100 words for each bibliography –Annotate the article–Research into human experience of this kind of place–Relevant organizational philosophy, goals, procedures, and policies related to your building type–Guidelines, standards or codes for design of this building type–Current design issues and trends, as reflected in the news, popular press, and professional journals of this building type
retail_land_use__neighborhood_satisfaction_and_the_urban_forest.pdf
retail_lighting_and_consumer_product_perception.pdf
the_hidden_ways_that_architecture_affects_how_you_feel.pdf
Unformatted Attachment Preview
Landscape and Urban Planning 74 (2006) 70–78
Retail land use, neighborhood satisfaction and the urban forest:
an investigation into the moderating and mediating
effects of trees and shrubs
Christopher D. Ellis∗ , Sang-Woo Lee1 , Byoung-Suk Kweon
Department of Landscape Architecture and Urban Planning, Texas A&M University, 311-A Langford Architecture Center,
College Station, TX 77845-3137, USA
Received 17 January 2003; received in revised form 23 August 2004; accepted 12 October 2004
Available online 19 January 2005
Abstract
This paper examines the relationship between retail land use and neighborhood satisfaction along with the moderating and
mediating effects of trees and shrubs. Neighborhood satisfaction has been related to a number of environmental factors including
land uses. However, no other research has reported the potential moderating and mediating effects of trees on these relationships.
This study included residents living in single-family housing located in typical suburban-type subdivisions with adjacent commercial strip development. Mail-in survey responses were geo-referenced to land parcel centroids, and compared to the amount
of retail land use, and tree and shrub cover existing within 1500 feet. Tree and shrub cover was measured using multi-spectral
satellite imagery classified with a normalized differences vegetation index (NDVI). Existing land use and parcel data were
acquired from the local city planning agency. Results indicate that the amount of tree and shrub cover within a 1500 ft radius
of single-family households significantly moderates and mediates the negative relationship between the amount of nearby retail
land use and neighborhood satisfaction. These results have important implications for urban planners and landscape architects.
Specifically, the findings suggest that communities should increase provisions for protecting and establishing trees and shrubs
in neighborhoods near retail land uses.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Neighborhood satisfaction; Quality of life; Land use planning; Urban development; Landscape architecture; Trees and shrubs
∗
Corresponding author. Tel.: +1 979 845 7857; fax: +1 979 862 1784.
E-mail addresses: [email protected] (C.D. Ellis), [email protected] (S.-W. Lee), [email protected] (B.-S. Kweon).
1 Present address: Program in Landscape Architecture, The University of Texas at Arlington, 601 W. Nedderman Drive, Arlington,
TX 76019-0108, USA.
0169-2046/$20.00 © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.landurbplan.2004.10.004
C.D. Ellis et al. / Landscape and Urban Planning 74 (2006) 70–78
1. Introduction
Current zoning practices separate residential, commercial, and industrial areas for the stated purposes of
protecting public health, safety, and welfare as well as
to promote a clean environment and better quality of
life. Regulations and standards for designing infrastructure within different zones such as the width of
right-of-ways, streets, sidewalks, and so on literally
shape the community environment in which we live,
work, and move about. In most communities these standards do not specify extensive requirements for planting trees and shrubs or other forms of vegetation except
to control for erosion or to prevent the obstruction of
driver sight lines at intersections. This is true despite
the growing evidence that urban forests and street trees
have been linked to numerous health, social, environmental, and economic benefits (Bolitzer and Netusil,
2000; Kweon et al., 1998; Tyrväinen, 1997; Ulrich,
1984).
Many designers and social scientists have questioned existing design standards and zoning practices
claiming they hurt communities by reducing social interaction, consuming fiscal and natural resources, and
increasing dependency on automobiles (Duany and
Plater-Zyberk, 1994; Katz, 1994; Krier, 1991). Among
prescriptions for better community design is the expectation that mixed land uses combined with walkable distances from residences to commercial nodes
will reduce automobile use and infrastructure costs,
while increasing social interactions (Achimore, 1993;
Audirac and Shermyen, 1994; Calthorpe, 1993). They
advocate a more “traditional” neighborhood structure
(pre-World War II) that is believed to support a greater
sense of community and therefore, a better quality of
life. These benefits have yet to be fully quantified. On
the contrary, some evidence exists that neighborhood
satisfaction is negatively related to non-residential zoning (Jirovec et al., 1984). Retail land use generates
more garbage and traffic than residential land use
and subsequently may lead to undesirable conditions
such as offensive odors, noise, and low visual quality. Kaplan (2001) reported that views of busy traffic, and structure dominated views from one’s home
have a negative influence on neighborhood satisfaction. If planners and designers are to advocate mixing land uses such as community retail in residential neighborhoods, then it is important to understand
71
how this can affect the quality of life of the residents.
Assuming retail land use is found to have a negative relationship with neighborhood satisfaction—yet
some benefits to mixed land uses are also found to
exist—then identifying environmental conditions that
could mitigate the negative effects of retail on satisfaction is essential. The amount of nearby trees and
shrubs may play a role in moderating or mediating
these negative effects. Researchers have investigated
the role of trees on neighborhood satisfaction. For example, Peterson (1967) found that harmony with nature is a preferred quality of neighborhood residents.
Kaplan (1983, 1985, 2001) reported that the most important factors in neighborhood satisfaction are the
availability of nearby trees, well-landscaped grounds,
and places for taking a walk. Perceptions of nearby nature also influence residential satisfaction in single and
multiple family housing clustered together (Talbot and
Kaplan, 1991). Other attributes that have been identified as important contributors to neighborhood satisfaction include upkeep and maintenance of homes
and yards (Lansing and Marans, 1969; Sirgy and Cornwell, 2002), sense of safety (Cook, 1988), place to live
(Ahlbrandt and Cunningham, 1979; Dahmann, 1981;
Lu, 1999), appearance of the neighborhood (Parkes et
al., 2002), and overall satisfaction (Galster and Hesser,
1981), as well as age, life cycle stage, income level,
family size, home ownership, etc. (Francescato et al.,
1987; Lu, 1999).
In addition to neighborhood satisfaction, links have
been made between natural elements such as trees and
shrubs, and other quality of life factors. According to
Cooper Marcus and Sarkissian (1986), a positive relationship between natural elements and psychological and social responses of people in urban settings
has been shown to be consistent in numerous psychological and social studies. Kweon et al. (1998) reported that spending time in green outdoor common
spaces is related to stronger social interaction among
residents. Hammitt (2002) found that people are attracted to urban forests and that they provide more
opportunities for needed privacy (voluntary physical
or psychological withdrawal) than can be achieved
at home or work. Several other studies report that
natural elements support public health and reduce
levels of violence and crime in the inner-city by aiding in the recovery from mental fatigue (e.g. Kaplan,
72
C.D. Ellis et al. / Landscape and Urban Planning 74 (2006) 70–78
1984; Kaplan and Kaplan, 1989; Kuo and Sullivan,
2001; Miles et al., 1998). In addition, Ulrich (1984)
and Ulrich et al. (1991) found that natural elements
speed recovery from surgery as well as aid in stress
reduction.
Urban nature has also been related to increases in
housing price. Based on the investigation of a hedonic
price model, Luttik (2000) found that attractive landscape types were shown to draw a premium of 5–12%
more in housing prices over less attractive environment
settings. Tyrväinen (1997, 2001) and Tyrväinen and
Väänänen (1998) also found that the proximity of watercourses and wooded recreation areas as well as a
higher proportion of total forested area in the housing
district had a positive influence on apartment prices.
Both trees and shrubs are expected to play a significant role in affecting the quality of life of community
residents. To understand these relationships, this paper attempts to answer the following questions: (1) is
there a relationship between nearby retail land use and
neighborhood satisfaction in single-family residential
neighborhoods, and (2) if a relationship exists, do trees
and shrubs have a moderating or mediating effect on
that relationship?
2. Methods
This study is a part of a larger study investigating
the effects of physical environments on residents’ wellbeing such as physical activity level, travel behavior,
and activity participation. In the study reported here,
the amount of retail land use within proximity (1500 ft)
of a residence was hypothesized to negatively affect
neighborhood satisfaction. In addition, the amount of
tree and shrub cover within that same proximity was
hypothesized to moderate and mediate the negative effects of retail land use.
2.1. Sampling
Our sampling was conducted among a total population of 9116 single-family households that exist in
suburban-type subdivisions within the City of College
Station, Texas. Eight hundred survey questionnaires
were mailed to a systematic random sample of these
single-family households. We randomly chose participants from environments with many trees, to envi-
ronments with few trees to insure the variability of
physical environments (Fig. 1). These sampling areas
(many trees/shrubs and few trees/shrubs) were determined by conducting an unsupervised classification on
a panchromatic aerial photo at 1 m resolution using
ERDAS Imagine. Among 800 survey questionnaires,
39 questionnaires (5%) were returned with a vacant
notice while 311 questionnaires (41%) were returned
with a valid street address. The address was essential
for comparing the perception data with the GIS data.
Not all respondents lived within 1500 ft of retail land
use. For this study, we used only a subset of questionnaires in which the calculated area of retail land use
within 1500 ft was greater than zero (122 total).
2.2. Participants
Of the 122 observations used in this study, 70 (57%)
participants were women and 47 (39%) were men with
five not responding on gender. The average age of the
participants was 43.3 years and ranged from 18 to 80
years old. The distribution of the highest educational
degree achieved was: 17 high school/GED (14%), 18
community college or technical school (15%), 43 college (34%), and 44 graduate degrees (36%). In terms of
total annual household income, 29 (28%) earned more
than $ 80 000, 32 (26%) ranged between $ 60 000 and
$ 80 000, 29 (24%) between $ 40 000 and $ 60 000, 15
(12%) between $ 20 000 and $ 40 000, and 13 (11%)
between $ 0 and $ 20 000. Home owners made up 101
(83%) of the participants, with 91 (75%) living at their
current location for less than 10 years. White was the
dominant ethnic group with 107 participants (88%).
The chosen sample is closely representative of the College Station, Texas, area (Table 1) with differences existing primarily in the income and education levels.
These differences are likely due to the sample being
drawn from single-family housing alone. Single-family
housing accounts for 39% of all housing types in College Station, Texas.
2.3. Measures
The measures used in this study include retail land
use, tree and shrub cover, and neighborhood satisfaction. Data on retail land use along with a complete
database of residential parcel boundaries were derived
from GIS files provided by the City of College Station
C.D. Ellis et al. / Landscape and Urban Planning 74 (2006) 70–78
73
Fig. 1. The typical environmental conditions tested.
Table 1
A demographic comparison between the study sample and College Station population
Item
Study sample
Gender
Female
Male
High school/GED
Comm. college
College
Graduate school
$ 0–20000
$ 20001–$ 40000
$ 40001–$ 60000
$ 60001–$ 80000
Over $ 80000
White
Education
Income
Race
Population (US census, 2000)
57%
39%
14%
15%
34%
36%
11%
12%
24%
26%
28%
88%
Female
Male
High school/GED
Comm. college
College
Graduate school
$ 0–25000
$ 25001–$ 50000
$ 50001–$ 75000
Over $ 75001
49%
51%
12%
24%
29%
30%
27%
21%
20%
33%
White
81%
74
C.D. Ellis et al. / Landscape and Urban Planning 74 (2006) 70–78
Geographic Information Services. The amount of retail
land use located within 1500 ft of the centroid of each
respondent’s home parcel was measured and recorded
in the database.
Data on the tree and shrub cover were derived from
satellite imagery. Four meter multi-spectral satellite
imagery was processed using a normalized difference
vegetation index formula (NDVI)1 and classified into
categories that included trees and shrubs. The amount
of tree and shrub cover located within 1500 ft of the
centroid of each respondent’s home parcel was measured and recorded in the database.
Neighborhood satisfaction was measured with eight
questions including “neighborhood as a place to live”,
“residents’ pride in their neighborhood”, “physical
condition of their neighborhood”, “safety”, “cleanliness”, “appearance of houses”, “quality of parks”, and
“overall satisfaction”. Participants responded by marking an “X” on a continuous graphic rating scale with
six equal intervals. The two ends of the scale were “Not
Satisfied” and “Very Satisfied”. Varimax rotation factor analysis of the eight items generated one factor.
The quality of parks question was excluded from the
factor due to low factor loading. The factor consists
of seven items and has internal consistency of 0.92.
The proportion of variance accounted for the factor is
63%. The neighborhood satisfaction score was calculated as the sum of the seven item scores from this factor with the possible range of 0–42 and was normally
distributed. Survey responses were address-matched to
College Station parcel boundaries in a geographic information system (GIS) to enable spatial analysis.
3. Results
3.1. Relationships between retail land use and
neighborhood satisfaction
The inter-correlations among the variables were calculated to determine if the hypothesized relationships
were present. The amount of retail land use within
close proximity (1500 ft) of a residence was hypoth1 NDVI is a radiometric measure of the amount, structure, and
condition of vegetation (Huete, 1988). The formula obtains the
grid cell values in the red (RED) and near infrared (NIR) image bands to calculate a vegetation cover image using the formula
(NIR − RED)/(NIR + RED).
esized to negatively affect neighborhood satisfaction.
In this case, retail land use was found to have negative
correlations with neighborhood satisfaction (r = −0.28,
P < 0.01), and trees and shrubs (r = −0.28, P < 0.01)
while neighborhood satisfaction was found to be positively related to trees and shrubs (r = 0.31, P < 0.01).
These relationships are summarized in Table 2. In addition, income was negatively related to retail land use
(r = −0.33, P < 0.01) and length of stay was positively
related to trees and shrubs (r = 0.23, P < 0.05). Neither
income nor length of stay showed a significant relationship with neighborhood satisfaction. Only level of
education had a positive relationship with neighborhood satisfaction (r = 0.22, P < 0.05).
3.2. Moderation effect of trees and shrubs on
the relationship between retail land use and
neighborhood satisfaction
A moderator can be understood as a third variable
that can change the direction and strength of the relationship between dependent and independent variables
(Baron and Kenny, 1986). In statistical terms, Baron
and Kenny (1986) explained that moderation can be
thought of as an interaction between an independent
variable and a third variable (i.e., moderator) to a dependent variable. The moderation effect of trees and
shrubs on the relationship between retail land use and
neighborhood satisfaction is tested by examining the
presence of path c in Fig. 2 (Baron and Kenny, 1986;
Cohen and Cohen, 1983, p. 56).
The amount of trees and shrubs within proximity to
residences was tested for a moderation effect on the relationship between the amount of nearby retail land use
and neighborhood satisfaction (Table 3). The amount of
trees and shrubs was divided into two categories: “few
trees and shrubs” and “many trees and shrubs”. This
was done by applying a median split on the 121 sample
observations using the trees and shrubs area variable.
In neighborhoods with few trees and shrubs (Model 1),
the amount of retail land use negatively affects neighborhood satisfaction (β = −0.37, P < 0.01) and shows
a moderate explanation of the variance (R2 = 0.14). In
neighborhoods with many trees and shrubs (Model 2),
the amount of retail land use does not significantly affect neighborhood satisfaction. The moderation effect
of trees and shrubs on the relationship between retail
land use and neighborhood satisfaction was determined
C.D. Ellis et al. / Landscape and Urban Planning 74 (2006) 70–78
75
Table 2
Inter-correlations among retail land use, neighborhood satisfaction, trees and shrubs, and some demographic variables
Variables
Retail land use
Neighborhood satisfaction
Trees and shrubs
Income
Level of education
Retail land use
Neighborhood satisfaction
Trees and shrubs
Income
Level of education
Length of stay
−0.28**
−0.28**
−0.33**
−0.13
−0.09
0.31**
0.18
0.22*
0.01
0.10
0.14
0.23*
0.45***
0.37**
0.22*
*
**
***
P < 0.05.
P < 0.01.
P < 0.001.
by comparing the slopes in Model 1 (B = −0.18) and
Model 2 (B = 0.21) using a t-test (tscore = 2.51, P < 0.01,
d.f. = 114) as described by Cohen and Cohen (1983, pp.
55–56). The significant t-value indicates a significant
difference between the two slopes and subsequently,
the amount of nearby trees and shrubs can be said to
moderate the negative effect of the amount of nearby
retail land use on neighborhood satisfaction.
3.3. Mediation effect of trees and shrubs on the
relationship between heterogeneous land use types
and neighborhood satisfaction
A variable can be considered a mediator to the extent that it accounts for the relationship between the
independent and dependent variables. A mediator differs from a moderator. While a moderator is treated
as a second independent variable, a mediator functions
as a mechanism through which the dependent variable
influences the independent variable. In other words,
“moderator variables specify when certain effects will
hold, while mediators speak to how or why such effects occur” (Baron and Kenny, 1986, p. 1176). In the
mediation model diagram (Fig. 3), path a represents a
hypothesized negative relationship between the amount
of retail land use and the amount of trees and shrubs
within 1500 ft of a residence. In a practical sense, it
implies that increases in the amount of retail land use
near a residence will reduce the amount of trees and
shrubs in that area. Path b refers to the positive function of trees and shrubs on neighborhood satisfaction,
and path c hypothesizes that the amount of retail land
use located near a home will decrease the resident’s
neighborhood satisfaction.
In order for the mediation effect of the model to hold,
the relationships stated above must be estimated and
tested separately (Baron and Kenny, 1986). That is, the
mediation effects of trees and shrubs on the relationship
between retail land use and neighborhood satisfaction
can only be sustained if the following three relationships are satisfied: (1) the amount of retail land use
affects the amount of trees and shr ...
Purchase answer to see full
attachment
