Land Use Policy 64 (2017) 233–244

Contents lists available at ScienceDirect

Land Use Policy

journal homepage: www.elsevier.com/locate/landusepol

Urban parks: Visitors’ perceptions versus spatial indicators

Gyula Kothencz ∗ , Thomas Blaschke
Department of Geoinformatics – Z GIS, University of Salzburg, Schillerstraße 30, 5020, Salzburg, Austria

a r t i c l e i n f o a b s t r a c t

Article history: Urban green spaces and their role in the quality of life of residents have been studied across multi-
Received 24 June 2016 ple disciplines, based on empirical measurements or qualitative studies – however, the relation, and its
Received in revised form 6 February 2017 strength, between spatial indicators of urban green spaces and visitors’ perceptions of green spaces are
Accepted 7 February 2017
less known. Addressing this knowledge deficit, the present research uses a Geographical Information
Available online 8 March 2017
System (GIS) to link subjective evaluations of the physical environment and objective spatial indicators,
to examine the correlation between the perceived and objective characteristics of five urban parks in the
Keywords:
city of Szeged, Hungary. A questionnaire survey was used to collect residents’ subjective perceptions of
Urban park
Urban green space
the parks, while objective – which is to say measurable and mappable – spatially explicit indicators of the
Subjective interpretation of the physical respective green spaces were calculated using GIS. The subjective evaluations of the parks were matched
environment to the objective indicators using a multiple regression analysis. The statistical analysis yielded two mod-
Objective, spatial indicators erate and two minor correlations between the human perceptions of the investigated green spaces and
Perceived quality of life the nine objective environmental indicators examined. These results showed that subjective evaluations
GIS and objective data reveal different aspects of the same reality. Therefore, the recommendations from this
study are to collectively use human perceptions and objective environmental indicators, both of which
are fundamental for adequately capturing the role of urban green spaces in quality of life.
© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).

1. Introduction
Urban society is largely detached from the natural environ-
ment. This triggers a great demand for the services provided by
urban green spaces. Urban green areas are essential contributors to
human health and quality of life (Lee and Maheswaran, 2011; Maas
et al., 2006), and they therefore provide a great scope for investi-
gating the human perceptions of the physical environment. Beyond
the most obvious aesthetical benefits (Chen et al., 2009; James et al.,
2009) served by urban green spaces, urban green contributes to the
recreation of inhabitants and to the maintenance of social cohesion
(Germann-Chiari and Seeland, 2004; Kaźmierczak, 2013; Tzoulas
et al., 2007). Other essential societal benefits, such as the advanta-
geous contribution of urban green to physical and mental health,
are also widely emphasised (Van Herzele and Wiedemann, 2003;
Wolch et al., 2014). Urban green surfaces effectively regulate the
microclimate of their surroundings (Oliveira et al., 2011; Takács
et al., 2014), and substantially contribute to air purification in cities
(Escobedo and Nowak, 2009). An indirect, but still essential service
∗ Corresponding author.
E-mail addresses: gyula.kothencz@sbg.ac.at (G. Kothencz),
thomas.blaschke@sbg.ac.at (T. Blaschke).
to human well-being is the shelter and habitat that green spaces
provide for biodiversity (Fontana et al., 2011).
The human perception of the ambient environment is subjec-
tive, and differs from person to person (Hernández-Morcillo et al.,
2013; Langemeyer et al., 2015); hence, the benefits derived from
urban green spaces and their objective properties are interpreted
individually. Therefore, the analysis of perceptual information
on green spaces is challenging, although not unprecedented. For
example, Baur et al. (2013) researched the public attitude about
urban nature parks of Portland, USA, based on cognitive, affective
and behavioural components, and on social networks. Irvine et al.
(2009) analysed the soundscape of green spaces in Sheffield, UK.
Jim and Chen (2006) investigated residents’ perceptions of ecosys-
tem services, and the condition and design preferences of green
spaces in Guangzhou, China.
In contrast to perceived benefits, the objective indicators of
urban green spaces capture well measurable and quantifiable phys-
ical attributes (Tsurumi and Managi, 2015). As their analysis is
relatively straightforward compared to subjective data genres,
objective indicators of green spaces are extensively researched
(Bowler et al., 2010) and utilised to support urban planning (Lakes
and Kim, 2012). For example, objective indicators of green spaces
are applied for biodiversity preservation (Fontana et al., 2011) or
microclimate regulation (Lin et al., 2015).

http://dx.doi.org/10.1016/j.landusepol.2017.02.012
0264-8377/© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
234 G. Kothencz, T. Blaschke / Land Use Policy 64 (2017) 233–244

Understanding residents’ perceptions of the services provided
by green spaces is equally important to quantitative data, and both
are crucial for the assessment of quality of life. Quality of life studies
have hitherto focused on either objective or subjective research,
but rarely analysed the relation of the two different information
domains. One of the few detailed studies using spatially explicit
methods found relatively low correlations (McCrea et al., 2006).
This finding corresponds with those from von Wirth et al. (2015),
who identified a weak relationship between subjectively assessed

Fig. 1. Geographic location of Szeged.

 G. Kothencz, T. Blaschke / Land Use Policy 64 (2017) 233–244
and objective quality of life indicators of the safety in public spaces
and the access to central urban facilities.
Studies exploring the linkage between perceived and objec-
tive attributes of urban green spaces are equally infrequent. In
the context of subjective and objective urban greenness, Rhew
et al. (2011) and Leslie et al. (2010) probed possible correlations
between values of the Normalized Difference Vegetation Index
(NDVI) and perceived greenness. Rhew et al. (2011) noted a strong
correlation between the perceived greenness, as reported by envi-
ronmental psychologists, and NDVI values. In contrast, based on
a survey amongst citizens of Warrnambool, Australia, Leslie et al.
(2010) found a lack of agreement between subjective neighbour-
hood greenness and the NDVI.
The knowledge deficit is even more apparent when the relation
between objective and subjective indicators of urban green spaces
is confounded by spatial indicators of the green spaces.
Considering all the aforementioned aspects in relation to urban
green area assessment, this study addressed the following knowl-
edge gap: the degree of relationship between visitors’ subjective
evaluations of urban parks and objective, spatially explicit indica-
tors of the same public green spaces remains to be explored.
Based on the outlined research gap, the objective of this study
was to investigate the degree of relationship between human
perceptions of urban green spaces and spatial environmental indi-
cators of these studied green spaces by matching perceived and
reported features of the parks to their spatial environmental indi-
cators.
Fig. 2. The studied parks.
235
2. Methods
2.1. Study area
The authors studied five parks of the city of Szeged, Hungary,
which is geographically situated in the Great Hungarian Plain
region and stretching across the River Tisza (Fig. 1). Szeged is the
economic, educational and cultural centre of the “Southern Great
Plain” NUTS 2 statistical region, and is home to 160,000 citizens.
The land cover of the urban area is highly diverse. The city centre
was built in a classic late 19th to early 20th century architecture.
The densely built up inner city districts around the historical cen-
tre are surrounded by five to ten storey apartment blocks in the
outer areas. The outermost districts are characterised by one or
two storey detached houses.
Within this urban construct, a variety of public green areas sup-
port the recreational demand of city dwellers. Our study focused
on five parks of Szeged (Fig. 2). The 20 ha Erzsébet liget (ER) is in
the vicinity of the city centre, and is mostly covered by lawn and
patches of woods, making it the most natural green space of the
studied areas. The surroundings are characterised by diverse archi-
tecture, ranging from spaciously placed (20th century, early 21st
century) detached and semi-detached houses with gardens, to 5–10
floor apartment blocks, to schools, student housings and sports
facilities. The Dugonics tér (DU) is a recently (2014) refurbished
square right in the city centre with 50% sealed surface to cater to
the needs of extensive pedestrian traffic. The Széchenyi tér (SZ) is
236 G. Kothencz, T. Blaschke / Land Use Policy 64 (2017) 233–244

the attractive main square of the city, largely covered by appeal-
ing greenery, flower beds and trees. The architecture around both
DU and SZ is dominated by late 19th early 20th century architec-
ture (neo-baroque, eclectic, etc.). The spacious Vér-tó square (VE),
of which a substantial part is covered by a lake, is situated near the
edge of the city, surrounded by 5–10 storey apartment blocks. The
fifth area, the Zápor-tó square (ZA) is characterised by similar envi-
ronmental and sociological settings as the VE, with a pond in the
middle, while being surrounded by blocks of flats. The differences
to the VE are the lower noise disturbance and the pleasant atmo-
sphere. While ER, DU and SZ attract the whole spectrum of Szeged’s
society, VE and ZA are almost exclusively visited by the middle class
and lower middle class population living in nearby housing.
Initially, surveys were also conducted in two additional parks, in
the Mátyás tér (MA) and the Búvár-tó (BU). Due to the unbalanced
gender distribution of the surveyed participants in these parks the
two areas were not included in the multiple regression analysis (for
details, see Section 2.7). Consequently, MA and BU are not described
in greater detail, and are only noted throughout the article when
relevant.
2.2. Hypotheses
We formulated a number of assumptions about the subjective
perceptions of urban parks. The hypotheses development was built
upon evidence from previous literature (Table 1). A survey was
designed to collect subjective data for testing the hypotheses. The
questions in the survey addressed these hypotheses, aiming to
reveal participants’ perceptions. Objective spatial indicators based
on remote sensing and GIS analyses were expected to correlate with
the aggregated survey responses (Table 1).
2.3. Source of subjective data: survey
A survey was designed to investigate park visitors’ subjective
evaluations of features and services of the five green spaces. Top-
ics of the survey were the personal judgement of the park, the
greenness, the accessibility, and the functions of the park. Partici-
pants were asked to respond to questions on a Likert scale, with
ratings between 1 (worst) and 5 (best). An early version of the
survey was tested with experts and non-experts and unsatisfac-

Table 1
Hypotheses on perceived and objective spatiality of urban parks.

Hypotheses Referring question(s) (Subjective evaluations) Verifying objective property (Objective spatial indicators)

1. Bigger sized parks are more appealing to
visitors (Nordh et al., 2009; Schipperijn
et al., 2010).
2. Greener parks are more appealing to
visitors (Nordh et al., 2009; Peschardt
et al., 2016).
How much do you like the area? Area of the parks
How much does the area satisfy the below
functions? => The look of the area, scenery.
How natural do you think the area is? Area weighted NDVI for the parks
Are there enough green areas here? Area of vegetated surfaces

How much does the greenness of the area

contribute to the quality of life?

How much does the area satisfy the below

3. Parks with greater ratio of green surfaces
are more appealing to visitors (Nordh
et al., 2009; Nordh and Østby, 2013;
Peschardt et al., 2016)
functions? => The look of the area, scenery.
How natural do you think the area is? Percentage of vegetated surfaces
Are there enough green areas here?
How much does the greenness of the area
contribute to the quality of life?

How much does the area satisfy the below

4. The presence of water surfaces makes
the green environment more appealing
to visitors (Nordh et al., 2009; Regan and
Horn, 2005).
5. The visibility and look of the buildings
surrounding the urban parks contributes
to the human perception of the park
(Dee, 2001; Nordh and Østby, 2013).
functions? => The look of the area, scenery.
How much do you like the area? Percentage of water surfaces
How much does the area satisfy the below
functions? => The look of the area, scenery.
How much do you like the area? Number of building units in a 50 m buffer zone around the parks
How much does the area satisfy the below Percentage of built up area in a 50 m buffer zone around the parks
functions? => The look of the area, scenery.
Average building height in a 50 m buffer zone around the parks

Standard deviation of building heights in a 50 m buffer zone

6. The building density around urban parks
influences the human perception of the
parks (Bonaiuto et al., 2003; Dee, 2001).
7. The typical architecture of the
surrounding buildings significantly
contributes to the personal opinion of
green area visitors (Berney, 2006).
8. Disturbance indicators are expected to
be significantly related to lower
perceptions of urban parks (Nordh and
Østby, 2013).
How much do you like the area?
How much does the area satisfy the below
functions? => The look of the area, scenery.
How much do you like the area?
How much does the area satisfy the below
functions? => The look of the area, scenery.
What are the disturbance factors at the area?
around the parks
Number of building units in a 50 m buffer zone around the parks
Percentage of built up area in a 50 m buffer zone around the parks
Site inspection
Number of disturbance indicators (non-spatial indicator)
 G. Kothencz, T. Blaschke / Land Use Policy 64 (2017) 233–244
tory questions were redesigned according to the experiences and
recommendations obtained from the pre-test. The amended ver-
sion was reviewed by an expert focus group, comprising scientists
from environmental psychology, geography and geoinformatics.
Ultimately, the survey was again reworked, and finally approved
by the expert focus group members. The final survey consisted of
30 questions.
Two field campaigns using a semi-random sampling method
were carried out in April and May of 2014 in the five parks. The
whole survey was conducted by one person interviewing visitors
in one park and then moving on to the next park. The survey of a
single green space was conducted throughout three to four consec-
utive days at different times each day to reach a wide spectrum of
visitors.
The survey targeted a confidence interval (margin of error)
below ±10%, with 95% confidence level, which required a sample
size of at least 100 answered questionnaires (Niles, 2015). To make
the data robust, the targeted sample size was a minimum of 200
questionnaires.
After the April survey, it became apparent that reaching 200
answers was not possible, as only one surveyor was available for
the field work for four working days in April and four in May. There-
fore, to gain the desired minimum sample size, an internet survey
with the same questions was set up, aside from the in-situ data col-
lection, and operated from May to July 2014. The link to the survey
was sent out to eighteen local non-governmental organizations and
clubs (sports clubs, senior clubs, the local Lions Club, an organisa-
tion for socially deprived citizens, etc.), thus reaching an expansive
and diverse range of the city’s society. The administrators of these
organisations distributed the questionnaire to their members, and
to anyone else the organisations are connected to or work with. It
was crucial to guarantee that the individuals sampled through the
internet survey were equal to the field survey participants in terms
of familiarity with the respective study area. Hence, respondents’
knowledge of the parks was ensured in the following way. Respon-
dents’ neighbourhood was elicited in the survey. Answers were
only used from residents of Szeged and its close agglomeration. As
the study areas are well known both in Szeged and its catchment,
this practice ensured that participants knew the respective parks.
During the two field campaigns, 125 samples were collected
(including MA and BU: n = 159), with response rates of 92% in April
and 88% in May. The web survey was completed by 130 persons.
Altogether, 255 surveys were completed for the five areas (289
including BU and MA). After data verification, and a spatial overlay
of the centroids of participants’ reported neighbourhoods with the
Pléiades scene (see Section 2.4), 227 (261 including BU and MA)
records proved to be sufficient for further processing.
2.3.1. Number of disturbance indicators
Participants were asked to note disturbances in each park in
their own words (e.g. traffic noise, bad quality pavement). The
number of noted disturbance indicators (Table 2) was numerically
summed up for each park, and the sum was also used as data input
for the regression analysis and to test hypothesis 8.
2.4. Source of objective data
2.4.1. Satellite imagery
A 120 km2 Pléiades satellite imagery (Astrium et al., 2014) scene
from the 30th of August 2014 was used to derive data to measure
the objective greenness of Szeged. The acquired bundle product
comprised a 2 m ground resolution multispectral image, including
red, green, blue and near-infrared bands, and a 0.5 m resolution
panchromatic image. As retaining the original pixel values was
central to the measurement, and the 2 m resolution was sufficient
for the planned work, the multispectral product was used with-
237
out panchromatic sharpening. This image was used to calculate
the NDVI, a widely-used measure of biomass based on the spectral
reflectance difference between the red and near-infrared spectra. A
higher quantity of vegetation and higher biodiversity within urban
settings has a positive relation to the quality of urban life (Carrus
et al., 2015). Being a measure of biomass, the NDVI has been widely
applied as an indicator of quality of life (Jun, 2006). The NDVI has
been used by a number of scholars (Balseviciene et al., 2014; Pereira
et al., 2013) to estimate greenery when investigating socially sig-
nificant issues.
2.4.2. Very high resolution aerial image
A 10 cm ground resolution colour aerial image (DPGG, 2011) was
used to map parks, vegetated areas, and water surfaces within the
green spaces.
2.4.3. Auxiliary GIS data
A cadastral layer of Szeged was used to calculate the number and
area of building units around the parks. A Szeged building height
layer (Sümeghy et al., 2011) enabled a very accurate representation
of building heights and their influence on the public perception of
the park amenities.
2.4.4. Area weighted NDVI for participants’ neighbourhoods
To capture the perceived greenness of the parks in comparison
to the respondents’ residential areas, the reported greenness dif-
ference between parks and neighbourhoods served as subjective
data input. The analogous objective indicator was calculated for
polygons of 50 m buffer zones around the centroids of participants’
reported neighbourhoods. A buffer size of fifty meters was chosen
according to an earlier study, in which the same buffer size even-
tuated in convincing results to describe area of green surfaces in
urban neighbourhoods (Chen et al., 2014).
2.5. Generation of objective green space indicators
Objective spatial indicators representing the physical properties
of the park environments were selected based on experiences from
earlier studies and on recommendations of our expert focus group.
The indicators were generated using GIS (except indicator 8.). The
indicators were expected to correlate with participants’ subjective
responses, according to our hypotheses (see Section 2.2).
1. Area of the parks: Polygons of the studied parks were created
based on the aerial image. Then the area of each park was calculated
(Nordh et al., 2009; Peschardt et al., 2016); Hypothesis 1.
2. Area weighted NDVI for the parks: Area weighted NDVI was
calculated for the polygons of the parks; Hypothesis 2.
3. Area of vegetated surfaces: Polygons of the vegetated sur-
faces within each park were digitised from the aerial image, and
the indicator was calculated for each park (Peschardt et al., 2016);
Hypothesis 2.
4. Percentage of vegetated surfaces: The indicator was calcu-
lated for the parks (Nordh et al., 2009; Nordh and Østby, 2013;
Peschardt et al., 2016) (Fig. 3); Hypothesis 3.
5. Percentage of water surfaces: The restorative function of
water surfaces is well known (Regan and Horn, 2005). The water
surfaces of the parks were digitised from the aerial image, and the
indicator was calculated; Hypothesis 4.
The perception of green spaces highly depends on the look of the
surroundings outside the park (Nordh and Østby, 2013; Peschardt
et al., 2016). We aimed to investigate this relation with indicators
6–9.
6. Number of building units in a 50 m buffer zone around
the parks: Kyttä et al. (2013) found that the building density in
buffer zones around the study participants’ homes contributes to
perceived environmental quality. This approach was adapted in
238 G. Kothencz, T. Blaschke / Land Use Policy 64 (2017) 233–244
Fig. 3. Percentage of vegetated surfaces within the parks and percentage of built up area in a 50 m buffer zone around the parks.
indicator 6. Fifty meter buffer zones were generated around the
parks. Based on the Szeged cadastral layer, the number of building
units overlaid by the buffer zones was summed up for each park;
Hypotheses 5 and 6.
The size of the buffer zones aimed for a high explanatory
potential of the buffer areas, and to keep the incorporated con-
founding factors low. Experiments with different buffer radiuses
(50 m, 100 m and 200 m) revealed that buffer zones as small as
50 m already included all typical buildings and non-built up areas
of the park surroundings, while expanding the radius did not add
any valuable information to the buffers. This finding was confirmed
by Veldhuizen et al. (2013), who suggested that “the explanatory
power is strongest for small buffers, and it declines with increas-
ing buffer size”. Eventually a buffer size of 50 m was chosen and
applied.
7. The percentage of built up area in a 50 m buffer zone
around the parks: The indicator was calculated for each park, based
on the area of buildings in the buffer zones of the respective parks
and the area of the zones (Fig. 3); Hypotheses 5 and 6.
8. Surrounding architecture: The view outside the green spaces
is largely dominated by the surrounding architecture, which can
therefore contribute to the perceived aesthetical qualities of the
study areas (Dee, 2001). To investigate the hypothesised relation,
information regarding the typical architecture surrounding the
green spaces was collected by site inspections; Hypothesis 7.
9 and 10. Average and standard deviation of building heights
in a 50 m buffer zone around the parks: Building heights in
Szeged are distinct markers of architectural styles. The studied
green spaces with lower surrounding building heights and lower
variations in the standard deviation of building heights are situ-
ated in pleasant late 19th − early 20th century architectural settings
(DU, SZ and partly ER). Meanwhile, architecture with greater build-
ing heights or with greater variation in the standard deviation of
building heights are indicators of blocks of flats (VE and ZA) built in a
less attractive socialist realist style. The calculation of the two indi-
cators was as follows: Building units overlaying the buffer zones
of the parks were selected from the Szeged building height layer.
Then, the two indicators were calculated; Hypothesis 5.
Table 2 lists the respective values for the calculated objective
spatial indicators of the five parks.
2.6. Integrating subjective and objective data
The 261 questionnaire records (including BU and MA), together
with all calculated objective spatial indicators, were joined to the
respective records of the attribute table of the neighbourhood cen-
troid layer, respectively to the parks in which the participants were
surveyed. The following process was applied to ensure optimal
matching of the sub-samples to the objective spatial indicators. The
first question of the web based survey was to specify which one
 G. Kothencz, T. Blaschke / Land Use Policy 64 (2017) 233–244 239

Table 2
Objective indicators of the studied parks.
Parks 1. Area of the 2. Area weighted 3. Area of 4. Percentage of 5. Percentage of 6. Number of 7. Percentage of 8. Surrounding 9. Average 10. Standard Number of
parks (ha) NDVI for the vegetated vegetated water surfaces building units in built up area in a architecturea building height deviation of disturbance
parks surfaces (ha) surfaces (%) (%) a 50 m buffer 50 m buffer zone in a 50 m buffer building heights indicatorsb
zone around the around the zone around the in a 50 m buffer
parks parks (%) parks (m) zone around the
parks (m)
Erzsébet liget 19.57 0.337316 16.78 85.75 0 45 10.19 B 12.88 6.94 42
Dugonics tér 0.60 0.232273 0.31 52.05 2.24 24 30.92 A 14.49 6.70 21
Széchenyi tér 2.97 0.314700 1.65 55.50 0 29 34.76 A 17.99 5.68 24
Vér-tó 5.87 0.180273 3.70 63.06 36.94 16 8.03 C 30.11 4.94 50
Zápor-tó 3.28 0.314616 2.38 72.65 21.59 20 16.50 C 13.38 12.34 25
a
Surrounding architecture: A) Dominated by late 19th early 20th century architecture (neo-baroque, eclectic, etc.). B) Mixed architecture: speciously placed (20th century,
early 21st century) detached and semi-detached houses (1–4 floors) with gardens; 5–10 floor apartment blocks; 2–10 floor educational, academic and public buildings. C)
Dominated by 5–10 floor apartment blocks (2nd half of the 20th century).
b
Reported by survey participants, therefore it is subjective.

of the five parks the participant wanted to characterise, by using a park. The described process enabled matching the sub-sample from
radio button right after the introductory text. The introductory text the web based survey to the objective spatial indicators by using
clearly informed respondents that their answers were to be made Table Join function in GIS software. Similarly, sub-samples from the
with reference only to the chosen park and none of the other areas. field survey were matched to the objective spatial indicators of the
This practice ensured that both the respondents and the authors parks.
knew for which of the five parks the answers were given. By using
the radio button, the name of the park was automatically writ-
2.7. Analytical process
ten in the respective record of the database of the online survey.
The name of the park in each record therefore uniquely identified
The collected subjective data and the calculated objective spa-
the study area for each respondent. Another unique identification
tial indicators were studied in comparison to each other during
code was generated for each answered questionnaire form. Thus
an extensive analytical process, to identify and quantify relation-
each questionnaire record included the unique identifier of the
ships between the perceived properties of the studied parks and
park and the unique identifier of the form. Records of the objec-
their environmental indicators. The overall workflow is depicted
tive spatial indicators were labelled with the name of the respective
in Fig. 4.

Fig. 4. The workflow of the applied methodology.

240 G. Kothencz, T. Blaschke / Land Use Policy 64 (2017) 233–244

Table 3
Factor loadings (F1–F3), communalities (h2 ), item means (M), item standard deviations (SD) and part-whole corrected selectivity of the 3-factor solution (rit ): F1 Impression
of green, healthy and recreational environment, F2 Impression of the state of the park, F3 Assessment of accessibility of the park by private or public transport.

Item F1a F2a F3a h2 M SD rit

How much does the area satisfy the below functions? 0.746 −0.133 0.198 0.613 3.765 1.072 0.675
Reduction of air pollution.

How natural do you think the area is? 0.745 0.006 0.077 0.562 3.765 0.992 0.646

How much does the area satisfy the below functions? A variety of plant and 0.716 −0.150 0.247 0.596 3.490 1.097 0.632
animal life.

Are there enough green areas here? 0.693 0.075 0.093 0.495 3.945 1.128 0.569

How much does the area satisfy the below functions? Cooler summer 0.693 −0.117 0.155 0.519 3.761 0.993 0.612
temperatures by the green area.

How much does the greenness of the area contribute to this quality of life? 0.597 0.195 −0.093 0.403 4.522 0.752 0.452

How quiet is the area in terms of traffic noise? 0.553 0.007 −0.286 0.387 3.216 1.029 0.389

How much does the area satisfy the below functions? Recreation. 0.523 0.103 0.327 0.391 4.224 0.799 0.457

How much do you like the area? 0.501 0.171 0.180 0.312 4.498 0.608 0.418

How green do you feel this area is compared to the area where you live? 0.479 −0.050 −0.032 0.233 3.231 1.266 0.365

Is the area well kept? −0.048 0.752 −0.087 0.575 3.935 0.974 0.480

Are there enough benches? −0.061 0.747 −0.027 0.562 3.273 1.165 0.519

Is the quality of paving of the sealed surfaces good enough? −0.020 0.681 0.192 0.500 3.665 1.139 0.490

How much does the area satisfy the below functions? The look of the area, 0.186 0.674 0.214 0.534 4.362 0.740 0.475
scenery.

How do you consider the accessibility of this area? By car. 0.139 −0.019 0.755 0.590 4.122 1.006 0.474

How do you consider the accessibility of this area? By public transport. −0.022 0.185 0.749 0.596 4.682 0.606 0.465

How do you consider the accessibility of this area? By bike. 0.158 0.055 0.659 0.462 4.659 0.613 0.426
a
F1: Impression of green, healthy and recreational environment; F2: Impression of the state of the park; F3: Assessment of accessibility of the park by private or public
transport. Values of factor loadings included into each factor are highlighted in bold.

The final output table from the GIS analysis served as the
input data for the statistical analysis. The initial step was to con-
vert nominal data to dichotomous type dummy variables. Then,
an explorative data analysis was conducted to check the data
set for errors and outliers. In the next step, principal component
analysis (PCA) was applied to reduce the number of variables
used in the questionnaire. In other words, based on the par-
ticipants’ answers, the questions addressing similar topics were
collected to principal common themes (conceptual data dimen-
sions, subjective assessment dimensions, or factor-item-pools).
The PCA identified a three-factor-solution to the inter-correlation
matrix of the questionnaire items (see Section 2.8 for details).
Then, a reliability analysis was conducted on the identified con-
ceptual data dimensions using the Cronbach’s Alpha method, a
coefficient that measures the internal consistency of the data
dimensions (Cronbach, 1951). A Pearson Chi-Square-Test (Pearson,
1900), a measure of data distribution, showed significant differ-
ences in the gender-distributions between visitors of the various
parks (2 (6,N = 261) = 13.54, p < 0.05). To eliminate the possibility of
a “gender-bias”, the parks BU and MA were excluded from the sub-
sequent multiple linear regression analysis, due to the prevalence
of female study participants in these two parks. After removing the
data of these parks from the study-sample, a further Pearson-Chi-
Square-Test no longer showed any significant gender-differences
between the parks (2 (4,N = 227) = 3.91, p = n.s.). The final analysis
was a multiple linear regression using the subjective assessment
dimensions as the dependent (“criterion”) variables (DV). This step,
together with the results, is described in Section 3.
2.8. Factor structure and reliability analysis
The KMO and Bartlett’s test of sphericity (2 (136,N = 261) = 1170.03,
p < 0.01), measuring the homogeneity of data variance (Snedecor
and Cochran, 1989), indicated that the data were suitable for per-
forming a factor analysis. An interpretable three-factor solution
was determined, with a total explained variance of 49.01%. Item for-
mulations, factor loadings, and communalities of the three-factor
solution (in descending order by factor loadings) are depicted in
Table 3.
Factor 1 was termed “Impression of green, healthy and recre-
ational environment” (explained variance: 24.05%), factor 2 as
“Impression of the state of the park” (explained variance: 13.04%),
and factor 3 as “Assessment of accessibility of the park by private
or public transport” (explained variance: 11.92%).
The factor dimensions were checked by item analysis (reliabil-
ity analysis) in regard to their measuring precision (reliability),
whereby all three-factor dimensions showed suitable reliability
(F1: Cronbach’s ␣ = 0.83, F2: Cronbach’s ␣ = 0.71, F3: Cronbach’s
␣ = 0.65). The suitability of the three-factor dimensions is supported
by earlier literature. Cronbach’s Alpha should be >0.80 when con-
structing psychological “tests”; however, in terms of questionnaire
data, Cronbach’s ␣ > 0.60 is acceptable (Bühner, 2004).
Scale type (interval) values for “Impression of green, healthy and
recreational environment”, “Impression of the state of the park”
and “Assessment of accessibility of the park by private or public
transport” were formed by calculating the mean, whereby higher
values implied a higher expression of the dimension. Table 4 depicts
the inter-correlation matrix of the calculated scale values.
Table 4 indicates that there is no correlation between the fac-
tors “Impression of the state of the park” and “Impression of green,
healthy and recreational environment”. There is only a small corre-
lation between the factors “Assessment of accessibility of the park
by private or public transport” and “Impression of green, healthy
and recreational environment”.
Finally, the Pearson’s r indicated that there is no significant corre-
lation between the factors “Assessment of accessibility of the park
by private or public transport” and “Impression of the state of the
park”.
 G. Kothencz, T. Blaschke / Land Use Policy 64 (2017) 233–244 241

Table 4
Scale inter-correlation matrix (Pearson’s r) of the three factors extracted (N = 261).
Impression of Impression of Assessment of
green, healthy and the state of the accessibility of the
recreational park park by private or
environment public transport
Impression of green, – 0.008 0.263*
healthy and
recreational
environment.
Impression of the state 0.008 – 0.118
of the park
Assessment of 0.263* 0.118 –
accessibility of the
park by private or
public transport
Note. * p < 0.01.
According to the reliability analysis described in this section, the
data proved to be solid and reliable enough to further analyse using
multiple regression analysis.
3. Results
3.1. Regression analysis
The three subjective assessment dimensions, excluding the
records of BU and MA, served as the dependent variables of
a multiple regression analysis. The objective spatial indicators
and the architecture surrounding the parks were used as inde-
pendent variables (predictors) in each regression model. First,
a stepwise regression model was used to isolate the rele-
vant predictors. Then the relevant predictors were concurrently
inputted to a second analysis, combined with bootstrapping, to
deliver robust regression estimates. Irrelevant independent vari-
ables above a threshold of p > 0.10 were excluded from the final
regression, where p < 0.05 was set as the statistically significant
probability for the entire model. Variance Inflation Factors (VIF)
measuring the multicollinearity effect (possible strong correlations
between independent variables) were calculated to identify possi-
ble interfering predictor-inter-correlations (which would lead to
effect-overestimations).
3.2. Computation and results of correlations for the “Impression
of green, healthy and recreational environment”
The final regression model for the variable “Impression of green,
healthy and recreational environment” was calculated. The crite-
rion variable, with a corrected R2 = .24, an explained variance of
24.9%, and F(2,213) = 35.39, p < 0.01, proved to be statistically signifi-
cant. The identified correlation between the criterion variable and
the correlating predictors is depicted in Table 5.
According to the results in Table 5, there is a moderate corre-
lation between the “Impression of green, healthy and recreational
environment” and the “Percentage of vegetated surfaces”.
A higher expression of the variable “Impression of green, healthy
and recreational environment” is associated with a higher propor-
tion of the objective indicator “Percentage of vegetated surfaces”.
Table 6
Beta values (Pearson’s r), t-values and VIF (collinearity statistics) for individual pre-
dictors for the dependent variable “Impression of the state of the park”.
Predictors ˇ∗∗ t p VIF
Percentage of built up 0.43 7.33 0.00 1.01
area in a 50 m buffer
zone around the
parks
Number of disturbance −0.25 −4.32 0.00 1.01
indicators
Note. ** p < 0.01.
Based on the result, our 3rd hypothesis (“Parks with a greater ratio
of green surfaces are more appealing to visitors”) is accepted.
3.3. Computation and results of correlations for the “Impression
of the state of the park”
The final regression model for the criterion variable “Impression
of the state of the park” was calculated. The p value proved to be
highly significant, with a corrected R2 = .26, an explained variance of
26%, and F(2,219) = 40.56, p < 0.01. The statistical characteristics and
correlations of the individual predictors to the dependent variable
identified by the final regression model are provided in Table 6.
According to the results in Table 6, there is a moderate cor-
relation between the “Impression of the state of the park” and
the “Percentage of built up area in a 50 m buffer zone around the
parks”. There is a low negative correlation between the indicators
“Impression of the state of the park” and “Number of disturbance
indicators”.
A higher expression of the dependent variable “Impression of
the state of the park” is correlated with a greater proportion of
the building density objective indicator (“Percentage of built up
area in a 50 m buffer zone around the parks”) and a lower number
of reported disturbance indicators (“Number of disturbance indi-
cators”). The moderate correlation with the indicator “Percentage
of built up area in a 50 m buffer zone around the parks” supports
hypothesis 6 (“The building density around urban parks influences
the human perception of the parks”); therefore hypothesis 6 is
accepted. The low negative correlation with the “Number of dis-
turbance indicators” supports the acceptance of hypothesis 8 (“The
disturbance indicators are expected to be significantly related to
lower perceptions of urban parks”).
3.4. Computation and results of correlations for the “Assessment
of accessibility of the park by private or public transport”
The final regression model was calculated for the dependent
variable “Assessment of accessibility of the park by private or pub-
lic transport”. The p value proved to be highly significant, with a
corrected R2 = .06, an explained variance of 7.6%, and F(3,223) = 6.14,
p< 0.01. The correlations for this criterion variable, identified by the
final regression model, are depicted in Table 7.
Table 7
Beta values (Pearson’s r), t-values and VIF (collinearity statistics) for individual pre-
dictors for the dependent variable “Assessment of accessibility of the park by private
or public transport”.

Table 5 Predictors ˇ* t p VIF

Beta values (Pearson’s r), t-values and VIF (collinearity statistics) for individual pre- Standard deviation of building −0.15 −2.29 0.02 1.01
dictors for the dependent variable “Impression of green, healthy and recreational heights in a 50 m buffer zone
environment”. around the parks
Number of building units in a 0.13 1.97 0.05 1.01
Predictors ˇ** t p VIF
50 m buffer zone around the
Percentage of vegetated surfaces 0.48 8.00 0.00 1.01 parks

Note. ** p < 0.01. Note. * p < 0.05.

242 G. Kothencz, T. Blaschke / Land Use Policy 64 (2017) 233–244
According to the results in Table 7, there is a low negative cor-
relation between the indicators “Assessment of accessibility of the
park by private or public transport” and the “Standard deviation of
building heights in a 50 m buffer zone around the parks”. There is a
low correlation between the indicators “Assessment of accessibil-
ity of the park by private or public transport” and the “Number of
building units in a 50 m buffer zone around the parks”.
A higher expression of the variable “Assessment of accessibility
of the park by private or public transport” is associated with the
objective indicator “Number of building units in a 50 m buffer zone
around the parks”, as well as with the indicator “Standard deviation
of building heights in a 50 m buffer zone around the parks”, to a
lesser extent.
3.5. Summary of identified correlations and accepted hypotheses
The statistical analysis revealed two moderate and two minor
correlations between the perceived park environments and the
nine objective indicators of the studied parks (Table 8). A low
correlation with the “Number of disturbance indicators” was also
identified. Beyond reflecting the identified correlations, Table 8
shows that three out of the eight hypotheses (detailed in Table 1)
were accepted through moderate or low correlations.
4. Discussion
City parks are essential contributors to the quality of urban life.
The importance of the societal and ecosystem benefits they provide
will proliferate in the future, and help to assist cities in tackling
the ever growing needs of urban society and the increasing envi-
ronmental challenges (e.g. extreme weather events). Methods for
analysing well-measurable objective features of the urban green
environment are vital for successfully coping with these challenges.
Beyond this factual information, the perceived urban green space
and the citizens’ satisfaction with and demand for services sup-
plied by urban parks are equally important. The inherently different
objective and subjective information reflect different aspects of
green spaces. Therefore, studies relying on either purely objective
or on purely subjective methods can be limited in their view. On the
contrary, combined analyses may increase the wealth of informa-
Table 8
Identified correlations and accepted hypotheses.
Objective indicators ␤-Beta values
(Pearson’s r)
Area of the parks (ha) – –
Area of vegetated surfaces (ha) – –
Percentage of vegetated surfaces (%) 0.48**
Area weighted NDVI for the parks – –
Percentage of water surfaces (%) – –
Number of building units in a 50 m buffer zone 0.13*
around the parks
Percentage of built up area in a 50 m buffer 0.43**
zone around the parks (%)
Average building height in a 50 m buffer zone – –
around the parks
Standard deviation of building heights in a −0.15*
50 m buffer zone around the parks
Number of disturbance indicators −0.25**
Note. * p < 0.05,** p < 0.01.
tion that can be incorporated in decision-making processes. This is
precisely the area that this study aimed to support by providing an
insight into different aspects of the same reality, revealed by a col-
lective assessment of subjective information, visitors’ perceptions
of urban parks, and spatially explicit, objective data on the same
parks. The results of the work are discussed here.
Results for “Impression of green, healthy and recreational envi-
ronment” are in agreement with previous scientific evidence.
Studies found the proportion of vegetated surfaces of urban parks
as being a key factor in visitors’ perceptions, and it has been proven
to promote the recreational use of green spaces (Nordh et al.,
2011; Nordh et al., 2009; Peschardt et al., 2016). The previous
evidence affirms the identified moderate correlation between the
“Impression of green, healthy and recreational environment” and
the “Percentage of vegetated surfaces”. Furthermore, it supports
the acceptance of our 3rd hypothesis (“Parks with a greater ratio of
green surfaces are more appealing to visitors”).
Results for the “Impression of the state of the park” suggest that
visitors’ impressions of the parks are influenced by the scene sur-
rounding the park and by the perceived disturbances. Concerning
the grounds and the environments of the five study areas, these
findings are logical. Parks DU and SZ are surrounded by aesthetically
appealing and densely built late 19th and early 20th century archi-
tecture. At the same time their visitors reported the lowest number
of disturbance indicators. The relatively sparsely built architecture
around ER, VE and ZA exhibits considerably lower aesthetic quali-
ties than that surrounding DU and SZ, and the former parks have a
high number of disturbance indicators. Our results are confirmed
by earlier studies, which found scenery – including building density
and building aesthetics around urban parks – and reported dis-
turbances as crucial determinants in human perception of urban
green spaces (Grahn and Stigsdotter, 2010; Mansor et al., 2012;
Nordh and Østby, 2013; Peschardt et al., 2016). Results of the afore-
mentioned studies also advocate for the acceptance of hypothesis
6 (“The building density around urban parks influences the human
perception of the parks”) and hypothesis 8 (“The disturbance indi-
cators are expected to be significantly related to lower perceptions
of urban parks”).
The possible argument for the results associated with “Assess-
ment of accessibility of the park by private or public transport”
Correlating subjective assessment Accepted hypotheses (from Table 1)
dimensions
–
–
Impression of green, healthy and 3. Parks with greater ratio of green
recreational environment surfaces are more appealing to visitors.
–
–
Assessment of accessibility of the –
park by private or public transport
Impression of the state of the park 6. The building density around urban
parks influences the human perception
of the parks.
–
Assessment of accessibility of the –
park by private or public transport
Impression of the state of the park 8. Disturbance indicators are expected
to be significantly related to lower
perceptions of urban parks.
 G. Kothencz, T. Blaschke / Land Use Policy 64 (2017) 233–244
stems from the urban peculiarities of Szeged. Study areas sur-
rounded by a greater number of buildings with a low variation in
height are likely to be more easily accessible. In regard to the pre-
dictor “Standard deviation of building heights in a 50 m buffer zone
around the parks”, there is little height deviation of buildings sur-
rounding ER, SZ and VE. In addition, the centre of Szeged, where DU
and SZ are situated, is densely built and boasts a high number of
buildings. Due to the size of ER, it is also surrounded by a high num-
ber of buildings. The accessibility of these parks is generally good
with all means of transport. Therefore, the aforementioned char-
acteristics influenced the surveyed population enough to have an
observable effect on the results. These outcomes are in line with
those from previous studies, which revealed the importance of
the built environment surrounding urban parks (Nordh and Østby,
2013) and its effect on human perceptions in general (Dee, 2001).
The importance of building density was a strong factor in the per-
ceived environmental quality of urban neighbourhoods (Bonaiuto
et al., 2003; Kyttä et al., 2013).
In the present study, nine spatially explicit objective indicators
and the disturbance indicator were examined and compared to the
three subjective assessment dimensions. Concerning the objective
indicators, two moderate and two minor correlations were found.
Another minor correlation was identified regarding the number
of disturbance indicators. The lack of strong correlations in this
research can be attributed to the different nature of the qualitative
and quantitative data genres, as they reflect various aspects of the
analysed urban parks. The results produced by our work are in line
with the outcomes of other studies that identified a lack of agree-
ment between subjective evaluations of the urban environment
and objective environmental indicators (Leslie et al., 2010; McCrea
et al., 2006; Van Herzele and Wiedemann, 2003). Our study thereby
suggests that decisions made based purely on objective data input
or solely on subjective information cannot reveal features of green
spaces as comprehensively as a mixed-methods approach. Conse-
quently, the authors emphasise the importance of the collective
application of qualitative and quantitative research instruments.
Such an approach will ensure the consideration of a wider spec-
trum of the physical environment, which would not be possible
using solely objective data or purely subjective information.
We identified two limitations of the study. First, there is a
potential for measurement bias in this study when integrating
sub-samples from personal interviews within the parks and from
the web survey. Online respondents’ knowledge of the parks was
ensured (see Section 2.3), and their answers were matched to
specific parks (see Section 2.6). Still, one needs to be aware of poten-
tial measurement bias. For example, participants of web surveys
respond to questions on the same park from different perspectives
than the participants who were physically interviewed within the
park during field surveys. Another aspect which potentially can
lead to measurement bias in mixed mode surveys is the absence of
the interviewer when respondents answer online surveys and the
presence of the interviewer when field survey participants respond
in person (Szolnoki and Hoffmann, 2013). These are typical issues
which cannot be fully eliminated from mixed mode data collec-
tion. Therefore, they can also contribute to measurement bias in
the current study. Concerning the second limitation of the work,
every city has its own character, which can affect the study results.
In the case of Szeged, confounding factors might be attributable to
the socialist type urban development (e.g. less attractive buildings
in the outer districts of the city, with an often very pleasant green
environment). Therefore, mitigating settlement-specific outcomes
would be central to future work. Since our methodology is adapt-
able to other urban areas, it would be advantageous to conduct the
same study in further cities. This would allow the authors to draw
a more comprehensive conclusion. In addition, further input from
the fields of geographic information science, sociology and environ-
243
mental psychology would be crucial to improve the understanding
of the fuzzy relation of the perceived environment and objective
environmental indicators.
5. Conclusion
This study provided an exploratory methodology to investigate
the link between measured spatially explicit indicators of urban
green spaces and their contributions to subjective quality of life,
as perceived and reported by their visitors. The statistical analy-
ses partly resulted in minor and moderate correlations, while the
majority of the input data yielded no agreement between the sub-
jectively observed and objectively measured characteristics of the
investigated areas. The links between objective measures of the
urban parks and their surroundings and related subjective evalua-
tions of the parks were not as strong as hypothesised. Nevertheless,
based on a rigid statistical analysis, the results still support the
authors’ opinion that it is important to improve the urban qual-
ity of life by taking into consideration both objective measures and
residents’ perceptions. The authors received very positive feedback
from many citizens in Szeged and from several institutions, and
believe that the simultaneous application of subjective and objec-
tive indicators, together with additional input from subsequent
studies to assess the citizens’ quality of life demands, can promote
informed decisions in urban planning and management to make
cities more liveable.
Acknowledgements
The authors thank Bernhard Brunner for his support in the sta-
tistical analysis. Thanks also go to our colleagues at the Department
of Climatology and Landscape Ecology and the Department of Phys-
ical Geography and Geoinformatics of the University of Szeged
for the data support and their help in the questionnaire survey.
We acknowledge Peter Ranacher and Michael Hodgson for their
valuable comments and suggestions. We thank the reviewer for
taking the time to read the manuscript and provide constructive
suggestions for improvements. We also wish to thank Isabella Mer-
schdorf and Helena Merschdorf for proofreading the article. This
research was funded by the Austrian Science Fund (FWF) through
the Doctoral College GIScience (DK W 1237-N23) at the University
of Salzburg.
References
Astrium, CNES, Airbus Defence and Space, 2014. Pléiades satallite imagery.
Astrium, and Airbus Defence and Space, Toulouse, France.
Bühner, M., 2004. Einführung in die Test- und Fragebogenkonstruktion. Pearson
Studium, München [etc.].
Balseviciene, B., Sinkariova, L., Grazuleviciene, R., Andrusaityte, S., Uzdanaviciute,
I., Dedele, A., Nieuwenhuijsen, M., 2014. Impact of residential greenness on
preschool children’s emotional and behavioral problems. Int. J. Environ. Res.
Public Health 11, 6757–6770.
Baur, J.W.R., Tynon, J.F., Gómez, E., 2013. Attitudes about urban nature parks: a
case study of users and nonusers in Portland, Oregon. Landscape Urban Plann.
117, 100–111.
Berney, R., 2006. Constructing the sustainable city: emerging public open spaces in
bogotá, Colombia. In: Mooney, P.F. (Ed.), CSLA/CELA 2006 Conference.
University of British Columbia, Vancouver, Canada, pp. 18–25.
Bonaiuto, M., Fornara, F., Bonnes, M., 2003. Indexes of perceived residential
environment quality and neighbourhood attachment in urban environments: a
confirmation study on the city of Rome. Landscape Urban Plann. 65, 41–52.
Bowler, D.E., Buyung-Ali, L., Knight, T.M., Pullin, A.S., 2010. Urban greening to cool
towns and cities: a systematic review of the empirical evidence. Landscape
Urban Plann. 97, 147–155.
Carrus, G., Scopelliti, M., Lafortezza, R., Colangelo, G., Ferrini, F., Salbitano, F.,
Agrimi, M., Portoghesi, L., Semenzato, P., Sanesi, G., 2015. Go greener, feel
better? The positive effects of biodiversity on the well-being of individuals
visiting urban and peri-urban green areas. Landscape Urban Plann. 134,
221–228.
244 G. Kothencz, T. Blaschke / Land Use Policy 64 (2017) 233–244
Chen, B., Adimo, O.A., Bao, Z., 2009. Assessment of aesthetic quality and multiple
functions of urban green space from the users’ perspective: the case of
Hangzhou Flower Garden, China. Landscape Urban Plann. 93, 76–82.
Chen, A., Yao, X.A., Sun, R., Chen, L., 2014. Effect of urban green patterns on surface
urban cool islands and its seasonal variations. Urban For. Urban Green. 13,
646–654.
Cronbach, L., 1951. Coefficient alpha and the internal structure of tests.
Psychometrika 16, 297–334.
DPGG, 2011. Szeged ortofotó 2011. University of Szeged, Department of Physical
Geography and Geoinformatics (DPGG), Szeged, Hungary.
Dee, C., 2001. Form and Fabric in Landscape Architecture: A Visual Introduction.
Spon Press London, New York.
Escobedo, F.J., Nowak, D.J., 2009. Spatial heterogeneity and air pollution removal
by an urban forest. Landscape Urban Plann. 90, 102–110.
Fontana, S., Sattler, T., Bontadina, F., Moretti, M., 2011. How to manage the urban
green to improve bird diversity and community structure. Landscape Urban
Plann. 101, 278–285.
Germann-Chiari, C., Seeland, K., 2004. Are urban green spaces optimally distributed
to act as places for social integration? Results of a geographical information
system (GIS) approach for urban forestry research. For. Policy Econ. 6, 3–13.
Grahn, P., Stigsdotter, U.K., 2010. The relation between perceived sensory
dimensions of urban green space and stress restoration. Landscape Urban
Plann. 94, 264–275.
Hernández-Morcillo, M., Plieninger, T., Bieling, C., 2013. An empirical review of
cultural ecosystem service indicators. Ecol. Indic. 29, 434–444.
Irvine, K.N., Devine-Wright, P., Payne, S.R., Fuller, R.A., Painter, B., Gaston, K.J., 2009.
Green space, soundscape and urban sustainability: an interdisciplinary,
empirical study. Local Environ. 14, 155–172.
James, P., Tzoulas, K., Adams, M.D., Barber, A., Box, J., Breuste, J., Elmqvist, T., Frith,
M., Gordon, C., Greening, K.L., Handley, J., Haworth, S., Kazmierczak, A.E.,
Johnston, M., Korpela, K., Moretti, M., Niemelä, J., Pauleit, S., Roe, M.H., Sadler,
J.P., Ward Thompson, C., 2009. Towards an integrated understanding of green
space in the European built environment. Urban For. Urban Green. 8, 65–75.
Jim, C.Y., Chen, W., 2006. Perception and attitude of residents toward urban green
spaces in Guangzhou (China). Environ. Manage. 38, 338–349.
Jun, B.-W., 2006. Urban quality of life assessment using satellite image and
socioeconomic data in GIS. Korean J. Remote Sens. 22, 325–335.
Kaźmierczak, A., 2013. The contribution of local parks to neighbourhood social ties.
Landscape Urban Plann. 109, 31–44.
Kyttä, M., Broberg, A., Tzoulas, T., Snabb, K., 2013. Towards contextually sensitive
urban densification: location-based softGIS knowledge revealing perceived
residential environmental quality. Landscape Urban Plann. 113, 30–46.
Lakes, T., Kim, H.-O., 2012. The urban environmental indicator Biotope Area
Ratio—an enhanced approach to assess and manage the urban ecosystem
services using high resolution remote-sensing. Ecol. Indic. 13, 93–103.
Langemeyer, J., Baró, F., Roebeling, P., Gómez-Baggethun, E., 2015. Contrasting
values of cultural ecosystem services in urban areas: the case of park Montjuïc
in Barcelona. Ecosyst. Serv. 12, 178–186.
Lee, A.C.K., Maheswaran, R., 2011. The health benefits of urban green spaces: a
review of the evidence. J. Public Health 33, 212–222.
Leslie, E., Sugiyama, T., Ierodiaconou, D., Kremer, P., 2010. Perceived and
objectively measured greenness of neighbourhoods: are they measuring the
same thing? Landscape Urban Plann. 95, 28–33.
Lin, W., Yu, T., Chang, X., Wu, W., Zhang, Y., 2015. Calculating cooling extents of
green parks using remote sensing: method and test. Landscape Urban Plann.
134, 66–75.
Maas, J., Verheij, R.A., Groenewegen, P.P., de Vries, S., Spreeuwenberg, P., 2006.
Green space, urbanity, and health: how strong is the relation? J. Epidemiol.
Community Health 60, 587–592.
Mansor, M., Said, I., Mohamad, I., 2012. Experiential contacts with green
infrastructure’s diversity and well-being of urban community. Procedia – Soc.
Behav. Sci. 49, 257–267.
McCrea, R., Shyy, T.-K., Stimson, R., 2006. What is the strength of the link between
objective and subjective indicators of urban quality of life? Appl. Res. Qual. Life
1, 79–96.
Niles, R., 2015. Statistics Every Writer Should Know. Robert Niles, Phoenix, USA.
http://www.robertniles.com/stats/. (Accessed: 6 January 2017).
Nordh, H., Østby, K., 2013. Pocket parks for people −A study of park design and use.
Urban For. Urban Green. 12, 12–17.
Nordh, H., Hartig, T., Hagerhall, C.M., Fry, G., 2009. Components of small urban
parks that predict the possibility for restoration. Urban For. Urban Green. 8,
225–235.
Nordh, H., Alalouch, C., Hartig, T., 2011. Assessing restorative components of small
urban parks using conjoint methodology. Urban For. Urban Green. 10, 95–103.
Oliveira, S., Andrade, H., Vaz, T., 2011. The cooling effect of green spaces as a
contribution to the mitigation of urban heat: a case study in Lisbon. Build.
Environ. 46, 2186–2194.
Pearson, K., 1900. X. On the criterion that a given system of deviations from the
probable in the case of a correlated system of variables is such that it can be
reasonably supposed to have arisen from random sampling. Philos. Mag. Ser. 5
(50), 157–175.
Pereira, G., Christian, H., Foster, S., Boruff, B., Bull, F., Knuiman, M., Giles-Corti, B.,
2013. The association between neighborhood greenness and weight status: an
observational study in Perth Western Australia. Environ. Health 12, 49.
Peschardt, K.K., Stigsdotter, U.K., Schipperrijn, J., 2016. Identifying features of
pocket parks that may be related to health promoting use. Landscape Res. 41,
79–94.
Regan, C.L., Horn, S.A., 2005. To nature or not to nature: associations between
environmental preferences, mood states and demographic factors. J. Environ.
Psychol. 25, 57–66.
Rhew, I.C., Vander Stoep, A., Kearney, A., Smith, N.L., Dunbar, M.D., 2011. Validation
of the normalized difference vegetation index as a measure of neighborhood
greenness. Ann. Epidemiol. 21, 946–952.
Sümeghy, Z., Gál, T., Unger, J., 2011. Szeged 3D városmodellje. Geodézia és
Kartográfia 63 (1), 13–16.
Schipperijn, J., Stigsdotter, U.K., Randrup, T.B., Troelsen, J., 2010. Influences on the
use of urban green space −A case study in Odense, Denmark. Urban For. Urban
Green. 9, 25–32.
Snedecor, G.W., Cochran, W.G., 1989. Statistical Methods, 8th ed. Iowa State
University Press, Ames.
Szolnoki, G., Hoffmann, D., 2013. Online, face-to-face and telephone
surveys—comparing different sampling methods in wine consumer research.
Wine Econ. Policy 2, 57–66.
Takács, Á., Kiss, M., Gulyás, Á., 2014. Some aspects of indicator development for
mapping microclimate regulation ecosystem service of urban tree stands. Acta
Climatol. Chorol. 47 (-48), 99–108.
Tsurumi, T., Managi, S., 2015. Environmental value of green spaces in Japan: an
application of the life satisfaction approach. Ecol. Econ. 120, 1–12.
Tzoulas, K., Korpela, K., Venn, S., Yli-Pelkonen, V., Kaźmierczak, A., Niemela, J.,
James, P., 2007. Promoting ecosystem and human health in urban areas using
Green Infrastructure: a literature review. Landscape Urban Plann. 81, 167–178.
Van Herzele, A., Wiedemann, T., 2003. A monitoring tool for the provision of
accessible and attractive urban green spaces. Landscape Urban Plann. 63,
109–126.
Veldhuizen, E.M., Stronks, K., Kunst, A.E., 2013. Assessing associations between
socio-economic environment and self-reported health in Amsterdam using
Bespoke Environments. PLoS One 8, e68790.
von Wirth, T., Grêt-Regamey, A., Stauffacher, M., 2015. Mediating effects between
objective and subjective indicators of urban quality of life: testing specific
models for safety and access. Soc. Indic. Res. 122, 189–210.
Wolch, J.R., Byrne, J., Newell, J.P., 2014. Urban green space, public health, and
environmental justice: the challenge of making cities ‘just green enough’.
Landscape Urban Plann. 125, 234–244.