Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 48
below-market-rate housing project that replaces a former surface park-and-ride lot. One report is
that one-third of Overlake residents regularly use the bus passes they received, with half of those
pass users indicating that they have increased their transit use since moving into the building
(Shelton and Lo, 2003). A survey with 40 returns suggested that roughly half the residents were
riding the bus on a regular basis. Resident parking use in the project garage, which also serves
park-and-ride, was observed to be 0.6 resident autos per dwelling unit (Prince et al., 2003,
Posthuma, 2003). An overview assessment has concluded that Overlake Station, by combining
"good quality TOD, good quality transit service and affordable housing," achieved success by
focusing on meeting the needs of those who cannot afford cars (Hendricks, 2005).
The Renton Transit Center and adjacent Metropolitan Place development are immediately
adjacent to Renton's traditional downtown. The TOD and transit center partially continue the
downtown street and sidewalk grid. In addition to the Metropolitan Place project, two other
residential apartment buildings have been constructed adjacent to the Transit Center, with 165
dwelling units. Other new development in the vicinity includes office space, condominiums, and
a large municipal parking garage with ground-floor retail. A little more than one-third of the
Metropolitan Place residents are reported to use their bus pass regularly. The residents are
anecdotally reported to be older and more predominantly empty-nesters than anticipated
(Shelton and Lo, 2003; Prince et al., 2003).
UNDERLYING TRAVELER RESPONSE FACTORS
A number of different factors that affect TOD resident, worker, and visitor travel decisions, along
with residence location choice decisions, are explored in this section. Of necessity, these factors are
laid out one at a time for discussion. The various influences on travel behavior choices are, how-
ever, decidedly interactive in nature. Moreover, these factors are not all transportation-related,
suggesting that it takes more than good transportation policy alone to develop high-quality and
effective TOD (Hendricks, 2006). The interactive nature of underlying traveler response factors
affecting TOD also poses an exceptional challenge to fully understanding the importance and role
of individual influences.
Land Use and Site Design
This discussion of land use and site design effects on TOD travel demand is organized to address
TOD-specific aspects of the "three D's" of density, diversity, and design from a TOD-supportive
perspective. Comprehensive coverage of the topic of traveler response to the "three D's" is pro-
vided by Chapter 15, "Land Use and Site Design."
TOD-Supportive Density
Higher development densities and correspondingly higher trip densities are associated with
TOD. Increased development density places more housing, jobs, and activities within the
same land area. The effective density of trip making in TOD may be further increased
relative to non-TOD by the transit-supportive practice of clustering the highest density TOD
components at or near the TOD's transit stops, rather than spread out evenly over the site.
The concentration of trip ends resulting from TOD-supportive density creates a larger
17-48
OCR for page 49
potential market for transit. Even assuming no increase in transit mode share as a result of density
in and of itself, higher-density development generates more transit ridership per unit of land
area--and thus per transit stop or station--than lower-density development.10
The added ridership potential of TOD-supportive densities can facilitate providing the cost-
effective, higher-quality transit service desirably associated with TOD. Although high density
may itself not directly cause significantly higher transit and walk mode shares, as explored in
Chapter 15 under "Response by Type of Strategy"--"Density," important second order effects of
high density can. These factors that boost transit ridership and walking include both the better
transit service density allows and lower vehicle ownership rates. In part for similar reasons, and
aided by the shorter travel distances involved, higher densities are also associated with greater
use of non-motorized transit access and egress modes--higher walk access mode shares in
particular. As a result of all these associations and related experience, many jurisdictions have
developed guidelines that call for increased TOD densities to achieve desired outcomes that
specifically include increased use of transit (Tumlin and Millard-Ball, 2003; Cervero et al., 2004).
(For more on TOD design to achieve objectives, see the "Transit Oriented Development Index"
presentation that concludes the "Related Information and Impacts" section.)
TOD-Supportive Diversity
More-diverse TOD projects in terms of land use mix offer the possibility of a greater proportion of
activities being conducted within the center and a corresponding reduction in motorized-travel
trip generation, as alluded to in Table 17-8 under "Response by TOD Dimension and Strategy"--
"Response to TOD by Land Use Mix." As with non-TOD development, diverse land use can enable
more needs to be satisfied on a single visit and allow internal walking trips to serve for visiting
multiple destinations. TODs with both jobs and housing can serve to balance the utilization of
transportation infrastructure, both highway and transit, and help create an all-day environment.
While such TODs can and do capture some commute trips internally, it should be noted that resi-
dence and workplace location decisions are not always contemporaneous or fully flexible (Cervero
et al., 2004).
To the extent that TOD leads occupants or visitors to arrive by transit in greater proportions than
at non-TOD development, the availability of automobiles to occupants and visitors is reduced. The
opportunity to meet needs within the TOD that land use mix affords makes this outcome more
acceptable, likely enhancing both the transit mode share and the prevalence of pedestrian travel
within the development. With implementation and study of more TOD examples should come
increased understanding of traveler response to different mixes of land use including the travel
behavior effects of potential synergies. In the meantime, research findings assembled in the
"Response by Type of Strategy"--"Diversity (Land Use Mix)" subsection of Chapter 15, "Land Use
and Site Design," may be judged largely relevant to TODs if applied with due appreciation of den-
sity and transit service level differences.
10
The relationship is not purely arithmetic. For example, with increased activity within a TOD may come a
greater internalization of travel (as discussed with respect to "TOD-Supportive Diversity"), and internal
trips will not be candidates for using the regional transit service. For an estimate of transit use increase in
response to TOD densification that is derived from actual experience, see the "Arlington, Virginia" exam-
ple under "Response by TOD Dimension and Strategy"--"Response to TOD by Primary Transit Mode"--
"Heavy Rail Transit."
17-49
OCR for page 50
In one study of rail mode of access and egress for development near rail stations in the San
Francisco Bay Area, a strong positive association was found between mixed station-area land
use and higher propensity to walk for access and egress trips. The study observed that people
are willing to walk farther to and from stations in denser, mixed-use settings than in areas with
large parking lots and low-density residential development. Conversely, people tended to forgo
walking to access stations surrounded by large surface parking lots and instead drive in
(Cervero et al., 1995).
TOD-Supportive Design
In TOD, buildings are concentrated within close proximity to the transit stop and particular
attention is paid to the pedestrian environment. The compact, pedestrian-friendly design of TOD
leads to higher transit usage and walking because of the underlying traveler responses to this
environment. In particular, the shorter walking distances encourage transit usage, the shorter
walking distances encourage walking for transit access, and the pedestrian-friendly design
encourages more walking overall.
Each of these factors is explored further in the subtopics below. Also highly relevant is the
pedestrian catchment area analysis reported in the case study, "Travel Findings for Individual
Portland, Oregon, Area TODs," under "Results." It points out that actual walking distances to
stations are greatly affected by street and pedestrian system layouts. Assessments of the street
systems within four TODs estimated that only 21 to 57 percent of the area within a 1/4-mile radius
of each station was actually within a 1/4-mile walk (Schlossberg et al., 2004).
Related general topic coverage is provided in Chapter 15, "Land Use and Site Design," and
Chapter 16, "Pedestrian and Bicycle Facilities." Within Chapter 15, see especially "Community
Design and Travel Behavior" and "Transit Supportive Design and Travel Behavior," both within
the "Response by Type of Strategy"--"Site Design" subsection. Also of special interest is the
Chapter 15 case study, "San Francisco East Bay Pedestrian Versus Auto Oriented Neighborhoods,"
discussed further below.
Walking Distance and Transit Usage. Nearly all trips made by transit involve at least one if not
two segments that require walking. Perhaps the single most important site design element when
it comes to influencing transit usage is the walking distance from the transit station to the front
door. Transit mode shares decline as distance from the transit station increases. This ridership
gradient is seen at both the home and non-home end of trips. Moreover, walking distances at the
home and non-home ends of trips work together to determine the likelihood of transit use.
The existence of a mode share gradient vis-à-vis walking distance is observed not only for both res-
idential development and office development, but also for both urban stations and suburban sta-
tions. Whether for residential or office, the transit mode shares observed at any given distance from
the station tend to be higher at urban stations than at suburban stations. Development integrated
with transit stations such as through direct pedestrian connections may receive an added boost in
transit mode shares (Cervero et al., 2004).
Observed differences among mode share gradients are partially a reflection of differences in
the specific developments that have been surveyed. Such differences include urban versus
17-50
OCR for page 51
suburban travel market differences, person-density differences, differing land use arrangements,
and differences in parking charges and commute transportation policies among office employers.
Table 17-24 presents a summary of available walk distance ridership gradient relationships from
California and Washington, DC.
Table 17-24 Summary of Walk Distance Ridership Gradient Relationships for Work Trips
Residential-Focused
Study Developments Office-Focused Developments
1992 California transit- Percent Rail = 32.24 - 0.0085*M Percent Rail = 1105*(M)-0.795
focused development R-Squared = 0.381 R-Squared = 0.381
study (Cervero, 1993) (Based on surveys at 27 projects) (Based on surveys at 18 projects)
2003 California TOD No discernable relationship. Percent Transit = 52.3 - 6.7*log(M)
travel characteristics (Based on surveys at 25 projects (Based on surveys at 10 projects
study (Lund, Cervero, within 0.5 miles of transit) within 0.5 miles of transit. See text for
and Willson, 2004a) discussion.)
1989 Washington, DC, Percent Transit = 66.52 - 0.0156*M CBD offices:
development-related R-Squared = 0.40 Percent Transit = 61.37 - 0.0076*M
ridership survey (JHK (Based on surveys at 18 buildings) R-Squared = 0.57
& Associates, 1989) (Based on surveys at 7 buildings)
Suburban offices:
Pct. Trans. = 27.16 - 0.0061*M - 0.84*D
R-Squared = 0.47
(Based on surveys at 40 sites)
2005 Washington, DC, Percent Rail = 54.15 - 0.0087*M Percent Rail = 35.38 - 0.0096*M
development-related R-Squared = 0.41 R-Squared = 0.25
ridership survey Percent Transit = 54.83 - 0.0071*M Percent Transit = 46.15 - 0.0121*M
(WMATA, 2006a) R-Squared = 0.24 R-Squared = 0.31
(Based on surveys of all trips at (Based on surveys of commute trips
18 sites) at 17 sites)
Notes: M = distance from station to building in feet. D = distance from building to CBD in miles.
Sources: As indicated in the "Study" column.
The ten office projects surveyed in the 2003 California TOD travel characteristics study provide a
striking example of the interplay between mode share, distance, and other different characteristics
of specific developments. Eight of the office developments surveyed were between 500 and 2,700 feet
of a rail station. None of these exhibited a rail/bus transit mode share of over 6 percent of workers
surveyed and there was no discernible relationship to distance among the corresponding eight data
points. The shape of the nearly asymptotic relationship reported in Table 17-24 is largely formed by
two statistical outliers, the state of California Department of Conservation building in downtown
Sacramento, reporting a 27 percent transit mode share, and the Great Western Building in downtown
Berkeley, in San Francisco's East Bay area, reporting a 17 percent share. Besides being just 165 feet
from a light rail transit (LRT) stop in the case of the Department of Conservation and 137 feet from
the nearest BART heavy rail transit (HRT) station entrance portal in the case of the Great
17-51
OCR for page 52
Western Building, these two buildings have the following similar contributing characteristics: high
workers per acre densities of 37.6 and 20.6, respectively; dense, mixed-use surroundings; parking
costs of over $100 per month; and no parking at the nearest rail stations (Lund, Cervero, and Willson,
2004a).
Analysis of a large survey sample resulting from trip reduction monitoring requirements in effect
during the mid-1990's in the San Francisco Bay Area provides findings that are not as project
dependent. From 1992 to 1995, more than 250,000 surveys were collected by the Bay Area Air
Quality Management District from over 1,100 work sites in Napa, Marin, San Mateo, Santa Clara,
Alameda, Sonoma, and Solano counties. Analysis of these survey findings clearly shows the drop-
off in transit mode share as distance between the work site and its rail station increases, and also
shows differences among rail transit modes. The differences among transit modes are muted once
BART HRT stations in higher density employment areas with paid parking are excluded. The
steepest transit share drop-off observed overall was that around Santa Clara County LRT stations.11
It appears that potential transit users are quite sensitive to walking distance, particularly around
stations located on the north end of the Santa Clara County system, where development is low
density and few shops or restaurants are nearby (Dill, 2003). Table 17-25 highlights these findings.
Table 17-25 Transit Mode Share by Distance of Work Site Location from Station
Transit Commute Mode Share by Station Type
HRT LRT CRR
BART
Distance of Excluding
Worksite from Oakland and Santa Clara
Station BART a Berkeley County Caltrain All
0.00 to 0.25 miles 33.6% (44) 6.2% (3) 5.9% (49) 7.0% (14) 19.8% (107)
0.25 to 0.50 miles 7.9% (22) 5.7% (13) 3.1% (56) 4.1% (39) 4.0% (117)
0.50+ miles -- -- -- -- 2.5% (929)
Notes: The numbers in parentheses are the number of worksites surveyed in the particular category.
a San Francisco work sites, with their prevalence of paid parking and high densities, were
not included in the survey or analysis. The additional exclusion (in the second of the two
BART columns of data) of Oakland and Berkeley, also characterized by prevalence of paid
parking and high densities, provides mode share by distance data for work sites where the
prevalent condition is free parking and lower densities (Dill, 2006a).
Source: Dill (2003).
Research in Toronto and Edmonton, Canada, has also illustrated the decline in transit mode
share as distance of development from the nearest rail station increases. Figure 17-3 shows
11 It was noted that the earlier 1992 California transit-focused development study obtained similar results for
BART HRT and Caltrain commuter rail (CRR), but somewhat higher transit mode shares among employ-
ees near Santa Clara County LRT stations. It was speculated that those outcomes related to the specific sites
selected for investigation (Dill, 2003). Findings from the 1992 effort (Cervero, 1993) were presented in the
subsection, "Response to TOD by Land Use Mix."
17-52
OCR for page 53
this distance-related decline in transit mode share for residential development in the Canadian
cities and for the earlier Washington, DC, region and California studies. Figure 17-4 shows the
same for office development. While the Canadian and California studies looked at rail transit trips,
the Washington study looked at all transit trips. It should also be noted that, because the studies
rely on relatively few specific projects for data collection, the results are influenced by the specific
characteristics of the projects selected.
Figure 17-3 Work trip rail mode share by distance from residential sites to station
Notes: The graphed 1989 Washington, DC, area shares are for all transit (rail and bus combined). California and
Canadian mode shares are for rail transit only. See last row of Table 17-24 for Washington, DC, 2005 mode
share gradients for Metrorail only and rail and bus combined.
Source: Cervero (1993).
Results from station-area analyses carried out using the 2000 San Francisco Bay Area travel survey
illustrate the interplay of walk distances at both ends of a trip. Mode shares were analyzed for per-
sons living and/or working within, or not within, 1/2-mile walks of rail transit stations or stops
and ferry terminals. Walking distances were measured along street-system approximations of the
true pedestrian network. The regionwide average home-based work (commute) trip transit shares
obtained for the four possible combinations of transit station proximity were:
· 42 percent transit for commutes with both residence and workplace within 1/2 mile of a
station/stop/terminal.
17-53
OCR for page 54
· 28 percent transit for commutes with the workplace but not the residence within 1/2 mile.
· 16 percent transit for commutes with the residence but not the workplace within 1/2 mile.
· 4 percent transit for commutes with neither the residence nor workplace within 1/2 mile.
Figure 17-4 Work trip rail mode share by distance to office sites from station
Notes: The graphed 1989 Washington, DC, area shares are for all transit (rail and bus combined). California and
Canadian mode shares are for rail transit only. See last row of Table 17-24 for Washington, DC, 2005 mode
share gradients for Metrorail only and rail and bus combined.
Source: Cervero (1993).
Even though there are other potential contributors to these mode share differentials besides walk-
ing distance, such as income (which tends to be lower close-in to San Francisco Bay Area stations),
the importance of home-work linkages via transit and location near stations is well illustrated by
these summary data (Metropolitan Transportation Commission, 2006).
Walking Distance and Transit Access/Egress Modes. The short walking distances from transit to
development entrances that pertain in the ideal TOD contribute not only to elevated transit mode
shares but also to high walk-access mode shares. The 2003 California TOD travel characteristics
study reported that over 90 percent of surveyed residents of station-area housing who were rail
commuters walked to access their rail station. The same study also surveyed station-area office
workers, albeit at different sites. At the workplace end of the trip, 78 percent of station-area
workers commuting by rail reported walking from the station to the workplace.
17-54
OCR for page 55
The design advantages of a TOD of course do not affect conditions at the opposite end of a trip
from the TOD. The same surveyed rail commuters--for whom transit access/egress share aver-
ages within the TOD are given above--also provided data for the egress/access modes at the other
end of their trip. These were vastly different for home station access shares. Only 33 percent of 187
surveyed TOD area workers that commuted using rail reported walking to their origin rail station.
Other modes they used to access the origin station were driving alone (51 percent), riding as an
automobile passenger (6 percent), riding a bus (8 percent), and bicycling (2 percent). TOD resi-
dents, on the other hand, reported about the same degree of walking from their non-home rail sta-
tion to the workplace (79 percent) as did surveyed station-area office workers (Lund, Cervero, and
Willson, 2004a).
Another instructive study, this one in the Chicago region, focused on commuter railroad (CRR)
travel by residents dwelling in the area around six high-ridership Metra stations. The case study
stations were selected to represent a geographical distribution, a variety of community types, and
differing service characteristics. A passenger survey was conducted at each of the stations with an
overall response rate of 32 percent. Most CRR riders traveling less than 1/2 mile to access the rail
station were found to be walking--82 percent on average. This walk access share was found to drop
sharply with increasing access distance. The six-station average shares for each access mode and
access-distance range are displayed in the upper half of Table 17-26. Distributions of station-access
trips across each access-distance range are shown for each access mode in the lower half of the table.
Table 17-26 Mode of Access Versus Distance Relationships for Six High-Ridership Metra
CRR Stations Combined
Drove Dropped
Alone Carpool Off Bus Walked Other
Mode of Access Shares by Access Trip Length (percentages calculated across each row)
0.00.5 mile 7.5% 0.4% 9.3% 0.0% 82.2% 0.7%
0.51.0 mile 33.4% 2.9% 16.4% 3.1% 41.3% 2.8%
1.02.0 miles 53.3% 3.2% 17.4% 14.2% 8.2% 3.8%
More than 2.0 miles 63.7% 5.9% 11.4% 16.4% 1.0% 1.6%
Overall Mode of Access 47.3% 3.8% 13.6% 11.0% 21.9% 2.2%
Access Trip Length Proportions for Each Mode of Access (percentages calculated down each column)
0.00.5 mile 2.2% 1.3% 9.5% 0.0% 52.4% 4.3%
0.51.0 mile 13.4% 14.3% 22.9% 5.4% 35.8% 23.9%
1.02.0 miles 29.9% 22.1% 33.8% 34.2% 10.0% 43.5%
More than 2.0 miles 54.5% 62.3% 33.8% 60.4% 1.8% 28.3%
Note: Access trip length was self-reported by survey respondents.
Source: S.B. Friedman & Company et al. (2000b).
The same study provides further information on the effect of station access distance on transit
use, CRR ridership in this instance. Rail transit mode market penetration was found to fall
sharply outside the 1/2-mile radius of each station. A number of characteristics were
identified that appeared to contribute to the high ridership of the six stations, most notably a
17-55
OCR for page 56
good pedestrian environment with a concentration of development around the stations including
stores (S.B. Friedman & Company et al., 2000a and 2000b). Table 17-27 presents the surveyed sta-
tion characteristics and ridership statistics, with the final entries highlighting the drop in market
penetration with increasing distance from the stations.
Table 17-27 Selected Attributes from Six High-Ridership Metra CRR Stations
Home- 103rd Glen Arlington Naper- Deer-
Attribute wood Beverly Ellyn Heights ville field Average
Miles to
Chicago CBD 23.5 12.8 22.4 22.8 28.5 24.2 22.4
Park-and-Ride
Spaces 560 346 666 1,178 1,339 641 788
Utilization 97% 91% 83% 78% 95% 85% 88%
Feeder Bus Routes 3 1 5 2 15 2 5
Weekday Trips a 1,578 969 1,889 2,579 4,040 1,279 2,056
Mode of Access
Drive alone 46% 41% 34% 50% 47% 66% 47%
Carpool 3% 6% 3% 4% 4% 3% 4%
Dropped off 17% 10% 15% 14% 14% 8% 14%
Bus 6% 6% 14% 3% 20% 1% 11%
Walked 26% 36% 32% 26% 13% 19% 22%
Other 2% 1% 2% 3% 3% 4% 2%
Residential b
Population 2,527 6,162 3,899 4,855 3,676 3,615 4,122
Households 1,146 1,857 1,690 2,480 1,528 1,538 1,707
Households/acre 6.4 15.6 9.9 12.3 9.3 9.2 10.4
Market Penetration (Percent of Households that use CRR)
0.00.5 mile 21% 15% 15% 15% 13% 10% 15%
0.51.0 mile 9% 2% 13% 7% 9% 7% 7%
1.02.0 miles 3% 1% 4% 4% 6% 4% 3%
Note: a Passenger boardings plus alightings from Metra's 1999 Metra Rail Service and Residential
Development Study.
b Within 1/2 mile of the station.
Source: S.B. Friedman & Company et al. (2000a and 2000b).
Pedestrian-Friendly Design and Walking and Transit Use. The quality of walk connections
has been shown to influence the distance people are willing to walk. A short walk made
difficult or unpleasant by adverse environmental conditions such as high-speed traffic or lack
of shade can seem longer while a long but pleasant or interesting walk can seem shorter.
It follows logically that quality of the pedestrian connections between the transit stop and the
front door of the development should be important to transit usage. In many TOD exam-
ples, special attention has been given to the pedestrian environment, including streetscape
17-56
OCR for page 57
improvements. It is generally held that the placement of parking lots, green spaces, and the build-
ings themselves can impact the pedestrian and transit friendliness and attractiveness of travel by
transit or walking (Arrington et al., 2002).
Results from development of an advanced travel demand model set for San Francisco County lend
support to the concept that the quality of walk connections to transit is positively related to tran-
sit use. Neighborhood vitality at the destination was found to have a strong positive relationship
to the choice of all non-auto modes examined (walk, bike, and transit) for most types of trips.
Adverse topology (steep gradients and barriers) was nearly as important. Connectivity at the des-
tination was also, for work trips, significantly and positively related to walk and transit choice
(Cambridge Systematics et al., 2002). The lesser importance in the San Francisco travel models of
connectivity, and the lack of significance of conditions at the trip origin, are likely artifacts of model
calibration with travel data from a city with limited pedestrian-friendliness contrasts. Few
city/county of San Francisco non-industrial areas have poor pedestrian connections and most
neighborhoods are basically pedestrian-friendly.
The calibration results for the San Francisco demand model are more extensively covered in
Chapter 15 under "Response by Type of Strategy"--"Site Design"--"Transit Supportive Design
and Travel Behavior"--"Pedestrian/Transit-Friendliness." The same subsection of Chapter 15,
along with the "Individual Urban Design Elements" subsection that precedes it, also presents other
bits of evidence that a relationship between quality of walk connections and transit use indeed
exists.
The 2003 California TOD travel characteristics study performed several analyses of collected sur-
vey data to explore the influence on transit usage of neighborhood design and streetscape attri-
butes specific to station-area developments. The analyses and findings were:
· Simple correlations between design attributes and transit usage for specific trips yielded hints
of positive relationships between more pedestrian friendly elements and greater transit usage,
but also produced some counter-intuitive correlations. The results were most consistent with
expectation for station-area office workers, showing modest positive relationships between
higher transit shares and densities of retail shops, street connectivity, sidewalks on at least one
side, street tree density, street light density, and frequency (shortness) of blocks, with a nega-
tive transit use relationship for street width.
· A multiple regression model of project-level transit mode choice versus project attributes
found greater street tree density, street furniture density, and crosswalk density to be
positively related to greater transit share, all else being equal. This model is presented in
Table 17-28.
· A disaggregate model of individual-level commute trip transit choice identified only
one neighborhood design variable as having statistical significance, namely, street con-
nectivity in the area around the work end of the commute trip. The research model identified
no such significance at the home end of the trip.12 Higher connectivity, measured in this
12 The low statistical significance of most neighborhood design variables examined in this assessment may well
have resulted from low variability in the design variable data set caused by its restriction (at the home end
of the trip) to TOD locations. For a relationship to be established, there must be sufficient differences in vari-
able values among observations.
17-57
OCR for page 58
instance in terms of the proportion of all intersections within the 1/2-mile radius of the station
that are four- or five-way (or more), suggests a more efficient pedestrian environment with less
indirectness of travel involved in walking. All else being equal, greater connectivity around the
workplace was related to an increased probability of transit choice. This research model is pre-
sented later in Table 17-30.
Table 17-28 Multiple Regression Model for Predicting Proportion of All Trips by Transit for
22 Rail-Based Housing Projects
Variable Coef. T-stat
Regional Accessibility
Relative Job Accessibility: Number of jobs that can be reached via transit
network within 60 minutes peak travel time divided by number of jobs that
can be reached via highway network within 60 minutes peak travel time. 1.306 2.317
Neighborhood Design / Station Provisions
Relative Parking Supply: Number of parking spaces at nearest station per
0.011 4.855
100 dwelling units within 1 mile of station.
Street Tree Density: Number of street trees along shortest route from
0.012 2.803
project to station per 1,000 ft walking distance.
Street Furniture Density: Number of street furniture items along shortest
0.016 2.972
route from project to station per 1,000 ft walking distance.
Crosswalk Density: Number of pedestrian crosswalks along shortest route
0.023 2.776
from project to station per 1,000 ft walking distance.
Socio-Demographic Control
Auto ownership levels: Average number of motorized vehicles per
household member 16 years old or older. -0.233 -1.763
Constant -0.079 -0.446
Notes: Based on data for the 22 projects with response rates deemed adequate. R-Squared is 0.811.
Source: Lund, Cervero, and Willson (2004a).
The fact that some neighborhood design variables appeared significant in project-level analysis but
"dropped-out" in individual-level analysis led the researchers to conclude that different individu-
als value neighborhood design and streetscape attributes differently. Such attributes were judged
more highly subjective than measures like travel time or distance. The lack of significance in the
individual-level analysis also led the researchers to urge that not too much be read into the results
of the multiple regression project-level model (Table 17-28) (Lund, Cervero, and Willson, 2004a).13
Pedestrian-friendliness and mixed land use within a TOD should contribute, additively to
station closeness, to walking as a mode-of-access to transit. Highly suggestive is the
Rockridge versus Lafayette comparative analysis provided within Chapter 15, "Land Use and
13 Additional cause for viewing the "Relative [station] Parking Supply" variable in Table 17-28 with special
caution is discussed further-on under "Parking Supply"--"Transit Parking" including footnote "15."
17-58
