Assessing Public Response

to Freeway Roadsides
Urban Forestry and Context-Sensitive Solutions

Kathleen L. Wolf

Social science methods can be used to assess how the public values context- roadside design? What is the driver’s experience of the roadside,
sensitive solutions. The roadside landscape is a public lands resource including urban forest elements?
that has many functions and provides many benefits. Diverse stakehold- This research report makes two contributions to transportation
ers may have varied expectations for roadside design. The urban forest planning. First, on the basis of results of two survey studies, it pro-
is often a contested component of the urban roadside. Two research sur- vides an empirical basis for decision making about vegetation in trans-
veys based on landscape assessment literature were used to assess and portation corridors. Second, it demonstrates methods for eliciting
quantify public preferences and perceptions with regard to trees in high- stakeholder response using design visualization techniques.
speed and freeway roadsides. One photo questionnaire was distributed
in urban areas nationally and the other in Washington State. To elicit
public attitudes about visual quality and community image, each survey BACKGROUND AND RESEARCH
included design visualizations constructed with digitally edited photo-
graphs. Research results were consistent across both studies. Respon- Stakeholders in both the transportation industry and community
dents judged images with increasing amounts of roadside vegetation, groups are collaborating to achieve greater design flexibility and
including trees, to have a higher amenity value. The presence of more the integration of community values in CSS. Highway projects once
extensive community greening was associated with positive consumer focused almost exclusively on faster travel times and safety and were
inferences and greater willingness to pay for goods and services. There perceived as having adverse impacts on communities through which
was little variation in responses across respondent demographics. Results they passed (2). Recent AASHTO and FHWA publications provide
provide an empirical basis for flexible highway design and promote plan- ideas, options, and examples of ways to design highways that are more
ning options for roadside urban forests that address multiple stakeholder environmentally friendly without compromising safety and mobility.
interests.
They stress the importance of eliciting early public participation, iden-
tifying community interests, and using creative thinking to achieve
Roadsides are an important landscape resource in the United States, community-friendly highway design (3).
comprising hundreds of thousands of acres. The public has ever-
increasing experiences of roadside landscapes. In recent decades each
American’s average miles of yearly travel, number of trips, and Landscape Assessment
mileage per trip have increased 60% to 85%. Much of that road time
is spent on urban high-speed roads. Americans currently travel about Landscape assessment studies have been used in natural resource
2.3 billion vehicle miles per day on approximately 61,000 mi of management since the 1960s to explore stakeholders’ perceptions and
urban freeways and highways (1). values with regard to specified landscapes. Data collection approaches
Expanded functions and benefits are expected for all urban pub- include surveys, on-site assessments, visitor-employed photography,
lic lands, including roadsides. Roadsides were historically dedicated and visitor evaluations. Assessments have evaluated public response
to providing clear zone and safety buffers. They are now designed to to urban residential settings and parks (4, 5) and various wildland land-
manage storm water, enhance air quality, perform as wildlife habitat, scapes (6), but have rarely been applied to the study of transportation
and provide visual amenities. landscapes.
The roadside urban forest provides extensive services and benefits. Roadside plants contribute to highway visual quality. In Califor-
Roadside design may include conservation of existing trees or incor- nia (7 ) people judged simulations of proposed roadside residential
porate new tree plantings. The policy and procedures of context- development for scenic quality. Drivers described roadside develop-
sensitive solutions (CSS) enable stakeholder input on freeway design. ment as “cluttered” and “ugly,” whereas “pleasant” and “beautiful”
How does a transportation agency gauge public response to freeway were descriptions of highly vegetated highway corridors. Van passen-
gers recorded attractiveness ratings for urban roadside views in Min-
nesota (8); highest values were awarded to road segments having
College of Forest Resources, University of Washington, Box 352100, Seattle, nature features and well-designed plantings and structural elements.
WA 98195. The presence of perimeter foreground vegetation, wildlife, open-
Transportation Research Record: Journal of the Transportation Research Board,
ness, and flowers were significant indicators of scenic beauty in a pre-
No. 1984, Transportation Research Board of the National Academies, Washington, dictive model derived from public ratings of primarily urban roadside
D.C., 2006, pp. 102–111. landscapes (9).

102
Wolf 103

Landscape assessment theory includes psychological dynamics of commercial office properties were about 7% higher for sites
that are associated with visual quality, such as affect, cognition, and having a good quality landscape that included trees (21).
behavior (10). For instance, roadside character can influence route
choice. Drivers chose a scenic parkway route to a shopping center
more often than a nonscenic expressway route, despite the parkway Research Program
route having more stops and taking more time (11). Drivers enjoyed
views of nature and reported feelings of relaxation while on the Environmental benefits of having trees on urban freeway roadsides
parkway route. are generally recognized. Sparse social science studies hint at the
psychosocial importance of the view from the road. Two studies were
conducted to learn more about public response to urban roadsides.
Nature Response Three questions guided the research:
• How do the urban forest and other landscape elements contribute
Stress is a public health issue. Stress response is documented for all
to the visual quality of the freeway roadside?
driving experiences, though intensity varies depending on road and
• How does roadside landscape affect place perceptions?
traffic conditions. Heart rate variability and blood pressure increase
• Are differences in drivers’ preferences and perceptions associated
when a person is driving in comparison to during nonactivity situa-
with demographic traits?
tions. Demanding driving conditions, such as on-ramps, off-ramps,
and roundabouts, increase the stress response (12).
Commuting may be one of the most stressful experiences of
METHODS
urban life. Increased blood pressure, higher illness rates, lowered
job satisfaction, absenteeism, and lower performance for cogni- Research data were obtained from two surveys detailed in Table 1.
tive tasks are all related to longer or more difficult commutes (13). The instruments were constructed to elicit public response on
Empirical studies have confirmed the restorative effects of passive roadside vegetation design on urban and semiurban high-speed
nature experiences (14, 15). One study specifically addressed the roadsides. Each survey contained variable sets to capture attitudes
effects of roadside character on stress response using physiologi- about aspects of roadside design and landscape. Response included
cal monitoring (16). Simulator scenarios containing built-up, visual image ratings and indicators of agreement to verbal state-
strip-mall style roadsides both slowed down and impeded driver ments. Demographic variables characterized driving behaviors and
recovery from introduced stressors. Subjects viewing roadside socioeconomic conditions.
nature scenes (forests or golf courses) returned to normal baseline Design visualization is used to share design proposals with trans-
conditions faster and had greater ability to cope with stressors. An portation clients. Built elements, as well as details of color and
“immunization effect” was discovered; exposure to a natural road- materials, can be depicted before major construction costs are com-
side setting decreased the magnitude of the stress response to a mitted. Engineers and laypersons can make better decisions about
later task. future transportation projects when considering what-if scenarios.
Trees and nature are also associated with positive cognitive Historically, artist renderings have been used for visualization; com-
appraisals of urban places. Urban scenes containing trees (particu- puterized digital tools now offer advantages of reduced cost and
larly large ones) are consistently highly preferred in urban landscape greater three-dimensional (3-D) precision (22).
assessment. Positive meanings and values are associated with the urban Simple computerized visualization approaches were used in each
forest (17, 18). Natural amenities influence perceptions of urban place of the surveys to gauge public evaluation of roadside elements. Each
and function (19). survey contained black-and-white images (because color reproduc-
Such nonconscious reactions have consequences for community tion was prohibitively expensive) used to prompt preference and
economics. In a series of studies of consumer response to downtown perceptual responses. By using a computer paint program, images
business districts (20), shoppers report increased patronage and pur- of vegetation were digitally copied from photographs and placed in
chasing behavior in districts having a quality urban forest. Rental rates base images to create composites showing hypothetical conditions.

TABLE 1 Content and Methods of Two Research Survey Instruments

Roadside Preference Survey Community Perception Survey

Content 36 images with preference rating response. Verbal variables Image sets for two community scenarios. Perceptual descriptor
about roadside visual features. Demographics variables. variables: two sets. Willingness-to-pay variables for goods and
services. Demographics variables.
Visualization images Six base images containing ground-level views of urban Base images digitally edited to depict two community scenarios,
freeway roadside conditions. Each edited for six land- of low and high levels of community greening. One aerial and
scape treatments. Response ratings: 1 (low)–5 (high) two ground-level images used to depict each community.
for visual quality.
Respondent sampling Geographic focus—residents of exurbs within the metro- Geographic focus: residents of metropolitan areas of Washington
politan areas of Seattle, Wash.; Minneapolis, Minn.; State: Bellingham, Olympia, Vancouver. Random selection of
Detroit, Mich.; and Baltimore, Md. Random selection licensed drivers.
of licensed drivers.
Survey return rates 3,000 surveys mailed, followed by one round of reminder 1,000 surveys mailed, followed by one round of reminder cards:
cards: 404 responses, 421 were undeliverable. 113 responses, 285 were undeliverable. Response rate of 16%.
Response rate of 16%.
104 Transportation Research Record 1984

(a) (b)

FIGURE 1 Scenes of (a) high preference (mean  4.52, 0.76 SD) and (b) low preference (mean  1.40, 0.78 SD).

RESULTS OF ROADSIDE PREFERENCES uses). The public associates trees and reduced views of built settings
with higher visual quality, indicated by a 3-point difference between
Image Preference and Categories high- and low-rated images. Dimensional analysis techniques were
then used to identify underlying response categories. Principal axis
Multistep analysis of preference ratings revealed the public’s value factor analysis with varimax rotation generated categories that were
of the roadside urban forest. Respondents were asked to indicate further defined based on a priori decision rules (10, 23).
how much they liked the content of each scene using a scale in which Five categories accounted for 57% of the total variable variance and
1 denoted “not at all” (low preference) and 5 indicated “very much” included 27 images. Category loadings are statistical indicators of the
(high preference). degree to which a variable fits in any category. Loadings range from 0,
All 36 images were sorted by preference means; Figure 1 depicts or low fit, to 1, or good fit. The threshold loading for inclusion in a cat-
high- and low-rated scenes. Scenes depicting roadsides having no egory was 0.4. New dependent variables were constructed by aggre-
trees received low ratings. Higher scores are associated with increased gating mean values across all category items for each respondent.
density of trees (and resulting screening of adjacent commercial land Figure 2 presents the categories, in ascending order of mean preference.

8 images
loadings .616 to .794
mean 1.56, 0.70 SD

(a)

2 images
loadings .590 to .640
mean 1.66, 0.77 SD

(b)

FIGURE 2 Image rating categories: (a) barren edge, (b) prominent buildings.
(continued)
Wolf 105

10 images
loadings .590 to .744
mean 2.71, 0.79 SD

(c)

2 images
loadings .497 to .674
mean 2.88, 0.86 SD

(d)

7 images
loadings .419 to .797
mean 3.87, 0.74 SD

(e)

FIGURE 2 (continued) Image rating categories: (c) ornamental frame,

(d ) tree buffer, and (e) tree screen.

The lowest-rated category, “barren edge,” displays little roadside Descriptive statistics were generated, followed by dimensional
vegetation. Adjacent commercial property uses (buildings, large analysis (Table 2).
products) are highly visible. Buildings dominate the roadside view The highest-rated category, “vegetation views,” has a mean that
in the next low-rated category. Higher ratings were given to “orna- is notably higher than the other four categories, demonstrating the
mental frame” because installations of shrubs and trees soften visual importance of a good-quality roadside landscape during the driving
obtrusiveness of built elements and screen ground-level views into the experience. The 3.64 rating for the second category, “built attrac-
commercial zone. Dispersed trees visually buffer midground build- tions,” is relatively high compared with public judgments of the
ings or products and create a greater sense of visual balance between built environment in previous urban studies (10, 24). Its high sum-
the built and natural elements of the scene in “tree buffer.” The cat- mary value indicates public interest in a driving environment that
egory with the highest ratings, “tree screen,” depicts opaque tree includes a blend of natural and built elements.
groves that obscure what lies beyond the right-of-way. The viewer The latter three categories specify attributes of the view beyond
detects buildings but cannot discern their commercial purpose. the road. Large signs are not preferred in roadside views. Bill-
boards, in particular, are judged to be detrimental to visual qual-
ity. MacGillivary (25) also found that as the density of billboards
Roadside View Features increases in a section of road, there is a corresponding decrease in
public opinion of visual quality.
Respondents also indicated their agreement on 20 statements about Category 4’s low rating indicates that drivers do pay attention to
roadside features, using a scale of 1 (not at all) to 5 (a great deal). the streaming landscape while driving. Finally, responses in the
106 Transportation Research Record 1984

TABLE 2 Roadside Features

Factor Categorya and Items Mean Rating St. Dev. Factor Loading

Category 1. Vegetation views 4.37 0.57

I like to see a variety of plants in the roadside. 4.43 0.73 .752
Having lots of plants creates a pleasing setting. 4.46 0.76 .628
Having lots of plants along the road is boring. 1.42 0.79 −.560b
The seasonal changes of roadside plants are interesting. 4.53 0.67 .545
I enjoy seeing large trees while driving. 4.52 0.67 .480
I often choose travel routes because they are scenic. 3.86 1.21 .415
Category 2. Built attractions 3.64 0.70
I find the views of nearby architecture to be interesting. 3.38 1.01 .585
There should be a blend of built and natural features near the road. 3.57 1.05 .559
I find views of neighborhoods to be interesting. 3.14 1.07 .548
I enjoy driving through communities with character. 4.33 0.83 .474
Category 3. Large signs 2.18 0.90
Large signs are appropriate for roadside advertising. 2.30 1.06 .635
Billboards and large signs should be regulated. 4.03 1.08 −.599b
Category 4. View attention 1.86 1.03
I don’t pay much attention to what’s beyond the paved edge. 2.04 1.19 .872
I rarely notice the view while driving. 1.77 1.20 .659
Category 5. Trees and safety 2.22 0.97
Trees can block or distract me from seeing signs. 2.39 1.19 .652
Trees near the road are a safety hazard. 2.00 1.10 .514

aPrincipal axis factoring with varimax rotation.

bInverse value used for category calculations.

“trees and safety” category suggest that trees are not considered to depicted roadside trees that framed views of the community beyond,
be safety hazards. This outcome merits further research as trees are as well as landscaped arterials and collectors within the community.
considered to elevate injury risk in run-off-the road accidents (26 ). Trees and vegetation were present in the other community but to a
lesser degree. There were no direct descriptors of social interactions.
Urban vegetation provides cues about social factors that may
RESULTS OF COMMUNITY PERCEPTIONS influence patronage behavior. Retailing studies have compared
“atmospherics” against consumer intentions and behavior. Indoor
Each survey respondent was asked to complete three sets of ver- environmental elements such as music, product layout, and store
bal variables for each of two community scenarios (Figure 3). Ini- lighting contribute to store image (27 ). In turn, store image influences
tial analysis was descriptive, followed by data reduction and then consumers’ perceptions (28). In urban outdoor settings both evalu-
comparisons between scenarios. ative appraisals (29) and affective response (30, 31) are boosted by
the presence of trees.

Place Perceptions Consumer Pricing

Two sets of Likert-scaled variables addressed what drivers thought Retailing studies find that certain store settings and positive staff attrib-
of each community as a potential place to shop and a general busi- utes produce higher price acceptability in consumers (32). Is it possible
ness environment. Ratings for the respective sets of 15 variables that the visual quality of a place has price behavior implications?
ranged from 1, indicating “strongly disagree,” to 7, specifying “strongly A set of response items assessed the nonmarket, nonutility values
agree,” with 4 as a neutral center. of trees in community settings using the contingent valuation method
Principal axis factor analysis was used for dimensional derivation (CVM). O’Doherty (33) regards CVM as a “monetized technique for
of both sets, using the same decision rules as for visual images. Three eliciting public preferences.” Contingent valuation surveys have
categories were derived for the “place to shop” variables, accounting been used to assess public willingness-to-pay (WTP) for urban and
for 67% of the total variable variance. Two categories emerged for rural resource use, conservation, and restoration of environmental
“business environment,” explaining 59% of the variance. damage.
Dummy variables were constructed by aggregating mean values by Respondents were asked to indicate the price they would be will-
category items and participants and then compared across scenarios ing to pay for each of eight items (Table 5). Outlier values were
(Tables 3 and 4) using paired samples T-tests (alpha = .05/3 = .015). removed, and means were calculated for each item by scenario.
Comparative results were consistent across perceptual categories, Response was consistent across all pricing variables and aligned
with higher ratings expressed for the “green” community. Business with perceptual response patterns. From low-cost convenience goods
patronage and shopping experience were judged more favorably in to larger purchases, greater WTP was reported for the community
the greener place. Commitment of merchants to community was also having a greener landscape character, including roadsides having
perceived to be greater in the green community. This is a remarkable trees. Results are consistent with studies that indicate consumers are
outcome because the presentations of community differed only in willing to spend 9% to 12% more for goods and services in central
the degree and character of landscape. Green community images business districts having a forested streetscape (20).
Wolf 107

(a)

(b) (c)

(d)

(e) (f)

FIGURE 3 Community scenarios using digitally edited images: (a–c) Community 1, little planning
for landscape or green space, and (d–f ) Community 2, planning for quality landscape and green
space was done.

RESPONDENTS population. Income was higher than the general populace, with
25% of reporting households having incomes of up to $35,000,
Roadside Preference Survey 43% indicating $35,000 to $75,000, and 32% earning more than
$75,000.
Respondents were somewhat younger than the U.S. population. When asked, “What is the size of the community you live in?”
Forty-five percent of respondents reported an age range in the 40s 22% of respondents indicated “large city” (>100,000 population),
or 50s, 32% were in their 30s or younger, and 23% were in their 60s 21% chose “small city” (20,000 to 100,000), and 42% indicated
or older. Gender distribution was approximately that of the U.S. “suburb of a large city.”
108 Transportation Research Record 1984

TABLE 3 Places to Shop

Factor Categorya and Items (Loading) Greenb Less Greenb p <c

Category 1. Local shopper appeal 5.34 (0.99) 4.75 (1.27) .001

I’d stop here for gas and convenience goods. (.57)
Local residents do much of their shopping here. (.56)
Category 2. Inviting to visitors 5.07 (1.12) 3.37 (1.37) .001
I’d stop here to see what shops are available. (.80)
This is appealing to tourists. (.67)
Category 3. Shopping experience 5.02 (1.03) 4.23 (1.17) .001
Business people are helpful and informative. (.79)
High-quality brands are available. (.72)
A wide selection of products and services is available. (.70)

aPrincipal axis factoring with varimax rotation.

bCategory mean (standard deviation).
cPaired comparisons t-tests.

TABLE 4 Business Environment Categories

Factor Categorya and Items (Loading) Greenb Less Greenb p <c

Category 1. Cooperative merchants 5.15 (0.83) 4.06 (0.92) .001

Public and private organizations work together. (.84)
Merchants care about the community. (.81)
Businesses sponsor community events. (.79)
Business people are community leaders. (.77)
Category 2. Fiscal health 3.27 (1.05) 3.89 (0.98) .001
Local business seems to be struggling financially. (−.54)
The crime rate is high. (−.66)

a
Principal axis factoring with varimax rotation.
b
Category mean (standard deviation).
c
Paired comparisons t-tests.

TABLE 5 Reported Product Pricing Between Community Scenarios

Product or Servicea Greenb Less Greenb % Diff. p <c

Take-out sandwich for lunch 4.93 (1.71) 4.49 (1.52) 10 .001

Flower bouquet 17.16 (10.97) 15.61 (10.64) 10 .001
Pair of sunglasses 18.87 (15.30) 15.78 (12.36) 20 .01
Sit-down dinner for 2 34.08 (12.65) 30.69 (12.49) 11 .001
Lightweight jacket 38.04 (15.90) 34.82 (13.85) 9 .001
Pair of sports shoes 47.07 (21.45) 43.81 (20.25) 7 .005
Motel room for 2, 1 night 62.78 (15.48) 57.09 (14.71) 10 .001
House (3 bedrooms, 2 baths) 145K (39.5K) 131K (36K) 11 .001

a
All means in U.S. dollars (1999).
b
Category mean (standard deviation).
c
Paired comparisons t-tests.

Respondents were also asked to recall their weekly driving habits. 3 h per week on high-speed roads, 27% reported 4 to 6 h, and 27%
Most drivers claimed to spend fewer than 10 h per week in a motor- reported 7 to 12 h. Increased time on freeways may be due to small-
ized vehicle (57%). In addition, 36% claimed 3 h or less per week city settings or the transportation configurations of Washington
on a high-speed road, compared with 28% for 4 to 6 h. State cities.
In regard to the size of the community of residence, more people
lived in small cities than did those involved in the preference survey.
Community Perceptions Survey Nine percent lived in a large city, and 75% claimed to live in a small
city. Rural dwellers numbered about 8%, and suburbs registered at
Sixty-nine percent of community respondents claimed to spend 10 h 5%. Household income was generally lower for this respondent
or less in a motorized vehicle per week. These respondents spent group; 31% reported up to $35,000 annually, 41% reported $35,000
proportionately more time on urban freeways; 27% reported up to to $75,000, and 23% reported greater than $75,000.
Wolf 109

Comparisons taining barren roadsides and views beyond of prominent buildings

or large products received the lowest ratings. Unfortunately they
Respondent characteristics were statistically compared with prefer- depict the visual condition of many urban roadsides. In the midrange
ence, perception, and pricing variables using one-way analysis of of preference are images with intermittent trees or low-growing
variance (ANOVA) and t-tests. Few relationships were identified, shrub masses or both. Plant materials create horizontally or vertically
indicating that people of diverse age, income, residence location, framed views of commercial uses, combining aesthetics with the
and driving behavior hold similar values in regard to the roadside opportunity to distinguish particular businesses adjacent to the road.
urban forest. Highlights are presented. Full tree screening is most appreciated by survey participants, which
Means comparisons were conducted on image preference cate- suggests underlying tensions in roadside management policy because
gories for driver subgroups (alpha = .05). No significant differences business interests insist on open views, whereas the public may prefer
across category means were detected for driver age, household total screening. Although some demographic traits are associated
income, time spent in a motorized vehicle each week, or time spent with variation in image response, the prevailing preference trend
per week in a vehicle on high-speed roads. across all respondents is greater appreciation of freeway roadsides
Considering roadside view categories, age and income differ- that contain trees.
ences were detected. Older drivers are more likely to regard trees as Nature-associated preference is consistent with previous studies
safety hazards (F = 4.17, df = 2, p < .02), but do not find large signs of many landscape settings. Generally, the presence of trees enhances
to be as serious an issue as do younger drivers (F = 7.22, df = 2, p < public judgment of visual quality (31, 35, 36 ). Kaplan (37 ) and
.001). Higher-income respondents (>$75,000) report paying more Schroeder (38) conclude that trees are highly valued components of
attention to the roadside (F = 5.38, df = 2, p < .005) and being less urban settings and that unkept nature in urban settings is less pre-
concerned about trees as safety hazards (F = 5.93, df = 2, p < .003). ferred than well-maintained nature. A three-point difference in
Of particular interest was the relationship of respondent income means between the highest- and lowest-rated scenes is a striking
to pricing response in the community survey. Budget bias can be an example of how plants can affect drivers’ judgments of visual qual-
issue in contingent valuation surveys. Do respondents, in reacting to ity. The degree of balance in built-to-nature scenes influences the
a hypothetical situation, take into account their ability to pay? Past public’s valuation. Modest and welcoming signage is judged to be
research suggests that if indicated willing-to-pay amounts are nom- suitable, whereas billboards and large signs are not acceptable in the
inal, the bias threat is minimal and probably comes into play only roadside. Respondents indicated psychological inferences based on
when major and costly programs or products are valued (34). Such landscape character. The “green” community was characterized as
conclusions were borne out by pricing for most goods. Only one being a more appealing place for shoppers, including positive mer-
item on the list—flower bouquet—was valued higher by individu- chant traits and product quality. The greener place was also judged
als having higher income at alpha = .05/8 = .006 (one-way ANOVA to be a more favorable environment for new businesses.
F = 4.832, df = 89, p < .004). This suggests that the CVM instrument What are the processes that underlie these response patterns?
is income-stable, making it valuable for future research. There has been little study of the mediating cognitive factors that
result in place judgments. The discipline of social psychology offers
insight, being defined as “the scientific study of how individuals think,
DISCUSSION OF RESULTS feel, and behave in regard to other people and how individuals’
thoughts, feelings, and behaviors are affected by other people” (39).
The purpose of the two studies was to provide empirical input for plan- Social perceivers assemble various bits of information and, medi-
ning and managing urban freeway roadsides. The roadside landscape, ated by perceiver dispositions, form impressions of others (40).
comprising land ribbons that crisscross the nation, may have a propor- Humans rely on perceived traits to infer a person’s behavior, evalu-
tionally greater impact on drivers than more expansive public lands, ate the person, and guide decisions about how to interact with a per-
because of the frequency and duration of driving activity. Multiple psy- son (41). Diverse information about a person is integrated to form a
chological responses were explored, starting with public preferences coherent impression. The observer uses subsequent information and
and attitudes concerning aesthetics and extending to cognitive assess- experience to confirm or modify the impression. Survey results sug-
ments of places that have varying degrees of landscape quality. Other gest that built settings evoke similar evaluative responses. Respon-
social science studies contribute knowledge about how roadside dent’s ratings extend beyond immediate physical traits, to include
elements influence driving cognitions, behavior, and physiology. inferences about social and psychological interactions. Social psy-
Public preferences for blends of built and natural elements in chological concepts of “social attribution” and “impression formation”
roadsides were systematically assessed. Results can be used to readily translate to interpretations of urban space and place.
address stakeholder conflicts. Although businesses adjacent to free-
ways (such as car dealerships) may desire “commercial windows”
for full property views, other stakeholder groups (such as neighbor- Opportunities and Limitations
hood associations) may request full screening. Using digital editing,
variations in the number and arrangement of trees (and associated A limitation of this research is the low survey return rate. Typically
vegetation) were presented in surveys to licensed drivers in urban a 20% to 25% return is expected for mail surveys of city residents
areas across the United States. (42), and response rates have declined more recently (43). The pos-
sibility of nonresponse bias must be considered. These research
findings, perhaps of limited generalizability due to respondent self-
Results and Theory selection, convey the positive public valuation of roadside urban
forests. Additional methodologies could improve response rates
Preference means are generally higher on image categories having (44). More direct delivery of surveys, including web-based choice
both increased quantities of vegetation and increased height and tools and placement of computer kiosks (at rest areas, for instance),
density of trees in relationship to commercial views. Scenes con- generates data for flexible design decisions.
110 Transportation Research Record 1984

Another concern is the degree to which one can generalize from on recommendation of the National Urban and Community Forestry
the landscape context of these studies, which contain coniferous for- Advisory Council. Additional project information can be found at
est plantings, to other landscape settings. Replicate studies are needed [Link]/[Link].
to demonstrate the degree to which these findings are applicable to
other landscape and geographic types.
Are static images adequate to represent dynamic drive-by expe- REFERENCES
riences? Budget limitations restricted survey materials to pencil-
and-paper formats. Additional research is needed to corroborate 1. Highway Statistics. FHWA, U.S. Department of Transportation, 1998.
response to still photographs versus views while driving at high 2. NCHRP Report 480: A Guide to Best Practices for Achieving Context-
Sensitive Solutions. TRB, National Research Council, Washington,
speeds, perhaps using computer or driving simulators. D.C., 2002.
Surveys do offer advantages that justify efforts to improve response 3. Moler, S. A. Hallmark of Context Sensitive Design. Public Roads, Vol. 65,
rates. Black-and-white photographs offer a low-cost approach to pro- No. 6, 2002, pp. 6–13.
viding visualizations for public response; color print technologies and 4. Schroeder, H. W. Householders’ Evaluations of Street Trees in Sub-
urban Chicago. In Managing Urban and High-Use Recreation Settings
3-D simulation are now more affordable. Studies have demonstrated (P. H. Gobster, ed.). General Technical Report NC-163. USDA Forest
that photo response procedures pose no problems in obtaining peo- Service North Central Forest Experiment Station, 1992.
ple’s reactions to outdoor settings (35, 45). Preferences for places 5. Dwyer, J. F., H. W. Schroeder, and P. H. Gobster. The Deep Signifi-
shown in photographs correlate highly with on-site responses (46). cance of Urban Trees and Forests. In The Ecological City: Preserving
and Restoring Urban Biodiversity (R. H. Platt, R. A. Rowntree, and
P. C. Muick, eds.). University of Massachusetts Press, Amherst, 1994.
6. Ribe, R. G. The Aesthetics of Forestry: What Has Empirical Preference
Design and Planning Research Taught Us? Environmental Management, Vol. 13, No. 1,
1989, pp. 55–74.
7. Evans, G. W., and K. W. Wood. Assessment of Environmental Aesthet-
Agencies and organizations (e.g., U.S. Forest Service, FHWA) ics in Scenic Highway Corridors. Environment and Behavior, Vol. 12,
have devised guidelines for managing visual quality in transporta- No. 2, 1980, pp. 255–273.
tion corridors. Most guidelines focus on rural and wildland settings; 8. Nassauer, J. I., and D. Larson. Aesthetic Initiative Measurement System:
few address urban and semiurban (e.g., suburbs, strip development) A Means to Achieve Context-Sensitive Design. In Transportation Research
Record: Journal of the Transportation Research Board, No. 1890, Trans-
contexts. Meanwhile, 80% of Americans now live in urban settings, portation Research Board of the National Academies Washington, D.C.,
and most of their travel occurs in built environments. This research 2004, pp. 88–96.
provides a scientific basis for urban visual guidelines. 9. Burley, J. B. Visual and Ecological Environmental Quality Model for
Respondents most preferred having greater views of trees versus Transportation Planning and Design. In Transportation Research Record
1549, TRB, National Research Council, Washington, D.C., 1996,
views of built elements. Yet, this design response is not always fea- pp. 54–60.
sible. In some climatic zones such a landscape approach is extremely 10. Kaplan, R., and S. Kaplan. The Experience of Nature: A Psychological
costly. Also, roadside designers must often juggle the demands of Perspective. Cambridge University Press, United Kingdom, 1989.
multiple (sometimes competing) stakeholder groups. Although drivers 11. Ulrich, R. S. Scenery and the Shopping Trip: The Roadside Envi-
ronment as a Factor in Route Choice. PhD dissertation. University of
may prefer screening vegetation, this practice may not be consistent
Michigan, Ann Arbor, 1974.
with safety conditions, expectations of adjacent property owners, or the 12. Rutley, K. S., and D. G. Mace. Heart Rate as a Measure in Road Layout
economic development plans of neighboring communities. Design. Ergonomics, Vol. 15, No. 2, 1972, pp. 165–173.
Midlevel responses on this study’s verbal and visual variables sug- 13. Novaco, R. W., D. Stokols, and L. Milanesi. Objective and Subjective
Dimensions of Travel Impedance as Determinants of Commuting
gest how to achieve compromise in design, as well as address safety
Stress. American Journal of Community Psychology, Vol. 18, 1990,
concerns about roadside fixed objects. For instance, “ornamental pp. 231–257.
frame” and “tree buffer” categories contain intermittent vegetation 14. Ulrich, R. S., R. F. Simons, B. D. Losito, E. Fiorito, M. A. Miles, and
masses (vertical and horizontal) that exclude the visual impact of M. Zelson. Stress Recovery During Exposure to Natural and Urban
complex ground-level elements (e.g., parking lots). Trees and veg- Environments. Journal of Environmental Psychology, Vol. 11, 1991,
pp. 201–230.
etation then serve as a green “frame” to momentarily focus the driver’s 15. Kaplan, S. The Restorative Benefits of Nature: Toward an Integrative
eye on adjacent elements. Placement of vegetation can be avoided in Framework. Journal of Environmental Psychology, Vol. 15, 1995,
those road segments having a high propensity for run-off-the road pp. 169–182.
accidents, such as the outside of curves. 16. Parsons, R. L., G. Tassinary, R. S. Ulrich, M. R. Hebl, and M. Gross-
man-Alexander. The View from the Road: Implications for Stress
Public input processes are used extensively in urban planning and Recovery and Immunization. Journal of Environmental Psychology,
resource management situations, to incorporate community values Vol. 18, 1998, pp. 113–140.
into policy, planning, and design of both built and natural environ- 17. Chenowith, R. E., and P. H. Gobster. The Nature and Ecology of Aes-
ments. This assessment project and others like it can advance the thetic Experiences in the Landscape. Landscape Journal, Vol. 9, No. 1,
1999, pp. 1–8.
understanding of public response to roadside settings, including the 18. Hull, R. B. How the Public Values Urban Forests. Journal of Arbori-
urban forest. Empirical data can support more comprehensive and culture, Vol. 18, No. 2, 1992, pp. 98–101.
defensible context-sensitive solutions. 19. Herzog, T. R. A Cognitive Analysis of Preference for Urban Nature.
Journal of Environmental Psychology, Vol. 9, 1989, pp. 27–43.
20. Wolf, K. L. Business District Streetscapes, Trees, and Consumer
Response. Journal of Forestry, Vol. 103, No. 8, 2005, pp. 396–400.
ACKNOWLEDGMENTS 21. Laverne, R. J., and K. Winson-Geideman. The Influence of Trees and
Landscaping on Rental Rates at Office Buildings. Journal of Arbori-
Appreciation is extended to the Landscape Architecture and Road- culture, Vol. 29, No. 5, 2003, pp. 281–290.
22. McDaniel, R. E. See It Before It’s Built. Public Roads, Vol. 65, No. 6, 2002,
side Management professionals of the Washington State Depart- pp. 28–31.
ment of Transportation for substantial contributions to this research. 23. Wolf, K. L. Freeway Roadside Management: The Urban Forest Beyond the
Projects were supported by a grant from the USDA Forest Service White Line. Journal of Arboriculture, Vol. 29, No. 3, 2003, pp. 127–136.
Wolf 111

24. Herzog, T. R., and T. A. Gale. Preference for Urban Buildings as 36. Ulrich, R. S. Human Responses to Vegetation and Landscapes. Land-
a Function of Age and Nature Context. Environment and Behavior, scape and Urban Planning, Vol. 13, 1986, pp. 29–44.
Vol. 28, No. 1, 1996, pp. 44–72. 37. Kaplan, R. The Role of Nature in the Urban Context. In Behavior and
25. MacGillivary, C. I. A Study of Billboards and Junkyards as Related to the Natural Environment (I. Altman and J. F. Wohlwill, eds.). Plenum,
Some Aspects of the Aesthetics of the Highway Environment. Joint High- New York, 1983.
way Research Project, Engineering Experiment Station, File No. 9-5-4, 38. Schroeder, H. W. Environment, Behavior and Design Research on
Project No. C-36-48D. Purdue University, West Lafayette, Ind., 1969. Urban Forests. In Advances in Environment, Behavior and Design,
26. Wolf, K. L., and N. J. Bratton. Urban Trees and Traffic Safety: Consid- Vol. 2 (E. H. Zube and G. T. Moore, eds.). Plenum, New York, 1989.
ering U.S. Roadside Policy and Crash Data. Arboriculture and Urban 39. Brehm, S. S., S. M. Kassin, and S. Fein. Social Psychology. Houghton
Forestry (in press). Mifflin, New York, 1999.
27. Zimmer, M. R., and L. L. Golden. Impressions of Retail Stores: A Con- 40. Leyens, J. P., and S. T. Fiske. Impression Formation: From Recitals to
tent Analysis of Consumer Images. Journal of Retailing, Vol. 64, 1988, Symphonie Fantastique. In Social Cognition: Impact on Social Psychol-
pp. 265–293. ogy (P. G. Devine, D. L. Hamilton, and T. M. Ostrom, eds.). Academic
28. Dodds, W. B., K. B. Monroe, and D. Grewal. The Effect of Price, Brand, Press, New York, 1994.
and Store Information on Buyer’s Product Evaluations. Journal of Mar- 41. Wyer, R. S., and A. J. Lambert. The Role of Trait Constructs in Person
keting Research, Vol. 28, 1991, pp. 307–319. Perception: An Historical Perspective. In Social Cognition: Impact on
29. Nasar, J. L. The Evaluative Image of the City. Sage, Thousand Oaks, Social Psychology (P. G. Devine, D. L. Hamilton, and T. M. Ostrom,
Calif., 1998. eds.). Academic Press, New York, 1994.
30. Sheets, V. L. Affect, Cognition, and Urban Vegetation: Some Effects of 42. Sommer, R., H. Guenther, and P. A. Barker. Surveying Householder
Adding Trees Along City Streets. Environment and Behavior, Vol. 23, Response to Street Trees. Landscape Journal, Vol. 9, No. 2, 1990,
No. 3, 1991, pp. 285–304. pp. 79–85.
31. Smardon, R. C. Perception and Aesthetics of the Urban Environment: 43. Elmendorf, W. F., and A. E. Luloff. Using Qualitative Data Collection
Review of the Role of Vegetation. Landscape and Urban Planning, Methods When Planning for Community Forests. Journal of Arboricul-
Vol. 15, 1988, pp. 85–106. ture, Vol. 27, No. 3, 2001, pp. 139–151.
32. Grewal, D., and J. Baker. Do Retail Store Environmental Factors Affect 44. Dillman, D. A. Mail and Internet Surveys: The Tailored Design Method.
Consumers’ Price Acceptability? An Empirical Evaluation. International Wiley & Sons, New York, 2000.
Journal of Marketing, Vol. 1, 1994, pp. 107–115. 45. Sullivan, W. C. Perceptions of the Rural–Urban Fringe: Citizen Prefer-
33. O’Doherty, R. Using Contingent Valuation to Enhance Public Participation ences for Natural and Developed Settings. Landscape and Urban Plan-
in Local Planning. Regional Studies, Vol. 30, No. 7, 1996, pp. 667–678. ning, Vol. 29, 1994, pp. 85–101.
34. Mitchell, R. C., and R. T. Carson. Using Surveys to Value Public 46. Stamps, A. E. Use of Photographs to Simulate Environments: A Meta-
Goods: The Contingent Valuation Method. Resources for the Future, Analysis. Perceptual and Motor Skills, Vol. 71, 1990, pp. 907–913.
Washington, D.C., 1989.
35. Kaplan, R. The Analysis of Perception via Preference: A Strategy for
Studying How the Environment Is Experienced. Landscape and Urban The Context-Sensitive Design–Solutions Task Force sponsored publication of this
Planning, Vol. 12, 1985, pp. 161–176. paper.