IMPACT OF BIKE LANES AND SIDEWALK IMPROVEMENTS IN NEW ORLEANS, LOUISIANA
A DISSERTATION SUBMITTED ON THE 15TH OF APRIL 2013
FOR THE DOCTOR OF PHILOSOPHY OF PUBLIC HEALTH DEGREE TO THE
DEPARTMENT OF GLOBAL COMMUNITY HEALTH AND BEHAVIORAL SCIENCES
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS OF
TULANE UNIVERSITY SCHOOL OF PUBLIC HEALTH AND TROPICAL MEDICINE
BY
TABLE OF CONTENTS
RESEARCH QUESTIONS AND HYPOTHESES ......................................................................... 17
METHODS ............................................................................................................................... 19
Setting: S. Carrollton neighborhood Study 1: Impact of new bike lane on cycling inS. Carrollton neighborhood
Research Questions Design Data collection Data analysis
Study 2: Intercept interviews of cyclists and pedestrians in S. Carrollton ne.ighborhood Research Questions Design Data collection Data analysis
Study 3: Impact of sidewalk improvements on walking in S. Carrollton neighborhood Research Questions Design Data collection Data analysis
RESULTS .................................................................................................................................. 31 Study 1: Impact of bike lane on cycling in S. Carrollton neighborhood Study 2: Intercept interviews of cyclists and pedestrians in S. Carrollton neighborhood Study 3: Impact of sidewalk improvements on walking inS. Carrollton neighborhood
DISCUSSION ............................................................................................................................ 41 Study 1: Impact of bike lane on cycling in S. Carrollton neighborhood Study 2: Intercept interviews of cyclists and pedestrians inS. Carrollton neighborhood Study 3: Impact of sidewalk improvements on walking in S. Carrollton neighborhood
2
Appendix II.
Baseline Survey
Follow-up Survey
Appendix Ill.
3
ABSTRACT
Overweight and obesity are major public health problems in the United States affecting
both children and adults. Regular exercise and a healthy diet can help people achieve a healthy
weight. Even small amounts of incremental exercise dispersed throughout the day are
beneficial to maintaining cardiovascular health. Active transportation, such as biking and
walking, to meet daily needs such as shopping and going to work or school is one way for
people to fit regular exercise into a daily schedule. Changing the built environment by adding
or improving sidewalks and bike lanes can encourage more people to become physically active.
Few rigorous studies, however, have examined the impact of bike lanes and improved
sidewalks on the amount of cyclists and pedestrians observed using those areas. The present
study evaluates the impact of bike lanes and sidewalk improvements in New Orleans, Louisiana,
by conducting counts of cyclists and pedestrians in areas where the improvements were made.
Counts took place before and after new bike lanes and sidewalks were constructed by the City
using standardized forms throughout eleven hours per day, five days per week, at one
intervention and two comparison streets. Intercept interviews were also conducted with both
cyclists and pedestrians five days per week {Tuesdays, Wednesdays, Thursdays, Saturdays and
Sundays) at both intervention and comparison count locations between the hours of 4:30 to
6:30p.m. The intercept interviews provided information about the original location,
destination and the reason for the trip. This research purported to answer the following
research questions: 1) Does adding exclusive bicycle lanes on existing major streets in New
Orleans significantly increase over baseline the number of people using those streets for
bicyciing; 2) What are the characteristics and preferences of pedestrians and cyclists using
4
improved bike lanes and sidewalks for transportation and recreation, and; 3) Does improving
streets and sidewalks on existing major streets in New Orleans significantly increase over
baseline the number of pedestrians observed using sidewalks in those places. Results of both
cyclist and pedestrian observations show significant increases in the number of people using
the intervention street for biking and walking. The results of the survey showed that
pedestrians and cyclists differed with respect to trip purpose, gender and car and bike access.
5
These studies detail the background, research questions, methods, results and
discussion for research relating the interaction between the built environment and physical
activity, serving ultimately as a means to impact the obesity epidemic. The scope and size of
the obesity epidemic are presented, as well as the link between obesity and the built
environment. Active transportation, i.e. biking and walking, will be reviewed as one potential
mechanism to lessen the epidemic's impact. Also, a conceptual framework to understand why
a person may or may not choose to bike is included.
The overall aim of this research is to determine the impact new bike lanes and sidewalk
improvements have made on levels of biking and walking in New Orleans. After Hurricane
Katrina, the federal government made funds available to the Louisiana Department of
Transportation and Development to repave roads submerged and damaged during the storm.
The "Louisiana Submerged Roads Program" pays for resurfacing of streets and some sidewalks,
while the City of New Orleans Department of Public Works is using funds from the sale of bonds
to pay for striping for bike lanes. This large scale infusion of infrastructure projects represents a
unique opportunity to study the impact of the changes to the built environment through a
natural experiment. Thus, several methods were employed to study the impact on cycling and
walking. Methods employed in this study included observations of cyclists and pedestrians on
intervention and comparison streets throughout the city of New Orleans. In addition, both
cyclists and pedestrians were interviewed at both intervention and comparison locations to
6
understand how the end-user perceives the new improvements and the impact of the
improvements on their behavior.
From the data collected, results indicated an increase of cyclists from baseline to follow­
up using the bike lanes for transportation or recreation. Since comparison streets were also
observed, we were able to determine that this difference was due to the presence ofthe bike
lanes and not simply a shift from another street. Second, we had hoped to see fewer adult
cyclists riding on the sidewalk and riding against traffic, because those two behaviors are unsafe
riding behaviors for cyclists, but we only saw that the lanes had an effect on the proportion of
cyclists riding in the correct direction. Finally, it was found, as expected, that the new lanes and
sidewalks brought more people to the area, and thus made it safer and more attractive for
pedestrians to walk; therefore, an overall increase in the number of pedestrians was observed
as well.
This research shows that upgrading pedestrian and bicycle facilities while completing
routine road resurfacing projects can make a significant impact in the number of people who
are outside walking and biking. If more routine road projects use a "complete streets"
approach upgrading spaces for all modes of transportation, cyclists and, pedestrians will benefit
along with those who are driving.
7
II. BACKGROUND AND SIGNIFICANCE
A. Overweight and Obesity
Daily physical activity and a healthy diet are essential components to achieving a healthy
weight and can prevent chronic diseases such as cancer, heart disease and stroke.1 Overweight
and obesity have become epidemics in the United States today. Some 68 percent of men and
women over 20 years of age living in the United States are either overweight or obese.2
Overweight and obesity are also affecting children, where 16.9 percent of children and
adolescents aged 2-19 are above the 95th percentile and 31.7 percent are above the 851 h
percentile.3 . Obese children and adolescents are at increased risk for obesity once they reach
adulthood.4 ' 5
There are many negative health risks that accompany overweight and obesity.
Overweight and obese adults are at higher risk for hypertension, type 2 diabetes, coronary
heart disease, stroke, gallbladder disease, osteoarthritis, sleep apnea, respiratory problems and
several cancers. 6 Overweight and obese individuals who lose weight can reduce blood
pressure, irrespective of being hypertensive.6 ' 7 Additionally, weight loss can lead to a reduction
in serum triglycerides and an increase in high-density lipoprotein (HDL} cholesterol.6 ' 7
The economic impact of obesity on medical costs is staggering. Previous estimates
showed that the cost of treating obesity among non-institutionalized adults was $86 billion per
year.8 A study published in 2010 that adjusts earlier estimates for participants' under­
representation of actual weight shows the cost to be higher than this.9 The newest estimated
8
cost of treating obesity among non-institutionalized, non-elderly adults is $168.4 billion per
year and treating obesity adds $2,800 per year to an obese person's medical bills.9
According to the New Orleans 2007 Behavioral Risk Factor Surveillance System (BRFSS)
telephone survey, 24.4 percent of adults were classified as obese and 28.6 percent were
overweight; because this is by self-report, obesity rates are likely higher than this.10 The sample
size for the BRFSS for New Orleans is not large enough to categorize these results by race, but
in the entire state of Louisiana, 39.7 percent of African-Americans were obese and 32.0 percent
were ·overweight.10 According to the 2007 Youth Risk Behavior Survey (YRBS), 17 percent of
New Orleans youth were overweight, compared to 14 percent nationally.11
B. Physical Activity
Physical activity alone can lead to modest reductions in weight and increase
cardiovascular fitness over time; therefore, physical activity is recommended as a part of a
comprehensive program to treat obesity and overweight. 6 Children and adolescents aged 6-17
should aim for at least 60 minutes of physical activity per day, with most of that activity being
moderate or vigorous. 12 Adults should aim to have 2 hours and 30 minutes a week of moderate
intensity activity over a week.12 The 2008 Physical Activity Guidelines for Americans also
prescribe that adults can have a shorter duration of physical activity, 75 minutes, if the activity
level is vigorous.12
Among adults, physical activity has important health benefits such as a lowered risk of
early death, stroke, heart disease, type 2 diabetes, high blood pressure and some cancers.12
9
Children also benefit from physical activity; in addition to maintaining a healthy body
composition, they experience enhanced muscular and cardiovascular fitness and improved
bone health.12
Youth and adults in New Orleans were less likely to be physically active than those
nationally: Forty-five percent of New Orleans youth had participated in at least 20 minutes of
vigorous physical activity on three or more of the past seven days compared to 69 percent
nationally. In New Orleans, 38 percent of adults met the recommended levels of physical
activity compared to 49 percent nationally.10 These findings were even more
pronounced among low and moderate income African Americans. A lack of physical activity is
one of the main factors contributing to obesity.10
C. Active Transportation
A comparison of both self-reported and objectively measured health and travel data for
14 countries, 50 U.S. states and 47 U.S. cities found that people who engage in active
transportation (i.e. biking and walking) have lower levels of obesity.13 These activities burn
calories and provide health benefits such as a decreased risk for cardiovascular disease and
weight gain.14 - 17 Even
1 short bouts of physical activity, such as walking to a store or school,
accumulated over the course of a day are an effective way to achieve recommended guidelines
of 30-minutes of physical activity per day to reduce blood pressure.18 Sustained active I
transportation has been found to be associated with healthy weight, particularly when children
begin walking or biking to school in kindergarten and continue through grade 2.19
10
Nationwide, the percentage of school age children who walk or bike to school decreased
68 percent from 1969 to 2001, while the number of children being driven to school has
increased.20 Parental perceptions about the route children take traveling from home to school
impact whether they allow their child to walk or bike.21 Concerns about crime and traffic safety
are associated with lower rates of children walking or biking to school.22 However, efforts to
encourage walking to school through increased programming has been found to be an effective
strategy, Safe Routes to School programs have been associated with higher walking rates. 23 ' 24
Improving the infrastructure for cycling may increase the proportion of adults and
children biking for transportation and/or recreation. Although walking and cycling can be
viewed as dangerous activities, due to the potential for collisions with motorists, evidence
exists that the health benefits from cycling exceed the risk collisions with motorized traffic.25
When more people are observed walking and biking, the less likely any of them will be injured
in traffic collisions.26 ' 27 The more cyclists observed leads to increased safety for all cyclists
because injury rates fall. 28 One possible explanation for this phenomenon is that when
motorists see more cyclists and pedestrians, they reduce their speed.29
New Orleans is well-suited for non-motorized transportation. It is flat and most
neighborhoods are within ten miles of the city center. Additionally, many neighborhoods are a
mix of both residential and commercial uses, giving residents destinations to walk or bike to in
their daily activities. Because 38 percent of residents do not own a car, many residents walk,
bike and/or use public transit to meet their needs.30 A greater proportion of New Orleans
residents bike or walk than residents nationwide (see Table 1.} New Orleans is also ranked 8th
11
in the estimated percent of all trips made by bicycle (0.6%) and by foot (8.0%).31 A greater
proportion of men in New Orleans bike (66% vs 34%) or walk (61% vs 39%) than women.31
Public health researchers and transportation planners have recently shared a common
goal: to increase the number of people using non-motorized transportation to carry out daily
activities. It is estimated that about half of all car trips are less than five miles, a distance easily
covered by bicycle; however,32 few trips are made by bicycle, less than one percent nationwide
(Table 1). Data from the 2001 Nationwide Household Travel Survey show that there is close to
one adult-sized bicycle per household in the U.S., however only 8 percent of adults reported
cycling in the previous week.33
D. Built Environment
There is an increasing amount of evidence that the built environment can promote
physical activity. Infrastructure such as mixed-land use where shops and public services are
dispersed within residential areas, increased housing density, availability of public transit, and
sidewalks, trails and bike lanes can increase the possibility that people will walk or bike to meet
their daily needs.34 - 37 The more attributes present, i.e. local shops, transit, sidewalks, bike
facilities, low-cost recreation facilities, the increased likelihood that an adult living near those
attributes will meet the recommended guidelines for physical activity.38 Local municipalities
can implement traffic calming devices, such as reducing the speed of vehicular traffic, adding
speed bumps and improving pedestrian visibility to enhance pedestrian safety.39 ,4o Lane widths
can be reduced so that cars travel at slower speeds and sidewalk extensions at intersections
can help pedestrians view cars before crossing streets.39 These calming devices have increased
12
the number of pedestrians observed after improvements took place and improved outdoor play
among children and adolescents.41
People are more physically active in neighborhoods that are well designed for walking
and bicycling.42 Building infrastructure for safer bicycling is one way to promote physical
activity.43 Having regular access to well-maintained sidewalks is associated with increased
walking.44 Conversely, being obese is associated with a lack of sidewalks, poor sidewalk quality,
and no nearby non-residential destinations.45 One cross-sectional study analyzing data from 43
large cities in the United States found that for every 1 per.cent increase in the length of on­
street bike lanes, there was an 0.31 percent increase in bike commuters.46 Experienced cyclists
prefer on-street bike facilities, such as bike lanes (where cyclists ride Or:J the street but are
separated from vehicle traffic by striping) and shared-lane markings (where cyclists ride on the
street, are not separated, indicated by a cyclist with two arrows above the cyclist), to off-street
bike paths that may take the cyclist out of their way.25 However, women cyclists prefer low­
traffic streets and felt less comfortable than men riding on off-street bike paths.47
E. Conceptual Framework:
To investigate the factors that promote the outcome of cycling for transportation or
recreation, it is helpful to use the conceptual model for bicycling purpose developed by Xing et
al. 48 This model, Figure 1, partitions factors that influence cycling into the individual, social­
environmental and physical environmental realms.
Individual factors, such as whether someone owns a bicycle, knows how to ride it and
enjoys riding play a substantial role in the outcome of whether someone rides a bicycle.49
13
Attributes of the physical-environment also play a strong role, such as land use, climate,
infrastructure and the built environment. 5° Social environmental factors, such as bicycling
culture, perceptions of safety, crime, and beliefs about physical activity, will also influence
• • 1 • so
dec1s1ons to b1cycle.
Previous research on factors affecting cycling have found that cyclists prefer bike lanes
to riding on open streets5 \ but it is unknown whether the bike lanes will attract people who are
new to cycling or if people are using the lanes to substitute for another mode of travel. This
project will contribute to that end by investigating whether cyclists would have ridden a bike
for their trip if there had been no lane present.
No before and after studies of bicycle lanes that collect demographic data on the user
type (age group, gender, raceL direction of travel (with or against traffic) in southern U.S. cities
have been found. Researchers in other well-known bicycling destinations such as Berkeley, CA
and Portland, OR have conducted count studies before and after installation of bike lanes;52 ' 53 54
however, only one of these studies included counts on nearby streets where there were no
lanes, and it was likely that cyclists were diverted to the new lanes from another street. 52
Additionally, none combine interviews with cyclists during periods when manual counts occur.
Manual counts also do not feature demographic data to examine the impact of new
infrastructure on levels of cycling.
Additionally, it is possible that an increased number of cyclists observed would lead to
increased pedestrian activity along bike routes. Pedestrians may feel safer and thus more likely
to go outdoors to be physically active if they see more cyclists outside using the new bike lanes.
14
In a study conducted in 2004-2007, Farley et al observed an increase in levels of physical
activity throughout a neighborhood after a schoolyard was opened afterschool and
weekends. 55 With more people outside being active, people may be more likely to consider the
area as a place that physical activity is welcome and even expected. This idea of creating "eyes
on the street" is not new, author and activist Jane Jacobs suggested that communities thrive
where there are a mix of uses connected by a transportation system that encourages people to
go outside and be seen.56
Unlike biking, which requires a bike and other equipment, walking is a form of physical
activity that is available to most people. Moreover, walking can be done for transportation
related purposes, such as going to work or school, or recreation and requires no special training
or instruction. Walking is frequently reported among adults meeting physical activity guidelines
as the preferred methods of physical activity. 57 Promoting walking as a means for increasing
physical activity has promise, even among subgroups that are at risk for inactivity. 58 Previous
research studying the impact of infrastructure changes such as building sidewalks, adding
crosswalks and reducing speeds, have found that pedestrian volume can increase-- a recent
publication by the Transportation Research Board examined four case studies of improvements
and fou·nd pedestrian volumes increased from 40-400 percent after improvements.59
Similar to biking there are factors influence the decision to walk for recreation or
transportation. The presence and condition of sidewalks can influence walking, as does the
distance and mix of features located along the route.60 Perceptions of the environment may be
influenced by personal characteristics such as motivation to walk, attitude towards walking for
15
transportation or recreation and prior experiences walking. 61 Additionally, aesthetics of the
surrounding environment is.also a factor, where walkers prefer places to walk with trees,
interesting buildings and natural landscapes. 62
F. Policy Implications
This research has several important policy implications locally. The City of New Orleans
adopted a "Complete Streets" policy for its Department of Public Works in December 2011 and
will implement this policy beginning December 2012. This policy would force city engineers to
design streets that meet the needs of all types of users: pedestrians, cyclists, motorists, and
people with mobility impairments. To chart the progress of this complete streets policy, City
Council members and DPW administrators have expressed that they would like to see what the
impact has been of their decisions to include non-motorists in road resurfacing projects thus
far. Additionally, a complete streets policy needs continued support at the grassroots level, and
there are many neighborhood organizations in the City of New Orleans that have not formed an
opinion about how bike lanes and sidewalks can improve their neighborhoods. This research
can help neighborhoods consider the benefits of active transportation as the debate whether to
support the DPW and City Council in enforcing a policy.
16
Ill. RESEARCH QUESTIONS AND HYPOTHESES
It is anticipated that this study will result in three publishable manuscripts based upon the
following research questions:
Study 1:
a. Does adding exclusive bicycle lanes on existing major streets in New Orleans significantly increase over baseline the number of people using those streets for bicycling?
a. Will there be an increase in the overall number of cyclists observed after the lanes are constructed?
b. Will the rate of women cyclists observed increase more than the rate of men per day?
c. Will there be increases in both the rates of white and black cyclists observed per day? d. Will the rate of children cyclists observed increase per day?
b. Does the installation of bike lanes on existing major streets in New Orleans contribute to a reduction in the number of people engaging in unsafe behavior such as cycling against traffic and on the sidewalk?
a. Will the proportion of people riding in the correct direction, e.g. with traffic increase?
b. Will the proportion of people riding on the street and not the sidewalk increase?
Study 2:
1) What are the characteristics and preferences of pedestrians and cyclists using improved bike lanes and sidewalks for transportation and recreation?
a. What are the demographic characteristics of respondents by mode of travel and purpose of trip?
b. Are there differences between cyclists and pedestrians with respect to the reason for their trip and distance traveled?
c. Are there differences between cyclists and pedestrians with respect to reaching a 3D­
minute threshold of daily physical activity? d. Do cyclists report that the presence of the bike lanes influenced their decision to ride a
bicycle on 5. Carrollton Ave? e. Are people who bike or walk likely to substitute an active mode of transit for a non­
active mode?
Study 3:
1) Does improving streets and sidewalks on existing major streets in New Orleans significantly increase over baseline the number of pedestrians observed using sidewalks in those places?
a. Will there be an increase in the overall number of pedestrians using the sidewalks in the same areas over the observation period?
b. Will the rate of women using the sidewalks per day in improved areas increase more than the rate of men per day?
c. Will there be increases in both the rates of whites and blacks observed using sidewalks in improved areas per day?
d. Will the rate of children using the sidewalks in improved areas increase per day? e. Will the proportion of people walking on the sidewalk and not on the street increase when
the new lanes and improved sidewalks are constructed?
18
Setting
The setting for all three studies is in New Orleans, LA, a city with a flat topography and a
temperate climate and well suited to cycling. Baseline (September 2009) and follow-up
observations (September 2010), took place at the same location mid-block on three parallel
streets: South Carrollton Avenue (the street with the n'ew bike lane), and Dublin and Short
streets (streets one block from and parallel to S. Carrollton without bike lanes, see Map 1 in
Appendix 1). S. Carrollton Avenue is an urban principal arterial street that is bounded by the
Mississippi River Levee on its west end and by Bayou St. John on its east end and connects
several neighborhoods. The observation area and survey location was between St. Charles Ave
and Claiborne Ave; two urban principal streets. The observation location was between two
neighborhoods that differ substantially on race, poverty status and access to vehicles.
Residents living northwest of S. Carrollton were 58 percent African American, 33 percent below
the poverty line, and 35 percent without access to a car. 63 Residents east of S. Carrollton were
28 percent African American, 26 percent below the poverty line, and 18 percent without access
to a car.63
The intervention included two types of infrastructure improvements, which happened
concurrently, the installation of new bike lanesand the enhancement of existing pedestrian
facilities. Both improvements occurred at the same time, performed by work crews contracted
by the State of Louisiana and the city of New Orleans. The one mile dedicated bike lane on S.
Carrollton Avenue is located between the travel lane and the parking lane and was completed
19
in June 2010. Bike lanes were striped on both sides ofthe road and are 5 feet wide. Sidewalks
existed on both S. Carrollton Avenue and the adjacent side streets, however, the quality of the
sidewalks on S. Carrollton varied from good to poor and there were few curb ramps. The
intervention included resurfacing the sidewalks on S. Carrollton and installing curb ramps at and
crosswalks at intersections. The sidewalk quality on the two adjacent side streets could be
characterized as poor, with numerous cracks, uneven surfaces, few curb ramps at intersections
and narrow width {see photos in Appendix 1). A May 2011 report of school zone infrastructure
in New Orleans found that 20 percent of sidewalks and 60 percent of curb ramps were poor or
missing in a two-block radius of the middle school located at the observation area. 64 There is
one 11ft-wide travel lane on either side of the road, separated by a 60ft-wide median. An
electric streetcar runs in the median, stopping approximately every 2 blocks to pick up and drop
off passengers. The posted speed limit on S. Carrollton is 35 mph and the average daily traffic
. reported by the metropolitan planning organization was 23,900 vehicles in 2008. 65 This section
of S. Carrollton includes a mix of land uses, including low-density residential and commercial.
Several destinations including public and private schools, churches, and businesses are located
along the corridor. Additionally a 5.6 acre neighborhood park is located at one end and the
Mississippi River levee and St. Charles Avenue are located at the other end of the corridor.
Methods Study 1: Impact of bike lane on cycling in S. Carrollton neighborhood
Research Questions
1} Does adding exclusive bicycle lanes on existing major streets in New Orleans significantly increase over baseline the number of people using those streets for bicycling?
a. Will there be an increase in the overall number of cyclists observed after the lanes are constructed?
20
b. Will the rate of women cyclists observed per day increase more than the rate of men per day?
c. Will there be increases in both the rates of white and black cyclists observed using the lanes per day?
d. Will the rate of children cyclists observed using the lanes increase per day? 2) Does the installation of bike lanes on existing major streets in New Orleans contribute to a
reduction in the number of people engaging in unsafe behavior such as cycling against traffic and on the sidewalk?
Design
a. Will the proportion of people riding in the correct direction, e.g. with traffic increase?
b. Willthe proportion of people riding on the street and not the sidewalk increase?
This quasi-experimental study includes pre-test and post-test observations of an
intervention street and a control area. The intervention being tested is the impact of striping
bike lanes on existing roads. This design is the most appropriate to test the effects of this
natural experiment as it is not possible to randomize people into using or not using the bike
lane for their choice of travel.
Data Collection
Using standardized forms adapted from an earlier study observing cyclists, pairs of
trained observers tallied the number of cyclists in September 2009 (baseline) and September
2010 (follow up) for 10 days: 6 weekdays and 4 weekend days for an 11-hour period from 7 am
to 6 pm each day (see Appendix 11). 66 Observers counted cyclists who crossed an imaginary
plane that ran from the sidewalk into the street on both sides of the road. The average
temperature over data collection was 79.4 and 78.1 degrees Fahrenheit in 2009 and 2010,
respectively. On days of heavy rain, observations ceased and another full day was completed
on the same day the following week. Data collected included the number of males, females,
21
adults and children riding a bicycle with traffic, against traffic and on sidewalks on S. Carrollton
Avenue and the two adjacent side streets, Short and Dublin Streets. Race of observed riders
was categorized as white, black or other by observers. Observers were certified for data
collection when their overall proportion of agreement was greater than 80 percent with a
trainer for location (street, neutral ground/median or sidewalkL gender and age group (adult
male, adult female, boy, giriL and race (white, black or other).
Procedures for observation of human subjects were approved and followed according to
guidelines established by the Institutional Review Board of Tulane University.
Data Analysis
Counts of male, female, adult, child, black, white and other cyclists were totaled for
each hour and day. Cyclists observed on Short and Dublin streets were summed and averaged
as the adjacent side streets where there was no bike lane. Means and standard deviations
were calculated for the number of people observed cycling before and after the bike lanes were
installed. We tested two main hypotheses: 1) more people would be observed cycling on the
intervention street after the bike lane was striped and 2} people would be more likely to ride
with traffic and in the bike lane instead of on the sidewalk. The unit of analysis was day for the
first hypothesis and individual cyclists for the second hypothesis. The outcome for the first
hypothesis was the number of people observed cycling. For the second hypothesis, the
outcome was binary, indicating appropriate vs inappropriate behavior (e.g. riding on the street
or sidewalk and with or against traffic). We used negative binomial regression to test the first
hypothesis and binary logistic regression to test the second. Separate models were constructed
22
for gender, age and race groups. We considered a P-value of 0.05 or less to be statistically
significant. Models included predictors such as time (pre/postL location (intervention or
adjacent streets) and location by time interactions. Additional interactions were included for
post-hoc analysis where appropriate. See below for a list of the models constructed:
Model 1: log(total neighborhood) =constant+ b1(time)
Model 2: log(S.Carrollton) =constant+ b1(1ocation) + b2(time) +b3 (1ocation*time)
Model 3: log (Side Streets)= constant+ b1(1ocation) + b2(time) +b3(1ocation*time)
Model 4: log( count)= constant+ b1(1ocation) + b2(time) +b3(1ocation*time) + b4(age) + b5(time*age) + b6(group*time*age)
Model 5: log(count) =constant+ bl(location) + b2(time) +b3 (1ocation*time) + b4(race) + b5(time*race) + b6(group*time*race)
Model 6: log( count)= constant+ b1(1ocation) + b2(time) +b3(1ocation*time) + b4(gender) + b5(time*gender) + b6(group*time*gender)
Model 7: log odds(street vs. sidewalk)= constant+ b1(1ocation) + b2(time) +b3(1ocation*time)
Model 8: log odds(with traffic vs. against)= constant+ b1(1ocation) + b2(time) +b3(1ocation*time)
Methods Study 2: Intercept interviews of cyclists and pedestrians in S. Carrollton neighborhood
Research Questions:
1) What are the characteristics and preferences of pedestrians and cyclists using improved bike lanes and sidewalks for transportation and recreation?
a. What are the demographic characteristics of respondents by mode of travel and purpose of trip?
b. Are there differences between cyclists and pedestrians with respect to the reason for their trip and distance traveled?
23
c. Are there differences between cyclists and pedestrians with respect to reaching a 3D­ minute threshold of daily physical activity?
d. Do cyclists report that the presence of the bike lanes influenced their decision to ride a bicycle on 5. Carrollton Ave?
e. Are people who bike or walk likely to substitute an active mode of transit for a non­ active mode?
Design:
This cross-sectional study consists of a survey to ask pedestrians and cyclists information
about their travel choices, preferences and amounts of physical activity. A pre-post survey was
considered, however, it was not possible to achieve a large enough sample size of both
pedestrians and cyclists at baseline.
Data collection:
This cross-sectional study took place on S. Carrollton Avenue in New Orleans, LA, from
mid-September to the end of October 2010. Surveys ~ook place at the same location mid-block
on three parallel streets: South Carrollton Avenue (the street with a new bike lane, sidewalks
and curb ramps), and Dublin and Short streets (streets one block off and parallel to S. Carrollton
Avenue without bike lanes, but both side streets have sidewalks and some curb ramps see
photos in Appendix 11).
Intercept interviews were conducted with both cyclists and pedestrians five days per
week (Tuesdays, Wednesdays, Thursdays, Saturdays and Sundays) at both S. Carrollton Avenue
and the two adjacent side streets between the hours of 4:30 to 6:30 p.m. using a survey initially
developed by Kevin Krizek from the University of Colorado at Boulder. 67 The survey was
24
adapted to include information about race, preferred cycling location and physical activity (see
Appendix II). These afternoon hours were included upon the recommendation of other
researchers using intercept interviews in other cities so that data could be comparable across
sites. Additionally, the hours selected would likely be either commuting time or immediately
after-work time. Thus people could be walking or biking to come home from work or to engage
in physical activity immediately following work. Lastly, it was thought that people may be more
likely to stop on their way home from work rather than their way to work. Interviewers
collected data from September 30, 2010 through October 12, 2010.
Variables collected included respondents' original location, destination and the reason
(e.g. transportation or recreation) for the trip. We also asked if respondents substituted this
trip for another type of travel, active or passive. Cyclists were asked if the bike lane specifically
drew them to that location. Respondents were also asked how many days per week they walk
or bike for transportation or recreation purposes and the number of minutes spent doing those
activities. Demographic information such as age group, gender and race was observed by the
interviewer.
Interviewers approached the first individual encountered (the first individual
encountered after completing the last survey and accompanying documentation information
about the observed individual) who was walking or cycling to participate in the survey.
Individuals 18 years or older were included. If a group of individuals (e.g., a platoon or group of
friends) were encountered by the interviewer, the individual with the most recent birthday
(month and day) was surveyed. A tally was kept of all individuals who declined to participate as
well as a record of their gender, race and mode of travel (walking or biking).
25
Interviewers received a minimum of three hours of classroom training and one hour of
practice interview time in the field with the project coordinator. Procedures for observation of
human subjects were followed according to guidelines established by the Institutional Review
Board ofTulane University.
Data analysis:
To analyze the survey data means and frequencies were calculated for all continuous
and categorical variables respectively. Trip distance for transportation trips was computed in
several steps. First, the starting intersection and ending location was imputed into Google maps
and the distance between the two locations was entered as the trip distance. This continuous
variable was then categorized into three separate distances of transportation trips: trips less
than or equal to one-quarter mile, trips greater than one-quarter but less than one mile and
trips one mile and greater. This approach was necessary because some respondents indicated
starting locations that yielded trip distances of several miles or more. These distances are
possible, as the respondent could have taken transit; therefore it was logical to collapse these
data into three categories that generally encompass trips for transportation in the literature.68
Respondents were also asked if they were combining this trip with another stop before
reaching their destination. To assess the amount of physical activity spent by walking or biking,
respondents were asked how many days per week and minutes per day they usually spend
walking and biking for transportation and recreation. Then, the number of minutes was
multiplied by the number of days spent doing that particular activity. A threshold of 30-
minutes of daily physical activity was used because reaching this minimum has important
26
health benefits. 12 Two sample t-tests and one-way analysis of variance were conducted to
assess bivariate relationships between continuous variables such as number of car trips and
categorical variables such as number of utilitarian bike trips per week. Chi-square tests of
independence to compare two categorical variables were also conducted. Results with p-values
less than 0.05 were deemed statistically significant.
Methods: Study 3 Observations of Pedestrians
Research questions
1) Does improving streets and sidewalks on existing major streets in New Orleans significantly increase over baseline the number of pedestrians observed using sidewalks in those places?
a. Will there be an increase in the overall number of pedestrians using the sidewalks in the same areas over the observation period?
b. Will the rate of women using the sidewalks per day in improved areas increase more than the rate of men per day?
c. Will there be increases in both the rates of whites and blacks observed using sidewalks in improved areas per day?
d. Will the rate of children using the sidewalks in improved areas increase per day? e. Will the proportion of people walking on the sidewalk and not on the street increase
when the new lanes and improved sidewalks are constructed?
Study design
This quasi-experimental study includes pre-test and post-test observations of an
intervention street and a control area. The intervention being tested is the impact of improving
the built environment for walking by repaving sidewalks and adding crosswalks and curb cuts to
intersections on existing roads. This design is the most appropriate to test the effects of this
natural experiment as it is not possible to randomize people into using or not using sidewalks
for their choice of travel.
27
Data Collection
Using standardized forms adapted from an earlier study observing cyclists, pairs of
trained observers tallied the number of pedestrians in September 2009 (baseline) and
September 2010 (follow up) for 10 days: 6 weekdays and 4 weekend days for an 11-hour period
from 7 am to 6 pm each day(see Appendix 111).66 Observers counted pedestrians who crossed
an imaginary plane that ran from the sidewalk into the street on both sides of the road. The
average temperature over data collection was 79.4 and 78.1 degrees Fahrenheit in 2009 and
2010, respectively. On days of heavy rain, observations ceased and another full day was
completed on the same day the following week. Data collected included the number of males,
females, adults and children walking in the street, neutral ground/median and on sidewalks on
S. Carrollton Avenue and the two adjacent side streets, Short and Dublin Streets. The adjacent
side streets did not include a median (see map 1 and photos 1-5). Race of observed pedestrians
was categorized as white, black or other by observers. Observers were certified for data
collection when their overall proportion of agreement was greater than 80 percent for location
(street, neutral ground/median or sidewalk), gender and age group (adult male, adult female,
boy, girl), and race (white, black or other).
Procedures for observation of human subjects were approved and followed according to
guidelines established by the Institutional Review Board of Tulane University.
Data Analysis
Counts of male, female, adult, child, black, white and other pedestrians were totaled for
each hour and day. Pedestrians observed on Short and Dublin streets were summed and
28
analyzed as the adjacent side streets where there were no sidewalk improvements and no bike
lane. Means and standard deviations were calculated for the number of people observed
walking before and after the sidewalks improvements and bike lanes were installed. We tested
two main hypotheses: 1} more people would be observed walking on the intervention street
after the sidewalk improvements and bike lane was striped and 2) people would be more likely
to walk on the sidewalks rather than in the street. The unit of analysis was day for the first
hypothesis and individual pedestrians for the second hypothesis. The outcome for the first
hypothesis was the number of people observed walking. For the second hypothesis, the
outcome was binary, indicating appropriate vs inappropriate behavior (e.g. walking in the street
or sidewalk). Negative binomial regression was used to test the first hypothesis and binary
logistic regression to test the second. Separate models were constructed for gender, age and
race groups. A P-value of 0.05 or less was considered to be statistically significant. Models
included predictors such as time (baseline/follow-upL location (intervention or adjacent
streets) and location by time interactions. Additional interactions were included for post-hoc
analysis where appropriate. Models used for analysis are listed below:
Modell: log(total neighborhood)= constant+ b1(time)
Model 2: log(S.Carrollton) =constant+ b1(1ocation) + b2(time) +b3(1ocation*time)
Model 3: log (Side Streets)= constant+ b1(1ocation) + b2(time) +b3(1ocation*time)
Model4: log(count) =constant+ b1(1ocation) + b2(time) +b3(1ocation*time) + b4(age) + b5(time*age) + b6(group*time*age)
Model 5: log( count)= constant+ b1(1ocation) + b2(time) +b3(1ocation*time) + b4(race) + b5(time*race) + b6(group*time*race)
29
Model 6: log{ count)= constant+ b!{location) + b2{time) +b3{1ocation*time) + b4{gender) + b5{time*gender) + b6{group*time*gender)
Model 7: log odds{street vs. sidewalk) =constant+ b1{1ocation) + b2{time) +b3{1ocation*time)
30
V. RESULTS
Study 1
Table 2 shows the number of cyclists observed riding on S. Carrollton Avenue and
adjacent side streets before and after a new bike lane was striped during both weekdays and
weekends. In the Carrollton neighborhood overall, an average of 62.5 (±28.8} cyclists at
baseline was observed. Of those people, there were more males than females, adults than
children and whites than blacks (Table 2}.
To determine if there was a change in the number of riders, a model using time
(pre/post} as a single predictor was examined. This model was significant; there was an overall
increase in the mean number of people cycling per day on all three streets combined from
baseline to follow-up [62.5 (±28.8} vs 110 (±109}; Z=8.97, p<O.OOO].
To test if the number of riders changed more on S. Carrollton than on the adjacent side
streets a model that included location (S. Carrollton or side streets}, time (pre and post}, and
location by time interaction was examined. The location by time interaction was significant
(Z=24.27, p<O.OOO}; therefore, additional analyses comparing the number of riders at each
location from baseline to follow-up was performed. More cyclists were observed on S.
Carrollton than the combined adjacent side streets [79.2 (±30.5} vs 54.4 (±24.1}, Z=8.50,
p<O.OOO] before the lanes were striped. The average number of cyclists increased from
baseline to follow-up on S. Carrollton [pre 79.2(±30.5}, post 257.1 (±50.9}; Z=23.37, p<O.OOO]
but decreased on the adjacent side streets (pre 54.4 (±24.1}, post 36.4 (±16.1}; Z= -10.79,
p<O.OOO].
31
The impact of the intervention on different subgroups, such as gender, race and age was
then compared. Location, time, location by time, demographic group as main effects in
negative binomial regression models was included. Also included were three first order
interactions of the demographic groups, location by demographic group (gender: male or
female; race: black or white; age: adult or youth), time by demographic group, and a three-way
interaction of time by location by demographic group.
There was an increase in both the number of male and female cyclists on S. Carrollton
and a decrease in the number of male and female cyclists on the side streets {Figure 2).
Additionally, males and females responded differently to the intervention of new bike lanes on
S. Carrollton, tested in the model which included a three-way interaction of location by time by
gender, (Z=- 2.53; p<O.Oll). This differential response can be seen in the ratio of post to pre
ridership by gender. The ratio for females was 4.69 while for males it was 3.12; thus, the
increase in cyclists was greater among females than males on S. Carrollton after the lane was
striped.
Both whites and blacks showed an increase in cycling after the lanes were striped along
with a decrease on the adjacent side streets (Figure 3). The three-way interaction of race by
location by time, however, was not significant (Z= -1.26, p=0.206). Therefore, the change
between racial groups was not significantly different by location, i.e. striping the bike lane did
not have a differential effect on whites than blacks.
32
The number of youth and adults cycling on 5. Carrollton increased after installation of
the bike lane (Figure 4). The three-way interaction term of age group, location and time was
significant (Z=-2.14, p=0.032L this is due in part to the decrease in adults on the side streets.
The proportion of people observed cycling on the sidewalk changed differentially for 5.
Carrollton and the adjacent side streets (Z=-3.87, p<O.OOO). There was no change in the
proportion of people observed riding in the street versus the sidewalk on S. Carrollton after the
bike lane was striped (pre 93%, post 93%; Z= -0.24, p=0.81). There was a significant decrease in
the number of people observed riding in the street on the side streets after the lane was
installed on S. Carrollton (pre 99.5%, post 97.8%; Z = -4.03, p<O.OOO).
After the bike lane was constructed on S. Carrollton, the proportion of riders observed
traveling with the direction oftraffic increased (pre 92.8%, post 95.6%; Z= 2.93, p<0.003) on the
intervention street. The proportion of people traveling with traffic on the side streets,
however, decreased (pre 96.6%, post 93.5%; Z=-3.05, p=0.002).
Study 2
Table 3 shows the demographic information of all respondents and cyclists and
pedestrians interviewed on 5. Carrollton Avenue in the late September/October 2010. The
average temperature was 78 degrees Fahrenehit over the survey data collection period.
Information on the number of people who refused to be interviewed is presented in table 4. A
total of 1A72 pedestrians were approched over the course of data collection. Five hundred
seventy-two surveys were completed and 890 pedestrians and cyclists refused, resulting in a
completion rate of 38.9 percent. The majority of refusals were from cyclists. The majority of
33
respondents were male, appeared to be under 31 years of age and white. Respondents
reported that their primary reason for their trip was to travel to or from a destination. Trip
start and end point data were collected and trip distance was derived with most trips being
greater than a distance of one mile. Most respondents had at least one working car and one
working bicycle in their household.
Table 3 also shows the proportion of cyclists and pedestrians by demographic factors of
gender, age group, race, trip purpose and car access. Of those interviewed, cyclists were more
likely to be male than female [70.3 vs 29.7; X2=24.9023, p<0.001], while pedestrians included an
equal proportion of males and females. There were differences in trip purpose, trip distance,
car access and bike access by mode of travel. Cyclists were more likely to be headed to or from
a destination than out for recreation [72.0 vs 28.0; X2=28.7601, p<0.001 ), while pedestrians
were equally likely to be out for recreation or going to or from a destination. Fewer cyclists
than pedestrians had access to a car [62.5 vs 78.0; X2=16.9681, p<0.001] and there were more
cyclists than pedestrians taking trips greater than one mile [68.9 vs 23.4; X2=64.1988, p<0.001].
Fewer pedestrians than cyclists reported having one or more bicycles in their household [68.3
vs 98.2; X2=91.6679, p<0.001].
Sixty percent (60.34% Cl: 56.4, 64.3) of all respondents indicated that they make
transportation trips by bike on a weekly basis. The mean number of weekly trips taken by car
was lower among those that reported biking for transit weekly (mean=7.2, sd=8.3} than among
those who reported walking for transit weekly (mean=9.2, sd=9.2); these findings were
statistically signifcant (t= 5.8674, p<0.001). Additional analysis ofthe number of days {1-7) a
34
bike trip for transportation was taken illustrated that the number of car trips per week is
significantly lower for respondents who bike for transportation more frequently F(6,243) = 7.64,
p<O.OOl. For instance, those who took one transportation bike trip per week, averaged 11.8 car
trips, while those who biked for transportation seven days averaged 3.5 car trips. However,
there was no significant difference in trips by car among those who walked for transportation
versus those who did not (t=0.9409, p=0.3473).
Table 5 presents demographic variables by trip purpose, differences in the table are
between those who were walking or biking to or from a destination versus those who were
walking or biking for recreation. Sixty-one percent of respondents reported biking and walking
to or from a destination at the time ofthe intercept interview [60.7 (destination) vs 39.3
(recreation)]. There were differences in the age groups of people out for transportation vs
those out for recreation, a greater proportion of those out for tranportation were in the less
than age 31 age group than those out for recreation [59.2 vs 45.4; X2= 10.2909, p=0.006].
Those who reported their trip was for transportation rather than recreation had less car access
compared to those who were out for recreation [58.8 vs 89.5; X2=63.0254, p<0.001]. People
out for transportation or recreation were more likely to be white than another racial group.
However, a higher proportion of blacks were out for transportation rather than recreation [16.4
vs 9.2; X2=7.5784, p=0.023]. People who were taking a transportation related trip were more
likely to report that they were combining the trip with another stop, than those who were out
for recreation to do so [43.2 vs 30.5; X2=9.3211, p=0.002].
35
Table 6 shows the perceptions of cyclists when asked about the bike lane. These
questions about the bike lane were not asked of pedestrians, only ofthose respondents who
were riding a bike at the time they were stopped for the survey. Forty percent of cyclists
surveyed responded that they only started riding on S. Carollton since the bike lane was
installed. All respondents overwhelmingly were positive about the bike lane, reporting that its
presence played a strong factor in their decision to ride on S. Carrollton. Fifty-one percent
indicated that the presence of the bioke lane was the most important reason why they choose
S. Carrollton for their trip and 88.8 percent either agreed or strongly agreed that the bike lane
influenced their decision to ride on S. Carrollton.
A third of respondents indicated that they would have driven or ridden in a car if they
were unable to bike or walk for their current utilitarian trip (see Table 7). Cyclists were more
likely to report choosing a walking trip if they could not have biked, while pedestrians were
more likely to report choosing driving or riding in a car (t=2.3099, p=0.0215). Table 8 presents
what cyclists new to riding on S. Carrollton Ave would have done if they could not have ridden
their bike. Forty-four percent indicated that they would have driven a car or ridden in a car
instead of biking. Twenty-one percent said that they would have walked.
Table 9 examines the prevlance of walking and cycling 30 minutes or more per day
among those surveyed. Sixty-nine percent indicated that they walk on average 30 minutes or
more per day and 76 percent of those surveyed indicated that they bike an average of 30
minutes or more a day. Among females, the cycling prevalence was signifcantly lower than the
whole sample (t=-2.0897, p<O.OS). Cycling prevalence of at least 30 minutes per day was higher
36
among respondents who did not have access to a car compared to those that did [88.2 vs
71.6;X2=7.3035, p=0.007].
Study 3
Table 10 shows the number of pedestrians observed on 5. Carrollton Avenue and
adjacent side streets during both weekdays and weekends before and after a new bike lane and
sidewalks were installed. At baseline, an average of 132.5(±124.2) pedestrians per day were
observed on these streets. Of those people, we observed more adults than children and whites
than blacks, but no significant difference in gender at baseline (see table 10}. More pedestrians
were observed on 5. Carrollton than the combined adjacent side streets [296.7 (±73.9) vs50.7
(±15.6), Z=48.76, p<O.OOO] at baseline.
Analysis to evaluate any change in number of observed pedestrians used "time"
(baseline and follow-up) as a single predictor in the first model. An overall increase in the mean
number of pedestrians per day was observed for all three streets combined (5. Carrollton and
side streets) from baseline to follow-up [132.5(±124.2) vs 185.4(±185.9); Z=4.08, p<0.001].
To test if the number of pedestrians changed more on 5. Carrollton than on the adjacent
side streets a model that included location (5. Carrollton or side streets), time (baseline and
follow-up), and location by time interaction was examined. Because the location by time
interaction was significant, (Z=5.75, p<0.001), additional analysis indicated that a greater
increase in pedestrians was observed on 5. Carrollton than on the adjacent side streets
combined from baseline to follow-up. The average number of pedestrians increased from
baseline to follow-up on 5. Carrollton [baseline 296.7(±73.9), follow-up 443.2 (±52.9); Z= 8.98,
37
p<0.001] and also increased on the adjacent side streets (pre 50.7 (±15.6), post 57.0 (±25.3); Z=
2.27, p=0.023].
Sub-Group Comparisons
Gender, race and age were sub-groups identified for analysis. The binomial regression model
included location, time and demographic group as main effects. Three first order interactions
of the demographic groups, location by demographic group (gender: male or female; race:
black or white; age: adult or youth), time by demographic group, and a three-way interaction of
time by location by demographic group were also included.
Gender
On S. Carrollton an increase in both male and female pedestrians was observed (Z= 1.86,
p=0.062), while on the adjacent side streets a small increase in the number of male pedestrians
was observed, but not female pedestrians (Figure 5). Males and females did not respond
differently to the sidewalk improvements and new bike lanes on S. Carrollton, tested in the
model which included a three-way interaction of location by time by gender, (Z= 1.86, p=0.062).
Since this interaction term was borderline significant, an additional regression model examined
the interaction of gender and time on S. Carrollton only, to determine if there were differences
between male and female pedestrians between baseline and follow-up. This interaction was
not significant; therefore, there were no differences between male and female pedestrians
prior to and after the new infrastructure improvements on S. Carrollton (Z=0.26, p=0.798).
38
Race
On S. Carrollton an increase in both white and black pedestrians was observed (Figure
6}. On the side streets, there was an increase in the number of white pedestrians but a
decrease in black pedestrians. The three-way interaction of race by location by time, was
significant (Z = -2.49, p=0.013}, illustrating that the change between racial groups was
significantly different by location, i.e. the improvements in the sidewalks and striping the bike
lane had a differential effect on whites than blacks. To determine which location had the most
differential effect, on S. Carrollton or the adjacent side streets, two additional analyses were
conducted. First, a model with an interaction term of race and time, limited only to S. Carrollton
was examined, but the interaction was not significant (Z=-1.30; p=0.194}. A second model with
the interaction term of race and time, but limited to the adjacent side streets, was used; this
interaction term was significant (Z=2.74; p=0.006}. Thus blacks and whites on the side streets
behaved differently than on S. Carrollton; on the side streets, the number of blacks decreased
while the number of whites increased. On S. Carrollton, the number of blacks and whites both
increased, however, the increase in the number of whites we observed was greater than the
number of blacks. The ratio of follow-up to baseline for blacks on S. Carrollton was 1.3 while
for whites it was 1.6; further explaining the borderline significant difference on the intervention
street for the different racial groups.
Age
The number of youth and adults walking on S. Carrollton increased after the sidewalk
improvements and installation of the bike lane (Figure 7). The three-way interaction term of
39
age group, location and time was significant (Z=4.70, p=O.OOO), and as in other subgroup
analyses, two additional models were created to test the effect of the intervention on S.
Carrollton and on the side streets. Both models included an interaction term for age group and
time, the interaction term was highly significant on the side street analysis (Z=-4.75; p=O.OOO),
but borderline significant on S. Carrollton {Z=1.9; p=0.058). The number of youth increased on
S. Carrollton, and decreased on the side streets; while the number of adults also increased on S.
Carrollton, but unlike the youth, also increased on the side streets (see figure 7).
Walking Location
The location of where pedestrians were observed walking, street or sidewalk, was also
examined (see table 11). The proportion of people observed walking in the street changed
differentially for S. Carrollton and the adjacent side streets (Z = -3.30, p= 0.001). There was a
slight change in the proportion of people observed walking on the sidewalk and neutral ground
versus the street on S. Carrollton after the improvements to the street (pre 98.8%, post 98.2%;
Z=2.06, p=0.039). There was a decrease in the number of people observed walking in the street
on the side streets after the lane was installed on S. Carrollton (pre 40.4%, post 33.2%; Z = 3.49,
p<O.OOO; see table 11}.
40
VI. DISCUSSION
Study 1
When bike lanes were striped in a racially and economically diverse and mixed-use
urban neighborhood in New Orleans, more people were observed riding their bicycles. More
people rode in the overall neighborhood after the lanes were striped; however, the increase in
cyclists was greatest on the street with the new bike lane. The decrease in cyclists on the side
streets suggests that some of those cyclists may have started using the dedicated bike lane. The
large increase on S. Carrollton suggests that even in a city not widely known for a cycling
culture, new bicycle lanes appeared effective in attracting new cyclists who were diverse in
race, sex and age. It is possible that large numbers of cyclists could act as role models, making
cycling more the norm and accepted.
Additionally, while the city of New Orleans has experienced some population growth
following the loss of residents after Hurricane Katrina, the growth in population is unlikely to be
the reason for this increase in ridership. The number of residents receiving mail in this
neighborhood remained unchanged over the course ofthis study. 69
Although all demographic subgroups experienced increases after the lane striping, the
differences were the most pronounced among women. More women were willing to ride on S.
Carrollton after the bike lane was striped; a finding consistent with other published studies
from New Orleans and elsewhere that indicated women prefer dedicated spaces to ride. 47 ' 66
However, the number of men observed in this study exceeded the number of women observed.
One potential explanation is that women tend to be more risk averse when it comes to cycling
41
and prefer off-street paths and bike boulevards (quiet residential streets with traffic calming
measures to encourage biking).47 ' 70 Yet there are few off-street paths and bike boulevards that
. I
are near destinations such as shopping or schools in New Orleans. Increasing bike boulevards
and other bike facilities separated from traffic may attract more women biking to meet their
daily needs.
Installing the bike lanes did increase the number of both whites and blacks using the (
lanes but did not differentially affect one racial group more than another. Overall there were
fewer blacks observed riding their bikes in this neighborhood than whites, even though this
neighborhood is racially diverse. Nationally, whites have the highest bicycle mode share amo,ng
all racial and ethnic groups, although cyclist rates are increasing for minorities-from 2001 to
2009 African Americans, Hispanics, and Asian Americans increased their total share of bike trips
from 16 percent to 23 percent.71 Furthermore, cycling is still predominately a white activity
nationally, whites make 77 percent of all bike trips in the United States, but account for only 66
percent ofthe population.72
There could be several reasons why the change in infrastructure alone did not bring
more black cyclists to the area. First, some of the socio-cultural factors that are present in the
conceptual diagram related to· bicycling (Figure 1) may not be present. Concerns about
violence, the lack of smooth streets on small neighborhood streets and the cost of a bicycle
may be some factors that limit cycling for this group. Furthermore, if few blacks are observed
riding their bikes, it is possible that other blacks see biking as an activity in which black people
do not engage. Programs (group rides, technical workshops, gear swaps) started by
42
organizations such as Black Women Bike could help address this phenomenon by encouraging
more blacks to ride bicycles through addressing barriers.
Few youth were observed riding a bicycle in this neighborhood at baseline and follow­
up. This is surprising, as observations took place near a school that included middle school
students. However, the students from this school are drawn from many neighborhoods
throughout New Orleans and they may live more than a mile or two away from school.
Additionally, parental concerns regarding crime may inhibit whether they allow their children to
bike to school. Lastly, there were no bike racks located at the school, and without a safe place
to lock their bike, children would probably not wish to leave it unguarded.
The new bike lane on S. Carrollton was not promoted through a special groundbreaking
or other event. This may have been a missed opportunity because there is evidence that
providing supportive environments for physical activity, in combination with promotional
activities, is an effective strategy.73 Promotional events such as periodically closing streets to
traffic (e.g. ciclovias) could allow people to sample safe places to bicycle and help build a
bicycling culture.74
The proportion of riders using the sidewalk instead of the street did not change on S.
Carrollton after the bike lane was installed. This is difficult to interpret because of migration of
some riders to S. Carrollton from adjacent streets. It is possible that there are riders who
moved from the side streets to S. Carrollton but were still more comfortable on the sidewalk
than in the bike lane.
43
More people chose to ride in the correct direction when the lane was implemented
perhaps due, in part, to the new symbols on the street that indicated the preferred direction
for riders. This is consistent with other findings showing that shared-lane markings and bike
lanes reduce the number of wrong-way riders. 66 ' 75
' 76 However, the present study is the first we
know of to measure this effect in bike lanes before and after striping.
Study 2
These intercept survey results of cyclists and pedestrians traveling along S. Carrollton
Avenue in New Orleans during the fall of 2010 found that cyclists and pedestrians differ with
respect to gender, trip purpose, car and bike access. Compared to pedestrians, cyclists were
more likely to be male and to be engaging in physical activity for utilitarian purposes, i.e.
traveling to or from a destination. Cyclists' trip distahce was also greater than pedestrians' and
they were more likely to combine stops during their trip and use their car less frequently. A
greater proportion of cyclists made a trip of one mile or more, while the majority of pedestrians
made trips of X mile to one mile.
The finding regarding male and female cyclists has been documented elsewhere, where
national transportation surveys indicated that cycling prevalence is nearly 3 times higher for
men than women.77 Women may be more risk-averse to using on-street bicycle facilities,
especially if there is minimal separation from traffic.78 Other factors such as physical comfort
and the perception that a car is necessary to perform daily utilitarian trips such as shopping or
bringing children to or from school, also may play a role.47
44
The standard that pedestrians in the United States will not walk more than a X mile for a
utilitarian trip is generally accepted in the literature, and is used as a rough guide of how close
people will walk to transit stops.79 Other researchers have challenged the X mile limit, stating
that pedestrians will walk longer and further depending upon trip purpose and neighborhood
conditions.80 ' 81 Seventy-nine percent ofthe pedestrians we surveyed were taking trips of
longer than X mile. A potential factor for this finding is the neighborhood of the study area: S.
Carrollton Avenue has access to a street car, several city bus lines, homes, schools, churches,
parks and businesses at either end of the street and throughout.
The majority of those in the survey indicated that if they could not walk or bike for their
utilitarian trip, they would have driven (see Table 7). However, fewer cyclists would have
driven than pedestrians. Possible explanations for this difference in choice of substitute travel
by mode could be that pedestrians may not feel comfortable biking or lack access to a bicycle. It
may be more difficult to convince pedestrians to ride a bike for a utilitarian trip than to
convince cyclists to walk for the same trip. Walking does not take any special equipment,
whereas to ride a bike, a bike is needed as is a place to park it once the destination is reached.
However, a proportion of cyclists, 44 percent who said that they only recently started biking on
the bike lane also indicated that they would have driven a car, had they been unable to bike.
This finding suggests that the presence of the bike lane was a strong factor as to the reason
people were cycling. This has implications for improving not only physical activity and the
population level, but also minimizing trips by car and improving air quality.67
45
A majority ofthe cyclists and pedestrians intercepted reported that they bike and walk
enough to reach a minimum of 30 minutes of physical activity per week. This proportion of
people who report reaching the 30 minute threshold is higher than indicated in other surveys
that included a national sample instead of one of only pedestrians and cyclists.77 However, as
mentioned previously, the location of the intercept surveys included many destinations that
would be attractive for both utilitarian and recreation type trips. Additionally, this finding
shows that people who walk and bike regularly are able to meet the recommended amount of
overall physical activity.
Study 3
When improvements to sidewalks and intersections were made in a racially and
economically diverse and mixed use neighborhood in New Orleans, more people were observed
walking. More people were observed in the overall neighborhood after the sidewalk
improvements took place; however, the greatest increase occurred on S. Carrollton Avenue, the
street with the new curb cuts, crosswalks, and repaved sidewalks. There was also an increase
in pedestrians walking on the side streets; however, this increase was small compared to the
increase observed on S. Carrollton. There was a 49 percent increase in pedestrians on S.
Carrollton, compared to a 14 percent increase on the side streets. The greater increase on the
street with the new pedestrian facilities and upgrades suggests that the infrastructure
improvements appear effective in attracting pedestrians to the corridor.
New Orleans has experienced some population growth following the loss of residents
after Hurricane Katrina, however the growth in population is unlikely to be the reason for the
46
increase in walking in this neighborhood. The number of residents receiving mail in the
Carrollton area did not change over the course ofthis study.69
The proportion of people walking on the sidewalk versus the street on S. Carrollton did
decrease, but this difference, while statistically significant is not meaningfully so. A majority of
those who were classified as pedestrians walking in the street were either walking in the street
or in the wide median that is between the lanes of traffic. The median is often used as a place
for exercise, and though there are streetcars that run approximately every 20 minutes, walkers
and runners were observed using the space for physical activity. (The streetcars are slow and
equipped with warning signals so that they pose little threat to pedestrians). The sidewalks at
follow-up were smoother and more level, but can pose difficulty for runners with strollers. The
curb placement at the intersections means that a runner with a stroller would have to make
several turns to come off the sidewalk, cross the street and enter the next sidewalk. The
median is grassy and uneven, which would be difficult to negotiate a stroller and would also
result in a bumpy ride for the child. Thus, some runners choose to use the bike lane as a place
to run with a stroller (see photo 3), and may explain why there are still some pedestrians who
prefer to use the street for physical activity even if given good quality sidewalks. This study did
not count pedestrians using the bike lane specifically, but future studies should consider
assessing whether on-street bike lanes are a location desired by pedestrians.
The number of youth using the sidewalks increased after the sidewalk improvements
and lanes were installed. This increase on S. Carrollton was greater than the decrease on the
side streets (96.8 vs. 44.2%). It should be noted that the observation area included a middle
47
school which draws from students throughout the city of New Orleans and experienced some
changes in enrollment, an additional 60 students attended the school between data collection
periods. However, there were no programs at the school that encouraged walking or biking to
school. It is possible that Safe Routes to School encouragement programs at local schools
would further improve the number of children seen walking in the neighborhood as the
infrastructure is now in place to support such activity safely.
Walking is positively associated with increased neighborhood density and mixed land
use,82 and this study confirms this association. This study also demonstrated that in places
where there are pedestrians present, small improvements to sidewalk infrastructure will yield
increases in the number of people using sidewalks.
48
Study 1
Some limitations of this study should be noted. The design of this study, though quasi­
experimental, is limited by the lack of comparison neighborhoods. There are comparison
streets, adjacent to the bike lane and sidewalk improvements; however, without comparing
levels of cycling in 2009 and 2010 in other neighborhoods, it is not possible to account for other
community-level factors that may be contributing to these changes over time. For example,
there are several new bicycle rental companies in New Orleans, as well as new bike parking
infrastructure. It is possible that these other environmental changes may be encouraging
people to ride more frequently. Future analysis will examine data from other researchers to
compare counts obtained across the city, using this methodology.
A second limitation is that observations were conducted three months after the bike
lane was finished. It is possible there could still be a novelty effect ofthe lanes in the three­
month period following construction. A third limitation was that race was categorized by casual
observation by the data collector. It is possible that individuals were categorized incorrectly. To
ascertain whether observers might be mis-categorizing race, we conducted a separate substudy
that compared observed race with reported race and found good overall agreement with a
kappa value of 0.813 (data not shownL providing a level of confidence to the method that was
used. Fourth, this study took place in one neighborhood, where car ownership was low and
access to walkable destinations was high. Other neighborhoods without similar features may
not see the pronounced increases in ridership that were seen in this study.
49
This study also has several strengths that should be noted. Observations were
conducted more hours per day (11 hours) and over more days (10 days) than other studies
evaluating bike lanes. The inclusion of two adjacent streets allowed us to address possible
displacement of riders from nearby streets to the street with the new bike lane. This study was
also the first study to count cyclists by age group, gender and race before and after a new bike
lane.
Study 2
This cross sectional survey is descriptive; therefore we cannot infer causation between
variables discussed. Furthermore, this is a survey of only cyclists and pedestrians, not of people
in the general population. It is not representative ofthe population at large, it can only be used
to describe what this sample of cyclists and pedestrians did and thought at the time of the
evaluation relative to their transportation and recreation trips. The results show that levels of
physical activity reported here are higher in this selective population.
Second, additional questions that asked respondents if they used public transportation
for a portion of their trip were not asked in this evaluation. This would have allowed further
categorization of trip length and could explain why some walking or cycling trips were much
longer than others.
Third, additional questions could have focused on how many bouts of physical activity
were of 10 minutes or more in duration. As some of the physical activity is transportation
related, it would be helpful to know whether transportation related physical activity trips are
50
meaningful for health. Assessing the length of each physical activity bout would assist in that
process.
The final additional questions could have asked pedestrians about their attitudes toward
the new sidewalk and bike lane improvements. The improvements to the sidewalks on S.
Carrollton included new crosswalks, curb ramps and repaving sections that were in disrepair.
Pedestrians may or may not have considered these improvements integral to the walking
environment of the streetscape.
The above sets of questions would add significant information to the literature
encompassing evaluation of bike lanes and should be considered for future research.
Study 3
Similar to study one, this study is limited in its design in that it did not conduct counts of
pedestrians in other neighborhoods. As mentioned previously, this potential for historical bias
is possible as there may be other environmental factors that might be encouraging people to
walk for transportation or recreation more frequently. An example of this may be that new
businesses might have opened between 2009 and 2010 that would become a destination for
pedestrians. Also, pedestrian observations were completed three months after the new bike
lane and sidewalk improvements. Similar to the cyclist observations, there could have been a
novelty effect of pedestrians being interested in the location because of the new
improvements. Furthermore, the increase in children observed could have been due to
changing enrollment at the local school near the observation site. The principal ofthe school
51
noted that an additional 60 children attended the school between the baseline and follow-up
observation periods. Thirdly, pedestrian observers could have miscategorized race when they
were observing race in the field.
52
VIII. CONCLUSIONS AND RECCOMMENDATIONS
Cycling for transportation or recreation is one low-cost way to improve physical activity.
According to an engineer83 from the New Orleans Department of Public Works, the total cost of
the bike lane was less than 1 percent ofthe total road resurfacing project. This research shows
that bike lanes are well-suited in diverse urban areas and that if bike lanes are built, people will
use them. More lanes should be constructed in areas where residents can travel to meet daily
needs by cycling. The results of this study indicate that planners, engineers and public health
practitioners should consider installation of bike lanes as a means to encourage physical
activity.
The survey confirmed what was observed in the count data of cyclists: cyclists indicated
that the new bike lanes were a strong factor in why they chose to ride a bike on S. Carrollton.
The popularity of the lanes is observed by the high increase in the counts along with the strong
survey results. Furthermore, cyclists interviewed in the intercept surveys were more likely to
be going to or from a destination than out for exercise. In New Orleans, a city with a significant
portion of the population without a car, it is likely that people without cars turn to bikes to
meet their needs. Building the infrastructure to support those who need to bike as well as
want to, is a smart decision to protect the safety of riders and promote health.
The lanes did bring a significant increase in the number of female cyclists, and that
increase was greater than the increase in the number of males. Female cyclists prefer riding on
streets with bike lanes to riding on streets without. Building more bike lanes and making sure
53
that they are connected to destinations such as stores and schools would likely increase the
number of females who feel comfortable cycling to meet their daily needs.
The increase in the pedestrian and cyclist counts on all three streets shows that more
people are active in this neighborhood overall. The numbers of observed pedestrians and
cyclists was higher on S. Carrollton where the new lane and sidewalk improvements took place.
Seeing more people outside being physically active in a city not currently known as one for
physical fitness is heartening. The Mayor of New Orleans, Mitch Landrieu, has announced a
goal for New Orleans to be a Fit City by 2018, where more citizens make healthy choices.
Observing more people outside and being active will help residents and visitors alike consider
New Orleans a place where everyone can be active outside.
An increased presence of people overall in the neighborhood can also be a promising
deterrent to crime and create natural surveillance that discourages criminal activity. 56 Seeing
more people outside being active, may encourage others to also go outside, because they feel
safer. Thus, a ripple effect where the upgrades to S. Carrollton may increase activity further
away from S. Carrollton as people travel to and from S. Carrollton on foot. This ripple effect
and it link to safety should be studied, potentially using crime mapping data.
As mentioned in the background section detailing the setting of this study, S. Carrollton
Avenue in uptown New Orleans, does include access to various types of public transit. There is
a streetcar that runs frequently on S. Carrollton itself, as well as, several bus lines that connect
where the streetcar line terminates. The role of public transit in encouraging physical activity
was not the central theme of this study; however, it is apparent through the data analysis of
54
trip distance that some of the respondents in the survey probably took transit for one portion
of their trip. The role of transit along coupled other improvements could be studied in future
research.
As noted in the background section, this research does have implications for policy at
the local level. The city of New Orleans has passed a complete streets policy and is now
implementing the policy, whereby any new street project will need to include accommodations
for all types of mode of transportation. This research has been cited by local and national
advocacy groups as they communicate with stakeholders about the need for more bike and
pedestrian infrast