of the
Seattle Bicycle Master Plan. Perhaps its two most defining attributes are that the levy
runs concurrently with the Plan and it mandates providing a percentage of funds to nonmotorized transportation projects. Just as notable was the passage of a self-imposed
levy, which indicates a wide degree of popular support beyond the control of any one
politician or group to block. Findings from the Seattle case suggests that dedicated
115
funding, in whatever form, is a necessary component of BMPs, the absence of which will
likely obstruct any implementation efforts.
In response to Hogwood and Gunn’s inclusion of the need for the ability to
enforce compliance in the face of opposition to a policy, interviews revealed a more
fundamental factor at play. The missing ingredient of political will was identified by
study participants, and its presence should be considered crucial for implementation
when dealing with opposition to projects the agency wishes to pursue in line with the
SBMP. The authority to enforce compliance takes place in the context of a pluralistic
system of governance where power shifts depend on the political will of elected officials
and the relative support of constituency groups. A well organized opposition or a Mayor
or City Council less amenable to championing the Plan and lacking the will to act on
agreed upon objectives could severely hamper the implementation process, all within
the same legal context. With this qualification in mind, it can be concluded that in the
Seattle case, SDOT has the ability to enforce compliance with the Plan due only to the
political will of actors such as the Mayor and City Council, themselves subject to the
political support of their constituents. This finding suggests that case studies should
consider whether the legal authority exists to follow through with the Plan, and if
political leaders that can exercise such authority are likely to support implementation
efforts in the face of adversity.
A convergence in opinions exists among study respondents that the adoption of
a Complete Streets policy was a critical piece of the policy framework necessary for
implementing the SBMP. The passage of the Complete Streets was the product of years
116
of advocacy and work with elected officials. Its adoption, first as a City Resolution, and
later as an Ordinance, could be seen as suggestive of a confluence of factors more
fundamental to the creation of a policy that mandates a divergence from a focus on
infrastructure provision primarily for the benefit of motor vehicles. While it is outside
the scope of this study to determine what these factors are, the adoption and
subsequent passage of the Complete Streets Ordinance by unanimous votes of City
Council suggests broad and sustained public support over a four year period, coupled
with an advocacy boasting the expertise, resources, and influence to affect the opinions
of elected officials. Its defining characteristic is institutionalizing the consideration of
bicycle facilities when any transportation projects are considered, ensuring that players
in the planning process do not push an underserved mode into the margins. The
possibility remains that cities may accomplish the goals of routine accommodation of
projects in a BMP without an explicit “Complete Streets” policy- cases studies should
consider the level of institutionalization of bicycle planning efforts in the absence of a
Complete Streets ordinance.
The support of constituency groups and the public was considered a final critical
factor that positively influences implementation. The inclusion of this factor in the
critical category is not surprising, and can be assumed to have an effect across the board
in many different aspects of implementation- notable are the passage of the Bridging
the Gap levy, the Complete Streets Ordinance, and as a foundational force in the
creation of political will among elected officials, all of which were identified above as
critical factors in implementation.
117
A clear policy champion was present in the form of Cascade Bicycle Club, whose
advocacy efforts were considered highly supportive of implementation efforts. By
rallying support around the Bridging the Gap levy, the Complete Streets Policy, and
creating political will, again, all found to be critical factors for implementation in this
study, Cascade Bicycle Club used the strength of its membership, its expertise, and its
connections with City staff to influence processes to provide a conducive environment
for the Plan to become adopted and realized on the ground. While the importance of
institutionalizing bicycle planning efforts in existing processes should be of paramount
importance, the Seattle case suggests that a well-positioned advocacy group can have
significant sway over events that will ultimately benefit implementation efforts overall.
Future research should consider the role that bicycle advocacy groups play in creating
support for Plan adoption, the technical assistance they provide to city staff and the
maintenance of support during actual policy execution.
The question of timing, defined by the author as contextual events that
influenced implementation that were not willed by policy actors, was found to influence
project implementation in the case of increased gasoline prices in the summer of 2008.
Interestingly, respondents discussed the influence in terms of increasing ridership,
which may have the indirect effect of increasing support for the SBMP as a greater
constituency has a stake in successful Plan outcomes. This relationship could not be
determined, yet the ability for events to affect the mood and subsequent support for
projects suggests a linkage between the implementation frameworks requirement for
timing and political support. The influence of timing may prove difficult to establish in
118
future case studies of bicycle facilities provision as the concept is broad enough to apply
to any contextual event that occurred during the scope of the study. The concept could
benefit from better operationalization more specific to bicycle use to reduce the
possibility of it being a catch all for factors not specifically outlined in the framework.
Environmental concern, either at the attitudinal level that may influence bicycle
use and possibly support for Plan implementation, or as embodied in adopted City
documents such as the Seattle Climate Action Plan, were determined to have no direct
effect on implementation outcomes. Interview responses do suggest an important
indirect effect in that environmental strategy documents may serve to frame the need
for bicycle facilities in terms of larger overarching values, thus building upon latent
constituencies of support. Despite the finding of a neutral effect, future research into
BMP facilities implementation should consider this effect not just for environmental
strategies and values, but also those related to public health, energy reduction, and
providing options for low-income individuals.
A lack of streets space, or the presence of competing uses that confined the
optimal project option, was considered a problematic factor influencing
implementation. This could result in the overdevelopment of one type of facility,
depending on a city’s geographical characteristics. The SBMP states that “Depending
upon an individual bicyclist’s level of experience, some types of bikeways are preferred
over others…new bicyclists tend to prefer off-road multi-purpose trails and quiet
neighborhood streets. More experienced bicyclists prefer on-road bicycle facilities such
as bike lanes, wide curb lanes, paved shoulders, etc.” (SBMP, 2007) As discussed earlier,
119
sharrows were widely considered confusing and not an ideal facility type for beginning
bicyclists in discussions during SBAB meetings. Attracting only those comfortable
sharing a lane with motor vehicles may ultimately serve to hamper implementation
outcomes by not providing the necessary environment to induce new cyclists to ride.
Responses suggest that the constraining effect of street space on project
implementation may be mitigated by an effort to obtain more detailed data regarding
constraints during plan development, considering both data on street characteristics
and potential conflicts with competing uses. The issue becomes less one of a lack of
street space and more one of having accurate data to reduce unseen conflicts later on
during implementation. The ability to get high quality front-end data will inevitably be
constrained by the plan development budget, time, and staff availability, but may allow
for smoother project provision by increasing awareness of issues inherent to contested
spaces, ultimately allowing for adjustment before the plan is adopted. Future research
should consider the extent to which high quality data is gathered at the front end and is
consistent with realities present at the time of actual implementation.
Funding of capital projects (specifically bridges and trails) was considered
problematic during the Short-Term Implementation Period primary due to the high
expense associated with their construction. While speculative, these large projects may
act as a potential barrier to full implementation of the Plan if resources are not
forthcoming, leaving critical gaps in the Bicycle Facilities Network. The Seattle case
demonstrates that these projects should be considered separately in order to avoid
funding issues where the construction of a larger project presents an opportunity costs
120
for developing on street bicycle facilities. The identification of this factor points to the
appropriateness of considering individual facility types and projects as the units of
analyses appropriate for case study research into factors influencing implementation.
The construction of large projects is likely to face a different set of barriers compared to
on-street bicycle lanes due to their size and complexity, and should receive a separate
treatment to ensure that important differences are not glossed over.
Responses regarding the importance of a single implementing largely
contradicted the requirement of Hogwood and Gunn that successful implementation
depends on the ability of an agency to operate in an environment with minimal decision
points outside of its control. Study findings regarding Ison and Rye’s (2003) requirement
for a flexible public process found a similar effect in the opposite direction, showing that
the need for public comment could serve to hamper project provision in the short-term.
Common to both of these was that the need to coordinate with other stakeholders and
the public initially slowed projects down, sometimes for extended periods of time, but
the end results was thought to be better implemented projects that generated public
trust in city processes, reducing opposition and creating buy-in as a result. Hogwood
and Gunn’s largely top-down approach for a single implementing agency does not take
into account the long-term benefits associated with coordination with multiple interests
that may allow for piggybacking on other projects, the avoidance of inefficiencies
between departments or groups, and increased data gathering that may impact the
quality of projects implemented. Interview responses suggested that streamlining data
gathering and the public process through survey tools, databases, and requirements for
121
coordinated stakeholder meetings may mitigate any detrimental effects on project
execution.
Assessing the effects of political stability on implementation outcomes proves
difficult in the Seattle case given the consistent political support for the Plan by elected
officials. Participants readily identified that the political support was important to
implementation outcomes, but stability could not be assessed as participants noted that
there were only minor changes at the city level and they were not able to assess the
importance of the effect on overall Plan efforts. While it is likely that political stability
played a part in Seattle’s successful execution of recommended projects in the SBMP,
future research efforts are needed to determine the impact of a change in executive
branch leadership or City Council vote composition on BMP facilities implementation. A
comparison of cities that experienced significant leadership changes to those that have
experienced relative stability may provide evidence for the effect of political stability on
facilities provision as envisioned in a BMP.
The role of data in Plan implementation outcomes was ambiguous, as Ison and
Rye’s inclusion of the requirement for monitoring outcomes provided little guidance on
what aspects of data to consider. The purpose of the interview question was to
determine whether the performance-based measures currently recommended for
assessing Plan implementation progress, or some other data, allowed for feedback that
assisted in developing bicycle facilities. The performance measure in the SBMP
monitors progress through a metric of bicycle facilities produced, with studies of
individual facilities outcomes being produced ad hoc, largely in response to public
122
opposition. While it is important to monitor overall progress in creating new facilities
and meeting Plan goals, future case studies could benefit from considering data
collection and monitoring more specific to bicycle planning that will allow for feedback
and course correction. Aspects of data found to be important by implementing staff
that should be specifically considered include creating a scheme to prioritize a list of
recommended projects, and the impact of metrics on project decisions. It may also be
appropriate to consider an agencies data needs and corresponding systems capacity in
its analysis of whether adequate resources exist. As respondents identified in the
Seattle case, the need for timely data and routine analysis that could just as easily be
handled by databases and survey software can be a drain on staff and expertise better
spent working on other aspects of project implementation. The consideration of
technology in the resources category could be an important addition the existing policy
implementation framework.
Several limitations to the study conclusions must be taken into account that are
directly related to the research design employed: a single case study analysis of data
from qualitative interviews of key participants. Foremost is that the study results are
necessarily limited by the single case research design, which is analogous to a single
experiment. The single case study design was necessary due to the scope of the
research project and resources available, and did not meet any of Yin’s (2003) criteria to
justify the appropriateness for the method, specifically that the case be critical, unique,
representative, revelatory or longitudinal. The possibility remains that contextual
events specific to the City of Seattle either before or during the study period are
123
responsible for implementation outcomes, or that an interplay of several factors were
necessary to produce the influence identified by study participants. Absent multiple
case studies that replicate the same study methodology for another city, it is impossible
to determine the extent to which the study results are determined by such an effect.
The choice of an implementation framework may also have driven the results in
important ways, as the categories in the framework influenced the construction of the
interview instrument and ultimately the contextual factors that participants focused on.
This does not invalidate the responses of interview participants, who were free to
identify factors and contradict or support the assumptions of the framework. The issue
remains, however, that factors specific to transportation planning or bicycling that were
not accounted for by the framework remain unidentified, and that a different
implementation framework and interview instrument may achieve different results.
The final interview question was intended to mitigate this possibility by allowing for the
respondent to consider any other factors thought to be important for implementation
outcomes. The author believes that overall the framework employed was successful in
identifying the major influences on SBMP implementation, with many of the responses
for the final non-structured interview question being readily categorized into the
existing implementation framework.
The choice of key participants in SBMP implementation was appropriate for the
research questions and study design, as responses needed to reflect the expertise and
insight of those most closely involved in the process. Yet this reliance on key
participants with a stake in the Plan’s ultimate success creates the possibility that
124
responses are biased toward participants’ downplaying aspects of the Plan that do not
meet expectations, or reflect professional assumptions that may not be borne out by
evidence. The need to keep interview participants identity’s confidential may again
mitigate this effect to some extent by allowing for an honest assessment of the
obstacles and shortcoming affecting Plan implementation without fear of contradicting
the attitudes of their peer groups. The corroboration of factors identified serves to
strengthen the study results by adding weight to the inherently subjective and
anecdotal nature of interview responses. Future case studies that incorporate individual
projects as the unit of analysis may be better suited for identifying specific groups
outside implementing staff that can corroborate or refute accounts based on their
unique perspectives.
Finally, the unit of analysis chosen, projects recommended in the SBMP as a part
of the Bicycle Facilities Network, may obscure important differences in implementation
between facilities types. Responses from study participants may be seen as supportive
of this view in that some bicycle facilities, such as sharrows, were not as constrained by
factors such as space, parking issues, conflicts with existing uses, or the need to go
through an extended public process. Capital projects such as bridges or multi-user trails
that are more complicated and expensive may be more problematic to implement in
terms of expertise, funding, opposition, and political will.
The identification of a wide range of factors influencing implementation of the
SBMP provides a starting point for analyzing other cases to provide support for the
findings of this study. Future research should consider a multiple case study design
125
stratified by project type across a range of cities of similar size and characteristics to
allow for a more robust theory on the influences on bicycle facilities implementation. A
more detailed look at individual projects can identify specific mechanisms of support
and opposition, as well as barriers that affect various facilities types, bringing to light
important differences that may further assist localities in identifying opportunities to
provide a more supportive environment for Plan and project execution.
In regard to influences on bicycle use, researchers should focus on study designs
that can provide additional support for the theory that the creation of bicycle facilities
will have a positive effect on ridership. Longitudinal studies of specific facilities that can
control for demographic changes are particularly well suited for this endeavor given that
they can suggest the direction of relationship and not just correlation. Such research
will assist advocates in making the case for increasing cycling should the data support
the creation of facilities to increase cycling, or allow for a better allocation of scarce
resources should findings show that self-selection is the driver of bicycle ridership. The
incorporation of ecological models that take into account attitudinal and sociodemographic factors will be crucial in providing evidence for the relationship, with
greater consistency in metrics and definitions across studies adding to the strength of
the findings.
Finally, a clear role is established for researchers in providing evidence that can
aid municipalities in managing the risk associated with developing a BMP. Generating
more successful BMPs through a front-end assessment of the strengths and weaknesses
of a cities’ ability to affect a positive implementation outcome will provide positive
126
modeling to other localities hesitant to commit to BMP development. The results
suggest that positive steps can be taken before adoption of a BMP to increase the
chances of successful implementation. Advocacy groups are well positioned to mobilize
supporters and lobby local officials to establish a supportive policy framework and
dedicated funding source well in advance of the policy implementation period. Only
through a combination of political will and the development of effective strategies
founded on verifiable theories of bicycle use and implementation can increases in
bicycling be expected in the coming years, supplanting short trips by automobile to
make a small, but necessary impact on climate change.
127
References
Alliance for Biking & Walking. (2010). Bicycling and Walking in the United States: 2010
Benchmarking Report. Washington D.C.: Alliance for Biking & Walking.
Baltes, M. (1996). Factors Influencing Nondiscretionary Work Trips by Bicycle
Determined from 1990 U.S. Census Metropolitan Statistical Area Data. Transportation
Research Record, 1538, 96-101.
Barnes, G., and Krizek, K.J. (2005). Estimating Bicycling Demand. Transportation
Research Record. 1939, Transportation Research Board Washington D.C., 45-51.
Bikes Belong Coalition. (2008). Bikes Belong Survey Report: Impact of Gas Prices on Bike
Sales. Bikes Belong Coalition Website. Retrieved June 12, 2011, from
http://www.bikesbelong.org/assets/documents/uploads/Bikes_Belong_Survey_Report(
2).pdf
Cervero, R. and Duncan, M. (2003). Walking, Bicycling, and Urban Landscapes: Evidence
from the San Francisco Bay Area. American Journal of Public Health, 93(9), 1478-1483.
Charles, P. (2005). Effective Implementation of Regional Transport Strategy: Traffic
Incident Management Case Study.” Urban Transport, 77, 609-618.
City of Seattle. (2006). Climate Action Plan. Retrieved June 12, 2011, from:
http://www.seattle.gov/archive/climate/docs/SeaCAP_summary.pdf
City of Seattle. (2007). Complete Streets Ordinance. Ordinance Number 122386.
Seattle City Clerk’s Online Information Resources. Retrieved June 12, 2011, from
http://clerk.ci.seattle.wa.us/~scripts/nph-brs.exe?d=CBOR&s1=115861.cbn.
&Sect6=HITOFF&l=20&p=1&u=/~public/cbor2.htm&r=1&f=G
de Geus, B., et al. (2008). Psychosocial and Environmental Factors Associated with
Cycling for Transport Among a Working Population. Health Education Research, 23(4),
697-708.
Dill, J. (2009). Bicycling for Transportation and Health: The Role of Infrastructure.
Journal of Public Health Policy, 30, S95-S110.
128
Dill, J., and Carr, T. (2003). Bicycle Commuting and Facilities in Major U.S. Cities: If You
Build Them, Commuters Will Use Them- Another Look. Transportation Research Record.
11828, Transportation Research Record, Washington, D.C., 116-123.
Federal Highway Administration. (2009). National Household Travel Survey. National
Household Travel Survey Website. Retrieved June 12, 2011 from http://nhts.ornl.gov/
Federal Highway Administration (2010). National Bicycling and Walking Study FifteenYear Status Report. Pedestrian and Bicycling Information Center. Retrieved June 12,
2011, from http://www.walkinginfo.org/ 15_year_report/
Gaffron, P. (2003). The Implementation of Walking and Cycling Policies in British Local
Authorities. Transport Policy, 10, 235-244.
Garrard, J., et al. (2008). Promoting Transportation Cycling for Women: The Role of
Bicycle Infrastructure. Preventive Medicine, 46, 55-59.
Heffter, E. (2010, April 20). Seattle City Council forms “Bike Caucus”. The Seattle Times.
Retrieved June 12th, 2011, from http://seattletimes.nwsource.com/
html/politicsnorthwest/2011653455_seattle_city_council_forms_bik.html
Heffter, E. (2010b, October 28). Politics, Friction Reshape Influential Cascade Bicycle
Club. The Seattle Times. Retrieved June 12th, 2011, from http://
seattletimes.nwsource.com/ html/localnews/2013287738_bikelobby29m.html
Hogwood, B. W. and Gunn, L.A. (1984). Policy Analysis for the Real World. New York:
Oxford University Press.
Hurdle, J. (2008, June 12). U.S. Cities Promote Bicycling as Gas Prices Soar. Reuters.
Retrieved June 12, 2011, from http://www.reuters.com/article/2008/06/12/us-usabicycles-idUSN1245324520080612?pageNumber=1
Ison, S and Rye, T. (2003). Lessons from Travel Planning and Road User Charging for
Policy Making: Through Imperfection to Implementation. Transport Policy, 10, 223-233.
Krizek, K.J., et al. (2009). Analyzing the Effect of Commute Mode Share Over Time.
Journal of Urban Planning and Development. Journal of Urban Planning and
Development, 135(2), 66-73.
Lagerwey, P. (2009). Creating a Roadmap for Producing and Implementing a Bicycle
Master Plan. The National Center for Bicycling and Walking. Retrieved June 12, 2011
from http:// www.bikewalk.org/pdfs/ BMP_RoadMap.pdf
129
Lavelle, Marianne. (2009). The Climate Lobby’s Nonstop Growth. Center for Public
Integrity. Retrieved June 12, 2011, from http://www.publicintegrity.org/
investigations/climate_change/articles/entry/1376/
Lindblom, M. (2007, November 11). Ambitious Bicycle Plan Wins City Council OK. The
Seattle Times. Retrieved June 12th, 2011 from http://seattletimes.nwsource.com/
html/localnews/2003995545_webbikeplan05m.html
Lindblom, M. (2010, May 26). Nickerson “Road Diet” Gives Councilman Rasmussen
Indigestion. The Seattle Times. Retrieved June 12th, 2011, from http://seattletimes.
nwsource.com/html/politicsnorthwest/2011961920_nickerson_road_diet_gives_coun.h
tml
Litman, T., et al. (2009). Pedestrian and Bicycling Planning Guide to Best Practices.
Victoria, Transport Policy Institute. Retrieved June 12, 2011, from
http://www.vtpi.org/nmtguide.doc
Merom, D., et al. (2003). An Environmental Intervention to Promote Walking and
Cycling- The Impact of a Newly Constructed Rail Trail in Western Sydney. Preventive
Medicine, 36, 235-242.
Moudon, A.V., et al. (2005). Cycling and the Built Environment, a US Perspective.
Transp. Res., Part D, 10, 245-261.
Nankervis, M. (1999). The Effect of Weather and Climate on Bicycle Commuting.
Transp. Res., Part A, 33, 417-431.
Nelson, A. C., and Allen, D.P. (1997). If You Build Them, Commuters Will Use Them.
Transportation Research Record. 1578, Transportation Research Record, Washington ,
D.C. 79-83.
Parkin, J., et al. (2008). Estimation of the Determinants of Bicycle Mode Share for the
Journey to Work Using Census Data. Transportation 35.1, 93-109.
Pikora, T., et al. (2003). Developing a Framework for Assessment of the Environmental
Determinants of Walking and Cycling. Soc. Sci. Med., 56, 1693-1703.
Plaut, P. (2005). Non-motorized Commuting in the U.S. Trans. Res., Part D, 10, 347-356.
Pucher, J., et al. (1999). Bicycling Renaissance in North America? Recent Trends and
Alternative Policies to Promote Bicycling. Transp. Res., Part A: Policy Pract., 33, 625654.
130
Pucher, J and Dijkstra, L. (2003). Promoting Safe Walking and Cycling to Improve Public
Health: Lessons from the Netherlands and Germany. American Journal of Public Health,
93:9, 1509-1516.
Pugliese, A. and Ray, J. (2009). Awareness of Climate Change and Threat Vary by Region.
Gallup. Retrieved June 12, 2011, from http://www.gallup.com/poll/124652/ awarenessclimate-change-threat-vary-region.aspx
Raderstong, J. (2008, July 6). More Commuters Pedaling to Work. The Seattle Times.
Retrieved June 12th, 2011, from http://seattletimes.nwsource.com/html/localnews/
2008035449_bikes06m0.html
Rietveld, P., and Daniel, V. (2004). Determinants of Bicycle Use: Do Municipal Polices
Matter? Transp. Res., Part A: Policy Pract. 38, 531-550.
Sabatier, P.A. and Mazmanian, D. (1979). The Conditions of Effective Implementation: A
Guide to Accomplishing Policy Objectives. Policy Analysis, 5, 481-504.
Saelens, B.E., et al. (2003). Environmental Correlates of Walking and Cycling: Findings
From the Transportation, Urban Design, and Planning Literatures. Annals of Behavioral
Medicine, 25(2), 80-91.
Sallis, J.F., et al. (1998). Environmental and Policy Interventions to Promote Physical
Activity. American Journal of Preventive Medicine, 15(4), 379-397.
Seattle Bicycle Advisory Board. (2007-2009). October 2007-December 2009 Meeting
Minutes of the Seattle Bicycle Advisory Board. Seattle Bicycle Advisory Board Website.
Retrieved June 12, 2011, from http://www.seattle.gov/sbab/minutes.htm
Seattle Department of Transportation. (2007). Seattle Bicycle Master Plan. Seattle
Department of Transportation Website. Retrieved June 12, 2011, from
http://www.cityofseattle.net/transportation/bikemaster.htm
Seattle Department of Transportation. (2009). Nickerson Rechannelization Facts and
Answers. Seattle Department of Transportation Website. Retrieved June 12, 2011, from
http://www.seattle.gov/transportation/ docs/NickersonRechannelizationFAQ.pdf
Seattle Department of Transportation. (2010). Bridging the Gap 3-year Annual Report .
Seattle Department of Transportation Website Retrieved June 12, 2011, from
http://www.cityofseattle.net/Transportation/ docs/btg/BTG3YearAnnualReportLETTER.pdf
131
Troped, P., et al. (2001). Associations Between Self-Reported and Objective Physical
Environmental Factors and Use of a Community Rail-Trail.” Preventive Medicine, 32,
191-200.
United Nations Educational, Scientific and Cultural Organization. (2009). UNESCO
Organizes International Conference on Broadcast Media and Climate Change. UNESCO
Website. Retrieved June 12, 2011, from http://portal.unesco.org/ci/en/ev.phpURL_ID=28769&URL_DO=DO_TOPIC&URL_SECTION=201.html
U.S. Census Bureau. (2010). U.S. 2010 Census. U.S Census Bureau Website. Retrieved
June 12, 2011, from http://2010.census.gov/2010census/data/
U.S. Energy Information Administration. (2009). Emissions of Greenhouse Gases in the
United States- 2008. U.S. Energy Information Administration Website. Retrieved June 12,
2011, from ftp://ftp.eia.doe.gov/pub/oiaf/ 1605/cdrom/pdf/ggrpt/057308.pdf
U.S. Energy Information Administration. (2011). West Coast Retail Gasoline Historical
Prices. U.S. Energy Information Administration Website. Retrieved June 12, 2011, from
http://www.eia.gov/oog/ftparea/wogirs/xls/pswrgvwrwc.xls
Xing, Y., et al. (2008). Factors Associated With Bicycle Ownership and Use: A Study of 6
Small U.S. Cities.” Presented at the 2008 Annual Meeting of the Transportation
Research Board.
Yin, R. (2003). Case Study Research Design and Methods. Thousand Oaks: Sage
Publications.
132
Appendices
Appendix A. Background: Goals and objectives of the Seattle Bicycle
Master Plan
The Seattle Bicycle Master Plan (SBMP) consists of two main goals that encompass
all activities of the City of Seattle related to bicycling, providing a basis for the Plan’s
recommendations. The first of these broad goals is to triple the share of bicycling in
Seattle for all trip purposes during the period of 2007-2017. The second goal is to
increase the overall safety of bicyclists in the City, with the aim of reducing the rate of
bicycle crashes by one third during the same period (2007-2017).
In service of the goals of increasing bicycle trips and safety in Seattle, specific action
items are organized under four broad objectives identified by the city that establish the
framework for the rest of the SBMP. These include:
•
•
•
•
Objective #1: Develop and maintain a safe, connected, and attractive network
of bicycle facilities throughout the city.
Objective #2: Provide supporting facilities to make bicycle transportation more
convenient.
Objective #3: Identify partners to provide bicycle education, enforcement, and
encouragement programs.
Objective #4- Secure funding and implement bicycle improvements.
It is with Objectives #1 and Objectives #4 that this paper is most concerned. The
effort to create a network of connected bicycle facilities in Seattle is central to
understanding the context of a local environmental intervention to increase cycling and
is therefore most relevant to the scope of a case study on implementation. Second,
Objectives #2 and #3, while important aspects of the Plan, are not directly related to the
scope of this paper and will therefore not be considered.
133
A diverse array of recommended facilities is provided for in the Seattle Bicycle
Master Plan’s Bicycle Facility Network (BFN), with individual treatments depending on
both environmental level factors (land use characteristics, available right of way space,
traffic volume, etc.) and individual level factors (rider preferences and level of
experience). The Plan cites the lack of facilities on arterial streets as preventing more
people from considering bicycling as a viable transportation option, and suggests that
improvements on these routes, which tend to be characterized by gentle grades, will
make conditions more comfortable and conducive to bicycle travel (Action 1.1).
Following this, nearly one-third (32%, see Table 5 below) of the total recommended
bicycle facilities are directed toward the creation of on-street bicycle and climbing lanes
on core arterial streets that offer the most direct routes to workplaces, shopping areas,
schools and transit hubs. Additional on-road facilities also contribute to the BFN,
including shared lane markings, paved shoulders, and shared bus-bike lanes, bringing
the total designated on-road bicycle facilities recommended to 295 miles of arterial
roadway, representing roughly two-thirds (65%) or all recommended facility miles.
Separate rights-of way from motorized traffic (i.e., off-road bicycle facilities and
recreational multi-use trails) represent another infrastructure component of the
Network objective (Action 1.2), with completion of The Urban Trails System, adopted in
the 2005 Seattle Department of Transportation (SDOT) Transportation Strategic Plan,
being a key action item. Several gaps in the existing Urban Trails System are identified
as requiring completion if a continuous and connected BFN is to be realized.
134
Table 5. Mile of Recommended Facilities
Miles of Recommended Facilities
Short-term
2007-2009
(includes
Total 2007-2016
Facility Type
Existing
existing)
(includes existing)
Bicycle lanes/climbing lanes
25.5
63.7
143.3
Shared lane pavement markings
0.3
54.2
110.5
Bicycle boulevards
0
7.6
18.1
Other on-road bicycle facilities
2.2
4.2
46.1
Signed local street connections
0
28.6
75.9
Multi-use trails
39.4
41.9
58.2
Other off-road bicycle facilities
0.2
1
2.6
TOTAL NETWORK
67.6 miles
201.2 miles
454.7 miles
Source: Seattle Bicycle Master Plan pg. 16 (SDOT, 2007)
Additional action items that serve to supplement or facilitate implementation of
bicycle facilities involve making operation improvements to complete connections
(Action 1.6) in the BFN and improving complex corridors and focus areas (areas with
right of way constraints, potential conflicts between multiple user groups, and multiple
alternatives for providing bicycle facilities, etc.) (Action 1.5). The need to improve
complex corridors is purposefully vague in the Plan, which may result from the fact that
a specific treatment needed for a given site will depend on a variety of factors. The Plan
cites the need to consider public input, trade-offs among other user groups, additional
design development, cost, and future opportunities as important considerations to take
into account in these highly contentious areas.
135
Appendix B. Seattle Bicycle Master Plan Thesis Interview Questions
1.) In what ways was the policy framework surrounding the Plan supportive/constricting
of projects recommended as a part of the Bicycle Facilities Network?
2.) How did the timing of the program influence implementation? (By timing, I am
referring to the larger contextual events from 2007-2009, not intentionally willed by
policy actors, which impacted plan implementation. For example, changes in economic
climate, demand for bicycle infrastructure due to increasing environmental
consciousness etc.?)
3.) Has political stability/change at the municipal/county/state level been important in
contributing to implementation success or failure? In what ways?
4.) What role did resources (staff, time, funding, expertise etc.) play in facilitating or
hampering the implementation of bicycle infrastructure projects?
5.) Have the performance-based measures currently recommended for assessing plan
implementation progress allowed for feedback that assisted in developing bicycle
facilities? If not, has other data accomplished this? Or, what data would have been
useful?
6.) What organizations/division/groups are required to be consulted before
implementation of a project can occur?
•
Overall, does the necessity of dealing with multiple agencies hamper the process
or does a multi-agency approach lead to more successful outcomes?
7.) What is the capacity of any one actor involved in the process (the agency, Mayor’s
office, City Council, etc.) to demand compliance if substantial opposition exists to a
recommended project?
•
•
If there is the capacity, has it been used?
What have been the consequences in cases where the opposition has been
overridden?
8.) In regard to the Bicycle Facilities Network, has there been agreement among groups
involved in how objectives of the Plan are to be achieved?
•
•
If yes, what allowed these conditions to persist?
If no, did it hinder facilities implementation, or was it an appropriate response to
changing conditions?
136
9.) Do one or more agreed upon policy champion(s) exist?
•
•
What is his or her role in influencing the BFN implementation process?
If no champion exists, can the respondent identify times when one would have
been useful?
10.) How have organizations involved with plan implementation ensured a flexible
attitude to public reaction?
•
If they have, has this hindered the plan in cases where there was strong
opposition or obstructionists? (This question is more about the institutionalized
processes that have been incorporated into the implementation process for
reacting to constituents, not about how constituents feel)
11.) What role did public trust (a transparent and inclusive process) and the support of
constituency groups play in Plan implementation? Was it significant enough to move
individual projects along more effectively than if it wasn’t there?
12.) During the course of your involvement with the Plan, what other factors not
covered earlier do you believe were important in determining both successful and
unsuccessful implementation outcomes?
137
Appendix C. Predictive Models in the Literature:
Attitudinal, Socio-demographic, Environmental and Policy Level Factors
Associated with Bicycle Use
positive and
significant
negative and
signficant
not applicable or
not signficant
+
< 0.05 signficance
-
0.05<p< 0.10 significance
n/a
ns
not significant
Sex
Independent
Model name
Variable
Bicycle-Choice Model
for predicting that a
Cervero and Duncan trip will be made by
Gender
(2003)
bicycle
Study
Associati
on
+
Coefficient
Moudon, et al (2005) Airline Model
gender (male)
+
Moudon, et al (2005) Network Model
Model of variation in
Parkin, Wardman,
use of the bicycle for
Page (2008)
England and Wales
gender (male)
Proportion of
employees who
are male
+
0.588
3.124 (odds
ratio)
3.196 (odds
ratio)
+
2.8284
Significa
Dependent variable
nce
probability of person n
choosing bicycle for travelling
between origin and
0.002 destination
# of times a week bicycling
0.01
for any reason
# of times a week bicycling
0.01
for any reason
(t-stat)
12.59
Plaut (2005)
Cyclists to work who
Dummy for female
own their home
-
-1.012
Plaut (2005)
Cyclists to work who
Dummy for female
rent their home
-
-1.768
proportion of individuals in
ward x cycling to work
The logit or log of the
probability of using biccyle
23.57 commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
38.05 commutting divided by the
Wald chi- probability of using car
square commuting.
Troped, et al (2001)
Self reported env.
Variable model
Sex (male)
+
(Odds Ratio)
1.91
pg 197 for
95% CI Use of a community rail trail
Troped, et al (2001)
GIS environmental
variable model
sex (male)
+
(Odds Ratio)
1.99
pg 197 for
95% CI Use of a community rail trail
138
Race
Independent
Associati
Variable
on
% of population
Baltes, M. (1996)
All MSAs
+
that is Asian
% of population
Baltes, M. (1996)
North Central
+
that is Asian
% of population
Baltes, M. (1996)
South
+
that is Asian
Not significant in any % of population
Baltes, M. (1996)
n/a
model
that is Black
% of population
Baltes, M. (1996)
North Central
that is non-White
% of population
Not significant in any that is of Hispanic
Baltes, M. (1996)
n/a
model
origin
% of population
between ages 16
Baltes, M. (1996)
Northeast
+
and 29
Bicycle-Choice Model
for predicting that a
Cervero and Duncan trip will be made by
African American
(2003)
bicycle
Study
Model name
Coefficient
0.098
0.489
0.288
/
-0.173
Significa
Dependent variable
nce
% of work trips in 1990 by
0.05
bicycle in each MSA
% of work trips in 1990 by
0.01
bicycle in each MSA
% of work trips in 1990 by
0.01
bicycle in each MSA
% of work trips in 1990 by
ns
bicycle in each MSA
% of work trips in 1990 by
0.05
bicycle in each MSA
/
ns
0.757
0.01
0.071
Moudon, et al (2005) Airline Model
race (white)
+
Moudon, et al (2005) Network Model
Model of variation in
Parkin, Wardman,
use of the bicycle for
Page (2008)
England and Wales
race (white)
Proportion of
population nonwhite
+
0.854
4.938 (odds
ratio)
3.626 (odds
ratio)
-
-1.1708
-
-0.703
Dummy if worker
is Non-White
-
-0.576
Proportion of nonnative residents
-
-0.625
n/a
/
n/a
/
ns
n/a
/
ns
Plaut (2005)
Plaut (2005)
Rietveld and Daniel
(2004)
Cyclists to work who Dummy if worker
own their home
is Non-White
Cyclists to work who
rent their home
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
white
not bike)
Model 3 (biked
frequently vs.
white
moderately)
Xin, Handy and
Buehler (2008)
Model 5 (Biked vs. did
White
not bike)
Xin, Handy and
Buehler (2008)
0.05
0.1
% of work trips in 1990 by
bicycle in each MSA
% of work trips in 1990 by
bicycle in each MSA
probability of person n
choosing mode I for travelling
between origin and
destination
# of times a week bicycling
for any reason
# of times a week bicycling
for any reason
(t-stat) - proportion of individuals in
11.93 ward x cycling to work
The logit or log of the
probability of using biccyle
4.98
commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
4.91
commutting divided by the
Wald chi- probability of using car
square commuting.
(t-value) - Share of bicycle use in
1.91
Dutch cities
Biked within last seven days
(vs. did not bike within the
ns
last seven days)
Biked frequently (vs. biked
moderately)
Biked within last seven days
(vs. did not bike within the
last seven days)
139
Age
Study
Model name
Moudon, et al (2005) Airline Model
Independent
Variable
Age
Moudon, et al (2005) Network Model
Parkin, Wardman,
Page (2008)
Age
Proportion of
employees in the
bands "16-24",
Model of variation in "25-34", "35-49',
use of the bicycle for "50-59", "60-64"
England and Wales
and "65-74"
Plaut (2005)
Cyclists to work who
Age
own their home
Rietveld and Daniel
(2004)
Cyclists to work who
rent their home
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Troped, et al (2001)
GIS environmental
variable model
Troped, et al (2001)
Xin, Handy and
Buehler (2008)
Self reported env.
Variable model
Model 1 (own a bike
vs. not own a bike)
Plaut (2005)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
not bike)
Model 3 (biked
frequently vs.
moderately)
Model 4 (own a bike
vs. not own a bike)
Model 5 (Biked vs. did
not bike)
Associati
on
(curviline
ar, see
page
254)
(curviline
ar, see
page
254)
Coefficient
Significa
Dependent variable
nce
?
# of times a week bicycling
for any reason
?
# of times a week bicycling
for any reason
proportion of individuals in
ward x cycling to work
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
commuting.
n/a
/
ns
n/a
/
ns
n/a
/
ns
Proportion of
young (15-19)
years
+
4.19
age
-
(Odds Ratio)
0.71
pg 197 for
95% CI Use of a community rail trail
Age (10 year
increase)
-
(Odds Ratio)
0.67
pg 197 for
95% CI Use of a community rail trail
Age
-
-0.029
0.01
age
-
-0.02
0.01
age
n/a
/
ns
Age
-
-0.029
0.01
Age
-
-0.019
0.05
age
(t-value) - Share of bicycle use in
2.10
Dutch cities
Owns a bike
Biked within last seven days
(vs. did not bike within the
last seven days)
Biked frequently (vs. biked
moderately)
Owns a bike (vs. not own a
bike)
Biked in previous 7 days (vs.
did not bike)
140
Marital Status
Moudon, et al (2005) Airline Model
martial status (vs.
never married)
+
Moudon, et al (2005) Network Model
martial status (vs.
never married)
n/a
2.519 (odds
ratio)
# of times a week bicycling
0.1
for any reason
not
significan # of times a week bicycling
for any reason
t
141
Education Level
Study
Xin, Handy and
Buehler (2008)
Troped, et al (2001)
Model name
Model 1 (own a bike
vs. not own a bike)
Self reported env.
Variable model
Troped, et al (2001)
GIS environmental
variable model
Xin, Handy and
Buehler (2008)
Independent
Variable
Associati
on
Coefficient
Education level
n/a
/
Educational level
n/a
/
educational level
+
(Odds Ratio)
2.19
educational level
+
0.231
educational level
n/a
/
Educational Level
n/a
/
Educational level
+
0.198
Significa
Dependent variable
nce
Owns a bike (vs. not own a
ns
bike)
ns
Use of a community rail trail
pg 197 for
95% CI Use of a community rail trail
Biked within last seven days
(vs. did not bike within the
0.01
last seven days)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
not bike)
Model 3 (biked
frequently vs.
moderately)
Model 4 (own a bike
vs. not own a bike)
Model 5 (Biked vs. did
not bike)
Plaut (2005)
Cyclists to work who
College graduate
own their home
n/a
/
Plaut (2005)
Cyclists to work who Dummy for
rent their home
college graduate
+
0.955
Plaut (2005)
Dummy if have
postgraduate
Cyclists to work who schooling beyond
own their home
BA
+
1.108
+
1.4
+
0.415
0.01
% of work trips in 1990 by
bicycle in each MSA
+
0.807
0.01
% of work trips in 1990 by
bicycle in each MSA
+
0.386
0.01
% of work trips in 1990 by
bicycle in each MSA
+
0.676
+
0.071
n/a
/
Plaut (2005)
Baltes, M. (1996)
Baltes, M. (1996)
Baltes, M. (1996)
Baltes, M. (1996)
Nelson and Allen
(1997)
Baltes, M. (1996)
Dummy if have
postgraduate
Cyclists to work who schooling beyond
rent their home
BA
% of pop. age 1824 enrolled in
All MSAs
school
% of pop. age 1824 enrolled in
West
school
% of pop. age 1824 enrolled in
North Central
school
% of pop. Age 1824 enrolled in
South
school
% college
students residing
Final model
in city
Not significant in any % of population in
model
high school
Biked frequently (vs. biked
moderately)
Owns a bike (vs. not own a
ns
bike)
Biked in previous 7 days (vs.
0.05
did not bike)
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
10.98 commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
17.54 commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
15.28 commutting divided by the
Wald chi- probability of using car
square commuting.
ns
% of work trips in 1990 by
bicycle in each MSA
% of commuters using
less than bicycles in their journey-to.10
work in city
% of work trips in 1990 by
ns
bicycle in each MSA
0.01
142
Income Level
Parkin, Wardman,
Page (2008)
Independent
Associati
Model name
Variable
on
Not significant in any Median family
n/a
model
income, 1990
% of families
below the poverty
West
level
index of
Model of variation in deprivation
use of the bicycle for income scoreEngland and Wales
English
index of
Model of variation in deprivation
use of the bicycle for income scoreEngland and Wales
Welsh
Plaut (2005)
Cyclists to work who Log of salary of
own their home
worker
Study
Baltes, M. (1996)
Baltes, M. (1996)
Parkin, Wardman,
Page (2008)
Plaut (2005)
Rietveld and Daniel
(2004)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Cyclists to work who
rent their home
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Model 1 (own a bike
vs. not own a bike)
Log of salary of
worker
Level of
disposable
income
Income
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
income
not bike)
Model 3 (biked
frequently vs.
income
moderately)
Model 4 (own a bike
Income
vs. not own a bike)
Xin, Handy and
Buehler (2008)
Model 5 (Biked vs. did
Income
not bike)
Coefficient
/
/
-2.22
-0.0159
-
-0.291
-
-0.291
Significa
Dependent variable
nce
% of work trips in 1990 by
ns
bicycle in each MSA
0.01
% of work trips in 1990 by
bicycle in each MSA
(t-stat) - proportion of individuals in
16.49 ward x cycling to work
(t-stat) - proportion of individuals in
5.39
ward x cycling to work
The logit or log of the
probability of using biccyle
21.00 commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
10.64 commutting divided by the
Wald chi- probability of using car
square commuting.
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
/
+
0.016
0.01
n/a
/
ns
n/a
/
ns
+
0.015
0.01
n/a
/
ns
Owns a bike
Biked within last seven days
(vs. did not bike within the
last seven days)
Biked frequently (vs. biked
moderately)
Owns a bike (vs. not own a
bike)
Biked within last seven days
(vs. did not bike within the
last seven days)
143
Employment & Workforce
Study
Model name
Moudon, et al (2005) Airline Model
Moudon, et al (2005) Network Model
Parkin, Wardman,
Page (2008)
Parkin, Wardman,
Page (2008)
Parkin, Wardman,
Page (2008)
Parkin, Wardman,
Page (2008)
Parkin, Wardman,
Page (2008)
Baltes, M. (1996)
Baltes, M. (1996)
Baltes, M. (1996)
Baltes, M. (1996)
Baltes, M. (1996)
Baltes, M. (1996)
Baltes, M. (1996)
Baltes, M. (1996)
Model of variation in
use of the bicycle for
England and Wales
Model of variation in
use of the bicycle for
England and Wales
Model of variation in
use of the bicycle for
England and Wales
Model of variation in
use of the bicycle for
England and Wales
Independent
Associati
Variable
on
weekly work
hours
weekly work
hours (less than
40, 40, more than
40 hours)
proportion in
higher managerial
& professional in
larger
organisation
proportion in
higher
+
professional
proportion in
intermediate
occupations
proportion in lower
managerial and
professional
Proportion of all
Model of variation in employees aged
use of the bicycle for 16-74 with higher
England and Wales
level qualifications
% of pop that is
All MSAs
unemployed
% of pop that is
West
unemployed
% of pop that is
South
unemployed
% of pop.
Employed in
All MSAs
agriculture
% of pop.
Employed in
West
agriculture
Not significant in any % of workers in
model
central city
% of workers
living in central
All MSAs
city
% of workers
Not significant in any working in place
model
of residence
Baltes, M. (1996)
% of workers
Not significant in any working outside
model
place of residence
% of population
employed in
Northeast
agriculture
% of population
employed in
Northeast
manufacturing
Baltes, M. (1996)
Not significant in any % of population in
model
the armed forces
Baltes, M. (1996)
Baltes, M. (1996)
Coefficient
0.643 (odds
ratio)
0.642 (odds
ratio)
-4.7724
5.7281
Significa
Dependent variable
nce
# of times a week bicycling
0.05
for any reason
0.01
# of times a week bicycling
for any reason
(t-stat) - proportion of individuals in
10.14 ward x cycling to work
(t-stat)
24.44
proportion of individuals in
ward x cycling to work
-2.4663
(t-stat) - proportion of individuals in
8.75
ward x cycling to work
-2.5041
(t-stat) - proportion of individuals in
11.32 ward x cycling to work
n/a
/
ns
+
0.431
0.01
+
0.566
0.01
+
0.508
0.01
proportion of individuals in
ward x cycling to work
% of work trips in 1990 by
bicycle in each MSA
% of work trips in 1990 by
bicycle in each MSA
% of work trips in 1990 by
bicycle in each MSA
+
0.103
0.05
% of work trips in 1990 by
bicycle in each MSA
+
0.236
0.05
n/a
ns
% of work trips in 1990 by
bicycle in each MSA
% of work trips in 1990 by
bicycle in each MSA
+
0.797
0.01
% of work trips in 1990 by
bicycle in each MSA
n/a
/
ns
% of work trips in 1990 by
bicycle in each MSA
n/a
/
ns
% of work trips in 1990 by
bicycle in each MSA
+
0.232
0.01
% of work trips in 1990 by
bicycle in each MSA
-
-0.224
0.05
% of work trips in 1990 by
bicycle in each MSA
n/a
/
ns
% of work trips in 1990 by
bicycle in each MSA
144
Household Characteristics
Independent
Associati
Variable
on
% of housing
units occupied by
the owner
Coefficient
Study
Model name
Baltes, M. (1996)
All MSAs
Plaut (2005)
Dummy if
Cyclists to work who commercial
own their home
properties nearby
n/a
/
Plaut (2005)
Dummy if
Cyclists to work who commercial
rent their home
properties nearby
n/a
/
Plaut (2005)
Cyclists to work who Dummy if green
own their home
area near unit
n/a
/
Plaut (2005)
Cyclists to work who Dummy if green
rent their home
area near unit
n/a
/
Plaut (2005)
Cyclists to work who Dummy if unit has
own their home
garage
-
-0.559
Plaut (2005)
Cyclists to work who Dummy if unit has
rent their home
garage
n/a
/
Plaut (2005)
Dummy if unit has
Cyclists to work who parking space
own their home
included
n/a
/
Plaut (2005)
Dummy if unit has
Cyclists to work who parking space
rent their home
included
n/a
/
Plaut (2005)
Log of home
Cyclists to work who owners insurance
own their home
premium
n/a
/
Plaut (2005)
Log of home
Cyclists to work who owners insurance
rent their home
premium
n/a
/
Plaut (2005)
Cyclists to work who Log of property
own their home
tax
n/a
/
-0.274
Significa
Dependent variable
nce
% of work trips in 1990 by
bicycle in each MSA
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
6.49
commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
0.01
145
Household Characteristicscontinued
Independent
Variable
Associati
on
Coefficient
n/a
/
Cyclists to work who Log of rental
rent their home
payment
+
0.312
Plaut (2005)
Cyclists to work who Log of rental
rent their home
payment
n/a
/
Plaut (2005)
Cyclists to work who Log of square foot
own their home
floor space of unit
n/a
/
Plaut (2005)
Cyclists to work who Log of square foot
rent their home
floor space of unit
n/a
/
Plaut (2005)
Cyclists to work who Log of value of
own their home
unit owned
-
-0.128
Plaut (2005)
Cyclists to work who # of bathrooms in
own their home
unit
n/a
/
Plaut (2005)
Cyclists to work who # of person in
rent their home
household
n/a
/
Plaut (2005)
Cyclists to work who # of persons in
household
own their home
n/a
/
Study
Model name
Plaut (2005)
Cyclists to work who Log of property
rent their home
tax
Plaut (2005)
Significa
Dependent variable
nce
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
1.97
commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
Wald chi- commutting divided by the
square probability of using car
2.103 commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
ns
commuting.
146
Household Characteristicscontinued
Associati
on
Coefficient
Cyclists to work who
Year unit built
own their home
-
-0.017
Plaut (2005)
Cyclists to work who
Year unit built
rent their home
-
-0.017
Plaut (2005)
Xin, Handy and
Buehler (2008)
Number of
Cyclists to work who bathrooms in the
rent their home
unit
Model 1 (own a bike
Household size
vs. not own a bike)
-
-0.59
n/a
/
n/a
/
-
-0.262
0.05
n/a
/
ns
Study
Model name
Plaut (2005)
Xin, Handy and
Buehler (2008)
Independent
Variable
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
Household size
not bike)
Model 3 (biked
frequently vs.
household size
moderately)
Xin, Handy and
Buehler (2008)
Model 5 (Biked vs. did
household size
not bike)
Significa
Dependent variable
nce
The logit or log of the
probability of using biccyle
17.83 commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
12.1
commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
2.38
commutting divided by the
Wald chi- probability of using car
square commuting.
Owns a bike (vs. not own a
ns
bike)
Biked within last seven days
(vs. did not bike within the
ns
last seven days)
Biked frequently (vs. biked
moderately)
Biked within last seven days
(vs. did not bike within the
last seven days)
147
Bicycle owership
Study
Model name
Moudon, et al (2005) Airline Model
Moudon, et al (2005) Network Model
Bicycle-Choice Model
for predicting that a
Cervero and Duncan trip will be made by
(2003)
bicycle
Xin, Handy and
Model 1 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 4 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 1 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 4 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 1 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 1 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 4 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 1 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 4 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 1 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 4 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 4 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 1 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 1 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 4 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 1 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 4 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 1 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 4 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 1 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 4 (own a bike
Buehler (2008)
vs. not own a bike)
Xin, Handy and
Model 4 (own a bike
Buehler (2008)
vs. not own a bike)
Independent
Variable
Associati
on
Significa
Dependent variable
nce
# of times a week bicycling
0.01
for any reason
owns a bicycle
owns a bicycle
(vs. not owning a
bicycle)
+
Coefficient
179.648 (odds
ratio)
+
163.204 (odds
ratio)
0.01
Number of
bicycles in
household
+
0.345
0
bikers poor
-
-0.264
0.05
Bikers poor
-
-0.201
0.1
bikers spend
-
-0.208
0.1
Bikers spend
-
-0.241
0.1
car ownership
n/a
/
ns
good health
+
0.361
0.01
good health
+
0.351
0.01
Like biking
+
1.166
0.01
Like Biking
+
1.17
0.01
like transit
-
-0.216
0.05
Like Transit
-
-0.231
0.05
like walking
n/a
/
ns
Like walking
n/a
/
ns
need car
-
-0.329
0.05
Need Car
-
-0.306
0.05
pro-exercise
-
-0.438
0.01
pro-exercise
-
-0.447
0.01
race (white)
+
0.507
0.1
race (white)
+
0.513
0.1
transit access
n/a
/
ns
transit access
n/a
/
ns
Household size
n/a
/
ns
# of times a week bicycling
for any reason
probability of person n
choosing mode I for travelling
between origin and
destination
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
Owns a bike (vs. not own a
bike)
148
Car Ownership
Independent
Variable
Study
Model name
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
car ownership
not bike)
Model 3 (biked
frequently vs.
car ownership
moderately)
Plaut (2005)
Cyclists to work who Dummy if own no
own their home
cars
Plaut (2005)
Cyclists to work who Dummy if own no
rent their home
cars
Plaut (2005)
Dummy if own
Cyclists to work who three or more
own their home
cars
Plaut (2005)
Dummy if own
Cyclists to work who three or more
rent their home
cars
Plaut (2005)
Cyclists to work who Dummy if own
own their home
two cars
Plaut (2005)
Rietveld and Daniel
(2004)
Cyclists to work who
rent their home
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Moudon, et al (2005) Airline Model
Dummy if own
two cars
Number of cars
per capita
exactly 1 car per
adult (vs. more
than 1 car)
exactly 1 car per
adult (vs. more
than 1 car)
Moudon, et al (2005) Network Model
Model of variation in
Parkin, Wardman,
use of the bicycle for Number of cars
Page (2008)
per employee
England and Wales
Associati
on
Coefficient
n/a
/
-
-1.602
0.1
1.048
10.56
Wald chisquare
3.061
178.6
Wald chisquare
-
-0.632
4.40
Wald chisquare
n/a
/
ns
-
-0.629
7.41
Wald chisquare
n/a
/
ns
-
-0.26
-
0.407 (odds
ratio)
0.05
# of times a week bicycling
for any reason
-
0.438 (odds
ratio)
0.1
# of times a week bicycling
for any reason
-
-0.9758
+
+
Significa
Dependent variable
nce
Biked within last seven days
(vs. did not bike within the
ns
last seven days)
Biked frequently (vs. biked
moderately)
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
commuting.
The logit or log of the
probability of using biccyle
commutting divided by the
probability of using car
commuting.
(t-value) - Share of bicycle use in
Dutch cities
1.95
(t-stat) - proportion of individuals in
22.90 ward x cycling to work
149
Car Ownership-continued
Study
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Independent
Model name
Variable
Model 2 (biked vs. did
need car
not bike)
Model 3 (biked
frequently vs.
need car
moderately)
Xin, Handy and
Buehler (2008)
Model 5 (Biked vs. did
Need car
not bike)
Model 1 (D.C.
Vehicles per
Dill & Carr (2003)
excluded)
household
Vehicles per
Dill & Carr (2003)
Model 3
household
Model 4 (D.C and
Vehicles per
Dill & Carr (2003)
NYC excluded)
household
Bicycle-Choice Model
Number of
for predicting that a
Cervero and Duncan trip will be made by
vehicles in
(2003)
household
bicycle
% of households
with no vehicle
Baltes, M. (1996)
All MSAs
available
Associati
on
Coefficient
n/a
/
ns
-
-0.389
0.05
n/a
/
ns
n/a
-0.698
0.208
% commuting by bicycle
n/a
/
ns
% commuting by bicycle
-
-1.52
0.02
-
-0.629
0
-
-0.259
0.01
Associati
on
Coefficient
-
-0.602
n/a
/
ns
-
-0.626
0.01
Associati
on
Coefficient
n/a
/
-
-0.237
0.05
n/a
/
2.965 (odds
ratio)
2.702 (odds
ratio)
ns
Significa
Dependent variable
nce
Biked vs. did not bike
Biked frequently (vs. biked
moderately)
Biked within last seven days
(vs. did not bike within the
last seven days)
% commuting by bicycle
probability of person n
choosing mode I for travelling
between origin and
destination
% of work trips in 1990 by
bicycle in each MSA
Interest in Walking
Independent
Variable
Study
Model name
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
like walking
not bike)
Model 3 (biked
frequently vs.
like walking
moderately)
Model 5 (Biked vs. did
Like walking
not bike)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Significa
Dependent variable
nce
Biked within last seven days
(vs. did not bike within the
0.01
last seven days)
Biked frequently (vs. biked
moderately)
Biked in previous 7 days (vs.
did not bike)
Transit Preference
Independent
Variable
Study
Model name
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
like transit
not bike)
Model 3 (biked
frequently vs.
like transit
moderately)
Xin, Handy and
Buehler (2008)
Model 5 (Biked vs. did
Like Transit
not bike)
Moudon, et al (2005) Airline Model
uses transit
+
Moudon, et al (2005) Network Model
uses transit
+
Significa
Dependent variable
nce
Biked within last seven days
(vs. did not bike within the
ns
last seven days)
0.01
0.1
Biked frequently (vs. biked
moderately)
Biked within last seven days
(vs. did not bike within the
last seven days)
# of times a week bicycling
for any reason
# of times a week bicycling
for any reason
150
Attitudes toward Cycling and Cyclists
Independent
Variable
Associati
on
Coefficient
+
1.252
+
1.045
0.01
+
1.26
0.01
lack of interest
-
(Odds Ratio)
0.45
0.003
bikers poor
-
-0.373
0.01
bikers poor
n/a
/
ns
-
-0.311
0.05
n/a
/
ns
(Odds Ratio)
0.61
0.078
n/a
/
ns
n/a
/
ns
n/a
/
ns
-
(Odds Ratio)
0.26
0.001
Study
Model name
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
like biking
not bike)
Model 3 (biked
frequently vs.
Like biking
moderately)
Model 5 (Biked vs. did
Like Biking
not bike)
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Xin, Handy and
Buehler (2008)
Parkin, Wardman,
Page (2008)
Model 2 (biked vs. did
not bike)
Model 3 (biked
frequently vs.
moderately)
Model 5 (Biked vs. did
not bike)
Model of variation in
use of the bicycle for
England and Wales
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Bikers poor
Probability of
acceptability of
cycling
Internal selfefficacy
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
bikers spend
not bike)
Model 3 (biked
frequently vs.
bikers spend
moderately)
Xin, Handy and
Buehler (2008)
Model 5 (Biked vs. did
Bikers spend
not bike)
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
lack of time
Significa
Dependent variable
nce
Biked within last seven days
(vs. did not bike within the
0.01
last seven days)
Biked frequently (vs. biked
moderately)
Biked in previous 7 days (vs.
did not bike)
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
study
Biked within last seven days
(vs. did not bike within the
last seven days)
Biked frequently (vs. biked
moderately)
Biked in previous 7 days (vs.
did not bike)
proportion of individuals in
ward x cycling to work
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
study
Biked within last seven days
(vs. did not bike within the
last seven days)
Biked frequently (vs. biked
moderately)
Biked within last seven days
(vs. did not bike within the
last seven days)
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
study
151
Individual's health/attitudes
Study
Model name
Independent
Variable
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Body image
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
good health
not bike)
Model 3 (biked
frequently vs.
good health
moderately)
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Associati
on
Coefficient
-
(Odds Ratio)
0.63
n/a
/
-
-0.333
0.05
n/a
/
ns
-
0.502 (odds
ratio)
0.05
Moudon, et al (2005) Network Model
Lack of skills and
health
attitude factor on
knowledge of
phys benefits
attitude factor on
knowledge of
phys benefits
Moudon, et al (2005) Airline Model
exercise at home
+
Moudon, et al (2005) Network Model
exercise at home
no (vs. sufficient)
vigourous phys
activity
no (vs. sufficient)
vigourous phy
sactivity
+
0.548 (odds
ratio)
2.134 (odds
ratio)
3.626 (odds
ratio)
-
Moudon, et al (2005) Airline Model
Moudon, et al (2005) Network Model
Moudon, et al (2005) Airline Model
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
pro exercise
not bike)
Model 3 (biked
frequently vs.
pro exercise
moderately)
Model 5 (Biked vs. did
Pro-exercise
not bike)
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Xin, Handy and
Buehler (2008)
Moudon, et al (2005) Airline Model
Moudon, et al (2005) Airline Model
Troped, et al (2001)
Troped, et al (2001)
Troped, et al (2001)
Physical wellbeing
moderate (vs.
sufficient) phys
active
moderate (vs.
sufficient)
physical activity
Self reported env.
Variable model
Temporary
illness/injury
Self reported env.
Variable model
GIS environmental
variable model
Long term
illness/injury
Long term
illness/injury
Significa
Dependent variable
nce
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.08
study
Biked within last seven days
(vs. did not bike within the
ns
last seven days)
Biked frequently (vs. biked
moderately)
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
study
# of times a week bicycling
for any reason
0.05
# of times a week bicycling
for any reason
# of times a week bicycling
for any reason
# of times a week bicycling
for any reason
0.231 (odds
ratio)
0.01
# of times a week bicycling
for any reason
-
0.225 (odds
ratio)
0.01
+
0.244
0.05
n/a
/
ns
+
0.247
0.05
n/a
(Odds Ratio)
1.05
0.845
0.527(odds
ratio)
0.1
# of times a week bicycling
for any reason
n/a
/
ns
# of times a week bicycling
for any reason
+
(Odds Ratio)
1.66
pg 197 for
95% CI Use of a community rail trail
-
(Odds Ratio)
0.43
pg 197 for
95% CI Use of a community rail trail
n/a
/
-
0.05
0.05
ns
# of times a week bicycling
for any reason
Biked within last seven days
(vs. did not bike within the
last seven days)
Biked frequently (vs. biked
moderately)
Biked in previous 7 days (vs.
did not bike)
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
study
Use of a community rail trail
152
External Support
Study
Model name
Independent
Variable
Associati
on
Coefficient
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
External selfefficacy
+
(Odds Ratio)
0.32
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Social support:
accompany
+
(Odds Ratio)
2.26
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Social support:
encourage
n/a
(Odds Ratio)
0.66
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Modeling
+
(Odds Ratio)
1.83
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Social norm
n/a
(Odds Ratio)
1.30
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Social influence
n/a
(Odds Ratio)
0.98
Significa
Dependent variable
nce
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.001 study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.012 study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.183 study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.043 study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.377 study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.94
study
153
Bicycle lanes and infrastructure
Study
Model name
Independent
Variable
Associati
on
Coefficient
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Bicycle lanes
(neighborhood)
n/a
(Odds Ratio)
0.70
Bicycle lanes
(road to work)
n/a
(Odds Ratio)
1.48
Bas de Geus et al
(2008)
Rietveld and Daniel
(2004)
Rietveld and Daniel
(2004)
Differences between
Cyclists and Noncyclists
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Bicycle network
Bicycle parks
Significa
Dependent variable
nce
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.13
study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.12
study
/
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
/
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
/
(odds ratio)
1.19 (0.85,
1.67)
Nelson and Allen
(1997)
Associations between
perceived
environmental
variables and physical
activity bheavior in the
Bicycle paths
US 1999-2000
bicycle pathways
per 100,000
Final model
residents in 1992
+
0.754+0.069x
Krizek, Barnes and
Thompson (2009)
Table 3- Bicycle
Commute Share in
Buffer Analysis
Areas, 1990-2000
Change in bicycle
commuters in
facility buffer
Battle Creek
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 3- Bicycle
Commute Share in
Buffer Analysis
Areas, 1990-2000
Change in bicycle
commuters in
facility buffer
Cedar LakeKenilworth
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 3- Bicycle
Commute Share in
Buffer Analysis
Areas, 1990-2000
Change in bicycle
commuters in
facility buffer ParkPortland
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 3- Bicycle
Commute Share in
Buffer Analysis
Areas, 1990-2000
Change in bicycle
commuters in
facility buffer
Phalen
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 3- Bicycle
Commute Share in
Buffer Analysis
Areas, 1990-2000
Change in bicycle
commuters in
facility buffer
Shepard
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 3- Bicycle
Commute Share in
Buffer Analysis
Areas, 1990-2000
Change in bicycle
commuters in
facility buffer
Summit
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
Brownson, et al
(2001)
Physical activity behavior
(defined as meeting public
health recommendations for
0.05
moderate or vigorous activity)
% of commuters using
less than bicycles in their journey-to.10
work in city
154
Bicycle lanes and infrastructurecontinued
Study
Model name
Independent
Variable
Associati
on
Coefficient
Significa
Dependent variable
nce
Krizek, Barnes and
Thompson (2009)
Table 3- Bicycle
Commute Share in
Buffer Analysis
Areas, 1990-2000
Change in bicycle
commuters in
facility buffer Univ.
of Minnesota
n/a
longitudinal
study
change in bicycle mode
0 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 2- Minneapolis
and St. Paul Bicycle
Commute Share,
1990-2000
Change in bicycle
commuters in
Minneapolis TAZs
in Buffer 1
+
longitudinal
study
change in bicycle mode
1 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 2- Minneapolis
and St. Paul Bicycle
Commute Share,
1990-2000
Change in bicycle
commuters in
Minneapolis TAZs
in Buffer 2
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
Change in bicycle
commuters in
Minneapolis, 19902000
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 2- Minneapolis
and St. Paul Bicycle
Commute Share,
1990-2000
Table 2- Minneapolis
and St. Paul Bicycle
Commute Share,
1990-2000
Change in bicycle
commuters in St
Paul, 1990-2000
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 2- Minneapolis
and St. Paul Bicycle
Commute Share,
1990-2000
Change in bicycle
commuters in St.
Paul TAZs in
Buffer 1
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 2- Minneapolis
and St. Paul Bicycle
Commute Share,
1990-2000
Change in bicycle
commuters in St.
Paul TAZs in
Buffer 2
/
longitudinal
study
change in bicycle mode
0 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 2- Minneapolis
and St. Paul Bicycle
Commute Share,
1990-2000
Change in bicycle
commuters in
zones outside
buffers
Minneapolis
+
longitudinal
study
change in bicycle mode
1 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Table 2- Minneapolis
and St. Paul Bicycle
Commute Share,
1990-2000
Change in bicycle
commuters in
zones outside
buffers St. Paul
+
longitudinal
study
change in bicycle mode
1 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
Change in bicycle
commuters trips
to major
employment/activi
Table 5- Major
ty centers
Destination Commute Downtown
Share
Minneapolis
+
longitudinal
study
change in bicycle mode
1 std dev share 1990-2000
Krizek, Barnes and
Thompson (2009)
155
Bicycle lanes and infrastructurecontinued
Independent
Variable
Study
Model name
Krizek, Barnes and
Thompson (2009)
Change in bicycle
commuters trips
to major
Table 5- Major
employment/activi
Destination Commute ty centers St.
Share
Paul
Change in bicycle
commuters trips
to major
Table 5- Major
employment/activi
Krizek, Barnes and Destination Commute ty centers Univ. of
Thompson (2009)
Share
Minnesota
Table 4- River
Trips crossing
Crossing Bicycle
south flowing
Krizek, Barnes and Commute Share,
portion of
Thompson (2009)
1990-2000
Mississippi River
Table 4- River
Trips originating
Crossing Bicycle
and terminating
Krizek, Barnes and Commute Share,
east of the
Thompson (2009)
1990-2002
Mississippi River
Table 4- River
Trips originating
Crossing Bicycle
and terminating
Krizek, Barnes and Commute Share,
west of the
Thompson (2009)
1990-2001
Mississippi River
Model of variation in
Parkin, Wardman,
use of the bicycle for Proportion of offPage (2008)
road route
England and Wales
Distance to
Self reported env.
bikeway (.25 mile
Troped, et al (2001) Variable model
increase)
distance to
bikeway via road
GIS environmental
network (.25 mile
Troped, et al (2001) variable model
increase)
distance to
Moudon, et al (2005) Airline Model
closest trail
distance to
Moudon, et al (2005) Network Model
closest trail
Associati
on
Coefficient
Significa
Dependent variable
nce
-
longitudinal
study
std dev (- change in bicycle mode
1)
share 1990-2000
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
+
longitudinal
study
change in bicycle mode
2 std dev share 1990-2000
+
longitudinal
study
change in bicycle mode
1 std dev share 1990-2000
+
12.5162
-
(Odds Ratio)
0.65
pg 197 for
95% CI Use of a community rail trail
-
(Odds Ratio)
0.58
-
0.801
-
0.728
pg 197 for
95% CI Use of a community rail trail
# of times a week bicycling
0.01
for any reason
# of times a week bicycling
0.01
for any reason
(t-stat)
18.72
proportion of individuals in
ward x cycling to work
156
Bicycle lanes and infrastructurecontinued
Study
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Independent
Variable
Miles of bike
lanes per square
Model 4 (own a bike mile (objectively
vs. not own a bike)
measured)
Miles of bike
lanes per square
Model 5 (Biked vs. did mile (objectively
not bike)
measured)
Model name
Proportion of road
that has a bicycle
or bus lane
Type 2 lanes per
Dill & Carr (2003)
square mile
Type 2 lanes per
Dill & Carr (2003)
square mile
Type 2 lanes per
Dill & Carr (2003)
Model 3
square mile
Model 4 (D.C and
Type 2 lanes per
Dill & Carr (2003)
NYC excluded)
square mile
Proportion of
Model of variation in cycle route that is
Parkin, Wardman,
use of the bicycle for adjacent to the
Page (2008)
England and Wales
road
subjectively
measured
presence of
cycling trails and
Moudon, et al (2005) Airline Model
lanes
subjectively
measured
presence of
cycling trails and
Moudon, et al (2005) Network Model
lanes
Parkin, Wardman,
Page (2008)
Model of variation in
use of the bicycle for
England and Wales
Model 1 (D.C.
excluded)
Model 2 (D.C.
excluded)
Associati
on
Coefficient
Significa
Dependent variable
nce
+
0.044
0.1
Owns a bike (vs. not own a
bike)
+
0.046
0.1
Biked in previous 7 days (vs.
did not bike)
n/a
/
ns
proportion of individuals in
ward x cycling to work
+
0.892
0.008
% commuting by bicycle
+
0.888
0.006
% commuting by bicycle
+
0.861
0.007
% commuting by bicycle
+
0.998
0.002
% commuting by bicycle
n/a
/
ns
proportion of individuals in
ward x cycling to work
+
1.704
0.01
# of times a week bicycling
for any reason
+
1.729
0.01
# of times a week bicycling
for any reason
Moudon, et al (2005) Airline Model
% of streets lined
with bicycle lanes
n/a
/
ns
# of times a week bicycling
for any reason
Moudon, et al (2005) Network Model
% of streets lined
with bicycle lanes
n/a
/
ns
# of times a week bicycling
for any reason
157
Weather/environment
Study
Parkin, Wardman,
Page (2008)
Rietveld and Daniel
(2004)
Model name
Model of variation in
use of the bicycle for
England and Wales
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Independent
Variable
Mean temperature
in degrees
centigrade
Average
temperature
Potential wind
speed
Basic wind speed
Model of variation in for sructural
Parkin, Wardman,
use of the bicycle for design for the
Page (2008)
England and Wales
district
number of days
during the year
Nelson and Allen
rain exceeds
Final model
(1997)
1/10th an inch
Model of variation in Total annual
Parkin, Wardman,
use of the bicycle for rainfall in
Page (2008)
England and Wales
millimetres
Semi-log linear
regression model,
explaining the share
Rietveld and Daniel of bicycle use in
rainfall (mm)
(2004)
cities
Dill & Carr (2003)
Model 4
Days of rain
Model 4 (D.C and
Dill & Carr (2003)
Days of rain
NYC excluded)
Model 1 (D.C.
Dill & Carr (2003)
Days on rain
excluded)
Bicycle-Choice Model
dark (before
for predicting that a
Cervero and Duncan trip will be made by
sunrise or after
(2003)
sunset)
bicycle
Total annual
hours of sunshine
for the year May
Model of variation in 2000 to April 2001
Parkin, Wardman,
use of the bicycle for for the weather
Page (2008)
England and Wales
region
Rietveld and Daniel
(2004)
Associati
on
Coefficient
+
0.0782
Significa
Dependent variable
nce
(t-stat)
7.87
proportion of individuals in
ward x cycling to work
/
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
/
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
/
n/a
ns
proportion of individuals in
ward x cycling to work
-
-0.008
% of commuters using
less than bicycles in their journey-to.10
work in city
-
-0.0006
(t-stat) - proportion of individuals in
17.40 ward x cycling to work
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
ns
% commuting by bicycle
n/a
n/a
/
/
-
-0.008
0.02
n/a
-0.005
0.206
-
-0.721
0.022
n/a
/
ns
% commuting by bicycle
% commuting by bicycle
probability of person n
choosing mode I for travelling
between origin and
destination
proportion of individuals in
ward x cycling to work
158
Busy Streets/Presence of
Automobiles
Study
Troped, et al (2001)
Troped, et al (2001)
Model name
Independent
Variable
Self reported env.
Variable model
GIS environmental
variable model
Busy street
barrier (no)
Busy street
barrier (no)
Moudon, et al (2005) Airline Model
Moudon, et al (2005) Network Model
Moudon, et al (2005) Airline Model
Moudon, et al (2005) Network Model
Semi-log linear
regression model,
explaining the share
Rietveld and Daniel of bicycle use in
(2004)
cities
Associati
on
Coefficient
Significa
Dependent variable
nce
+
(Odds Ratio)
2.01
pg 197 for
95% CI Use of a community rail trail
n/a
(curviline
problems- auto
ar, see
faciliites in
page
neighborhoods
254)
(curviline
problems- auto
ar, see
faciliites in
page
neighborhoods
254)
(curviline
ar, see
problems- auto in
page
neighborhood
254)
(curviline
ar, see
problems- auto in
page
neighborhood
254)
/
ns
Use of a community rail trail
/
0.1
# of times a week bicycling
for any reason
/
0.1
# of times a week bicycling
for any reason
/
0.05
# of times a week bicycling
for any reason
/
0.05
# of times a week bicycling
for any reason
Motorised traffic
noise
n/a
/
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Traffic danger
(neighborhood)
n/a
(Odds Ratio)
1.02
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Traffic danger
(road to work)
n/a
(Odds Ratio)
1.30
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
n/a
(Odds Ratio)
1.27
-
-0.0373
Parkin, Wardman,
Page (2008)
Traffic safety
(neighborhood)
Transport demand
intensity
Model of variation in (employees
use of the bicycle for dvided by road
England and Wales
length)
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.93
study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.26
study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.36
study
(t-stat) - proportion of individuals in
17.74 ward x cycling to work
159
City Level/Population Size
Study
Rietveld and Daniel
(2004)
Rietveld and Daniel
(2004)
Baltes, M. (1996)
Parkin, Wardman,
Page (2008)
Rietveld and Daniel
(2004)
Baltes, M. (1996)
Rietveld and Daniel
(2004)
Rietveld and Daniel
(2004)
Rietveld and Daniel
(2004)
Model name
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Independent
Variable
Coefficient
Significa
Dependent variable
nce
n/a
/
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
-
-0.000829
(t-value) - Share of bicycle use in
3.90
Dutch cities
n/a
/
ns
% of work trips in 1990 by
bicycle in each MSA
+
0.0001
(t-stat)
9.11
proportion of individuals in
ward x cycling to work
-
-0.00669
n/a
/
Safety level
(number of victims
of serious
accidents per 100
million bicyclekilometres)
+
0.0109
(t-value) Share of bicycle use in
1.83
Dutch cities
School for Higher
Vocational
Training
+
0.0742
(t-value) Share of bicycle use in
2.32
Dutch cities
/
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
Human activity
indicator
Population
(thousands)
Population
Not significant in any density (persons
model
per square mile)
Population
Model of variation in density
use of the bicycle for (population
England and Wales
divided by area)
Semi-log linear
Density of Human
regression model,
explaining the share Activity
(addresses per
of bicycle use in
square kilometre)
cities
Not significant in any Inverse of MSA
model
population
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Associati
on
Presence of a
university
n/a
(t-value) - Share of bicycle use in
3.00
Dutch cities
% of work trips in 1990 by
ns
bicycle in each MSA
160
Attitudes of
restraint/environmental
attitudes
Study
Model name
Bas de Geus et al
(2008)
Xin, Handy and
Buehler (2008)
Differences between
Cyclists and Noncyclists
Model 4 (own a bike
vs. not own a bike)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
not bike)
Model 3 (biked
frequently vs.
moderately)
Model 5 (Biked vs. did
not bike)
Model 1 (own a bike
vs. not own a bike)
Xin, Handy and
Buehler (2008)
All models
Independent
Variable
Significa
Dependent variable
nce
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.029 study
owns a bike (vs. does not
ns
own a bike)
Biked within last seven days
(vs. did not bike within the
0.01
last seven days)
Associati
on
Coefficient
Ecologicaleconomic
awareness
+
(Odds Ratio)
1.71
limit driving
n/a
/
limit driving
+
0.312
limit driving
+
0.606
0.01
Limit driving
+
0.313
0.01
limit driving
n/a
/
ns
Environmental
concern
n/a
/
ns
Associati
on
Coefficient
-
-1.392
(t-stat) - proportion of individuals in
50.93 ward x cycling to work
-
-0.745
(t-value) - Share of bicycle use in
10.76 Dutch cities
n/a
/
+
(Odds Ratio)
1.9
n/a
-7.796
Biked frequently (vs. biked
moderately)
Biked in previous 7 days (vs.
did not bike)
owns a bike (vs. does not
own a bike)
Bike ownership, biking in
previous 7 days, and biked
frequently
Slope
Study
Parkin, Wardman,
Page (2008)
Rietveld and Daniel
(2004)
Troped, et al (2001)
Independent
Variable
Proportion 1km
Model of variation in squares with
use of the bicycle for slope 3% or
England and Wales
steeper
Semi-log linear
regression model,
explaining the share
relief (hills and
of bicycle use in
slopes)
cities
Self reported env.
Steep hill
Variable model
Model name
GIS environmental
steep hill barrier
variable model
(no)
Bicycle-Choice Model
for predicting that a
Cervero and Duncan trip will be made by
Slope (rise/run)
(2003)
bicycle
Troped, et al (2001)
Significa
Dependent variable
nce
ns
Use of a community rail trail
pg 197 for
95% CI Use of a community rail trail
probability of person n
choosing mode I for travelling
between origin and
0.187 destination
161
Geographical factors
Associati
on
Coefficient
Cyclists to work who Dummy for West
own their home
Coast
+
1.216
Plaut (2005)
Cyclists to work who Dummy for West
rent their home
Coast
+
0.517
Plaut (2005)
Dummy if live in
Cyclists to work who central city of the
own their home
MSA
+
0.131
Plaut (2005)
Dummy if live in
Cyclists to work who central city of the
rent their home
MSA
+
0.178
Plaut (2005)
Cyclists to work who Dummy if live in
own their home
rural area of MSA
-
-1.03
Plaut (2005)
Cyclists to work who Dummy if live in
rent their home
rural area of MSA
-
-0.818
Plaut (2005)
Dummy if live in
Cyclists to work who secondary urban
own their home
area in MSA
-
-0.517
-
-0.802
-
-0.683
Study
Model name
Plaut (2005)
Plaut (2005)
Baltes, M. (1996)
Independent
Variable
Dummy if live in
Cyclists to work who secondary urban
rent their home
area in MSA
% of population
living in central
All MSAs
city
Significa
Dependent variable
nce
The logit or log of the
probability of using biccyle
34.15 commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
5.01
commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
0.259 commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
0.301 commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
5.8 Wald commutting divided by the
chiprobability of using car
square commuting.
The logit or log of the
probability of using biccyle
1.59
commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
3.56
commutting divided by the
Wald chi- probability of using car
square commuting.
The logit or log of the
probability of using biccyle
4.68
commutting divided by the
Wald chi- probability of using car
square commuting.
0.05
% of work trips in 1990 by
bicycle in each MSA
162
Trip specific variables
Study
Rietveld and Daniel
(2004)
Rietveld and Daniel
(2004)
Parkin, Wardman,
Page (2008)
Parkin, Wardman,
Page (2008)
Cervero and Duncan
(2003)
Cervero and Duncan
(2003)
Cervero and Duncan
(2003)
Cervero and Duncan
(2003)
Cervero and Duncan
(2003)
Baltes, M. (1996)
Rietveld and Daniel
(2004)
Rietveld and Daniel
(2004)
Model name
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Independent
Variable
directness of trip
delays
Proportion in the
Model of variation in distance band
use of the bicycle for "2km to less than
England and Wales
5km"
Proportion in the
Model of variation in distance band
use of the bicycle for "5km to less than
England and Wales
20km"
Bicycle-Choice Model
for predicting that a
Trip distance
trip will be made by
(miles)
bicycle
Bicycle-Choice Model
for predicting that a
trip will be made by
Shop purpose
bicycle
Bicycle-Choice Model
for predicting that a
Recreation/entert
trip will be made by
ainment purpose
bicycle
Bicycle-Choice Model
for predicting that a
trip will be made by
Weekend trip
bicycle
Bicycle-Choice Model
for predicting that a
trip will be made by
Social purpose
bicycle
Not significant in any
model
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
% of travel time to
work <10 minutes
One riding behind
the other
Priority
Associati
on
Coefficient
Significa
Dependent variable
nce
n/a
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
-
-0.6916
-
-1.6556
-
-0.291
0.443
+
0.602
0.226
+
n/a
0.861
(t-stat) - proportion of individuals in
8.53
ward x cycling to work
(t-stat) - proportion of individuals in
20.49 ward x cycling to work
probability of person n
choosing mode I for travelling
between origin and
0.001 destination
probability of person n
choosing mode I for travelling
between origin and
0.256 destination
probability of person n
choosing mode I for travelling
between origin and
0.001 destination
probability of person n
choosing mode I for travelling
between origin and
0.301 destination
probability of person n
choosing mode I for travelling
between origin and
0.002 destination
ns
% of work trips in 1990 by
bicycle in each MSA
n/a
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
163
Trip specific variablescontinued
Study
Independent
Variable
Associati
on
Coefficient
Rietveld and Daniel
(2004)
Model name
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Destinations food
shops
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Destinations other
shops
(Odds Ratio)
0.75
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
Destinations work
(Odds Ratio)
0.77
Rietveld and Daniel
(2004)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
not bike)
Model 3 (biked
frequently vs.
moderately)
Model 4 (own a bike
vs. not own a bike)
Model 5 (Biked vs. did
not bike)
Model 1 (own a bike
vs. not own a bike)
Moudon, et al (2005) Airline Model
Moudon, et al (2005) Network Model
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
slowdowns
Turns off
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
n/a
+
(Odds Ratio)
0.60
safe destinations
+
0.617
safe destinations
+
0.338
safe destinations
n/a
safe destinations
+
safe destinations
presence of
destinations
(grocery stores
and schools)
presence of
destinations
(grocery stores
and schools)
n/a
Facilities for
cyclists at the
workplace
Significa
Dependent variable
nce
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.058 study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.278 study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.293 study
Biked within last seven days
(vs. did not bike within the
0.01
last seven days)
ns
Biked frequently (vs. biked
moderately)
Owns a bike (vs. does not
own a bike)
Biked in previous 7 days (vs.
did not bike)
Owns a bike (vs. does not
own a bike)
# of times a week bicycling
for any reason
0.01
ns
0.601
0.01
-
0.702
0.1
-
0.718
0.1
+
(Odds Ratio)
0.28
0.001
# of times a week bicycling
for any reason
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
study
164
Neighborhoods- Origins and Destinations
Independent
Associati
Model name
Variable
on
Bicycle-Choice Model
for predicting that a
Cervero and Duncan trip will be made by
Land-use diversity
n/a
(2003)
factor, origin
bicycle
Smaller total area
of convenience
store parcels
Moudon, et al (2005) Airline Model
within 3km buffer
Smaller total area
of convenience
store parcels
Moudon, et al (2005) Network Model
within 3km buffer
Bicycle-Choice Model
Pedestrian/bike
for predicting that a
Cervero and Duncan trip will be made by
friendly design
n/a
(2003)
factor, origin
bicycle
Retail/service
density: # of
Bicycle-Choice Model retail/service jobs
for predicting that a
per net comercial
Cervero and Duncan trip will be made by
acre within 1 mile
n/a
(2003)
bicycle
of orgiin
more parcels
within the closest
(NC10) (office fast
food hospital
Moudon, et al (2005) Airline Model
+
clinic)
more parcels
within the closest
(NC10) (office fast
food hospital
Moudon, et al (2005) Network Model
+
clinic)
Study
Bas de Geus et al
(2008)
Xin, Handy and
Buehler (2008)
Differences between
Cyclists and Noncyclists
Crime
(neighborhood)
Coefficient
0.156
Significa
Dependent variable
nce
probability of person n
choosing mode I for travelling
between origin and
0.112 destination
0.01
0.784
0.01
0.234
0.122
# of times a week bicycling
for any reason
probability of person n
choosing mode I for travelling
between origin and
destination
0.005
0.114
probability of person n
choosing mode I for travelling
between origin and
destination
1.16
0.1
1.238
0.05
(Odds Ratio)
0.63
0.14
0.694
0.05
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
transit access
not bike)
Model 3 (biked
frequently vs.
Transit Access
moderately)
n/a
Xin, Handy and
Buehler (2008)
Model 5 (Biked vs. did
transit access
not bike)
+
0.664
0.05
Bas de Geus et al
(2008)
Differences between
Cyclists and Noncyclists
n/a
(Odds Ratio)
0.83
0.494
Bus, tram or
metro stop
+
# of times a week bicycling
for any reason
0.822
ns
# of times a week bicycling
for any reason
# of times a week bicycling
for any reason
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
study
Biked within last seven days
(vs. did not bike within the
last seven days)
Biked frequently (vs biked
moderately)
Biked within last seven days
(vs. did not bike within the
last seven days)
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
study
165
Neighborhoods- Origins and Destinations- continued
Study
Model name
Independent
Variable
low-income
neighborhood
(proportion of
households within
Bicycle-Choice Model 1 mile of orgin
for predicting that a
and destination
Cervero and Duncan trip will be made by
with annual
(2003)
bicycle
incomes<$25,00
Bicycle-Choice Model
for predicting that a
Cervero and Duncan trip will be made by
Land-use diversity
(2003)
factor, destination
bicycle
Bicycle-Choice Model
Pedestrian/bike
for predicting that a
Cervero and Duncan trip will be made by
friendly design
(2003)
factor, destination
bicycle
Associati
on
Coefficient
Significa
Dependent variable
nce
n/a
-1.657
0.175
n/a
0.056
0.57
+
0.193
0.088
Associati
on
Coefficient
probability of person n
choosing mode I for travelling
between origin and
destination
probability of person n
choosing mode I for travelling
between origin and
destination
probability of person n
choosing mode I for travelling
between origin and
destination
Road Quality
Study
Rietveld and Daniel
(2004)
Model name
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Independent
Variable
Parkin, Wardman,
Page (2008)
Xin, Handy and
Buehler (2008)
Pavement
vibrations
Proportion nonModel of variation in principal roads
use of the bicycle for with negative
England and Wales
residual life
Proportion
Model of variation in principal roads
use of the bicycle for with negative
England and Wales
residual life
Model 1 (own a bike
Biking comfort
vs. not own a bike)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
biking comfort
not bike)
Model 4 (own a bike
Biking Comfort
vs. not own a bike)
Xin, Handy and
Buehler (2008)
Model 5 (Biked vs. did
Biking Comfort
not bike)
Model 3 (biked
frequently vs.
bike comfort
moderately)
Parkin, Wardman,
Page (2008)
Xin, Handy and
Buehler (2008)
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
-
-0.783
-
-0.3493
+
0.72
n/a
+
n/a
n/a
Significa
Dependent variable
nce
0.718
(t-stat) - proportion of individuals in
8.25
ward x cycling to work
(t-stat) - proportion of individuals in
3.87
ward x cycling to work
Owns a bike (vs. not own a
0.05
bike)
Biked within last seven days
(vs. did not bike within the
ns
last seven days)
Owns a bike (vs. not own a
0.05
bike)
Biked within last seven days
(vs. did not bike within the
ns
last seven days)
ns
biked frequently (vs. biked
moderately)
166
Policy Level
Study
Rietveld and Daniel
(2004)
Model name
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Dill & Carr (2003)
Model 1 (D.C.
excluded)
Dill & Carr (2003)
Model 2
Rietveld and Daniel
(2004)
Dill & Carr (2003)
Rietveld and Daniel
(2004)
Rietveld and Daniel
(2004)
Rietveld and Daniel
(2004)
Rietveld and Daniel
(2004)
Rietveld and Daniel
(2004)
Model 4 (D.C and
NYC excluded)
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Independent
Variable
Municipal budget
municipal
incentives
State spending
per capita on
bke/pedestrian
State spending
per capita on
bke/pedestrian
State spending
per capita on
bke/pedestrian
Associati
on
Coefficient
Significa
Dependent variable
nce
n/a
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
0.771
0.144
% commuting by bicycle
n/a
0.427
0.328
% commuting by bicycle
+
1.021
0.047
% commuting by bicycle
Parking costs
(policy)
(eurocents per
hour)
+
0.0522
(t-value) Share of bicycle use in
4.13
Dutch cities
Speed (compared
with the car)
(policy)
+
0.03392
(t-value) Share of bicycle use in
4.41
Dutch cities
Stop frequency
(policy) (cyclists
stops per
kilometre)
-
-0.0499
(t-value) - Share of bicycle use in
3.63
Dutch cities
Degree of
satisifcation (with
bicycle policies,
provisions, etc.)
+
0.0509
(t-value) Share of bicycle use in
3.50
Dutch cities
Proportion of VVD
voters (main
liberal party)
-
-0.753
(t-value) - Share of bicycle use in
3.27
Dutch cities
167
Employment & Workforce
Study
Model name
Baltes, M. (1996)
Not significant in any
model
Parkin, Wardman,
Page (2008)
Model of variation in
use of the bicycle for
England and Wales
Baltes, M. (1996)
Not significant in any
model
Bicycle-Choice Model
for predicting that a
Cervero and Duncan trip will be made by
(2003)
bicycle
Independent
Associati
Variable
on
% of females age
16 and over in the
n/a
work force
proportion of
small employers
& own account
workers
Percent of males
age 16 and over in
n/a
the work force
Employment
accessibility:
number o jobs (in
10 000s) within 5
n/a
miles of origin
Coefficient
/
-4.1446
Significa
Dependent variable
nce
ns
(t-stat) - proportion of individuals in
13.63 ward x cycling to work
ns
-0.017
% of work trips in 1990 by
bicycle in each MSA
0.106
% of work trips in 1990 by
bicycle in each MSA
probability of person n
choosing mode I for travelling
between origin and
destination
Self Selection
Xin, Handy and
Buehler (2008)
Model 1 (own a bike
vs. not own a bike)
Xin, Handy and
Buehler (2008)
Model 2 (biked vs. did
not bike)
Model 3 (biked
frequently vs.
moderately)
Model 4 (own a bike
vs. not own a bike)
Model 5 (Biked vs. did
not bike)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
Xin, Handy and
Buehler (2008)
self selection
n/a
ns
self selection
+
self selection
n/a
ns
self selection
n/a
ns
self selection
n/a
0.524
0.508
0.05
0.1
owns a bike (vs. does not
own a bike)
Biked within last seven days
(vs. did not bike within the
last seven days)
biked frequently (vs. biked
moderately)
owns a bike (vs. does not
own a bike)
Biked in previous 7 days (vs.
did not bike)
Other Factors
Rietveld and Daniel
(2004)
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Semi-log linear
regression model,
explaining the share
of bicycle use in
cities
Bas de Geus et al
(2008)
Bas de Geus et al
(2008)
Rietveld and Daniel
(2004)
Rietveld and Daniel
(2004)
Bas de Geus et al
(2008)
Parkin, Wardman,
Page (2008)
Baltes, M. (1996)
Gain of time
Hindrance
frequency (policy)
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
n/a
-
-0.0126
Insurance
premium
n/a
Differences between
Cyclists and Noncyclists
Psychosocial
n/a
(Odds Ratio)
1.25
Differences between
Cyclists and Noncyclists
External
obstacles
n/a
(Odds Ratio)
1.03
Differences between
Cyclists and Noncyclists
Model of variation in
use of the bicycle for
England and Wales
Not significant in any
model
Crime (road to
work)
Dichotomous
variable for nonmapped wards
Inverse of MSA
land area
n/a
(Odds Ratio)
1.06
+
0.9376
n/a
(t-value) - Share of bicycle use in
Dutch cities
2.22
multicolli
nearity or Share of bicycle use in
Dutch cities
ns
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.388 study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.916 study
Cycling for Transport at least
once a week to work in the
last 6 months prior to start of
0.8
study
(t-stat)
18.78
ns
proportion of individuals in
ward x cycling to work
% of work trips in 1990 by
bicycle in each MSA
168
