Attenuating subjective crowding through beauty: an online study on the interaction between environment aesthetics, typology and crowdedness

Abstract

This paper reports on an online study to investigate the interaction between the environment typology, and crowdedness with aesthetic perceptions of the environment.

Full text

Graphical Abstract
Attenuating subjective crowding through beauty: an online study on the interaction
between environment aesthetics, typology and crowdedness
Panagiotis Mavros 1, Zi Liang Ngoi, Stephanie Kirk, An Shu Te, Jascha Grübel, Leonel Aguilar,
Agnieszka Olszewska-Guizzo, Dominique Makowski1
1Corresponding authors. PM: panos.mavros@telecom-paris.fr; present address: i3, CNRS, Télécom Paris, Institut
Polytechnique de Paris, France. DM: d.makowski@sussex.ac.uk; School of Psychology, University of Sussex, UK

Highlights
Attenuating subjective crowding through beauty: an online study on the interaction
between environment aesthetics, typology and crowdedness
Panagiotis Mavros 2, Zi Liang Ngoi, Stephanie Kirk, An Shu Te, Jascha Grübel, Leonel Aguilar,
Agnieszka Olszewska-Guizzo, Dominique Makowski1
•Study on joint influence of environment typology (urban/green), beauty, and crowding
•Created large (480) international database of dynamic stimuli of video walk-throughs
•Online study of stimuli with emotion ratings, motivation, environmental appraisals
•Found positive effect of environmental beauty, order both for urban and green space
•Moderate pedestrian crowding increased the desirability of both urban and green space
2Corresponding authors. PM: panos.mavros@telecom-paris.fr; present address: i3, CNRS, Télécom Paris, Institut
Polytechnique de Paris, France. DM: d.makowski@sussex.ac.uk; School of Psychology, University of Sussex, UK

Attenuating subjective crowding through beauty: an online study on the
interaction between environment aesthetics, typology and crowdedness
Panagiotis Mavros 1a,c, Zi Liang Ngoib, Stephanie Kirkb, An Shu Teb, Jascha Grübele, Leonel
Aguilarf, Agnieszka Olszewska-Guizzog, Dominique Makowski1b,d
aFuture Cities Laboratory, Singapore-ETH Centre, CREATE campus, 1 CREATE Way, #06-01 CREATE
Tower, 138602, Singapore
bNanyang Technological University, , Singapore
ci3, CNRS, Télecom Paris, Institut Polytechnique de Paris 19 Pl. Marguerite Perey, 91120, France
dSchool of Psychology University of Sussex UK
eCenter for Sustainable Future Mobility Chair of Geoinformation Engineering Chair of Cognitive Science \& Game
Technology Center ETH Zurich Switzerland
fData Science Service and Systems Group \& Chair of Cognitive Science ETH Zurich Switzerland
gNeuroLandscape Poland
Abstract
Past research suggests a positive impact of green (compared to built) spaces on multiple facets of
cognition and emotion. However, how crowding (high density of other people) affects this relationship,
and whether it is modulated by the aesthetic appeal of the environment, remains unclear. We compiled
a library of 480 video-clips of first-person walking through diverse urban and green environments with
varying degrees of crowding, across 30 cities in 20 countries. These were rated by 377 participants on
emotional reactions, willingness to visit and other environmental appraisals.
Our results underline the interaction between environmental aesthetics and crowding in people’s ap-
praisal of various environments. Specifically, aesthetically pleasing (i.e. more beautiful) environments
evoke more positive emotions and higher willingness to visit, regardless of their level of crowding.
Less aesthetically pleasing environments presented an inverse-u relationship with crowdedness, where
valence and willingness to visit ratings peaked under moderate levels of crowding.
As many cities worldwide grapple with higher population densities and increased pedestrian flows,
these results highlight the importance of aesthetics and design to foster psychological well-being. The
validated library of videos is made available open-access on zenodo.
Keywords: urban design, architecture, aesthetics, crowding, restoration, green space
1. Introduction
Imagine walking down a beautiful, tree-lined boulevard, teeming with pedestrians - people strolling,
shopping, or enjoying the day. How would you feel? Would you like to visit such a place again? Now
consider walking through a dull, empty street – what does it look like? How would your mood and
desire to visit such a place change? It is increasingly well understood that various physical features
can modulate people’s psychological reactions and preferences towards a given environment (Coburn,
2019). Although there is extensive research on the negative psychological effects of crowding, such as
increased stress, cognitive load, and negative affect (Blut and Iyer, 2020; Schneider et al., 2023), in
some cases a certain degree of crowding may enhance the subjective experience of an environment.
For instance, in urban planning it is considered that "the presence of people in a space is interpreted
as indicative that a public space is functioning and healthy" (Gehl, 2013; Honey-Rosés et al., 2021).
However, the way crowding and environmental characteristics, such as its type or beauty, interact with
one another remains unclear.
1Corresponding authors. PM: panos.mavros@telecom-paris.fr; present address: i3, CNRS, Télécom Paris, Institut
Polytechnique de Paris, France. DM: d.makowski@sussex.ac.uk; School of Psychology, University of Sussex, UK

Extensive research during the past decades, has consistently reported that walking, being or even
looking at green spaces, compared built environments, brings multiple cognitive, attentional, and
psychological benefits (Wicks et al., 2022; Moll et al., 2022). Research focused into the psychological
effects of the built environment itself, for example of different urban designs or architectural styles,
also suggests that various design features can improve both the psychological effects, as well as the
aesthetic preferences of various urban environments. The architectural design itself can evoke different
psychological reactions, where architectural styles more rich in information (higher fractal dimension)
tend to be evaluated as more pleasing (Vartanian et al., 2015; den Berg et al., 2016; Patuano, 2017).
Nature-related features, such as the number of trees (Parsons and Daniel, 2002; Hands and Brown,
2002), presence of flowers (White and Gatersleben, 2011) and water features (Faggi et al., 2013; Voelker
and Kistemann, 2013), but also other types of interventions such as pedestrianisation (Bornioli et al.,
2018) also tend to improve psychological reactions.
In everyday life, spaces are often encountered in the presence of others – especially in pedestrianised
urban settings or in popular parks. However, in many studies in the field of environmental psychology,
the stimuli are comprised of images or videos where environments are presented empty of people, as
part of the effort to isolate the environmental effects and reduce confounding factors. Intuition, as
well as research on crowding (Engelniederhammer et al., 2019; Mavros et al., 2022) alike suggest that
the co-presence of people might influence our subjective experience, but it remains unclear whether
it can also influence on the valuation of the place itself. Thus, understanding the social influence on
the psychological effects of spaces is important given the the need of city authorities to anticipate and
design for increasing urban densities in many parts of the world.
From a methodological perspective, one challenge of environmental studies is to generate dynamic
stimuli such as videos which capture the properties of diverse environments. Previous studies have
achieved this at the architectural scale (e.g., Gregorians et al., 2022). To accomplish this for urban
environments we turned to the increasing number of ’walking’ videos that are available online (for
instance in YouTube) which capture longer or shorter walks from a first-person perspective without
any edits. This type of online material presents an under-explored source for dynamic stimuli for
environmental research with wide geographic coverage.
In this article, we address these questions regarding the affective and motivational impact of the
design qualities of different environments (urban built, or green) as well as human activity, specifically
pedestrian crowding. We conducted an online, environmental preference study (N=377), showing
participants 480 videos of walking compiled from different urban and green spaces of varying aesthetic
qualities and levels of crowding, and asking them to report their psychological reaction and provide
environmental appraisals. Our findings suggest multiple interactions between subjective crowding,
environmental perceptions such as beauty and emotional responses to places.
2. Background
2.1. Urban and green spaces
Extensive research during the past decades has established their positive psychological effects of
being exposed to green, compared to urban, spaces including the improvement of mood, attention
and cognition (for a recent review see Norwood et al., 2019), enhancing creativity (Sharam et al.,
2023), promoting restoration (Hartig et al., 2003), social interaction (Maas et al., 2009) and physical
activity (Ribeiro et al., 2015). These positive effects tend to hold across different kinds of green spaces,
from forests to urban green (parks), having views of greenery and trees (Ulrich, 1984; Olszewska-Guizzo
et al., 2018), sitting (Olszewska-Guizzo et al., 2022), or walking (Aspinall et al., 2013). Several theories
have been proposed to explain this phenomenon. The biophilia hypothesis (BH) posits that due to
our evolutionary history, humans have an innate preference for nature (Kellert and Wilson, 1995).
Ulrich’s stress-reduction theory (SRT) posits that access to green space fosters recovery from mental
(cognitive, psychological) stressors (Ulrich et al., 1991), while the attention restoration theory (ART)
postulates that natural and built environments differ in the way they recruit our visual attention:
urban settings evoke activity related, directed attention (e.g. traffic) which is taxing our attentional
capacities whereas nature recruits our attention through multifractal patterns, a phenomenon called
soft fascination which restores our psychological resources (Kaplan, 1995).
2

Studies have found evidence for both theories, for example, exposure to scenes of nature promotes
stress reduction (de Kort et al., 2006; Velarde et al., 2007) and restores attentional capacity (Berto,
2005). Conversely, exposure to built environment is associated with less positive and more negative
emotions (Yao et al., 2021). Furthermore, numerous neuroscientific studies have shown that exposure
to green spaces, compared to urban, are associated with different patterns of brain activity indicative
of lower arousal and reduced attentional load, such as increased power in the alpha (8-13Hz) and theta
(4-8Hz) frequency bands (e.g. Grassini, 2022; Mavros et al., 2022; Olszewska-Guizzo et al., 2021). Other
factors associated with built/urban environments, such as air pollution (Chen et al., 2018; Buoli et al.,
2018), traffic noise (Sygna et al., 2014), and biological (Hoisington et al., 2019) also have negative
effects.
ample research has focused on this ’urban vs nature dichotomy’ (Karmanov and Hamel, 2008), it
is increasingly recognised that aesthetic, design, social and other qualities of both green/natural and
urban/built environments can also modulate the subjective experience, emotional reactions and envi-
ronmental preferences of people. Due to increasing urbanisation worldwide, researchers are interested
to understand what aspects of the built environment itself can improve mental and physical wellbeing
(Saarloos et al., 2011; Guzman et al., 2021; Buttazzoni et al., 2021).
2.2. Environmental aesthetics
It is not only the typology (green or urban) of environments that triggers psychological reactions but
also their quality and aesthetics. In terms of green space quality, Olszewska-Guizzo (2023) developed
the Contemplative Landscape Model (CLM), which scores green spaces on seven dimensions consid-
ering among others diverse landscape and design elements, biodiversity, colour and lighting. They
found that spaces that score higher in CLM have a higher capacity to alleviate stress and promote
attention restoration (Olszewska-Guizzo et al., 2022). When the aesthetic value of urban stimuli is
matched with that of green stimuli, differences in affect disappear, although attentional effects persist
(Meidenbauer et al., 2020). Johansson et al. (2016) observed that perceived urban design qualities,
specifically complexity, aesthetic, upkeep and presence of greenery are significant in predicting intention
to choose and avoid a certain walking route, while this effect was mediated by affective experience (i.e.,
valence). Higher quality of public open spaces, for example having activities and better morphological,
architectural and urban design elements, is associated with improved mental health (Wei et al., 2021).
Various other factors influence aesthetic perception of the environment, but also mood and behaviour,
for instance, colours (Satariano, 2021; Jalil et al., 2012), daylight (Beute and de Kort, 2014).
At the architectural scale of buildings and interior spaces, multiple designed features influence our
emotions. Spaces that are more open, have higher ceilings and curvilinear elements are rated as more
attractive, pleasant and comfortable (Vartanian et al., 2013, 2015; Presti et al., 2022). Coburn et al.
(2019) proposed that there are three psychological dimensions underlying these psychological reactions.
Fascination, or the degree of engagement and interest in an environment and tends to be linked closely
to the level of complexity, coherence which captures the ease of comprehension of an environment, and
hominess which captures the degree that a spaces reminds us of familiar or personal spaces. These
dimensions influence the desire to visit these environments, expressed as an approach or avoidance
decision (Coburn, 2019), and are also positively correlated with valence and arousal (Gregorians et al.,
2022).
Similarly, in the case of outdoor urban environments, multiple design and aesthetic qualities are
found to influence psychological reactions. Taller buildings are often associated with more negative
affect (Mazumder et al., 2020), whereas enclosure, the visual framing of an environment when it is
bounded by building facades, is associated with better ratings for spaces (Stamps, 2010) and increased
pedestrian trips (Ewing and Handy, 2009) The visual complexity of buildings facades, which can be
assessed by its fractal dimension, is associated with higher aesthetic preference and scenic beauty rating
of both interior and exterior urban environments (Coburn et al., 2019; Yusufzyanova, 2020). Visual
complexity can achieved by architectural design or through the inclusion of green elements (White
and Gatersleben, 2011). More generally, fractal patterns have a positive relationship with aesthetic
perception of the environment (Marchand et al., 2014; Weinberger et al., 2021; Zhang et al., 2018).
Studies using EEG have found that fractal patterns evoke increase in both in parietal beta frequency
and in frontal alpha frequency, which are indicative of increased attention and wakeful relaxation
3

respectively (Hagerhall et al., 2008, 2015). This may also partially explain the restorative effects of
green environments, for example the mere presence of an indoor plant, or office window looking out at
nature is associated with significantly less stress than their counterparts lacking botanical decor, with
a view of a more plain cityscape, in healthcare, workplace or residential settings (Ulrich, 1984; Chang
and Chen, 2005; Olszewska-Guizzo et al., 2018)
The aesthetic qualities of architectural and urban environments are inextricably linked with human
experience. Recent studies reported that buildings rated by participants as more beautiful were also
perceived as more pleasant (Coburn et al., 2019) and have positive psychological outcomes (Johansson
et al., 2016, 2011). Conversely, environments that are rated as unattractive are associated with more
negative affect (Marchand et al., 2014; Yao et al., 2021), anxiety and depression (Beemer et al., 2021).
While environmental beauty, is positively correlated with valence, it is, however, uncorrelated from
arousal (Gregorians et al., 2022). This line of inquiry weaves together strands of research from envi-
ronment psychology and the emerging field of Neuroaesthetics, investigating the cognitive, attentional
and brain processes underpinning aesthetic experiences and, in particular, the experience of beauty.
According to Chatterjee, the experience of beauty is the result of an interaction between sensory-
motor, emotion-valuation, and meaning-knowledge processing of stimuli, that jointly contribute to
our aesthetic perception (Chatterjee and Vartanian, 2014). These findings are consistent with long
established ideas that beauty and pleasure are closely connected. The exprience of beauty, however, is
more nuanced; in arts and architecture beautiful artefacts can sometimes stir unpleasant emotions and
discomfort, reminding us of the notion of the sublime (Brielmann and Pelli, 2019; Brielmann et al.,
2022; Dewey, 1934). In this context, it is worth investigating how static and transient elements of the
environment influence our subjective experiences within it.
2.3. Psychological effects of crowding
The notion of crowding is a complex psychological construct, that can be defined as the ratio of
elements (typically people) in a given amount of space. For instance, in vision research, crowding
typically focuses on the negative effects of irrelevant stimuli (called flanker) on the perception of a
target stimulus (e.g. Levi, 2008). In retail studies, researchers differentiate between human crowding
which refers to the number of individuals and the rate of social interaction, and spatial crowding which
is defined as the (lack of) available space due to various obstructions (e.g. furniture, Machleit et al.,
1994). In transport and urban settings crowding can also be a dynamic phenomenon, referring to
the flow of people or vehicles moving through a passage, such as a gantry, a street, or other space.
Importantly, (Stokols, 1972) postulated that there is distinction between social density, defined as
number of people per unit area, and the subjective judgements of crowding which also depend on the
interaction between individuals (Stokols, 1972; Shelby et al., 1989).
Historically, concerns about overcrowding in residential settings led to studies on the social and
psychological consequences of high population density. Studies in the 1960s and 70s, found detrimental
effects of crowding both in animal (Calhoun, 1962) as well as in human communities (Galle et al., 1972).
Various studies have also found that more crowded settings are associated with a decline in cognitive
performance (Saegert, 1973; Levi, 2008) and have strong influence on emotions (Blut and Iyer, 2020).
More recently, research on crowding has mainly focused on its effects in transport, retail, tourism,
and residential settings. In line with earlier work, researchers often find negative effects of crowding
on passenger (see review by Tirachini et al., 2013; Schneider et al., 2023) and customer satisfaction
(Machleit et al., 2000), mood (d’Astous, 2000; Elbachir and Chenini, 2017; Machleit et al., 2000), and
mental well-being (Regoeczi, 2008; Mangrio and Zdravkovic, 2018; Wang and Liu, 2022), even in the
case of marine tourism (Bentz et al., 2015; Needham et al., 2018).
Research suggests that although spatial crowding has a negative effect that is linear, the psycholog-
ical effects of human crowding may have an inverted-U shape (Cheng et al., 2021; Eroglu et al., 2005;
Mehta et al., 2013). In other words moderate levels of human crowding are preferable compared to
either too low or too high levels. One explanation for this is that increased spatial crowding restricts
space available for movement and potentially a sense of control, and thus evoke negative reactions (Hui
and Bateson, 1991).
On the other hand, human crowding can be to a certain degree desirable, and its effect can be
moderated by spatial and social parameters. As Beermann and Sieben (2022) note "the specific social
4

situation in the box appears to play a more important role than merely the number of people waiting
there". This could be explained by theories such as optimal social contact (Altman, 1975; Sundstrom
and Altman, 1976) and arousal theory (Wohlwill, 1974) which suggest that individuals desire and seek
optimal levels of arousal in certain settings. Indeed, in the urban planning literature, the presence of
other people in urban spaces is typically associated with positive effects (Gehl, 2013; Whyte, 1980).
Recent work has started looking into the effects of human co-presence on the experience of green
and urban spaces. Zapata and Honey-Rosés (2022) found gender effects of the increased (co)presence of
people in a public space; whereas men chose to stay longer, women stayed less in the public space. Neale
et al. (2021) found that in nature, restorative effects persist with and without other people (walkers),
but the study involved only 7 other people (confederates) which was relatively low. Using a visual
discrete choice experiment, Campagnaro et al. (2020) found that with increasing number of pedestrians
in a park, general preference and (expected) stress reduction reduces, but conversely sense of safety
increases. Kyle et al. (2022) conducted a survey to understand social norms related with crowding in
a natural setting (lake). Their analysis shows that people’s emotional and behavioural responses to
encountering a crowded settings depend on their attitudes (i.e. preferences for) towards crowding, but
also their expectations for a particular time and setting. In another recent study, Gregorians et al.
(2022) also examined the effect of people presence in walk-through videos of architectural space but
did not find an effect of the presence of people on measures on self-reported emotion or environmental
appraisals.
Several of these studies crowding is generally low and the presence of people is treated as a dichoto-
mous variable, i.e. coded as with or without people, rather than counting pedestrians. Studies which
consider larger density of people, as in central business districts, tend to find more clear effects (Schnei-
der et al., 2023). Li et al. (2019) found that high levels of crowding did not impact wayfinding strategies
but influenced walking behaviours. In a study comparing high versus low crowds in a downtown urban
environment, Mavros et al. (2022) found that participants reported more negative emotions and higher
arousal, accompanied with higher indices of physiological arousal and EEG markers of cognitive load.
2.4. Methodological considerations
There are three methodological considerations that emerge in this literature. Firstly, as noted by
Gregorians et al. (2022), many of the studies mentioned above rely on static images of real or artificially
generated environments, whereas urban environments are often encountered in motion, during walk-
ing or being on a moving vehicle. In response, more recent studies have emphasised the use of video
walk-throughs to emulate in laboratory conditions the dynamic aspects of environmental exposure
(e.g. Gregorians et al., 2022; Mavros et al., 2022) . Second, psychological reactions to the environ-
ment can be measured in affective, cognitive-perceptual and motivational terms. Self-reported tests of
emotion following the pleasure-arousal model (a. Russell, 2009) or the positive-negative affect model
(Watson et al., 1988), and can be accompanied with behavioural task (e.g. the stroop or variations of
the digit span tasks). Participants are also asked to rate multiple environmental factors (Kort et al.,
2003; Gregorians et al., 2022). More recently, researchers have started examining motivational as-
pects of environmental preferences, which can be conceptualised as an approach or avoidance decision
(Vartanian et al., 2013, 2015; Coburn, 2019). This concept relates with the notion of willingness to
visit or willingness to walk that is used in the planning literature to describe the propensity of urban
populations to walk for leisure or transport in different environments (Ewing and Handy, 2009; Ameli
et al., 2015; Johansson et al., 2016; Sevtsuk and Basu, 2022). Through these different measurement
approaches we can assess the multifaceted effects of different environments on human psychology and
behaviour. Lastly, one challenge in the investigations of crowding is to establish objective numbers
of people in different spaces; this is especially difficult when considering moving pedestrians. Novel
techniques that have been recently developed, such as unsupervised machine-learning, can be used to
extract such information from images and videos alike (e.g., Cao et al., 2019). This allows to study
the precise effects of increasing levels of crowding, to identify tipping points when the perception of
crowding shifts from desirable to neutral or undesirable, and also to investigate whether the design of
spaces can moderate this effect.
5

2.5. The present study
Weaving together the three strands of literature on environmental restoration, aesthetics and social
co-presence (or crowding), our study aims to understand whether the inverse-U relationship between
human crowding and its effects on human’s emotional and behavioural reactions (Nouri, 2011; Cheng
et al., 2021; Eroglu et al., 2005; Mehta et al., 2013) is moderated by environmental typology (urban
vs green) and aesthetics. Given also that pedestrian activity is considered positive in the planning
literature, but negative in other cases, its effect on emotions, attention and restoration is unclear.
Further, it is unclear if there is constant effect or a tipping point after which reactions to the presence
of people change, and if these reactions are influenced by typology (urban or green) and by aesthetic
qualities.
The primary objective of the present study is to addresses these questions through a large, online,
environmental preference study. Second, in line with other recent studies(Gregorians et al., 2022),
we aim to establish a pool of of video walk-throughs (dynamic stimuli) through real-world urban
environments, varying in terms of aesthetics, crowding and typology, which can be used to study
urban experience. We consider both objective social density (i.e. people count) as well as the subjective
judgement of perceived crowding to understand how psychological experience is moderated by spatial
and environmental factors. Here we take advantage of state-of-the-art machine learning methods to
automatically derive pedestrian counts in the video stimuli. To the best of our knowledge, this study is
the first to investigate the joint effects of crowding and beauty (aesthetics) on the emotional experience
and perceptions of different environments. Our research questions can be summarised in the following
hypotheses:
H1 There is an interaction between aesthetic judgement (beauty) and crowding on their effect on
emotional experience; i.e., highly aesthetic (beautiful) environments are perceived positively re-
gardless of the level of pedestrian activity, whereas perceptions of less aesthetic environments are
more positive when there moderate levels of crowding;
H2 There is an inverse-U relationship between levels of pedestrian crowding and affective experience,
i.e., moderate levels of pedestrian crowding lead to more positive emotions.
H3 Environment Typology (green or urban) moderates the effect of social density on aesthetic judge-
ments and affective experience, i.e., more pedestrian activity leads to a larger decrease in affective
ratings in green spaces, compared to urban spaces.
H4 Environment beauty moderates the effect of social density on affective experience.
H5 Personality traits explain inter-individual differences in the effects of crowding on aesthetic judge-
ments of green and urban spaces.
3. Materials and Methods
3.1. Data and code availability
The study was pre-registered on the Open Science Framework (osf.io), including the sample size,
hypotheses and analyses plan. The preregistration file, the preprocessed (i.e., cleaned) data and analysis
scripts are available on OSF project page (https://osf.io/pwz6g/).
3.2. Participants
Data from a total of 377 participants (Mean age = 30.8, SD = 10.0, range: [18, 67]; Sex: 49.3%
female) were included in analysis, while data from 31 participants were excluded from analysis due
to attention or streaming issues (see Section 3.7.1). All participants were recruited via the online
platform prolific.co (see Table 1), with recruitment criteria requiring that they complete the study
from a desktop computer, and that they are willing to turn-on their webcam (used solely to assess their
attention on the task with the webgazer.js plugin; no videos were recorded and data are not analysed
in this article). The sample size was determined to 320 in order to ensure at least 20 participants rated
each stimulus (see below, 16 sets of 30 stimuli). Due to randomisation, technical issues, and participant
6

failure in attention checks, data collection continued until each set of stimuli was watched at least by
20 participants. The study was approved by the Ethics Committee of the Nanyang Technological
University (IRB-2022-188) . All methods described in this section were performed in accordance with
the Declaration of Helsinki, and with the relevant guidelines and regulations.
Table 1: Participant characteristics
Sex
Total Female Male
n = 377 n = 186 n = 191
Age
18–29 229 (60.7%) 112 (60.2%) 117 (61.3%)
30-49 125 (33.2%) 61 (32.8%) 64 (33.5%)
50-64 22 (5.8%) 12 (6.5%) 10 (5.2%)
65+ 1 (0.3%) 1 (0.5%) 0 (0%)
Education
Architecture 17 (4.5%) 9 (4.8%) 8 (4.2%)
Arts 50 (13.3%) 26 (14%) 24 (12.6%)
Other 310 (82.2%) 151 (81.2%) 159 (83.2%)
Ethnicity
Asian 31 (8.2%) 15 (8.1%) 16 (8.4%)
Black 54 (14.3%) 33 (17.7%) 21 (11%)
Mixed 55 (14.6%) 28 (15.1%) 27 (14.1%)
Other 24 (6.4%) 10 (5.4%) 14 (7.3%)
White 213 (56.5%) 100 (53.8%) 113 (59.2%)
Employment
Other (e.g., unemployed, retired) 209 (55.4%) 106 (57%) 103 (53.9%)
Working/Studying 168 (44.6%) 80 (43%) 88 (46.1%)
Upbringing Environment
Other 189 (50.1%) 86 (46.2%) 103 (53.9%)
Urban 188 (49.9%) 100 (53.8%) 88 (46.1%)
Current Environment
Other 126 (33.4%) 62 (33.3%) 64 (33.5%)
Urban 251 (66.6%) 124 (66.7%) 127 (66.5%)
Approval rate
99.5 (1.0) 99.5 (1.0) 99.5 (1.0)
3.3. Study Design
We followed a repeated measures partial incomplete block study design (Figure 1, see also prereg-
istration https://osf.io/pwz6g/). The 480 stimuli were pre-allocated in 16 sets of 30, which were
balanced to ensure diversity of environments (see below). Participants were randomly assigned to
watch 1 set of 30 video stimuli and their presentation was randomised to reduce potential order effects.
Each stimulus was watched and rated by a minimum 20 participants.
3.4. Materials
A database of a total of 480 video stimuli was curated from publicly available online sources of
first-person perspective, unedited, ‘walking’ videos, i.e., a walk through in a real-world environment
(Figure 2) from .
The original videos can sometimes be more than 1 hour long, covering hundreds of meters and
many streets in one or more neighbourhoods of a town or city. Stimuli for the study were created by
extracting sub-clips between 20 and 60 seconds. These sub-clips were chosen so that the environments
varied in terms of (i) environment typology (408 urban and 72 green space), (ii) levels of crowding
(ranging from empty to very crowded), aesthetic qualities, as well as geographic origin. The original
7

Debrief
Experiment Procedure
+
Willingness to
walk
Arousal &
Valence
Restoration
Presence
Sign-up in
Prolific.co
Informed
Consent
Questionnaires
& Survey
Supplementary
Task*
Repeats X 31 (demo + 30 videos)
Instructions &
eye-tracking
calibration
Stimuli (videos)
presentation
Experiment timeline
Env. appraisals
480 videos
30 videos
16 sets
Kmeans &
dissimilarity
Figure 1: Schematic diagram of study design and experimental procedures.
Figure 2: Still frames extracted from 16 out of the 480 video walk-throughs that were used as stimuli. The videos,
obtained from public sources, originate from 31 different places (cities, parks and forests) across 20 countries.
8

videos were edited to extract sections of linear movement with minimal changes in camera direction
(panning) – in other words the person is walking forward and the camera looks forward in the direction
of movement. Finally, the videos were displayed for a maximum duration of 30 seconds so that
environmental exposure is consistent across stimuli. Stimuli are made available on a public repository
2when their license allows redistribution, to enable further research.
The stimuli pool of 480 videos was then subdivided into 16 sets of 30 video clips, that we tried to
balance using the following pipeline: we first performed image processing, using the python libraries:
cv2 (opencv.org) and pyllusion (realitybending.github.io/Pyllusion/), on all frames from each
video, extracting an average score in terms of Number of Faces, Colorfulness, Contrast, Perceived
Luminance and Entropy. We then applied K-Means clustering on the five metrics, resulting in 3
clusters, and partitioned the total pool into 16 sets of 30 videos so that each set has an equal proportion
of videos from each cluster. During data collection, the 16 sets of videos remained consistent, but their
order of presentation was shuffled to reduce potential order effects.
During data analysis, we repeated the crowd detection in each video, using the more advanced
OpenPose library (Cao et al., 2019). OpenPose is a robust, scalable, online library for pose-detection in
larger crowds which allowed us to estimate the number of people detected in each frame. We computed
the average pedestrian count per video, which we used in later analyses as an index of “objective
crowding” and we refer to as ’Pedestrian counts’ to differentiate from self-reported, or subjective,
crowding.
3.5. Item ratings & questionnaires
Immediately after the presentation of each video, reactions to stimuli were captured with two sets
of ratings. First, participants were asked to rate their (i) ‘willingness to walk’, (ii) their psychological
reaction using Russell’s ‘circumplex model of affect’ (a. Russell, 2009) (they were asked to rate their
experience in terms of valence - negative to positive - and arousal - calm to excited - on a 9-point
bipolar scale), and (iii) their sense of presence (with a single item; "I had a sense of being there in
that place") as a manipulation check. (iv) To relate our findings with research on restorative theory,
we also included a single item from the perceived restorativeness scale as suggested by Lombardi and
Ciceri (2021); "Spending time there would give me a good break from my day-to-day routine".
Second, participants were asked to complete a brief environment appraisal for each stimulus. Seven
bipolar adjective items were selected from the semantic differential scale for environmental evaluation
(Kort et al., 2003): boring-interesting, narrow-wide, ugly-beautiful, chaotic-ordered, unfamiliar-familiar,
unattractive-scenic, and empty-crowded.
After watching all the stimuli, participants also completed the mini-IPIP6 (Sibley et al., 2011),
measuring Extraversion, Agreeableness, Conscientiousness, Neuroticism and Intellect/Imagination, and
Honesty-Humility with 7 point Likert scales, a short-form version of the stimulus screening ability scale
(Mehrabian, 1977) to measure the participants’ reaction towards distracting stimuli (associated with
the need for privacy in workplaces and psychological reactions towards perceived crowding; see Oldham,
1988; Maher and von Hippel, 2005), one item from the noise sensitivity scale (“I get used to most noises
without much difficulty"; Benfield et al., 2014), and the short form versions of the social phobia (SPS)
and the short social interaction anxiety (SIAS) scales (Peters et al., 2012). Finally, participants were
also asked to rate their general attitudes towards walking in crowded environments and, because the
study was conducted during the COVID-19 pandemic, they were also asked to rate how concerned
they generally are about contracting COVID-19 when they are walking within crowded places (see
questionnaire items in the Appendix).
3.6. Procedure
The experiment procedure was implemented in jsPsych (de Leeuw, 2015), and the full code for
the task is available at https://github.com/pmavros/nice_online_study/. Data collection was
performed using the platform https://cognition.run and participant management was performed
using Prolific.co, a platform and subject pool for online experiments which is consider to offer more
210.5281/zenodo.7923847
9

transparency and ethical treatment of participants, e.g., in terms of minimum payment (Palan and
Schitter, 2018) and its participant pool is providing more reliable and attentive responses to studies
(Peer et al., 2022). Upon providing informed consent, participants were randomly assigned to one of
16 lists of videos and were shown the 30 videos walkthroughs in a randomized order. The experiment
consisted of three main phases: (i) instructions and calibration of the eye-tracking (used as an attention
check and not analysed in this article), (ii) a demo trial and 30 trials of presentation and rating of
video walk-throughs, and (iii) completing a personality and demographic survey. In line with Prolific’s
guidelines, participants in our study were compensated at a rate of GBP 9 (approx US$10) per hour,
independent of their geographic location. Data collection was performed over several dates and at
different times of the day, as an effort to reduce the geographic bias of respondents.
3.7. Data Analysis
Data analysis was carried out using Python and R (version 4.2.0, R Core Team, 2022), with the
help of the tidyverse (Wickham et al., 2019) and the easystats (Makowski et al., 2020) collections of
packages. The full reproducible analysis script is available at
3.7.1. Preprocessing steps and data quality
A series of checks to ensure data quality. First, participants who abandoned before completing all
stages (technical and other reasons) were removed. Second, we performed attention checks based on the
‘browser window’ state, i.e. which detected whether the participants had other windows open in their
computer (they were informed in advance, and provided a reminder halfway through the session that
we keep track of their focus). We only included all trials where participants had the browser window
(i.e., the study tab) in ’focus’ and in full-screen; all other trials were removed. Data from participants
who were found to be distracted for more than 2 stimuli were removed completely. We also removed
participants with low variance in responses (i.e. average standard deviation across stimuli less than
0.75), as well as participants who had difficulties in the video playback (i.e., streaming); specifically, if
the video buffering exceeded the video duration (i.e., more than 30 seconds of buffering), as this could
impact the viewing experience and subsequent response. Two participants that experienced repeatedly
video playback less than video duration were also removed. In summary, out of 408 participants
who completed the study, 31 were removed after the preprocessing steps described above, leaving 377
included for analysis.
3.7.2. Statistical approach
Statistical analyses were performed using generalised mixed models. Since we hypothesised that
the effect of crowding has an inverse-U shape, we modelled pedestrian counts and perceived crowding
using second-order polynomial terms (see Supplementary materials for analysis code). In the next
section we present descriptive statistics for the variables measured, perform a manipulation check, and
report statistical models for subjective crowding, perceived beauty, psychological reactions, as well as
the effect of personality traits on preferences for crowded environments.
4. Results
4.1. Descriptives of responses to stimuli
After removing participants with low attention, low effort, or streaming difficulties (see Sec-
tion 3.7.1), we analysed 11117 trials from 377 participants, corresponding to ≈29.5 trials per par-
ticipant. Each video was seen on average 23.2 ±1.88 times. Table 2 shows the number of valid
responses for a total of 11117 trials, the mean and standard deviation for each dimension of emotional
responses and environmental appraisals. As an initial exploratory analysis, Bayesian t-tests, which
are more robust to multiple comparisons, show decisive evidence for a difference between the green
and urban space typologies. Further, a partial correlation matrix (Figure 3) shows clear associations
between the measures themselves, in particular positive correlations between emotional dimensions of
valence and arousal, willingness to walk, restoration and presence with the beauty, structure, width
and interest evoked by an environment. We can also observe a negative correlation between subjective
crowding, and valence, beauty and restoration; in addition this relationship changes depending on the
typology. These relationships are examined in more depth below.
10

Table 2: Descriptive statistics of emotional reactions and environmental appraisals of the stimuli, grouped by typology.
Urban (408 stimuli) Green (72 stimuli)
Unique (#) Mean SD Unique (#) Mean SD BF*
WillingnessToWalk 9289 3.63 1.73 1620 4.82 1.45 324.02
Arousal 9380 3.90 1.50 1639 3.27 1.89 110.38
Valence 9380 4.75 1.45 1639 5.81 1.27 367.35
Restoration 9459 3.25 1.92 1658 4.85 1.48 487.29
Presence 9458 4.23 1.58 1658 4.71 1.44 63.18
Beauty 9457 4.41 1.53 1657 5.89 1.27 647.00
Crowdedness 9456 4.41 1.79 1658 2.98 1.82 427.16
Width 9456 4.21 1.74 1656 5.11 1.81 180.22
Structure 9453 4.38 1.66 1656 5.47 1.41 304.94
Scenic 9453 4.24 1.61 1658 5.82 1.34 665.64
Interest 9456 4.34 1.80 1657 5.35 1.59 221.00
Familiarity 9455 4.10 1.76 1658 4.50 1.82 31.53
(*) Bayes Factor of a bayesian t-test
Figure 3: (A) Overview of average ratings for each video stimulus. (B) Correlation Matrix between self-reported measures
of emotion and environmental appraisals for videos of urban (B1) and green (B2) environments.
11

4.2. Manipulation check
As a manipulation check, we asked participants to report their sense of presence while watching
each video. On a scale from 1 to 7, the average presence score was 4.3, reflecting the fact that this study
was conducted online and participants viewed the stimuli (videos) on a computer screen. However,
we also observed an interaction between levels of crowding, perceived beauty and sense of presence
(Figure 4). Participants reported a higher sense of presence in environments considered beautiful, and
sense of presence increased with higher levels of crowding for environments considered less beautiful.
12

4.3. Determinants of subjective crowding
Figure 4: (A) Joint effects of beauty, typology, and environmental characteristics on perceptions of crowding. (B) Joint
effects of typology and crowding on perceptions of environmental beauty. (C) Joint effects of beauty, typology, and
crowding on psychological reactions. Note that lines show model estimates, whereas 2D density layer (background)
shows raw values.
Psychological perceptions of crowding are a composite of the actual number of people in a loca-
tion, together with spaciousness, and other characteristics of the physical and social environment. To
examine which environmental and psychological factors influence the degree to which an environment
is considered crowded (Figure 4 A), we fitted linear mixed effects models (LMM), specifying subjective
crowding as the outcome (dependent variable), and specifying as fixed effects typology (green vs urban)
and log-transformed pedestrian counts (see definition in Methods). We also included participants as
random effects and included random slopes for the effects of interest. We fitted four models, one for
each environmental appraisal: beauty,scenery,structure, and width. Table 3 shows results of the LMM
fitted to each appraisal (separate models). There is a significant effect for average pedestrian counts
and familiarity with subjective crowding. Further we found a statistically significant negative effect
of each of the four factors on subjective crowding, in other words environments that are considered
13

Table 3: Linear mixed-effects models of Subjective Crowding.
Model 1 Model 2 Model 3 Model 4
Term Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI
(Intercept) 1.642*** [1.463, 1.821] 1.544*** [1.374, 1.714] 1.851*** [1.684, 2.017] 3.760*** [3.562, 3.957]
Ped.Counts [log1p] 1.850*** [1.760, 1.939] 1.879*** [1.797, 1.962] 1.844*** [1.766, 1.923] 1.180*** [1.093, 1.267]
Beauty -0.035+ [-0.072, 0.003]
Typology [green] 0.003 [-0.300, 0.307] 0.091 [-0.195, 0.377] -0.119 [-0.327, 0.088] -0.871*** [-1.135, -0.607]
Familiarity 0.028*** [0.014, 0.043] 0.026*** [0.011, 0.040] 0.039*** [0.024, 0.053] 0.055*** [0.042, 0.068]
Ped.Counts [log1p] ×Beauty -0.028** [-0.047, -0.008]
Ped.Counts [log1p] ×Typology [green] 0.565*** [0.460, 0.670] 0.582*** [0.477, 0.687] 0.593*** [0.494, 0.692] 0.662*** [0.567, 0.758]
Beauty ×Typology [green] -0.054* [-0.104, -0.004]
Scenic -0.009 [-0.045, 0.028]
Ped.Counts [log1p] ×Scenic -0.038*** [-0.056, -0.019]
Scenic ×Typology [green] -0.075** [-0.122, -0.027]
Width -0.079*** [-0.113, -0.046]
Ped.Counts [log1p] ×Width -0.040*** [-0.057, -0.024]
Width ×Typology [green] -0.045* [-0.081, -0.009]
Structure -0.416*** [-0.454, -0.378]
Ped.Counts [log1p] ×Structure 0.048*** [0.031, 0.066]
Structure ×Typology [green] 0.094*** [0.051, 0.137]
Num.Obs. 11107 11104 11105 11102
R2 Cond. 0.631 0.632 0.646 0.689
R2 Marg. 0.565 0.564 0.575 0.621
wider and more structured, scenic, and beautiful, are perceived as less crowded. In addition, we found
significant interactions between pedestrian counts, beauty, scenicness, structure and width. In other
words, for the same number of people, ratings of subjective crowding are lower for environments that
are rated as more spacious, ordered, scenic and beautiful.
4.4. Determinants of perceived environmental beauty
We then looked at the parameters associated with higher ratings of environmental beauty. We
fitted non-linear mixed models (NLMM) specifying beauty as the outcome (dependent) variable, and
fixed effects for typology and familiarity. We fitted four models, one for each environmental appraisal:
subjective crowding, structure, scenicness, and width; all were specified as second-degree polynomials to
test for an inverse-U effect. We also included participants as random effects and included random slopes
for the effects of interest. Table 4 show the model results. We found a consistent effect of environmental
typology, where videos of green spaces are rated as more beautiful. We found statistically significant
effects for structure, scenicness and spaciousness, however, the models suggest these relationships are
linear (2nd degree polynomial is not significant). There was also a small but significant positive effect
of familiarity, i.e., environments that look more familiar were considered more beautiful. Importantly,
we find significant negative effect of crowding and perceptions of beauty, which is amplified as crowding
increases (Figure 4B).
14

Table 4: Non-linear mixed effects models of Perceived Beauty.
0 Model 1 Model 2 Model 3 Model 4
Term Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI
(Intercept) 3.861*** [3.764, 3.957] 4.006*** [3.919, 4.094] 4.469*** [4.422, 4.516] 4.071*** [3.983, 4.159]
Subj. Crowding [1st degree] -0.972 [-5.336, 3.391]
Subj. Crowding [2nd degree] -17.489*** [-20.562, -14.416]
Typology [green] 1.328*** [1.245, 1.410] 1.087*** [1.010, 1.165] 0.243*** [0.189, 0.297] 1.178*** [1.107, 1.249]
Familiarity 0.133*** [0.117, 0.150] 0.113*** [0.098, 0.128] 0.033*** [0.024, 0.042] 0.099*** [0.084, 0.115]
Subj. Crowding [1st degree] ×Typology [green] -18.689*** [-27.473, -9.904]
Subj. Crowding [2nd degree] ×Typology [green] 9.603* [1.638, 17.567]
Structure [1st degree] 55.409*** [51.183, 59.636]
Structure [2nd degree] 4.805** [1.783, 7.827]
Structure [1st degree] ×Typology [green] 0.514 [-8.989, 10.017]
Structure [2nd degree] ×Typology [green] -3.159 [-11.461, 5.143]
Scenic [1st degree] 137.104*** [134.614, 139.594]
Scenic [2nd degree] -2.104* [-3.992, -0.216]
Scenic [1st degree] ×Typology [green] 1.391 [-4.896, 7.679]
Scenic [2nd degree] ×Typology [green] -4.448+ [-9.594, 0.698]
Width [1st degree] 59.713*** [55.215, 64.210]
Width [2nd degree] 5.092*** [2.085, 8.099]
Width [1st degree] ×Typology [green] -17.449*** [-24.457, -10.440]
Width [2nd degree] ×Typology [green] 1.021 [-5.801, 7.842]
Num.Obs. 11107 11102 11104 11105
R2 Cond. 0.366 0.519 0.797 0.539
R2 Marg. 0.136 0.195 0.688 0.193
4.5. The joint psychological effect of crowding and beauty
15

Table 5: Non-linear mixed effects models of psychological reaction.
0 WillingnessToWalk Valence Arousal Restoration Presence
Term Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI
(Intercept) 0.020 [-0.080, 0.120] 1.622*** [1.539, 1.705] 2.929*** [2.802, 3.056] -0.399*** [-0.520, -0.279] 2.490*** [2.353, 2.627]
Typology [green] 0.023 [-0.271, 0.318] -0.013 [-0.253, 0.227] -0.024 [-0.396, 0.348] 0.575*** [0.248, 0.902] 0.062 [-0.209, 0.333]
Familiarity 0.051*** [0.038, 0.065] 0.065*** [0.054, 0.076] 0.007 [-0.009, 0.024] 0.007 [-0.008, 0.021] 0.151*** [0.138, 0.163]
Typology [urban] ×Subj. Crowding [1st degree] -6.052+ [-12.498, 0.395] -8.555** [-13.828, -3.281] 13.632** [5.459, 21.805] 1.918 [-5.321, 9.157] 15.067*** [9.065, 21.068]
Typology [green] ×Subj. Crowding [1st degree] -23.512+ [-48.895, 1.870] -41.439*** [-62.274, -20.605] 6.878 [-25.421, 39.177] -22.054 [-50.641, 6.533] 15.103 [-8.575, 38.781]
Typology [urban] ×Subj. Crowding [2nd degree] -26.946*** [-33.409, -20.482] -30.770*** [-36.058, -25.482] 13.767*** [5.571, 21.962] -15.490*** [-22.749, -8.230] 3.827 [-2.192, 9.845]
Typology [green] ×Subj. Crowding [2nd degree] -30.196* [-54.042, -6.349] -44.388*** [-63.912, -24.864] -2.553 [-32.816, 27.710] -25.095+ [-51.919, 1.729] -16.595 [-38.815, 5.626]
Typology [urban] ×Beauty 0.771*** [0.756, 0.787] 0.650*** [0.637, 0.663] 0.201*** [0.182, 0.221] 0.823*** [0.805, 0.840] 0.255*** [0.240, 0.269]
Typology [green] ×Beauty 0.757*** [0.708, 0.806] 0.653*** [0.613, 0.694] 0.087** [0.025, 0.149] 0.776*** [0.721, 0.830] 0.250*** [0.204, 0.295]
Typology [urban] ×Subj. Crowding [1st degree] ×Beauty 0.767 [-0.627, 2.161] 0.974+ [-0.166, 2.113] 7.536*** [5.770, 9.302] -1.402+ [-2.967, 0.163] -2.505*** [-3.802, -1.207]
Typology [green] ×Subj. Crowding [1st degree] ×Beauty 1.179 [-3.212, 5.569] 5.060** [1.459, 8.661] 3.839 [-1.743, 9.422] 1.488 [-3.451, 6.427] -2.955 [-7.046, 1.137]
Typology [urban] ×Subj. Crowding [2nd degree] ×Beauty 3.111*** [1.706, 4.516] 4.637*** [3.488, 5.787] -3.858*** [-5.640, -2.077] 0.961 [-0.617, 2.539] -0.502 [-1.810, 0.806]
Typology [green] ×Subj. Crowding [2nd degree] ×Beauty 2.653 [-1.411, 6.716] 6.075*** [2.747, 9.402] -0.259 [-5.417, 4.899] 2.748 [-1.822, 7.318] 1.934 [-1.851, 5.720]
Num.Obs. 10899 11009 11009 11107 11106
R2 Cond. 0.640 0.664 0.311 0.645 0.625
R2 Marg. 0.545 0.563 0.127 0.497 0.118
16

Following the separate analyses of the factors which influence perceptions of crowding and beauty,
we looked at the joint effect of beauty and crowding on psychological reactions: willingness to walk,
valence, arousal, restoration, interest and presence. We fitted NLMMs, specifying each reaction as the
outcome variable, adding polynomial predictors for subjective crowding, and fixed effects for typology
and beauty. Table 5 shows model estimates and performance. In Figure 4(B) it can be observed
that spaces that are considered as more beautiful, are consistently rated higher in willingness to visit,
leading to more positive emotions, excitement, sense of presence and sense of restoration. Green
spaces are rated as more lower in arousal, but for the other variables after taking beauty into account
ratings of valence, willingess to walk, and presence are similar. Further, we can observe an inverse-U
relationship between levels of crowding and valence, willingness to walk, and restoration. Presence
appears modulated by the interaction of beauty and typology, but not crowding.
4.6. Individual differences in preference for crowding
Last, to understand individual difference in preferences for crowding, we analysed self-reported
preference for crowded environments is related with participant characteristics, personality traits and
participants’ own responses to the different stimuli. Overall, participants tend to dislike crowded envi-
ronments (mean 1.2, on a scale 0 to 4). We performed we fitted a linear mixed effects model specifying
willingness to walk as the outcome variable, and fixed effects for average pedestrian count, and envi-
ronmental beauty, random effects for participants and random for average pedestrian count. From this
model we then extracted the random intercept for each individual, which conceptually corresponds
to their average response to pedestrian crowds, as well as the random slope, which corresponds to
whether participants respond more positively or more negatively when pedestrian crowds increase.
These values were then compared with their self-reported preference for crowded environments using
Bayesian Pearson correlations. We found decisive evidence for a positive relationship between random
intercept and crowd preference (Rho = 0.24 , CrI=[0.15, 0.33], BF > 1000) as well as decisive evidence
for a positive relationship between random slope and crowd preference (Rho = 0.35, [0.25, 0.43], BF
> 1000). We then compared the self-reported preference for crowds (see Section 3.5), with age, sex,
environment of upbringing and current residence (urban vs other), and personality traits (IPIP6). We
found that individuals who self-reported higher preference for being in crowded environments, tend to
score higher in extraversion, lower in the honesty-humility scale, higher in stimulus screening ability
and also low in noise sensitivity (Table 6). We did not observe any differences explained by sex or
whether people grew up or currently live in a city.
17

Table 6: Determinants of Crowd preference
Parameter Median 95% CI BF
Bayesian Pearson Correlations
Crowd Prefs ~ Extraversion 0.272 [ 0.179 - 0.364 ] > 1000***
Crowd Prefs ~ Honesty -0.270 [ -0.358 - -0.177 ] > 1000***
Crowd Prefs ~ NSS -0.237 [ -0.334 - -0.144 ] > 1000***
Crowd Prefs ~ SSA 0.194 [ 0.097 - 0.289 ] 160.99***
Crowd Prefs ~ SIAS -0.118 [ -0.224 - -0.021 ] 1.83
Crowd Prefs ~ Neuroticism -0.110 [ -0.209 - -0.011 ] 1.14
Crowd Prefs ~ SPS -0.102 [ -0.201 - -0.000 ] 0.861
Crowd Prefs ~ Openness -0.038 [ -0.140 - 0.058 ] 0.159
Crowd Prefs ~ Agreeableness 0.036 [ -0.058 - 0.138 ] 0.153
Crowd Prefs ~ Conscientiousness -0.029 [ -0.128 - 0.068 ] 0.144
Crowd Prefs ~ Age -0.019 [ -0.120 - 0.082 ] 0.130
Bayesian T-tests
Crowd Prefs ~ Sex -0.145 [ -0.314 - 0.026 ] 0.431
Crowd Prefs ~ Upbringing Env. -0.049 [ -0.231 - 0.125 ] 0.135
Crowd prefs ~ Current Env. -0.079 [ -0.266 - 0.104 ] 0.168
5. Discussion
The present study investigated how the presence of pedestrians in urban or natural environments, in
other words levels of crowding, as well as the aesthetic qualities of these spaces, influence psychological
and environmental perceptions. Extensive research during the last decades points to psychological
benefits of being or walking in natural, compared to urban environments, and has also identified positive
psychological effects of high quality architectural, urban, and natural spaces. There are however
divergent effects of crowding, depending on the spatial, behavioural and social context. Furthermore,
it remains unclear what is the joint effect of crowding and aesthetics – i.e., encounters with places of high
or low aesthetic qualities under different levels of crowding. In addition, although a lot of research on
the psychological influence of urban and/or natural environments, has studied static exposure to those
environments (e.g., sitting in park or a room), here we turn our attention to the psychological effects
of pedestrian experience – in other words how we experience different urban or green environments
during walking.
To address these research questions, we compiled a library of 480 videos comprised of first-person
perspective walk-throughs that depict different outdoor environments, sampled from online sources to
ensure an international coverage of urban and natural environments from diverse spatial, social, and
cultural contexts (30 cities and parks from 20 countries). We then conducted an online study (N=379),
asking each participant to watch a subset of 30 videos and then provide ratings of their subjective
experience, as well as ratings of qualitative environmental characteristics. In line with our hypotheses,
we find interactions between perceptions of crowding, environmental aesthetics (perceptions of beauty)
and typology. These findings have important implications for our understanding of the relationship
between the design quality of green and urban environments, human behaviour and well-being. In the
following sections we discuss our findings, organised by theme, the limitations, policy implications and
future research directions.
5.1. Determinants of subjective crowding
We first examine which factors influence perceptions of crowding in pedestrian settings. In line with
our expectations, we found that subjective crowding is positively correlated with pedestrian counts,
and also that for the same number of people more spacious environments are perceived as less crowded
supporting the notion of social density. We also found that environmental typology also contributes to
the relationship between objective and subjective crowding, specifically the same amount of persons in
18

the scene is rated as more crowded in green spaces. Notably, we observed that when adjusting for the
number of people present, spaces rated as more beautiful, scenic and ordered, are also perceived as less
crowded. Previous literature suggests that the presence of people, in the context of cities, has divergent
associations, which are in some cases negative, such as crowding in public transportSchneider et al.
(2023), and in other cases positive, for instance when considering the vibrancy of streets Gehl (2013).
Here we find that environmental typology and aesthetics also have a moderating role, an observation
that calls for a more nuanced study of crowding, and how to mitigate its effects.
5.2. Environmental perceptions
We asked participants to rate environments for beauty,scenicness,structure, and interest. In
line with recent studies, all these dimensions were positively correlated (e.g. Coburn et al., 2019;
Gregorians et al., 2022). Intuitively, we found that ratings of scenicness were closely correlated with
those of beauty, and green environments were generally perceived as more scenic. This is consistent
with past research linking scenic beauty with both the presence of green space and botanical features
(Yusufzyanova, 2020). Higher beauty ratings were also positively correlated with structure and width.
This is inline with previous studies showing that individuals consider more pleasing and prefer more
spacious indoor and outdoor environments (Vartanian et al., 2015), and relates with the notion of
enclosure (Csanády, 2019; Presti et al., 2022). Ratings of how structured an environment is were found
to be positively correlated with beauty, regardless of environment typology. This is consistent with
previous research, where orderliness predicted preference for urban parks (Olszewska-Guizzo et al.,
2022), wetland landscapes (Dobbie, 2013). This preference may partly explain the general preference
for historical architectural styles (Brielmann et al., 2022; Mouratidis and Hassan, 2020). Furthermore,
it is likely that a clear organization facilitates navigation and wayfinding which may also contribute to
more positive perceptions (Csanády, 2019).
Notably, we found an inverse-U relationship between subjective crowding and beauty. In other
words, environments that are rated as more beautiful are those which have moderate levels of subjective
crowding. In the contrary, empty and overly crowded environments, particularly when combined with
a narrowed and unstructured features are considered less beautiful. Past research has also linked
extremely crowded tourist sites, environments as having a negative effect on the perceived attractiveness
of immediate environment (Yin et al., 2020). Furthermore, environments with too few or too many
people may evoke a lack of safety or control in an environment (Altman, 1975), and thus perceived
as less aesthetically appealing. Moreover, for the same level of subjective crowding, green spaces were
considered more beautiful. This is inline other findings that the presence of trees on urban streets
appeared to moderate the perception of pleasantness and friendliness (e.g., Basu et al., 2022). More
research is needed to tease apart these relationships and the potential implications of designing high-
density urban environments which are also aesthetically pleasing.
5.3. Determinants of psychological reaction to places
Based on previous literature (e.g. Cheng et al., 2021; Eroglu et al., 2005; Mehta et al., 2013;
Wohlwill, 1974; Altman, 1975)), we hypothesised the effect of pedestrian activity on affective experi-
ence to have an inverse U-shaped relationship. Supporting this, our results show that environmental
factors, including crowding and aesthetic qualities, influence affective experiences. We found a complex
relationship between crowding and valence. For less beautiful green or urban spaces, there appears to
be an inverse-U relationship, whereas beautiful environments were related to positive emotions regard-
less of crowding. Valence ratings were correlated with willingness to walk, which is in line with studies
showing that people avoid unpleasant routes (Johansson et al., 2016). Urban environments were con-
sistently rated higher in terms of arousal, i.e. as more exciting, whereas green environments were rated
as more calm. In contrast, for green spaces, arousal was uncorrelated from valence, in line with other
recent findings Gregorians et al. (2022) which suggests a dissociation between how exciting or calm
an environment is, and the pleasantness of spatial experience. In contrast, for urban spaces there was
a correlation between arousal and valence – exciting spaces evoke more positive emotions. Further,
for urban environments, we observed a positive relationship between subjective crowding and arousal.
This is inline with Wohlwill’s earlier arousal theory who postulated that arousal increases along with
crowding density, and can even lead to hyper-arousal Wohlwill (1974). However, this effect was not
19

observed for green environments, which suggests potentially different psychological mechanisms behind
the evaluation of green and urban environments in terms of crowding.
Overall, in terms of restorativeness, for similar levels of subjective crowding, green spaces were rated
as more restorative, which is in line with previous studies (Berman et al., 2008; Kaplan, 1995; Ulrich
et al., 1991). However, when taking into account the effects of beauty, we noted that urban spaces
that are rated high in terms of beauty achieve comparable levels of restorativeness ratings, which
supports earlier findings by Meidenbauer et al. (2020). For beautiful environments, restorativeness
declined steadily with increasing levels of subjective crowding. By contrast, an inverted-U effect was
observed for environments rated lower in terms of beauty; these were rated most restorative when they
were depicting moderate levels of crowding, and this effect was stronger for urban scenes. This finding
suggests that human co-presence, and crowding, contributes to some facets of the psychological process
of restoration, which is inline with the CLM where sense of solitude moderates restoration (Olszewska-
Guizzo et al., 2022). Given that in this study we only a single item of the perceived restorativeness
scale (’being away’), more research is needed to clarify the role of subjective crowding and aesthetics
have on restoration and typology.
Initially included as a manipulation check, participants reported a moderate sense of presence, as
would be expected from a desktop study, However, we also found that moderate levels of crowding
increased the sense of presence, which is in line with previous studies that observed a positive effect
of co-presence on sense of presence, for example in virtual places (Nowak and Biocca, 2003). Notably,
this effect was more pronounced for less aesthetically pleasing environments, suggesting that people
feel more drawn and immersed while watching beautiful environments, whereas for environments that
are considered less pleasing ugly, increased crowding can enhance presence and immersion.
Finally, participants indicated greater willingness to walk in environments with moderate levels of
subjective crowding, and more importantly, both typology and beauty showed significant interactions
in this relationship. This is in line with several past studies which also found association between an
aesthetically pleasing environment and pedestrian activity (Ball et al., 2001; Inoue et al., 2010; Chan
et al., 2021), and higher preference to walk in urban environments described as having picturesque
scenery, and architecturally beautiful structures (Ferrer et al., 2015; Davies et al., 2012), whereas
individuals are less keen on walking in areas with dilapidated shop-fronts (Middleton, 2009). The results
imply that environment aesthetic can also act in mitigating the aversive effect of higher crowding. In the
light of current efforts to increase walking, this suggests that improving the environmental aesthetics of
walking routes – especially those with higher pedestrian flows – alongside other measures like greening,
can enhance and encourage pedestrian experience.
5.4. Typology
Overall, we found that environmental typology, i.e., green versus urban spaces, played an important
role on emotional reactions and environmental appraisals. Green environments were systematically
rated as more beautiful than urban, which is consistent with past literature(Weinberger et al., 2021;
Marchand et al., 2014; Zhang et al., 2018). Nevertheless, for both typologies we found moderating
effects of aesthetic qualities as well as presence of people. For instance, urban environments considered
as beautiful received emotional or willingness to visit ratings that were highly similar with those of
green spaces. Beautiful urban spaces were considered more pleasant and rated higher for willingness
to walk than green environments of low beauty.
In previous studies, crowding negatively impacted the pleasure derived with being in green recre-
ational spaces (Arnberger, 2012; Arnberger and Eder, 2015). Studies have also shown that context
or frequency of visits can moderate these effects. Arnberger and Eder found that green space visitors
had similar preferences for their environment, except for when they sought stress relief, in which case
visitor density became an important factor in their experience and selection of environment. Another
study by Sharp et al. (2015) showed that regular visitors, compared to first-time or irregular visitors
expect and thus are less negatively impacted by the crowding. In our study, we found an additional
moderating factor. After controlling for levels of familiarity with each scene, green spaces with higher
levels of beauty were less influenced by subjective crowding, compared to green spaces with lower lev-
els of beauty, in terms of evoking positive emotions and a desire to visit. Similarly, for urban spaces,
aesthetic qualities, i.e., higher beauty, appear to mitigate the negative effects of increased crowding.
20

5.5. Individual differences
We hypothesised that personality traits will be associated with inter-individual differences in aes-
thetic judgements of green versus urban spaces, and our results support this hypothesis. We measured
personality traits directly with a custom questionnaire to obtain self-reported preference for crowds, as
well as indirectly by assessing emotional reactions (valence and willingness-to-walk ratings) for crowded
environments, in other words whether they felt more pleasant and wished to visit places with higher
levels of crowding. First of all, we found that direct and indirect measures of preference for crowds
were highly correlated, suggesting that both approaches assess a common underlying preference. We
then compared self-reported preference for crowds with personality traits, and found that self-reported
preference for crowded environments correlated positively with extraversion, and negatively with noise
sensitivity, honesty-humility, and stimulus screening ability. We found no effect of gender, or whether
participants grew up or live in an urban versus non-urban environment. In other words, overall people
are aversive towards crowded place, but more extroverted people enjoy being in more crowded envi-
ronments, whereas people who are less cooperative or they are more sensitive to stimuli and noise tend
to dislike more crowded environments.
5.6. Limitations and future work
This study presents a number of limitations. First of all, exposure to the different environments
was simulated through short 30-second videos viewed on a desktop computer without sound. Thus
participants’ emotional reactions and behaviour (e.g., gaze) may differ from physically walking within
these environments (Foulsham et al., 2011; Lin et al., 2020), and having a complete multi-sensory
experience which includes auditory (including noise pollution, e.g. Zare Sakhvidi et al., 2018) and
olfactory information (Bentley et al., 2022). Second, our stimuli present environments to a certain
degree, but there exist other more extreme cases of pedestrian crowding that we could not find adequate
videos for our testing; future work should examine reactions to higher numbers of pedestrians. Third,
despite our efforts to mitigate and control for differences in the presentation of stimuli (e.g. video
quality) or participant attention (see Methods), due to the nature of an online study, it is still possible
that some of these issues influenced the results. Last, in the present study we did not assess factors that
may modulate and moderate feelings of safety, (e.g. such as graffiti or litter Beenackers et al., 2013;
Kweon et al., 2021). Future studies should address these methodological issues and also investigate
more specific features and aspects of the scenes pedestrians are exposed to for determining their impact
on affect, restoration, etc, in urban and green spaces.
5.7. Implications for design
•Strive for high aesthetic quality for spaces which accommodate large pedestrian crowds
•Prioritise green spaces of high aesthetic value to enhance positive psychological effects
•Use strategies to limit crowding in green spaces to ensure restorative psychological effects
5.8. Conclusion
In conclusion, this study investigated the joint effects of environmental aesthetics, typology and
crowding on the psychological experience of urban and green spaces. A database with 480 dynamic
stimuli (video walkthroughs) with international coverage of 30 cities / 20 countries was established, and
used in a large online study (N = 377) following a screen-based environmental exposure paradigm. Our
results show environmental perceptions, specifically beauty and order, have an attenuating effect on
the subjective experience of crowding. We also found an inverse-U relationship between crowding and
psychological outcomes (positive emotions, arousal, willingness to walk), highlighting that moderate
levels of pedestrian activity can increased the desirability of both urban and green spaces.
These results advance our understanding of person-environment interactions, and contribute to
the ongoing debate about the psychological impact of urban environments. Our main finding that
aesthetics increased the positive emotional experience and desirability of urban spaces, as well as
makes these spaces more restorative has implications not only for environmental psychology but also
for design and planning. Our results also suggest that to a certain extent environmental aesthetics can
21

mitigate subjective crowding. In other words for the same number of people, spaces that are designed
to higher aesthetic standards are associated with improved psychological reactions. This finding has
potential implications for the design of public spaces and pedestrian facilities. For instance, spaces
that are expected to accommodate large pedestrian crowds, aesthetic designs can mitigate some of the
psychological effects. Making aesthetic quality a mandate for the design of spaces, especially those
that accommodate large numbers of pedestrians, may also bring psychological and other benefits.
Future work should investigate subconscious cognitive and emotional processes that take place during
the experience of urban and green spaces, positive and negative effects of pedestrian co-presence for
emotional wellbeing promotion, as well as strive for more ecologically valid experimental designs, for
instance outside the lab.
6. Acknowledgements
We would like to acknowledge the creators of the original videos that were used to generate the stim-
uli (Youtube accounts): @4kseoulwalkingtour400, @Yi1080, @BangkokUnmasked, @ASimpleVideoSG,
@ADTravel, @NordinWalks, @letsgowalkingtour, @elmundotravel7396, @revisualsla, @poptravelorg,
@Bomani2007.
This research was supported by a Intra-CREATE Seed Collaboration Grant (NRF2021-ITS008-
0010) from the National Research Foundation, Prime Minister’s Office, Singapore, under its Campus
for Research Excellence and Technological Enterprise (CREATE) programme.
Appendix A. Stimuli location
Stimuli were sourced from video clips filmed in 20 countries (Austria, Canada, Czech Republic,
France, Germany, Ghana, Greece, Ireland, Italy, Korea, Malaysia, Netherlands, Singapore, Spain,
Sweden, Switzerland, Thailand, Turkey, UK, USA), and more 31 different locations/cities. Specifically:
Amsterdam (18 clips), Bangkok (4), Cuxhaven (8), Dublin (31), Esquimalt (4), Georgetown (15),
Heraklion (3), Istanbul (13), Jurong Lake (2), Kaiserslautern (13), Kakum National Park (2), Lille (23),
Liverpool (18), London (39), Manheim (11), Milan (17), New York (11), Nottingham (9), Palisades
Interstate Park (2), Prague (7), Salem, Oregon (3), Salzburg (10), Santa Monica (10), Seoul (18),
Singapore (65), Stockholm (18), Usti Nam Labem (22), Valencia (14), Vancouver (27), Vienna (27),
Zurich (18).
Appendix B. Additional questionnaires
Appendix B.1. Preference for crowds
To assess the degree to which people might enjoy being in the presence of pedestrian crowds during
different situations, the following three questions were devised, presented in a randomised order, and
rated on a 5 point scale from Strongly disagree to Strongly agree. The score was calculated as the
average of the three responses:
•I like being in crowded places.
•During my leisure time, I enjoy walking downtown among pedestrian crowds.
•During my commute, I like being in a crowded public transport.
Appendix B.2. Concern for Covid-19
To assess the degree to which people might be concerned about contracting Covid-19, the following
four questions were devised, presented in a randomised order, and rated on a 5 point scale from Strongly
disagree to Strongly agree. The score was calculated as the average of the four responses:
•Contracting COVID-19 is something that worries me when I am in a crowd.
•When I am in public spaces, I pay a lot of attention if people are wearing their masks.
•Because of COVID, I avoid going to busy locations like downtown, shop, or malls.
•I think when you are outdoors there is not a lot of risk to contract COVID. [Reversed]
22

References
a. Russell, J. (2009). Emotion, core affect, and psychological construction. Cognition Emotion,
23:1259–1283.
Altman, I. (1975). The environment and social behavior: privacy, personal space, territory, and crowd-
ing. ERIC.
Ameli, S. H., Hamidi, S., Garfinkel-Castro, A., and Ewing, R. (2015). Do better urban design qualities
lead to more walking in salt lake city, utah? Journal of Urban Design, 20(3):393–410.
Arnberger, A. (2012). Urban densification and recreational quality of public urban green spaces—a
viennese case study. Sustainability, 4(4):703–720.
Arnberger, A. and Eder, R. (2015). Are urban visitors’ general preferences for green-spaces similar to
their preferences when seeking stress relief? Urban Forestry & Urban Greening, 14(4):872–882.
Aspinall, P., Mavros, P., Coyne, R., and Roe, J. (2013). The urban brain: analysing outdoor physical
activity with mobile eeg. British journal of sports medicine, 1:1–7.
Ball, K., Bauman, A., Leslie, E., and Owen, N. (2001). Perceived environmental aesthetics and conve-
nience and company are associated with walking for exercise among australian adults. Preventive
medicine, 33(5):434–440.
Basu, N., Oviedo-Trespalacios, O., King, M., Kamruzzaman, M., and Haque, M. M. (2022). The
influence of the built environment on pedestrians’ perceptions of attractiveness, safety and security.
Transportation research part F: traffic psychology and behaviour, 87:203–218.
Beemer, C. J., Stearns-Yoder, K. A., Schuldt, S. J., Kinney, K. A., Lowry, C. A., Postolache, T. T.,
Brenner, L. A., and Hoisington, A. J. (2021). A brief review on the mental health for select
elements of the built environment. Indoor and Built Environment, 30(2):152–165.
Beenackers, M. A., Kamphuis, C. B., Mackenbach, J. P., Burdorf, A., and van Lenthe, F. J. (2013). Why
some walk and others don’t: exploring interactions of perceived safety and social neighborhood
factors with psychosocial cognitions. Health education research, 28(2):220–233.
Beermann, M. and Sieben, A. (2022). Waiting behavior and arousal in different levels of crowd density:
A psychological experiment with a “tiny box”. Journal of Advanced Transportation, 2022.
Benfield, J. A., Nurse, G. A., Jakubowski, R., Gibson, A. W., Taff, B. D., Newman, P., and Bell, P. A.
(2014). Testing Noise in the Field: A Brief Measure of Individual Noise Sensitivity. Environment
and Behavior, 46(3):353–372.
Bentley, P. R., Fisher, J. C., Dallimer, M., Fish, R. D., Austen, G. E., Irvine, K. N., and Davies, Z. G.
(2022). Nature, smells, and human wellbeing. Ambio, pages 1–14.
Bentz, J., Rodrigues, A., Dearden, P., Calado, H., and Lopes, F. (2015). Crowding in marine environ-
ments: Divers and whale watchers in the azores. Ocean & Coastal Management, 109:77–85.
Berman, M. G., Jonides, J., and Kaplan, S. (2008). The cognitive benefits of interacting with nature.
Psychological science, 19(12):1207–1212.
Berto, R. (2005). Exposure to restorative environments helps restore attentional capacity. Journal of
environmental psychology, 25(3):249–259.
Beute, F. and de Kort, Y. A. (2014). Salutogenic effects of the environment: Review of health protective
effects of nature and daylight. Applied psychology: Health and well-being, 6(1):67–95.
Blut, M. and Iyer, G. R. (2020). Consequences of perceived crowding: A meta-analytical perspective.
Journal of Retailing, 96(3):362–382.
23

Bornioli, A., Parkhurst, G., and Morgan, P. L. (2018). Psychological wellbeing benefits of simulated
exposure to five urban settings: an experimental study from the pedestrian’s perspective. Journal
of Transport and Health, 9:105–116.
Brielmann, A. A., Buras, N. H., Salingaros, N. A., and Taylor, R. P. (2022). What happens in your
brain when you walk down the street? implications of architectural proportions, biophilia, and
fractal geometry for urban science. Urban Science, 6(1):3.
Brielmann, A. A. and Pelli, D. G. (2019). Intense beauty requires intense pleasure. Frontiers in
psychology, 10:2420.
Buoli, M., Grassi, S., Caldiroli, A., Carnevali, G. S., Mucci, F., Iodice, S., Cantone, L., Pergoli, L.,
and Bollati, V. (2018). Is there a link between air pollution and mental disorders? Environment
international, 118:154–168.
Buttazzoni, A., Parker, A., and Minaker, L. (2021). Investigating the mental health implications of
urban environments with neuroscientific methods and mobile technologies: A systematic literature
review. Health & Place, 70:102597.
Calhoun, J. B. (1962). Population density and social pathology. Scientific American, 206(2):139–149.
Campagnaro, T., Vecchiato, D., Arnberger, A., Celegato, R., Re, R. D., Rizzetto, R., Semenzato, P.,
Sitzia, T., Tempesta, T., and Cattaneo, D. (2020). General, stress relief and perceived safety
preferences for green spaces in the historic city of padua (italy). Urban Forestry and Urban
Greening, 52.
Cao, Z., Hidalgo Martinez, G., Simon, T., Wei, S., and Sheikh, Y. A. (2019). Openpose: Realtime
multi-person 2d pose estimation using part affinity fields. IEEE Transactions on Pattern Analysis
and Machine Intelligence.
Chan, E. T., Schwanen, T., and Banister, D. (2021). The role of perceived environment, neighbourhood
characteristics, and attitudes in walking behaviour: Evidence from a rapidly developing city in
china. Transportation, 48(1):431–454.
Chang, C.-Y. and Chen, P.-K. (2005). Human response to window views and indoor plants in the
workplace. HortScience, 40(5):1354–1359.
Chatterjee, A. and Vartanian, O. (2014). Neuroaesthetics. Trends in cognitive sciences, 18(7):370–375.
Chen, S., Oliva, P., and Zhang, P. (2018). Air pollution and mental health: evidence from china.
Technical report, National Bureau of Economic Research.
Cheng, H., Liu, Q., and Bi, J.-W. (2021). Perceived crowding and festival experience: The moderating
effect of visitor-to-visitor interaction. Tourism Management Perspectives, 40:100888.
Coburn, A. (2019). Buildings, beauty, and the brain: psychological responses to architectural design.
PhD thesis, University of Cambridge.
Coburn, A., Kardan, O., Kotabe, H., Steinberg, J., Hout, M. C., Robbins, A., MacDonald, J., Hayn-
Leichsenring, G., and Berman, M. G. (2019). Psychological responses to natural patterns in
architecture. Journal of Environmental Psychology, 62:133–145.
Csanády, P. (2019). Architectural space density: The effect of enclosure. Symmetry: Culture and
Science, 30(1):43–58.
d’Astous, A. (2000). Irritating aspects of the shopping environment. Journal of Business Research,
49(2):149–156.
Davies, N. J., Lumsdon, L. M., and Weston, R. (2012). Developing recreational trails: Motivations for
recreational walking. Tourism Planning & Development, 9(1):77–88.
24

de Kort, Y. A., Meijnders, A. L., Sponselee, A. A., and IJsselsteijn, W. A. (2006). What’s wrong
with virtual trees? restoring from stress in a mediated environment. Journal of environmental
psychology, 26(4):309–320.
de Leeuw, J. R. (2015). jspsych: A javascript library for creating behavioral experiments in a web
browser. Behavior Research Methods, 47(1).
den Berg, A. E. V., Joye, Y., and Koole, S. L. (2016). Why viewing nature is more fascinating and
restorative than viewing buildings: A closer look at perceived complexity. Urban Forestry and
Urban Greening, 20:397–401.
Dewey, J. (1934). Art as Experience. Allen & Unwin, London, UK.
Dobbie, M. F. (2013). Public aesthetic preferences to inform sustainable wetland management in
victoria, australia. Landscape and Urban Planning, 120:178–189.
Elbachir, S. and Chenini, A. (2017). How crowding influences emotional, perceptual and behavioural
reactions in store. Economics and Management.
Engelniederhammer, A., Papastefanou, G., and Xiang, L. (2019). Crowding density in urban environ-
ment and its effects on emotional responding of pedestrians: Using wearable device technology
with sensors capturing proximity and psychophysiological emotion responses while walking in the
street. Journal of Human Behavior in the Social Environment, 29:630–646.
Eroglu, S. A., Machleit, K., and Barr, T. F. (2005). Perceived retail crowding and shopping satisfaction:
the role of shopping values. Journal of business research, 58(8):1146–1153.
Ewing, R. and Handy, S. (2009). Measuring the unmeasurable: Urban design qualities related to
walkability. Journal of Urban design, 14(1):65–84.
Faggi, A., Breuste, J., Madanes, N., Gropper, C., and Perelman, P. (2013). Water as an appreciated
feature in the landscape: a comparison of residents’ and visitors’ preferences in buenos aires.
Journal of Cleaner Production, 60:182–187.
Ferrer, S., Ruiz, T., and Mars, L. (2015). A qualitative study on the role of the built environment for
short walking trips. Transportation research part F: traffic psychology and behaviour, 33:141–160.
Foulsham, T., Walker, E., and Kingstone, A. (2011). The where, what and when of gaze allocation in
the lab and the natural environment. Vision research, 51(17):1920–1931.
Galle, O. R., Gove, W. R., and McPherson, J. M. (1972). Population density and pathology: What
are the relations for man? evidence from one city suggests that high population density may be
linked to" pathological" behavior. Science, 176(4030):23–30.
Gehl, J. (2013). Cities for people. Island press.
Grassini, S. (2022). Virtual reality assisted non-pharmacological treatments in chronic pain manage-
ment: a systematic review and quantitative meta-analysis. International Journal of Environmental
Research and Public Health, 19(7):4071.
Gregorians, L., Velasco, P. F., Zisch, F., and Spiers, H. J. (2022). Architectural experience: clarifying
its central components and their relation to core affect with a set of first-person-view videos.
bioRxiv.
Guzman, V., Garrido-Cumbrera, M., Braçe, O., Hewlett, D., and Foley, R. (2021). Associations
of the natural and built environment with mental health and wellbeing during covid-19: Irish
perspectives from the greencovid study. The Lancet Global Health, 9:S20.
25

Hagerhall, C., Laike, T., Kuller, M., Marcheschi, E., Boydston, C., and Taylor, R. (2015). Human
physiological benefits of viewing nature: Eeg responses to exact and statistical fractal patterns.
Nonlinear dynamics, psychology, and life sciences, 19(1):1–12.
Hagerhall, C. M., Laike, T., Taylor, R. P., Küller, M., Küller, R., and Martin, T. P. (2008). Investiga-
tions of human eeg response to viewing fractal patterns. Perception, 37(10):1488–1494.
Hands, D. E. and Brown, R. D. (2002). Enhancing visual preference of ecological rehabilitation sites.
Landscape and Urban Planning, 58(1):57–70.
Hartig, T., Evans, G. W., Jamner, L. D., Davis, D. S., and Gärling, T. (2003). Tracking restoration
in natural and urban field settings. Journal of environmental psychology, 23(2):109–123.
Hoisington, A. J., Stearns-Yoder, K. A., Schuldt, S. J., Beemer, C. J., Maestre, J. P., Kinney, K. A.,
Postolache, T. T., Lowry, C. A., and Brenner, L. A. (2019). Ten questions concerning the built
environment and mental health. Building and environment, 155:58–69.
Honey-Rosés, J., Anguelovski, I., Chireh, V. K., Daher, C., van den Bosch, C. K., Litt, J. S., Mawani,
V., McCall, M. K., Orellana, A., Oscilowicz, E., Sánchez, U., Senbel, M., Tan, X., Villagomez,
E., Zapata, O., and Nieuwenhuijsen, M. J. (2021). The impact of covid-19 on public space: an
early review of the emerging questions–design, perceptions and inequities. Cities and Health,
5:S263–S279.
Hui, M. K. and Bateson, J. E. (1991). Perceived control and the effects of crowding and consumer
choice on the service experience. Journal of consumer research, 18(2):174–184.
Inoue, S., Ohya, Y., Odagiri, Y., Takamiya, T., Ishii, K., Kitabayashi, M., Suijo, K., Sallis, J. F., and
Shimomitsu, T. (2010). Association between perceived neighborhood environment and walking
among adults in 4 cities in japan. Journal of epidemiology, 20(4):277–286.
Jalil, N. A., Yunus, R. M., and Said, N. S. (2012). Environmental colour impact upon human behaviour:
A review. Procedia-Social and Behavioral Sciences, 35:54–62.
Johansson, M., Hartig, T., and Staats, H. (2011). Psychological benefits of walking: Moderation by
company and outdoor environment. Applied psychology: health and well-being, 3(3):261–280.
Johansson, M., Sternudd, C., and Kärrholm, M. (2016). Perceived urban design qualities and affective
experiences of walking. Journal of Urban Design, 21(2):256–275.
Kaplan, S. (1995). The restorative benefits of nature: Toward an integrative framework. Journal of
environmental psychology, 15(3):169–182.
Karmanov, D. and Hamel, R. (2008). Assessing the restorative potential of contemporary urban
environment(s): Beyond the nature versus urban dichotomy. Landscape and Urban Planning,
86:115–125.
Kellert, S. R. and Wilson, E. O. (1995). The biophilia hypothesis. Island press.
Kort, Y. A. D., Ijsselsteijn, W. A., Kooijman, J., and Schuurmans, Y. (2003). Virtual laboratories:
Comparability of real and virtual environments for environmental psychology. Presence: Teleop-
erators and Virtual Environments, 12:360–373.
Kweon, B.-S., Rosenblatt-Naderi, J., Ellis, C. D., Shin, W.-H., and Danies, B. H. (2021). The effects of
pedestrian environments on walking behaviors and perception of pedestrian safety. Sustainability,
13(16):8728.
Kyle, G., Landon, A., and Schuett, M. (2022). Crowding, coping and place attachment in nature.
Current Psychology.
26

Levi, D. M. (2008). Crowding-an essential bottleneck for object recognition: A mini-review. Vision
Research, 48:635–654.
Li, H., Thrash, T., Hölscher, C., and Schinazi, V. R. (2019). The effect of crowdedness on human
wayfinding and locomotion in a multi-level virtual shopping mall. Journal of Environmental
Psychology, 65.
Lin, W., Chen, Q., Jiang, M., Tao, J., Liu, Z., Zhang, X., Wu, L., Xu, S., Kang, Y., and Zeng,
Q. (2020). Sitting or walking? analyzing the neural emotional indicators of urban green space
behavior with mobile eeg. Journal of Urban Health, 97(2):191–203.
Lombardi, D. B. and Ciceri, M. R. (2021). Dealing with feeling crowded on public transport: The
potential role of design. Environment and Behavior, 53:339–378.
Maas, J., Van Dillen, S. M., Verheij, R. A., and Groenewegen, P. P. (2009). Social contacts as a possible
mechanism behind the relation between green space and health. Health & place, 15(2):586–595.
Machleit, K. A., Eroglu, S. A., and Mantel, S. P. (2000). Perceived retail crowding and shopping
satisfaction: what modifies this relationship? Journal of consumer psychology, 9(1):29–42.
Machleit, K. A., Kellaris, J. J., and Eroglu, S. A. (1994). Human versus spatial dimensions of crowding
perceptions in retail environments: A note on their measurement and effect on shopper satisfaction.
Marketing Letters, 5(2):183–194.
Maher, A. and von Hippel, C. (2005). Individual differences in employee reactions to open-plan offices.
Journal of Environmental Psychology, 25:219–229.
Makowski, D., Ben-Shachar, M., and Lüdecke, D. (2020). The easystats collection of r packages.
Github.
Mangrio, E. and Zdravkovic, S. (2018). Crowded living and its association with mental ill-health among
recently-arrived migrants in sweden: a quantitative study. BMC research notes, 11(1):1–5.
Marchand, G. C., Nardi, N. M., Reynolds, D., and Pamoukov, S. (2014). The impact of the classroom
built environment on student perceptions and learning. Journal of Environmental Psychology,
40:187–197.
Mavros, P., Wälti, M. J., Nazemi, M., Ong, C. H., and Hölscher, C. (2022). A mobile eeg study on the
psychophysiological effects of walking and crowding in indoor and outdoor urban environments.
Scientific Reports, 12.
Mazumder, R., Spiers, H. J., and Ellard, C. G. (2020). Exposure to high-rise buildings negatively
influences affect: evidence from real world and 360-degree video. Cities & Health, pages 1–13.
Mehrabian, A. (1977). A questionnaire measure of individual differences in stimulus screening and
associated differences in arousability. Environmental Psychology and Nonverbal Behavior, 1(2):89–
103.
Mehta, R., Sharma, N. K., and Swami, S. (2013). The impact of perceived crowding on consumers’
store patronage intentions: Role of optimal stimulation level and shopping motivation. Journal of
Marketing Management, 29(7-8):812–835.
Meidenbauer, K. L., Stenfors, C. U., Bratman, G. N., Gross, J. J., Schertz, K. E., Choe, K. W., and
Berman, M. G. (2020). The affective benefits of nature exposure: What’s nature got to do with
it? Journal of Environmental Psychology, 72:101498.
Middleton, J. (2009). ‘stepping in time’: walking, time, and space in the city. Environment and
planning A, 41(8):1943–1961.
27

Moll, A., Collado, S., Staats, H., and Corraliza, J. A. (2022). Restorative effects of exposure to nature
on children and adolescents: A systematic review. Journal of Environmental Psychology, 84.
Mouratidis, K. and Hassan, R. (2020). Contemporary versus traditional styles in architecture and
public space: A virtual reality study with 360-degree videos. Cities, 97:102499.
Neale, C., Lopez, S., and Roe, J. (2021). Psychological restoration and the effect of people in nature
and urban scenes : A laboratory experiment. Sustainability, 13.
Needham, M. D., Szuster, B. W., Lesar, L., Mora, C., and Knecht, D. P. (2018). Snorkeling and scuba
diving with manta rays: encounters, norms, crowding, satisfaction, and displacement. Human
Dimensions of Wildlife, 23(5):461–473.
Norwood, M. F., Lakhani, A., Maujean, A., Zeeman, H., Creux, O., and Kendall, E. (2019). Brain
activity, underlying mood and the environment: A systematic review. Journal of Environmental
Psychology, 65:101321.
Nouri, P. (2011). Desirable pedestrian density. Master’s thesis.
Nowak, K. L. and Biocca, F. (2003). The effect of the agency and anthropomorphism on users’ sense
of telepresence, copresence, and social presence in virtual environments. Presence: Teleoperators
and Virtual Environments, 12:481–494.
Oldham, G. R. (1988). Effects of changes in workspace partitions and spatial density on employee
reactions: A quasi-experiment. Journal of applied psychology, 73(2):253.
Olszewska-Guizzo, A. (2023). Neuroscience for Designing Green Spaces: Contemplative Landscapes
(1st ed.). Routledge.
Olszewska-Guizzo, A., Escoffier, N., Chan, J., and Puay Yok, T. (2018). Window view and the brain:
Effects of floor level and green cover on the alpha and beta rhythms in a passive exposure eeg
experiment. International Journal of Environmental Research and Public Health, 15(11):2358.
Olszewska-Guizzo, A., Mukoyama, A., Naganawa, S., Dan, I., Husain, S. F., Ho, C. S., and Ho,
R. (2021). Hemodynamic response to three types of urban spaces before and after lockdown
during the covid-19 pandemic. International Journal of Environmental Research and Public Health,
18(11):6118.
Olszewska-Guizzo, A., Sia, A., Fogel, A., and Ho, R. (2022). Features of urban green spaces associated
with positive emotions, mindfulness and relaxation. Scientific Reports, 12(1).
Palan, S. and Schitter, C. (2018). Prolific. ac—a subject pool for online experiments. Journal of
Behavioral and Experimental Finance, 17:22–27.
Parsons, R. and Daniel, T. C. (2002). Good looking: in defense of scenic landscape aesthetics. Landscape
and Urban Planning, 60(1):43–56.
Patuano, A. (2017). The fractal dimensions of landscape images as predictors of landscape preference.
PhD thesis.
Peer, E., Rothschild, D., Gordon, A., Evernden, Z., and Damer, E. (2022). Data quality of platforms
and panels for online behavioral research. Behavior Research Methods, 54:1643–1662.
Peters, L., Sunderland, M., Andrews, G., Rapee, R. M., and Mattick, R. P. (2012). Development of a
short form Social Interaction Anxiety (SIAS) and Social Phobia Scale (SPS) using nonparametric
item response theory: The SIAS-6 and the SPS-6. Psychological Assessment, 24(1):66–76.
Presti, P., Ruzzon, D., Avanzini, P., Caruana, F., Rizzolatti, G., and Vecchiato, G. (2022). Measur-
ing arousal and valence generated by the dynamic experience of architectural forms in virtual
environments. Scientific reports, 12(1):1–12.
28

R Core Team (2022). R: A Language and Environment for Statistical Computing. R Foundation for
Statistical Computing, Vienna, Austria.
Regoeczi, W. C. (2008). Crowding in context: An examination of the differential responses of men and
women to high-density living environments. Journal of health and social behavior, 49(3):254–268.
Ribeiro, A. I., Pires, A., Carvalho, M. S., and Pina, M. F. (2015). Distance to parks and non-residential
destinations influences physical activity of older people, but crime doesn’t: a cross-sectional study
in a southern european city. BMC public health, 15(1):1–12.
Saarloos, D., Alfonso, H., Giles-Corti, B., Middleton, N., and Almeida, O. P. (2011). The built
environment and depression in later life: the health in men study. The American Journal of
Geriatric Psychiatry, 19(5):461–470.
Saegert, S. (1973). Crowding: Cognitive overload and behavioral constraint. Environmental design
research, pages 254–260.
Satariano, B. (2021). Recreating a therapeutic blue urban space through the architectural restoration
of the triton fountain in valletta, malta. Cities & Health, pages 1–12.
Schneider, A., Vollenwyder, B., Krueger, E., Miller, D., Thurau, J., and Elfering, A. (2023). Mobile
eye tracking applied as a tool for customer experience research in a crowded train station. Journal
of Eye Movement Research, 16.
Sevtsuk, A. and Basu, R. (2022). The role of turns in pedestrian route choice: a clarification. Journal
of transport geography, 102:103392.
Sharam, L., Mayer, K., and Baumann, O. (2023). Design by nature: The influence of windows on
cognitive performance and affect. Journal of Environmental Psychology, 85:101923.
Sharp, R. L., Sharp, J. A., and Miller, C. A. (2015). An island in a sea of development: An examination
of place attachment, activity type, and crowding in an urban national park. Visitor Studies,
18(2):196–213.
Shelby, B., Vaske, J. J., and Heberlein, T. A. (1989). Comparative analysis of crowding in multiple
locations: Results from fifteen years of research. Leisure sciences, 11(4):269–291.
Sibley, C. G., Luyten, N., Purnomo, M., Mobberley, A., Wootton, L. W., Hammond, M. D., Sengupta,
N., Perry, R., West-Newman, T., Wilson, M. S., et al. (2011). The mini-ipip6: Validation and
extension of a short measure of the big-six factors of personality in new zealand. New Zealand
Journal of Psychology, 40(3).
Stamps, A. E. (2010). Effects of permeability on perceived enclosure and spaciousness. Environment
and Behavior, 42:864–886.
Stokols, D. (1972). On the distinction between density and crowding: some implications for future
research. Psychological review, 79(3):275.
Sundstrom, E. and Altman, I. (1976). Interpersonal relationships and personal space: Research review
and theoretical model. Human Ecology, 4(1):47–67.
Sygna, K., Aasvang, G. M., Aamodt, G., Oftedal, B., and Krog, N. H. (2014). Road traffic noise, sleep
and mental health. Environmental research, 131:17–24.
Tirachini, A., Hensher, D. A., and Rose, J. M. (2013). Crowding in public transport systems: effects
on users, operation and implications for the estimation of demand. Transportation research part
A: policy and practice, 53:36–52.
Ulrich, R. S. (1984). View through a window may influence recovery from surgery. Science (New York,
N.Y.), 224:420–1.
29

Ulrich, R. S., Simons, R. F., Losito, B. D., Fiorito, E., Miles, M. A., and Zelson, M. (1991). Stress
recovery during exposure to natural and urban environments. Journal of environmental psychology,
11(3):201–230.
Vartanian, O., Navarrete, G., Chatterjee, A., Fich, L. B., Gonzalez-Mora, J. L., Leder, H., Modroño,
C., Nadal, M., Rostrup, N., and Skov, M. (2015). Architectural design and the brain: Effects
of ceiling height and perceived enclosure on beauty judgments and approach-avoidance decisions.
Journal of environmental psychology, 41:10–18.
Vartanian, O., Navarrete, G., Chatterjee, A., Fich, L. B., Leder, H., Modroño, C., Nadal, M., Rostrup,
N., and Skov, M. (2013). Impact of contour on aesthetic judgments and approach-avoidance deci-
sions in architecture. Proceedings of the National Academy of Sciences, 110(supplement_2):10446–
10453.
Velarde, M. D., Fry, G., and Tveit, M. (2007). Health effects of viewing landscapes–landscape types in
environmental psychology. Urban forestry & urban greening, 6(4):199–212.
Voelker, S. and Kistemann, T. (2013). Reprint of:“i’m always entirely happy when i’m here!” urban
blue enhancing human health and well-being in cologne and düsseldorf, germany. Social science
& medicine, 91:141–152.
Wang, X. and Liu, T. (2022). Home-made blues: Residential crowding and mental health in beijing,
china. Urban Studies, page 00420980221101707.
Watson, D., Clark, L. A., and a Tellegen (1988). Development and validation of brief measures
of positive and negative affect: the panas scales. Journal of personality and social psychology,
54:1063–70.
Wei, Z., Jiejing, W., and Bo, Q. (2021). Quantity or quality? exploring the association between public
open space and mental health in urban china. Landscape and Urban Planning, 213:104128.
Weinberger, A. B., Christensen, A. P., Coburn, A., and Chatterjee, A. (2021). Psychological responses
to buildings and natural landscapes. Journal of Environmental Psychology, 77:101676.
White, E. V. and Gatersleben, B. (2011). Greenery on residential buildings: Does it affect preferences
and perceptions of beauty? Journal of environmental psychology, 31(1):89–98.
Whyte, W. H. (1980). The Social Life of Small Urban Spaces. Project for Public Spaces.
Wickham, H., Averick, M., Bryan, J., Chang, W., McGowan, L., François, R., Grolemund, G., Hayes,
A., Henry, L., Hester, J., Kuhn, M., Pedersen, T., Miller, E., Bache, S., Müller, K., Ooms, J.,
Robinson, D., Seidel, D., Spinu, V., Takahashi, K., Vaughan, D., Wilke, C., Woo, K., and Yutani,
H. (2019). Welcome to the tidyverse. Journal of Open Source Software, 4:1686.
Wicks, C., Barton, J., Orbell, S., and Andrews, L. (2022). Psychological benefits of outdoor phys-
ical activity in natural versus urban environments: A systematic review and meta-analysis of
experimental studies. Applied Psychology: Health and Well-Being, 14:1037–1061.
Wohlwill, J. F. (1974). Human adaptation to levels of environmental stimulation. Human Ecology,
2(2):127–147.
Yao, W., Chen, F., Wang, S., and Zhang, X. (2021). Impact of exposure to natural and built environ-
ments on positive and negative affect: a systematic review and meta-analysis. Frontiers in Public
Health, 9.
Yin, J., Cheng, Y., Bi, Y., and Ni, Y. (2020). Tourists perceived crowding and destination attrac-
tiveness: The moderating effects of perceived risk and experience quality. Journal of Destination
Marketing & Management, 18:100489.
30

Yusufzyanova, D. N. (2020). The role of aesthetic quality in the preference and use of green open spaces:
a case study on Grandview and Victoria parks in Vancouver, Canada. PhD thesis, University of
British Columbia.
Zapata, O. and Honey-Rosés, J. (2022). The behavioral response to increased pedestrian and staying
activity in public space: A field experiment. Environment and Behavior, 54:36–57.
Zare Sakhvidi, F., Zare Sakhvidi, M. J., Mehrparvar, A. H., and Dzhambov, A. M. (2018). Environmen-
tal noise exposure and neurodevelopmental and mental health problems in children: a systematic
review. Current environmental health reports, 5(3):365–374.
Zhang, W., Tang, X., He, X., and Chen, G. (2018). Evolutionary effect on the embodied beauty of
landscape architectures. Evolutionary Psychology, 16(1):1474704917749742.
31