Place attachment and lived values in Western Australian communities

Abstract

Substantial research exists on attachments people have to places and the phenomena and objects they value. However, insights on how values vary between different locations and across demographics and how place attachment differs between rural and urban areas are more limited. These understandings are needed to design meaningful adaptation strategies for people and communities at risk from climate change. This study examines attachment to place and things people value in eight communities in Western Australia, using a survey with 403 participants. Results showed that residents across the rural communities shared similar values, but that the values of urban communities were differentiated socioeconomically. Contrary to our hypothesis, place attachment was not stronger among the rural compared to the urban sites. The findings point to the importance of incorporating place-based, lived values and needs, particularly from less affluent residents, into inclusive adaptation planning.

Full text

Place attachment and lived values in Western Australian communities
A. Kelly
a,b,*
, P. Tschakert
c
, C. Lawrence
d
, P. Horwitz
e
, C. Bourgault
f
, N. Ellis
1
a
School of Geography, Earth and Atmospheric Sciences, The University of Melbourne, Australia
b
Melbourne Climate Futures, Australia
c
School of Media, Creative Arts and Social Inquiry, Curtin University, Australia
d
School of Psychological Science, The University of Western Australia, Australia
e
Centre for People, Place and Planet, Edith Cowan University, Australia
f
School of Humanities, The University of Western Australia, Australia
ARTICLE INFO
Keywords:
Sense of place
Rural communities
Urban communities
Climate change adaptation
ABSTRACT
Substantial research exists on attachments people have to places and the phenomena and objects they value.
However, insights on how values vary between different locations and across demographics and how place
attachment differs between rural and urban areas are more limited. These understandings are needed to design
meaningful adaptation strategies for people and communities at risk from climate change. This study examines
attachment to place and things people value in eight communities in Western Australia, using a survey with 403
participants. Results showed that residents across the rural communities shared similar values, but that the
values of urban communities were differentiated socioeconomically. Contrary to our hypothesis, place attach-
ment was not stronger among the rural compared to the urban sites. The ndings point to the importance of
incorporating place-based, lived values and needs, particularly from less afuent residents, into inclusive
adaptation planning.
1. Introduction
The role of place-based values in climate change adaptation.
Strategies designed to reduce exposure to climatic hazards and to
lessen the vulnerability of populations, sectors, and systems often start
with an assessment of the threat of harm from changing climatic con-
ditions and extreme events. Less attention has been given to what people
at risk value as an entry point to adaptation thinking and planning
(Henrique & Tschakert, 2023). Moreover, place values are often over-
looked in climate-related responses such as designing local mitigation
strategies and enhancing resilience or resourcefulness. Yet, as Karlsson
and Hovelsrud (2015) argued, understanding what people value is
fundamental for identifying and deliberating the various trade-offs that
communities and individuals make when undertaking adaptive actions.
Indeed, what people value in the places they live in ought to be the
starting point for such projects and, as recommended by Barnett et al.
(2016), the cornerstone of a socially engaged science of loss in the
context of climate change.
People tend to be strongly attached to the places in which they live
(e.g., Devine-Wright, 2013; Lewicka, 2011). Such attachment often
inuences people’s actions, for example when making decisions about
migrating when facing threats (Farbotko & McMichael, 2019). People
value a variety of phenomena and objects in their local area and homes,
ranging from facilities and services to lifestyle and local histories
(Graham et al., 2018). Yet, such place values are often inadequately
addressed in land-use decision making (Brown et al., 2020; Brown &
Fagerholm, 2015) and are viewed as ‘intangibles’ in discussions of
climate-related loss and damage (Tschakert et al., 2019).
The main aim of this study is to investigate how place attachment
and ‘lived values’ vary between different locations and across de-
mographic groups in Western Australia. Our focus on Western Australia
is deliberate, avoiding the lure of ‘charismatic but remote places’ such as
the Arctic or Pacic atolls (Devine-Wright, 2013). The results presented
here involve eight communities along a ~400 km transect, from the
capital city, Perth, eastward, with differential experiences of res,
droughts, excessive heat, and river and coastal erosion. By investigating
how values vary across these urban and rural communities, as well as
associated socioeconomic levels, the study aims to identify which vari-
ables are signicant predictors of place attachment. Our premise is that
understanding spatial and demographic variation in what people value
* Corresponding author. School of Geography, Earth and Atmospheric Sciences, The University of Melbourne, Melbourne, Australia.
E-mail address: adkelly@student.unimelb.edu.au (A. Kelly).
1
No Afliation.
Contents lists available at ScienceDirect
Applied Geography
journal homepage: www.elsevier.com/locate/apgeog
https://doi.org/10.1016/j.apgeog.2024.103424
Received 13 March 2024; Received in revised form 30 July 2024; Accepted 14 September 2024
Applied Geography 172 (2024) 103424
Available online 1 October 2024
0143-6228/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (
http://creativecommons.org/licenses/by/4.0/ ).

in the places they call home is vital to designing adaptation pathways to
climate change and other stressors and disturbances.
1.1. Overview of place values
1.1.1. Values
The extent to which values vary depending on urban or rural context
has been partially explored (e.g., Crameri, 2014; Hinds & Sparks, 2008),
yet is becoming even more relevant given the uneven boundaries be-
tween urban, peri-urban, and rural areas. Across these elds of inquiry,
values are recognised as a signicant factor in determining interactions
with place, but there is a lack of coherence between different
approaches.
Graham et al. (2013), having reviewed the concept of values across
the elds of climate change adaptation, human geography, social impact
assessment, and decision analysis, found that most research focuses on
the objects or phenomena to which people assign value. Hence, they
proposed the concept of ‘lived values’ instead, dened as “valuations
that individuals make, in isolation or as part of a group, about what is
important in their lives and the places they live. These valuations may be
articulated verbally or expressed through everyday activities” (p.15).
This conceptualisation highlights social values that may be affected by
social or environmental changes. The advantage of a lived values lens is
that it captures what an individual actually values in practice rather
than the motivational cognitive domains investigated in psychology that
are hypothesised to drive behaviour. Graham et al. (2013) offered a
non-exhaustive list of 33 lived values under ve categories of health,
safety, belonging, esteem, and self-actualisation.
More recently, Graham et al. (2018) suggested that the large number
and diversity of lived values across communities may act as a barrier for
climate change adaptation. Despite strong agreement on the most
important lived values across the marginal rural communities in Victo-
ria, Australia, that they studied (slow pace of life, peacefulness, prox-
imity to water, and the natural environment), differences in some values
(e.g. housing affordability, being respected, and business prospects)
became apparent when viewed across population clusters (e.g.
socially-engaged circumstantial sea-changers versus regional retirees).
Understanding such divergences in lived values, as the authors suggest,
is essential for evaluating and negotiating the suitability of options for
fair climate change adaptation and shared governance. Overall, the
concept of ‘lived values’ provides a practical framework for making
sense of the ways in which values inform interactions with place.
1.1.2. Place
Place generally refers to local environments given meaning by the
people and communities that form bonds with them (Hess et al., 2008).
For many, the archetypical place is the home, typically a relatively small
place of residence, seen as “a symbol of continuity and order, rooted-
ness, self-identity, attachment, privacy, comfort, security and refuge”
(Lewicka, 2011, p. 211). Home and place are often conceptualised
differently among indigenous peoples, and can for example incorporate
the implied familial obligation to care for many-species relations
including the ground (Wooltorton et al., 2017). In non-Indigenous
contexts, such as in settler farming communities in Australia, many
‘globally-engaged’ family farmers constantly renegotiate relationships
with place, farming, and the farm business to form complex and
ever-shifting emotional attachments to the farm in the context of
neoliberal governance regimes (Cheshire et al., 2013).
1.1.3. Place attachment and sense of place
While there is no uniform and cross-disciplinarily accepted denition
of place attachment in the literature, a working consensus sees place
attachment as “an affective bond or link between people and specic
places” (Hidalgo & Hern´
andez, 2001, p.274). Such an affective bond
tends to be treated as a multidimensional construct, although with little
agreement on what the dimensions are (Anton & Lawrence, 2014). In
practice, studies on place attachment have tended to differentiate be-
tween functional or physical attachment, and emotional or symbolic
attachment, often distinguished as place dependence and place identity,
respectively (e.g. Anton, 2015).
Another important place concept in geography is that of ‘sense of
place’, which is seen as the meaning imbued in a place by people and the
attachments they have with it (Tuan, 1977). Other authors have viewed
it as an emotional bond with place (Brown & Raymond, 2007). While
due to disciplinary divides and inconsistent denitions, sense of place is
sometimes used interchangeably with place attachment (Lin & Lock-
wood, 2014; Williams & Vaske, 2003), but it has also commonly been
conceptualised as encompassing all of place attachment, identity and
dependence while also incorporating place meanings (Jorgensen &
Stedman, 2001; Wartmann & Purves, 2018). Sense of place has been a
signicant motivator of environmental action and has been shown to
shape adaptive behaviour (Adger et al., 2013; Masterson et al., 2019;
Murphy et al., 2019). Enqvist et al. (2019) demonstrates the analytical
potential of sense of place through their nding that communities and
individuals with similar levels of place attachment but differing place
meanings can exhibit different approaches and preferences to environ-
mental stewardship.
Demographic variables such as age or gender are useful predictors of
place attachment (Lewicka, 2011). The most consistent predictor of
place attachment in the literature is home ownership (e.g., Brown et al.,
2003; Mesch & Manor, 1998). Place attachment tends to strengthen the
longer a person resides in a place (Hess et al., 2008), although Scannell
and Gifford (2010) found that only the social dimension of place
attachment correlated to the length of residence.
By the same token, being older generally correlates with higher place
attachment (e.g., Lewicka, 2010), but not in all cases (e.g., Anton &
Lawrence, 2014). Other potential predictors of place attachment include
socioeconomic status and education levels (Lewicka, 2011).
Empirical research, particularly in psychology, has tended to neglect
the role the physical environment in the study of people-place bonds and
as predictors of place attachment (Lewicka, 2011). However, within
geography and related disciplines the mapping of place attachment has
been extensively researched in recent years (e.g Brown et al., 2015;
Maguire & Klinkenberg, 2018). Research has shown the mapping of
‘landscape values’ (or ‘place values’) to provide a good, albeit not
identical, spatial proxy for mapping place attachment within a given
region (see Brown et al., 2015; Brown & Raymond, 2007). Such land-
scape value mapping approaches are argued to provide a way of
reconciling what is considered important by individuals (‘held values’)
and what is meaningful and important to the individual in a given
landscape (‘assigned values’) to give a geographically-explicit under-
standing of what is valued in a landscape and where such values occur
(see Brown & Donovan, 2014). This has allowed for detailed in-
vestigations into the compatibility of residents’ place values and pro-
posed land-use changes (e.g., Brown & Reed, 2012), as well as
associations between landscape values and perceived climate change
risks (Raymond & Brown, 2011). As with place attachment, sense of
place has also been used as a management tool to engage communities.
For example, Li et al. (2023) found that adaptive co-management that
incorporates a focus on sense of place is necessary in the Wulingyuan
World Heritage Site in China. However, sense of place has been seen as
an important cultural ecosystem service in itself and not just a tool to aid
in management (Ryeld et al., 2019; Wartmann & Purves, 2018).
Gajardo et al. (2023) found that sense of place was the most valued
cultural ecosystem service in a recent study in the Philippines.
Place attachment is generally seen as benecial, but it is not without
potential drawbacks. It is tied to higher awareness of environmental
risks but lower understandings of social risks (Quinn et al., 2019). It can
be correlated with mistrust of governance but also to increased resil-
ience and questioning of poor governance (Clarke & Murphy, 2023).
Bonaiuto et al. (2016) concluded that those who are strongly attached to
place: perceive natural environmental risks but underestimate their
A. Kelly et al.
Applied Geography 172 (2024) 103424
2

likely effects; are unwilling to relocate when confronted by natural di-
sasters; and are more likely to return following disasters. Strong place
attachment and the need to keep alive cultural and spiritual values, as
shown particularly for older Pacic Island populations, explains why
relocation options are rejected, despite the threat of rising sea levels
(Farbotko & McMichael, 2019).
1.2. Knowledge gaps and contribution
There has been extensive investigation into how values vary between
different cultures (e.g. Gardner et al., 1999; Peng et al., 1997). However,
very few papers have investigated how values, and especially lived
values, vary across demographics and different locations within a single
region (Graham et al., 2014). Studies on lived values have so far only
focused on coastal communities (e.g., Graham et al., 2018; Kreller,
2020; Ramm et al., 2017). There is only limited research on the differ-
ences in place attachment between both urban and rural areas (Belanche
et al., 2021), even though it is easy to envision that they would neces-
sitate different approaches to planning. Here, the work by Anton and
Lawrence (2014) provides rare insight, with a focus on Western
Australia, illustrating that residents of rural areas were generally more
attached to their homes and local areas than residents of urban areas.
Finally, although predictors of place attachment have been extensively
researched, tracing replicability of common predictors in different
contexts remains a challenge.
In order to address these gaps, this study with a focus on eight
communities in Western Australia aimed to assess, based on a detailed
online survey: a) how what people value vary across the locations and
demographics; and b) how place attachment differs between rural and
urban areas. For this purpose, we tested the following hypotheses.
H1.The types of lived values prioritised by respondents will not be different
in urban and rural communities. If this could be shown to be false, we
planned to investigate whether respondents in rural communities: a)
emphasize environmental values of their local area (as per Hinds &
Sparks, 2008) more; and b) have stronger attachment to place than re-
spondents from the urban communities (as per Anton & Lawrence,
2014).
H2.The types of lived values prioritised will not differ between communities
with higher and lower socio-economic levels. To test this hypothesis, we
were particularly interested in: a) whether safety and health categories
are more likely to be prioritised by respondents in communities with
lower socio-economic levels; and b) whether the self-actualisation and
belongingness categories are more likely to be prioritised in commu-
nities with higher socio-economic levels.
2. Methodology
2.1. Study area
This study was undertaken in 2018 in Western Australia along an
approximately 400-km West-East transect extending from the state
capital of Perth through to the Wheatbelt town of Southern Cross and
divided into four segments: the Swan Coastal Plain, the Perth Hills, the
Western Wheatbelt, and the Central and Eastern Wheatbelt combined
(Fig. 1). These segments represent communities at increasing distances
from Perth, with urban, peri-urban, and rural characteristics. For the
subsequent analysis, the communities in the Swan Coastal Plain and
Perth Hills were classied as urban as they are both within the Perth
metropolitan boundaries whereas the Wheatbelt communities were
classied as rural (WA Gov, 2009).
Two communities were chosen from each segment for a total of eight
communities, with one community in the segment pair having a higher
socio-economic status than the other (Table 1).
2.2. Materials and procedure
Data was collected via an online survey in Qualtrics (with additional
hard copies made available) with a total of 33 questions relating broadly
to the demographic backgrounds of participants, lived values, and place
attachment. Here, we focus on four major questions relating directly to
lived values and place attachment (Table 2).
The two lived values questions addressed phenomena or objects that
people could value in their local area, which an earlier question indi-
cated could be, for example, their shire or district. First, respondents
were asked to select what they valued most about the place they lived in
from a list provided based on urban and rural virtues developed by
Crameri (2014), or to add their own. Secondly, respondents were asked
to indicate whether they agreed or disagreed with fourteen statements
on a ve-point Likert scale, inspired by those discussed in Graham et al.
(2013) and Tschakert et al. (2017).
The two questions on place attachment, were adapted from Anton
(2015) which built upon work by Brown and Raymond (2007) and
others. The two questions, addressing respondent’s attachment on the
scales of place of residence and local area respectively, each contained
eleven questions measured on a ve-point Likert scale. Of these eleven
Fig. 1. Map of study area in Western Australia, with four segments and two communities or suburbs (south of Perth) per segment.
A. Kelly et al.
Applied Geography 172 (2024) 103424
3

questions, the rst six addressed place identity and the next ve place
dependence. Place identity was interpreted as the symbolic or emotional
aspect of place attachment and place dependence as the functional or
physical aspect (Anton, 2015). A composite score for the place identity
dimension and place dependence dimension were calculated by aver-
aging the Likert scale values for each relevant sub-question following
Anton (2015).
Approximately 10,000 iers were sent out to the eight communities
via Australia Post through the unaddressed mail service, targeting res-
idential addresses. These iers contained a brief overview of the
research project and a link to the Qualtrics survey. Additionally online
community groups and buy-and-sell pages were contacted to raise
awareness of the survey, and iers and survey hard-copies were pro-
vided to local libraries.
2.3. Participants
Any resident of the eight communities above the age of 18 was
eligible to participate. A participant information form was provided at
the beginning of the survey, and consent was given through taking the
survey. In total, 403 completed surveys were submitted, with partici-
pants per community ranging from 31 participants in Southern Cross to
102 in Darlington (Table 3). With the exception of Southern Cross, the
majority of respondents from each of the communities were female
(≥65%). This represents a higher proportion than the actual community
gender distributions (44.3–53.2%) (ABS, 2016a).
The median and mean age in the sample cannot be directly compared
with the 2016 Australian census data as the age categories used did not
sufciently overlap. With the exception of Northam, the community
samples all had a higher percentage of respondents aged over 55
compared with the ABS (2016b) data (Table 3). This suggests that the
respondent samples were older than the community populations, which
was also the case in similar studies that have used surveys (e.g., Anton &
Lawrence, 2014; Graham et al., 2014).
2.4. Statistical analysis
Due to the small sample sizes, communities with similar character-
istics based off the location, socioeconomic status and the prioritised
lived values were aggregated together for the statistical analysis. The
responses to the lived values Likert-scale questions were compared
across the aggregated groups, and across socio-economic levels. Income
and employment were used as proxies for socio-economic status (ABS,
2011). The data derived from Likert-scale questions was in an ordinal
form which is decidedly non-normal. The histogram of the data residuals
showed the non-normal distribution, further conrmed by a
Shapiro-Wilks test for normality of residuals. As such, the Kruskal-Wallis
H Test was used instead of an ANOVA test for assessing possible dif-
ferences between groups. A signicance level of 0.05 was used for all
statistical tests, but with Bonferroni adjustments for multiple
comparisons.
In the case of signicant differences between groups, the Dunn’s
post-hoc test with Bonferroni adjustments was used. This analysis option
is in line with the recommendation by Ruxton and Beauchamp (2008)
for cases with pairwise multiple comparisons without assuming the data
is normally distributed. Additionally, Breusch-Pagan tests for hetero-
scedasticity were undertaken.
Responses to the two questions on place attachment required two
different approaches to test place attachment predictors on the two di-
mensions of place identity and place dependence, and on the scales of
place of residence and local area. As an analogue to the statistical
approach for the lived values Likert question, we rst undertook an
exploratory approach and identied which variables were potentially
signicant predictors of place attachment. Second, we used multivariate
linear models to assess which variables account for the variability of
place attachment conditional on statistically controlling for other
factors.
The nal model was selected using an automated-approach to
Table 1
Geography, demographics, sectors, and socioeconomic disadvantage of the eight locations.
Geographical
segment
Suburb/
community
City/Shire Main sectors/industries Distance from
Perth
Population (2016
census)
Index of Relative Socioeconomic
Disadvantage
b
Swan Coastal Plain Attadale Melville residential, small business, education 9 km 5720
a
10
Swan Coastal Plain Willagee Melville residential, small business, education 15 km 4356
a
3
Perth Hills Darlington Mundaring tourism, arts, hospitality 20 km 3656 10
Perth Hills Kelmscott Armadale residential, small business, and light
industrial
23 km 10,462 3
Western Wheatbelt Toodyay Toodyay tourism, arts, small business, small
holder farming
85 km 6566 3
Western Wheatbelt Northam Northam tourism, immigration detention
centre, hospital
96 km 6548 2
Central Wheatbelt Merredin Merredin broadacre agriculture, small business 318 km 2850 3
Eastern Wheatbelt Southern Cross Yilgarn broadacre agriculture, mining 370 km 680 2
a
2006 census.
b
Expressed in deciles ranging from the least disadvantaged at 10, to the most disadvantaged (The Australian Bureau of Statistics (ABS), 2016a).
Table 2
Main survey questions. A full copy of the survey including the response options
for every question is attached in the Supplementary Material.
Question Answer Type Relevant
Sources
15. What do you (and your family)
value most about the place
where you live? [Please tick all
that matter to you]:
Tickboxes Crameri
(2014)
16. How strongly do you agree/
disagree with the following
statements? [Please tick]
5-Point Likert scale with the
options: Strongly Agree,
Disagree, Neutral, Agree,
Strongly Agree
Graham et al.
(2013)
Tschakert
et al. (2017)
18. Based on your understanding
of “home” from the question
above, how strongly do you
agree or disagree with the
following statements? These
eleven statements are related to
your place of residence (your
house, incl. Garden and farm,
where applicable). [Please tick]
5-Point Likert scale with the
options: Strongly Agree,
Disagree, Neutral, Agree,
Strongly Agree
Anton (2015)
Brown and
Raymond
(2007)
19. Based on your understanding
of “home” from the question
above, how strongly do you
agree or disagree with the
following statements? These
eleven statements are related to
your local area (including your
neighbourhood, your
community, where applicable).
[Please tick]
5-Point Likert scale with the
options: Strongly Agree,
Disagree, Neutral, Agree,
Strongly Agree
Anton (2015)
Brown and
Raymond
(2007)
Williams and
Vaske (2003)
A. Kelly et al.
Applied Geography 172 (2024) 103424
4

narrow down from a general-specic model using the Akaike Informa-
tion Criterion (AIC) (see Wagenmakers & Farrell, 2004). The model was
evaluated by applying the Shapiro-Wilks, Breusch-Pagan and Ramsay’s
RESET tests for normality, heteroscedasticity and omitted variable bias
respectively.
3. Results
3.1. Lived values
Residents in the eight Western Australian communities demonstrate
a wide range of lived values linked to their local area. Those that they
value most are shown in Table 4. The evidence did not support hy-
pothesis H1a that environmental aspects of the local area would be
valued more in rural communities. In fact, there was no clear urban-
rural divide in valuing natural aspects.
Instead, there was a stronger division along socio-economic lines,
with all four communities associated with higher socio-economic status,
but none of their paired communities, selecting ‘Beauty of the natural
environment’ as part of their top ve lived values. Further, in Attadale
and Darlington (both high socio-economic standing), the natural envi-
ronment and family values (being a good place to bring up children)
were ranked highly. In contrast, in Willagee and Kelmscott, communities
with lower socio-economic status, safety and health lived values such as
affordability and general services and infrastructure were given highest
priority.
In the rural communities, values representing belongingness (see
Table 4) were prioritised throughout, with at least three belongingness
values in the top ve for all four communities. Not surprisingly, country
town atmosphere was valued highly. A sense of community and having
friendly neighbours were in the top value priorities of respondents from
three out of four of the rural communities.
The pattern visible in Table 4 suggests that there are three groupings
or clusters of communities with similar results. The rst grouping con-
tains Attadale, Darlington and Toodyay—the urban localities with high
socio-economic levels combined with Toodyay which is characterised by
a high percentage of respondents living on rural lifestyle properties or
bush blocks within commuting distance from Perth —where access to
and enjoyment of the beauty of nature and processes of self-actualisation
are given priority. This nding is consistent with hypothesis H2b,
although Toodyay’s socioeconomic status is lower than that of Attadale
and Darlington. The second grouping consists of Willagee and Kelms-
cott—the urban suburbs with lower socioeconomic status—where
health and safety lived values were the most prioritised. The third
grouping consists of the four rural towns and communities of Northam,
Merredin, and Southern Cross, all of which show a distinct emphasis on
values of belongingness. This is consistent with H1, that the types of
lived values prioritised by respondents will differ between urban and
rural communities.
However, these groupings neither support H1a (that the environ-
ment is more highly valued in rural communities) nor H1b (that
attachment to place is stronger in rural communities).
Further, while the rural communities in this third grouping have a
comparable socio-economic standing to the urban communities in the
Table 3
Participant characteristics across the eight localities.
Community Sample
Size
Median Age
Range
Percentage >55 years of
age (sample)
Percentage >55 years of
age (ABS)
Gender ratio (%
male/female)
Median Weekly
Income (ABS)
Unemployment
(ABS)
Attadale 42 36–55 years 40% 29% 31:69 $920 5.8%
Willagee 42 36–55 years 36% 25% 26:74 $629 8.1%
Darlington 102 36–55 years 36% 32% 32:68 $765 4.6%
Kelmscott 34 56–70 years 53% 29% 35:65 $608 10%
Toodyay 55 56–70 years 64% 47% 33:67 $490 10%
Northam 42 36–55 years 33% 33% 21:79 $579 9.5%
Merredin 55 36–55 years 36% 28% 31:69 $641 5.9%
Southern
Cross
31 36–55 years 39% 31% 45:55 $668 9.4%
Note: ABS data is for state suburbs from the Australian Bureau of Statistics, 2016a. Median Weekly Income is for individuals aged 15+.
Table 4
Prioritised lived values of residents. Percentages indicate the proportion of respondents who ticked each option.
Lived Value Swan Coastal Plain Perth Hills Western Wheatbelt Central Wheatbelt
Att. Wil. Dar. Kel. Too. Nor. Mer. Sou. Total
Beauty of the natural environment 76% 43% 95% 41% 89% 48% 55% 48% 68%
Access to nature 60% 36% 82% 53% 73% 55% 55% 42% 61%
Friendly neighbours and helpful citizens 52% 50% 62% 47% 60% 50% 56% 68% 57%
Community spirit/sense of community 55% 52% 59% 41% 71% 55% 53% 55% 56%
Family values (a good place to bring up children) 76% 36% 65% 29% 25% 45% 58% 55% 51%
Country town atmosphere 10% 7% 46% 26% 84% 86% 71% 68% 51%
Slow pace/laid-back lifestyle 21% 21% 60% 32% 75% 50% 45% 52% 48%
Close-knit community (people know each other) 33% 38% 56% 21% 56% 38% 49% 65% 47%
Safety/low crime 52% 19% 61% 26% 67% 12% 40% 52% 45%
Low levels of pollution 26% 19% 50% 38% 58% 38% 40% 35% 41%
Affordability 26% 71% 11% 65% 33% 52% 47% 45% 38%
Health and medical facilities 36% 48% 11% 62% 33% 45% 51% 32% 35%
General services/infrastructure 48% 57% 11% 59% 27% 40% 40% 6% 32%
Education facilities 48% 31% 30% 26% 13% 24% 40% 29% 30%
Shopping/restaurants/entertainment 50% 50% 5% 44% 13% 19% 31% 13% 24%
Transportation 45% 33% 7% 59% 9% 7% 16% 10% 20%
Employment opportunities 12% 12% 2% 18% 15% 33% 45% 39% 19%
Family histories and connections 17% 12% 13% 18% 11% 21% 35% 39% 19%
Beauty of the built environment 26% 7% 13% 6% 18% 2% 4% 10% 11%
Multi-cultural environment 0% 17% 4% 12% 0% 7% 9% 3% 6%
Other 5% 10% 3% 9% 5% 0% 5% 10% 5%
City atmosphere 7% 5% 0% 0% 0% 0% 0% 6% 2%
A. Kelly et al.
Applied Geography 172 (2024) 103424
5

second group, they did not prioritise health and safety lived values. As
such, it appears that hypothesis H2a only held true for urban
communities.
3.1.1. Aggregation and statistical analysis
As per the methodology, the communities were aggregated as per the
groupings outlined above to account for the low size of different sam-
ples. Toodyay was kept separate from the Darlington and Attadale ag-
gregation due to its lower socioeconomic status and rural classication.
Merredin, despite being less socioeconomically disadvantaged accord-
ing to the index in Table 2, could be grouped with Northam and
Southern Cross due to similar reported employment and median income
results coupled with similar lived value priorities shown in Table 4.
In terms of aggregated location, there was a statistically signicant
difference for ten out of the thirteen lived values assessed (Table 5). This
suggests that there were genuine differences in how much these phe-
nomena were valued between communities.
Based on the means, all ten of the lived values for which there was a
statistically signicant difference were emphasised as more important in
the two urban communities with higher socioeconomic status than their
lower socioeconomic status neighbours (Table 6). These included many
self-actualisation and belongness values, which supports hypothesis
H2b, at least for urban communities. There were no clear differences
between rural communities whether aggregated or not.
Comparing urban to rural communities, the urban high socioeco-
nomic status communities emphasised six lived values more than the
rural low socioeconomic status communities. At the same time, the rural
communities emphasised several lived values more than the urban low
socioeconomic status communities. This suggests that there are differ-
ences between the lived values of urban and rural communities of
similar socioeconomic status, with the rural communities valuing
several more than their urban counterparts. Overall, the priority given to
lived values varied signicantly across the different communities.
3.2. Place attachment
For place dependence at the scale of place of residence, the signi-
cant predictors were location, age, living with children, and length of
residence (Kruskal-Wallis test) (Table 7). For place identity at the same
scale, the predictors were the same with the addition of home ownership
Table 5
Location and lived values (Kruskal-Wallis H-Test).
Lived Value Breusch-Pagan
Test Statistic
Chi-Square
Statistic
df
Lifestyle in local area 8.341* 60.706
a
3
Contribution of home and local area to
identity
1.904 17.380
a
3
Plants and animals in local area 3.908 56.041
a
3
Contribution of home and local area to
mental wellbeing
13.985
a
35.547
a
3
Culturally signicant sites important to
the respondent
7.367 12.020
a
3
Being able to travel easily within local
area
9.749* 2.630 3
Environmental benets provided by
local area
0.122 37.259
a
3
Value local community 2.950 24.29
a
3
Being able to practice religion within
local area
8.342* 6.948 3
Being able to act in accordance with
culture within local area
0.618 4.310 3
Social bonds in home and local area 16.010
a
8.351* 3
Devoted time, energy and money to
create a home
4.383 10.886* 3
Pride in local area 7.597 45.030
a
3
Note: Bonferroni adjustments were used.
a
* Signicant at the 1% and 5% level.
Table 6
Comparisons of location with what people value most (Dunn post-hoc test).
Comparison Mean Standard
Error
z-test
Statistic
Lifestyle in local area
Rural Low – Urban Low 4.37,
4.07
0.07, 0.10 3.127*
Toodyay – Rural Low 4.73,
4.37
0.08, 0.07 3.152
a
Toodyay – Urban Low 4.73,
4.07
0.08, 0.10 5.428
a
Urban High– Rural Low 4.75,
4.37
0.05, 0.07 4.565
a
Urban High– Urban Low 4.75,
4.07
0.05, 0.10 7.105
a
Contribution of home and local area to identity
Toodyay – Urban Low 4.11,
3.63
0.12, 0.12 2.770*
Urban High– Urban Low 4.16,
3.63
0.07, 0.12 3.863
a
Plants and animals in local area
Toodyay – Rural Low 4.38,
3.89
0.09, 0.08 3.548
a
Urban High – Rural Low 4.38,
3.89
0.07, 0.08 5.228
a
Urban High – Urban Low 4.38,
4.01
0.07, 0.11 3.086*
Contribution of home and local area to mental wellbeing
Toodyay – Urban Low 4.44,
4.04
0.10, 0.09 3.132*
Urban High – Rural Low 4.58,
4.10
0.05, 0.08 4.607
a
Urban High – Urban Low 4.58,
4.04
0.05, 0.09 5.094
a
Culturally signicant sites
Rural Low – Urban Low 3.44,
2.99
0.08, 0.13 2.932*
Urban High – Urban Low 3.49,
2.99
0.08, 0.13 3.296
a
Environmental benets provided by local area
Urban High – Rural Low 4.50,
4.06
0.06, 0.06 5.482
a
Urban High – Urban Low 4.50,
4.09
0.06.0.08 4.522
a
Value local community
Toodyay – Urban Low 4.31,
3.82
0.11, 0.10 4.308
a
Urban High – Rural Low 4.35,
4.08
0.06, 0.07 2.793*
Urban High – Urban Low 4.35,
3.82
0.06, 0.10 4.569
a
Social bonds in home and local area
Rural Low – Urban Low 4.27,
3.84
0.06, 0.12 2.825*
Devoted time, energy and money to
create home
  
Toodyay – Urban Low 4.51,
4.07
0.10, 0.11 2.859*
Pride in local area
Toodyay – Urban Low 4.31,
3.54
0.10, 0.13 4.196
a
Rural Low – Urban Low 4.00,
3.54
0.08, 0.13 2.867*
Urban High – Rural Low 4.40,
4.00
0.07, 0.08 3.994
a
Urban High – Urban Low 4.40,
3.54
0.07, 0.13 6.351
a
Note: Only the signicant results are shown. Bonferroni adjustments were used.
‘Urban High’ was the aggregation of Attadale and Darlington, ‘Urban Low’ that
of Willagee and Kelmscott, and ‘Rural Low’ the grouping of Northam, Merredin
and Southern Cross.
a
* Signicant at the 1% and 5% level.
A. Kelly et al.
Applied Geography 172 (2024) 103424
6

and employment, although the latter was excluded from further
consideration due to heteroscedastic issues. At the local area scale,
location and length of residence were signicant predictors of place
dependence, and the same with the addition of age and home ownership
were predictors of place identity.
The urban higher socioeconomic communities had more pronounced
place attachment than the lower urban socioeconomic communities
across all dimensions except for place of residence place identity
(Table 8). In terms of the differences in place attachment between rural
and urban communities, Toodyay consistently scored higher than the
urban lower socioeconomic communities, but not higher than the urban
higher socioeconomic communities, which had a higher local area place
identity score than the rural lower socioeconomic communities. How-
ever, there were signicant differences in local area place attachment
between the rural and urban communities of lower socioeconomic
standing.
In terms of age and place attachment, the results conrm what we
would expect. Increased age correlated with greater place attachment
overall (Table 9). At the same time, the percentage of respondents
residing in their birth community declined with increasing age (65%
aged 18–25; 39% aged 26–35, 21% aged 36–55, 22% aged 56–70, and
14% aged>70). Taking age and years of residence together showed that
≤5 years of residence was correlated with lower place attachment, in
both dimensions and scales, compared to those with residency duration
of ≥30years. Moreover, renters had lived in their home for less time
(77% for <5 years) than home-owners (29% for <5 years).
The scores in the linear models showed how much variation in
respondent scores was explained for place identity and place depen-
dence on the scale of place of residence (17% and 14% respectively), and
on the scale of local area (16% and 13% respectively) (see Table 10).
For place identity at place of residence, the signicant variables were
age and gender, once other factors had been controlled for. Age and
whether a respondent had been brought up in their community emerged
as signicant variables for place of residence place dependence, and for
local area place identity. Location was signicant for place dependence
for both place of residence and local area, and length of residence was
signicant for place identity and dependence in a local area. Additional
models with the community names (location) replaced with ‘urban’ and
‘rural’ for the eight communities, coupled with t-tests, also yielded no
signicant differences.
Overall, while the Kruskal-Wallis and Dunn tests suggested that local
area place attachment was greater among the rural than the urban lower
socioeconomic communities, the linear models showed no clear
Table 7
Place attachment and demographics (Kruskal-Wallis H-test).
Demographic df Chi-Square Statistic
Place of Residence Place Identity Place of Residence Place Dependence Local Area Place Identity Local Area Place Dependence
Location 3 13.262** 13.748** 27.933** 20.847**
Age 4 24.754** 21.251** 17.921* 8.906
Gender 1 2.566 0.323 1.076 0.374
Level of Education 4 2.891 7.543 4.095 3.710
Employment 8 23.188* 17.847 13.150 9.842
Income 4 3.732 5.550 1.362 3.338
Children in Home 1 10.033* 11.347** 0.068 0.298
Home Ownership 3 28.796** 9.750 17.911** 9.027
Property Type 5 30.099** 36.418** 23.018** 28.232**
Residence Length 4 17.772* 18.233* 20.479** 19.354**
Birth Community 1 0.003 0.486 4.972 2.410
Note: ** * Signicant at the 1% and 5% level.
Table 8
Signicant pair-comparisons for location with place attachment (Dunn post-
hoc).
Place attachment
factor and scale
Comparison of
location
Means Standard
Error
z-test
statistic
Place of residence –
Place identity
Toodyay – Rural
Low
4.33,
3.96
0.11, 0.08 2.975*
Place of residence –
Place identity
Toodyay – Urban
Low
4.33,
3.90
0.11, 0.10 3.244
a
Place of residence –
Place dependence
Toodyay – Urban
Low
3.56,
2.94
0.13, 0.12 3.329
a
Place of residence –
Place dependence
Urban High –
Urban Low
3.39,
2.94
0.08, 0.12 3.086*
Local Area – Place
identity
Rural Low –
Urban Low
3.59,
3.27
0.08.0.10 2.730*
Local Area – Place
identity
Toodyay – Urban
Low
3.76,
3.27
0.11, 0.10 3.181
a
Local Area – Place
identity
Urban High –
Rural Low
3.90,
3.59
0.07, 0.08 2.780*
Local Area – Place
identity
Urban High –
Urban Low
3.90,
3.27
0.07, 0.08 5.170
a
Local Area – Place
dependence
Rural Low –
Urban Low
3.03,
2.63
0.09, 0.10 2.973*
Local Area – Place
dependence
Toodyay – Urban
Low
3.17,
2.63
0.14, 0.10 2.942*
Local Area – Place
dependence
Urban High –
Urban Low
3.26,
2.63
0.08, 0.10 4.515
a
Note: Only the signicant results are shown. Bonferroni adjustments were used.
‘Urban High’ was the aggregation of Attadale and Darlington, ‘Urban Low’ that
of Willagee and Kelmscott, and ‘Rural Low’ the grouping of Northam, Merredin
and Southern Cross.
a
* Signicant at the 1% and 5% level.
Table 9
Signicant paired comparison of age classes with place attachment (Dunn post-
hoc).
Place Attachment
Factor and Scale
Comparison of
age groups
Means Standard
Errors
z-test
statistic
Place of residence –
Place identity
18-25 with
56–70
3.51,
4.27
0.21, 0.07 −3.715
a
Place of residence –
Place identity
18-25 with 71+3.51,
4.23
0.21, 0.10 −3.079*
Place of residence –
Place identity
26-35 with
56–70
3.69,
4.27
0.15, 0.07 −3.647
a
Place of residence –
Place dependence
18-25 with
56–70
2.76,
3.50
0.23, 0.08 −3.155*
Place of residence –
Place dependence
18-25 with 71+2.76,
3.53
0.23, 0.13 −2.920*
Place of residence –
Place dependence
26-35 with
56–70
2.84,
3.50
0.14, 0.08 −3.534
a
Place of residence –
Place dependence
26-35 with 71+2.84,
3.53
0.14, 0.13 −3.063*
Local Area – Place
identity
18-25 with
36–55
3.03,
3.70
0.19, 0.06 −3.116*
Local Area – Place
identity
18-25 with
56–70
3.03,
3.75
0.19, 0.08 −3.520
a
Local Area – Place
identity
18-25 with 71+3.03,
3.87
0.19, 0.11 −3.669
a
Note: Only the signicant results are shown. Bonferroni adjustments were used.
a
* Signicant at the 1% and 5% level.
A. Kelly et al.
Applied Geography 172 (2024) 103424
7

differences between place attachment among members of urban and
rural communities. The differences in place attachment that did emerge
between the eight localities were more strongly related to socioeco-
nomic variables, and in the linear models were mainly due to re-
spondents from Darlington scoring higher on relevant measures than
respondents elsewhere, and respondents from Willagee scoring lower.
Hence, the ndings suggest that a rural-urban dichotomy is inadequate
for describing variations in place attachment and that other factors may
play a more signicant role.
4. Discussion
4.1. Urban-rural and socio-economic differences
Having established that there were indeed differences between urban
and rural communities with respect to lived values, closer attention to
environment-related lived values is warranted (H1a). The relatively
high importance placed on ‘Beauty of the natural environment’ mirrored
ndings from other studies in Australia, including the highly prioritised
values of the environment and scenery in Lakes Entrance (Graham et al.,
2014), and Kingston Beach (Ramm et al., 2017), both coastal commu-
nities deliberately relocated to by many residents. Similarly, partici-
pants in our study appeared more likely to choose to live in areas with
desirable environmental or landscape characteristics, if they could
afford it. Such aesthetic landscape choices were not limited to rural
areas as they were also apparent in urban areas near the Swan River.
The fact that environmental aspects of the local area were valued
alike in urban and rural areas, rather than more strongly in the latter,
may well be tied to the historical valuations of the Western Australian
Wheatbelt as an area prized primarily in instrumentalist and economic
terms. An observable difference could well exist in other parts of this
area of the continent, such as the far southwestern corner, and therefore
match results reported elsewhere in the literature, including those by
Hinds and Sparks (2008) in the UK. In our case, as discussed below, the
combination of access to natur e and socio-economic status emerged as
more important.
Probing further to test whether respondents from rural communities
will have stronger attachment to place than respondents from the urban
communities (H1b), our results, once more, did not unequivocally
support the hypothesis. Our study found that the rural communities
exhibited higher place attachment to their local area, but not their place
of residence, than their urban counterparts with similar socioeconomic
characteristics, but that this relationship did not hold for the linear
models which adjusted for confounding factors such as age and length of
residence. As such, contrary to the ndings of Anton and Lawrence
(2014) whose focus was on rural areas in the far south-western corner of
Western Australia, much appreciated as tourism and lifestyle destina-
tions, our study did not detect stronger place attachment in the agri-
cultural areas of the Wheatbelt investigated, but it did indicate that
among communities of similar socioeconomic standing that local area
place attachment is more likely to be higher than lower in the rural
communities compared with urban. This was further conrmed through
linear model runs using urban and rural for the location variable, which
with Darlington omitted, showed a signicant difference in place
attachment between urban and rural areas for local area place identity.
These differing results with and without Darlington included may be due
to the inherent complexity of place identity, as shown by Belanche et al.
(2021)’s work in Spain that further subdivided place identity into af-
fective, evaluative and cognitive components and found only the rst
two differed between urban and rural areas.
It is conceivable that a focus on farm businesses, as promoted under
neoliberal governance regimes (Cheshire et al., 2013), undermines
strong emotional place attachment among the rural (farming) partici-
pants from the Wheatbelt. Alternatively, our communities in the Perth
Hills and the Western Wheatbelt (Darlington and Toodyay respectively)
that did show higher place attachment may in fact better reect a
rurality of the far South-West, characterised by properties similar to the
bush blocks and rural lifestyle properties. The purchase of a bush
Table 10
Place attachment models.
Category Variable PI POR PD POR PI LA PD LA
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. Coef. Std. Err.
Intercept Intercept 3.4899 0.7947** 2.1108 2.1108* 2.7231 0.2148** 2.6704 0.1595**
Age 26–35 years 0.1847 0.2102 0.2864 0.2864 0.5696 0.2060**  
Age 35–55 years 0.4002 0.1979* 0.6520 0.6520** 0.7237 0.1873**  
Age 56–70 years 0.5612 0.1989** 0.8067 0.8067** 0.7556 0.1979**  
Age 71 years +0.5930 0.2145** 0.8648 0.8648** 0.9158 0.2190**  
Gender Male −0.2060 0.0865* −0.1453 0.1043 −0.1538 0.0902  
Living with Children Yes −0.1402 0.0993 −0.2255 0.1189    
Home Ownership Other −0.2886 0.5986      
Home Ownership Owning 0.1692 0.5692      
Home Ownership Renting −0.3745 0.5730      
Property Type Bush Block 0.4028 0.5579 0.9397 0.6742    
Property Type Commercial agricultural 0.2265 0.6000 0.7076 0.7265    
Property Type Lifestyle rural 0.3890 0.5552 1.007 0.6717    
Property Type Other −0.3277 0.5807 −0.0150 0.7052    
Property Type Urban 0.0520 0.5497 0.4389 0.6663    
Born in Community Yes 0.1366 0.0931 0.2650 0.1122* 0.3673 0.0997** 0.20532053 0.1090
Location Darlington     0.3362 0.1497* 0.41074107 0.1705*
Location Kelmscott     −0.3512 0.1877 −0.2243 0.2149
Location Merredin     0.0044 0.1658 −0.0097 0.1903
Location Northam     −0.1652 0.1770 0.0031 0.2038
Location Southern Cross     0.0621 0.1931 0.36613661 0.2210
Location Toodyay     0.1090 0.1682 0.2755 0.1912
Location Willagee     −0.4773 0.1783** −0.3956 0.2049
Length of Residence 6–10 years     0.2529 0.1166 0.2717 0.1310*
Length of Residence 11–20 years     0.2909 0.1103** 0.2428 0.1204*
Length of Residence 21–30 years     0.4489 0.1570** 0.5957 0.1664**
Length of Residence 30+years     0.2227 0.1775 0.5269 0.1824**
R-Squared 0.2036 0.1655 0.1970 0.13101310
Adjusted R-squared 0.1727 0.1398 0.1615 0.0421042
F-Ratio 6.596** 6.443** 5.555** 4.898898**
Note: ** * Signicant at the 1% and 5% level. Model diagnostics are in Appendix A.
A. Kelly et al.
Applied Geography 172 (2024) 103424
8

lifestyle block may serve as an indicator of a desire to attach to a place as
suggested by the word ‘lifestyle’. At the same time, the stronger
attachment of respondents living in bush blocks could also be due to the
immediate threat of bushres, forcing residents to be more aware of
place (Anton & Lawrence, 2014).
As for predictors of place attachment, home ownership, length of
residence, and age emerged as the most prominent ones. Home owner-
ship was a signicant predictor of place identity, but not place depen-
dence. This concurred with the ndings of Belanche et al. (2021) that
home owning residents display higher levels of place identity. This may
be due, as Anton and Lawrence (2014) argue, to renters tending to reside
for shorter periods of time in a place which results in less time to
incorporate a place into their identity structure. Most older studies did
not adequately assess the physical or functional dimension of place
attachment and hence may have missed the relevance of place identity
(e.g., Brown et al., 2003; Mesch & Manor, 1998). The prominent role of
home ownership as a determinant of place attachment likely reects a
particular western mindset typical for settler colonies, and the suspected
absence of Indigenous participants in our survey. This bias is overlaid
with uneven socio-economic realities between owners and renters.
Age emerged as the most consistent signicant predictor of place
attachment, appearing in all models except for the local area place
dependence model. Age had been identied as a signicant predictor in
many studies (e.g., Hidalgo & Hern´
andez, 2001; Lewicka, 2010) but not
consistently (e.g., Anton & Lawrence, 2014).
Overall, the models accounted for between 10% and 17% of varia-
tion in place attachment, using a conventional R2 metric. As measured
by the F-test, there is evidence in the sample to indicate that in all cases
this is a statistically signicant t at the 1% signicance level. While
these are adequate results for models in the social sciences (e.g., Falk &
Miller, 1992), the relatively low percentages of variation explained by
the models could be a reason why the literature is often unclear about
different predictors of place attachment and why conicting results have
been found in the past.
Finally, with respect to our second line of inquiry (H2)—whether the
types of lived values prioritised differ between communities with higher
and lower socio-economic levels —our ndings for the urban commu-
nities were clear and in alignment with the existent albeit still small
literature on lived values (e.g. Graham et al., 2014; Ramm et al., 2017).
The prioritised lived values across the four urban communities broadly
reected Maslow’s (1943) hierarchy of needs with lived values: the
safety and health values were prioritised in the communities with lower
socio-economic status (H2a) while lived values associated with
self-actualisation and belongingness were ranked higher in communities
with higher socio-economic status (H2b). This result is not surprising as
safety and health are often taken for granted by more afuent residents.
However, these trends did not hold for the rural communities, which all
highly prioritised belongness lived values such as sense of community
and friendly neighbours, and between the paired rural communities of
different socioeconomic levels there were no notably different patterns
of prioritised lived values.
4.2. Relevance for climate change adaptation
The literature is clear that a better understanding of what people
value (Barnett et al., 2016; Karlsson & Hovelsrud, 2015) and how they
are attached to places (Reid et al., 2020) is essential for determining
socially-salient preparedness efforts and responses to threats from
climate change and other incremental and rapid-onset dynamics. Our
ndings show that there are important commonalities in lived values
that, as a start, can inform more inclusive planning for climate change
adaptation that is likely to garner community buy-in, despite differences
within and beyond communities. For instance, scenery and aesthetics
are crucial to residents in coastal or riverine communities and hence
would make seawalls or other structural adaptation strategies politically
very difcult. By the same token, priority needs to be given to values
that are prominent in less afuent communities, as well as specic needs
of residents such as those on disability pensions. Such place-specic
attention to lived values and attachment to place highlights the cen-
tral role of affordable and accessible general and health infrastructure
during crisis events. Our ndings suggest that re-establishing commu-
nity ties and environmental rehabilitation efforts may be as important to
affected residents as nancial compensation for experienced losses in
the face of disasters like the 2019/2020 eastern Australian bushres.
We are cognizant of the fact that surveys, such as the one used in this
study, are snapshots in time. They are best suited as one of several ways
to engage urban and rural communities. In our case, the online survey
provided the starting point, followed by face-to-face interviews,
participatory mapping activities, and workshops. These additional in-
quiries are vital to understand how people negotiate difcult trade-offs
between the many things they value, given the prospect of more extreme
climatic events and intensifying trends and the realisation that not
everything can be saved. Inquiries that combine a variety of methods
help addresses the knotty issue outlined by Henrique and Tschakert
(2023) and McNamara and Jackson (2019) of how value prioritisations
and place attachment, including those of vulnerable populations, may
drive trade-offs between acceptable and intolerable losses related to
climate change while allowing space for loss and grief.
4.3. Limitations and future directions
Arguably the largest limitation of this study is related to the size and
composition of our samples. Firstly, our survey results are not repre-
sentative of the whole communities, and the communities were not
necessarily representative of other similar communities in this part of
Western Australia. For example, due to the voluntary nature of partici-
pation in the survey, more older than younger residents make up the
sample population, arguably because they have more spare time to
participate or were more enticed by the subject matter. Moreover, there
were distinctly more women than men among the participants, perhaps
more intrigued by the topic. Secondly, our samples were smaller than
desired, with Southern Cross and Kelmscott notably having less than 40
respondents. This reduced the test power of our statistical tests and
meant we were less able to determine whether there were signicant
differences between the individual communities unless they were
aggregated together. Hence, the results should be treated as a guideline
to what a signicant proportion of respondents from a community
values rather than a comprehensive summary of what is in fact valued.
Another issue was that the socioeconomic status of the ‘higher’ so-
cioeconomic rural communities of Toodyay and Merredin were not
substantially different than their paired ‘lower’ socioeconomic com-
munities of Northam and Southern Cross as shown through the partici-
pant characteristics in Table 3, and hence little could be concluded
about differences in place attachment and lived values between rural
communities with higher and lower socioeconomic levels.
As for future directions, we recommend that studies further examine
links between home ownership and place attachment. This would pro-
vide insight on whether our nding that home ownership does not
signicantly predict place attachment is an anomaly in the literature or
whether a two-factor approach to place attachment is necessary to un-
derstand home ownership as a predictor. Moreover, research on rural-
urban differences in place attachment should incorporate more
detailed physical predictors of place attachment instead of focussing on
a categorical rural-urban dichotomy. This would provide more clarity as
to differing place attachment between urban and rural regions in
otherwise similar locations.
Finally, studies comparing more sites within different Australian
and/or international settings are warranted, particularly to further
investigate place-based drivers as well as communalities across diverse
geographic and cultural contexts. We strongly recommend that such
scholarship also considers lived values and place attachment in Indige-
nous communities as they are likely to diverge from our ndings
A. Kelly et al.
Applied Geography 172 (2024) 103424
9

conned to settler communities.
5. Conclusion
A place-based approach to place attachment, grounded in lived
values, is an important rst step to identify what people value and what
could be potentially lost due to climate change. Our research showed
that the value priorities of different communities can be predicted based
on geographical and socioeconomic characteristics of the community
across urban and rural locales. At the same time, place attachment varies
signicantly and cannot be simply distilled without engaging with a
community from the ground up, including via online surveys as under-
taken here to establish a baseline from which to expand. Intertwined
with lived values and place attachment are experiences of climatic
threats such as bushres and the hierarchy of differential needs between
disadvantaged and afuent residents. Understanding such complexities
makes it possible for researchers to contribute to socially-salient climate
change and disaster management efforts that put people’s values and
place-based priorities at the centre of inclusive adaptation planning.
Funding
The author(s) disclosed receipt of the following nancial support for
the research, authorship, and/or publication of this article: This work
was supported by the Australian Research Council ARC Discovery
Project DP180103700.
CRediT authorship contribution statement
A. Kelly: Writing – review & editing, Writing – original draft, Vali-
dation, Software, Resources, Project administration, Methodology,
Investigation, Formal analysis, Data curation. P. Tschakert: Writing –
review & editing, Supervision, Project administration, Investigation,
Funding acquisition, Conceptualization. C. Lawrence: Validation, Su-
pervision, Methodology, Investigation, Funding acquisition, Formal
analysis, Data curation, Conceptualization, Writing – original draft. P.
Horwitz: Funding acquisition, Conceptualization, Writing – original
draft. C. Bourgault: Funding acquisition, Conceptualization, Writing –
original draft. N. Ellis: Resources, Methodology, Writing – original
draft.
Declaration of competing interest
Declarations of interest: none.
Acknowledgements
We would like to thank Paul Plummer for his guidance during the
statistical analysis, and Karen Paiva Henrique for her contribution
during meetings and sharing of literature.
Appendix BSupplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.apgeog.2024.103424.
Appendix A
Linear model diagnostics:
Ramsey’s rest test for functional form indicated that all four models
were suitable (F (2, 385) =1.801, p =0.167), (F(2,388) =1.539, p =
0.216), (F(2, 383) =1.540, p =0.216), (F(2, 382) =0.417, p =0.660).
However, the Shapiro-Wilks normality test suggested that the data in the
rst three models was not normal (W =0.955, p=<0.01), (W =0.992, p
=0.024), (W =0.978, p <0.001), but that for local area place depen-
dence the data was normal (W =0.995, p =0.273). Heteroscedasticity
was an issue for the place of residence place identity model (BP(15, 387)
=26.365, p =0.034), but not for the other three models (BP(12,390) =
7.169, p =0.846), (BP(17, 385) =21.903, p =0.1889), (BP(18, 384) =
6.953, p =0.991).
Only the local area place dependence model had to be reselected due
to having both property type and location variables included.
References
Adger, W. N., Barnett, J., Brown, K., Marshall, N., & O’brien, K. (2013). Cultural
dimensions of climate change impacts and adaptation. Nature Climate Change, 3(2),
112–117.
Anton, C. E. (2015). Home sweet home: An examination of the relationship between place
attachment and place-protective actions. Crawley, WA: Doctoral dissertation,
University of Western Australia. https://api.research-repository.uwa.edu.au/portalf
iles/portal/9651536.
Anton, C. E., & Lawrence, C. (2014). Home is where the heart is: The effect of place
residence on place attachment and community participation. Journal of
Environmental Psychology, 40(1), 451–461.
Australian Bureau of Statistics. (2011). Information paper: Measures of socioeconomic
status. http://www.ausstats.abs.gov.au/Ausstats/subscriber.nsf/0/367D380060
5DB064CA2578B60013445C/$File/1244055001_2011.pdf.
Australian Bureau of Statistics. (2016a). Socio-economic indexes for areas. http://www.
abs.gov.au/websitedbs/censushome.nsf/home/seifa.
Australian Bureau of Statistics. (2016b). QuickStats. http://www.abs.gov.au/websitedbs
/censushome.nsf/home/quickstats?opendocument&navpos=220.
Barnett, J., Tschakert, P., Head, L., & Adger, W. N. (2016). A science of loss. Nature
Climate Change, 6(11), 976–978.
Belanche, D., Casalo, L. V., & Rubio, M. A. (2021). Local place identity: A comparison
between residents of rural and urban communities. Journal of Rural Studies, 82,
242–252.
Bonaiuto, M., Alves, S., De Dominicis, S., & Petruccelli, I. (2016). Place attachment and
natural hazard risk: Research review and agenda. Journal of Environmental
Psychology, 48, 33–53.
Brown, G., & Donovan, S. (2014). Measuring change in place values for environmental
and natural resource planning using public participation GIS (PPGIS): Results and
challenges for longitudinal research. Society & Natural Resources, 27(1), 36–54.
Brown, G., & Fagerholm, N. (2015). Empirical PPGIS/PGIS mapping of ecosystem
services: A review and evaluation. Ecosystem Services, 13, 119–133.
Brown, B., Perkins, D. D., & Brown, G. (2003). Place attachment in revitalising
neighbourhood: Individual and block levels of analysis. Journal of Environmental
Psychology, 23(3), 259–271.
Brown, G., & Raymond, C. (2007). The relationship between place attachment and
landscape values: Towards mapping place attachment. Applied Geography, 27(2),
89–111.
Brown, G., Raymond, C., & Corcoran, J. (2015). Mapping and measuring place
attachment. Applied Geography, 57, 42–53.
Brown, G., & Reed, P. (2012). Values compatibility analysis: Using public participation
geographic information systems (PPGIS) for decision support in national forest
management. Applied Spatial Analysis and Policy, 5(4), 317–332.
Brown, G., Reed, P., & Raymond, C. M. (2020). Mapping place values: 10 lessons from
two decades of public participation GIS empirical research. Applied Geography,
Article 102156.
Cheshire, L., Meurk, C., & Woods, M. (2013). Decoupling farm, farming and place:
Recombinant attachments of globally engaged farmers. Journal of Rural Studies, 30,
64–74.
Clarke, D., & Murphy, C. (2023). Incremental adaptation when transformation fails: The
importance of place-based values and trust in governance in avoiding maladaptation.
Journal of Environmental Psychology, 88, Article 102037.
Crameri, K. M. (2014). Deconstructing place identity: A socio-spatial exploration of
contemporary Ballarat, Australia. Crawley, WA: Doctoral dissertation, University of
Western Australia. https://api.research-repository.uwa.edu.au/portalles/portal
/5299090/Crameri_Kim_2014.pdf.
Devine-Wright, P. (2013). Think global, act local? The relevance of place attachments
and place identities in a climate changed world. Global Environmental Change, 23(1),
61–69.
Enqvist, J. P., Campbell, L. K., Stedman, R. C., & Svendsen, E. S. (2019). Place meanings
on the urban waterfront: A typology of stewardships. Sustainability Science, 14,
589–605.
Falk, R., & Miller, N. (1992). A primer for soft modelling. The University of Akron Press.
Farbotko, C., & McMichael, C. (2019). Voluntary immobility and existential security in a
changing climate in the Pacic. Asia Pacic Viewpoint, 60(2), 148–162.
Gajardo, L. J., Sumeldan, J., Sajorne, R., Madarcos, J. R., Goh, H. C., Culhane, F., …
Creencia, L. (2023). Cultural values of ecosystem services from coastal marine areas:
Case of Taytay Bay, Palawan, Philippines. Environmental Science & Policy, 142,
12–20.
Gardner, W. L., Gabriel, S., & Lee, A. (1999). “I” value freedom, but “we” value
relationships: Self-construal priming mirrors cultural differences in judgement.
Psychological Science, 10(4), 321–326.
Government of Western Australia. (2009). Perth Metro region. http://www.drd.wa.gov.
au/publications/Documents/Metropolitan_Perth_LGA_boundaries.pdf.
A. Kelly et al.
Applied Geography 172 (2024) 103424
10

Graham, S., Barnett, J., Fincher, R., Hurlimann, A., & Mortreux, C. (2014). Local values
for fairer adaptation to sea-level rise: A typology of residents and their lived values in
lakes entrance, Australia. Global Environmental Change, 29, 41–52.
Graham, S., Barnett, J., Fincher, R., Hurlimann, A., Mortreux, C., & Waters, E. (2013).
The social values at risk from sea-level rise. Environmental Impact Assessment Review,
41(1), 45–52.
Graham, S., Barnett, J., Mortreux, C., Hurlimann, A., & Fincher, R. (2018). Local values
and fairness in climate change adaptation: Insights from marginal rural Australian
communities. World Development, 108, 332–343.
Henrique, K. P., & Tschakert, P. (2023). Everyday limits to adaptation. Oxford Open
Climate Change, 2(1), kgab013.
Henrique, K. P., Tschakert, P., Bourgault du Coudray, C., Horwitz, P., Krueger, K. D. C., &
Wheeler, A. J. (2023). Navigating loss and value trade-offs in a changing climate.
Climate Risk Management, 35, Article 100405.
Hess, J. J., Malilay, J. N., & Parkinsons, A. J. (2008). Climate change: The importance of
place. American Journal of Preventive Medicine, 35(5), 468–478.
Hidalgo, M. C., & Hern´
andez, B. (2001). Place attachment: Conceptual and empirical
questions. Journal of Environmental Psychology, 21(3), 273–281.
Hinds, J., & Sparks, P. (2008). Engaging with the natural environment: The role of
affective connection and identity. Journal of Environmental Psychology, 28, 109–120.
Jorgensen, B. S., & Stedman, R. C. (2001). Sense of place as an attitude: Lakeshore
owners attitudes toward their properties. Journal of Environmental Psychology, 21(3),
233–248.
Karlsson, M., & Hovelsrud, G. K. (2015). Local collective action: Adaptation to coastal
erosion in the Monkey river Village, Belize. Global Environmental Change, 32, 96–107.
Kreller, A. (2020). Transforming fair decision-making about sea-level rise in cities: The
values and beliefs of residents in Botany Bay, Australia. Environmental Values, 30(1),
7–42.
Lewicka, M. (2010). What makes neighbourhood different from home and city? Effects of
place scale on place attachment’. Journal of Environmental Psychology, 30(1), 35–51.
Lewicka, M. (2011). Place attachment: How far have we come in the last 40 years?
Journal of Environmental Psychology, 31(3), 207–230.
Li, J., Stoffelen, A., Meijles, E., & Vanclay, F. (2023). Local people’s sense of place in
heavily touristied protected areas: Contested place meanings around the
Wulingyuan World Heritage Site, China. Landscape and Urban Planning, 237, Article
104792.
Lin, C. C., & Lockwood, M. (2014). Forms and sources of place attachment: Evidence
from two protected areas. Geoforum, 53, 74–81.
Maguire, B., & Klinkenberg, B. (2018). Visualization of place attachment. Applied
Geography, 99, 77–88.
Maslow, A. H. (1943). A theory of human motivation. Psychological Review, 50(4),
370–396.
Masterson, V. A., Enqvist, J. P., Stedman, R. C., & Teng¨
o, M. (2019). Sense of place in
social–ecological systems: From theory to empirics. Sustainability Science, 14,
555–564.
McNamara, K. E., & Jackson, G. (2019). Loss and damage: A review of the literature and
directions for future research. Wiley Interdisciplinary Reviews: Climate Change, 10(2),
e564.
Mesch, G. S., & Manor, O. (1998). Social ties, environmental perception, and local
attachment. Environment and Behavior, 30(4), 504–519.
Murphy, A., Enqvist, J. P., & Teng¨
o, M. (2019). Place-making to transform urban
social–ecological systems: Insights from the stewardship of urban lakes in Bangalore,
India. Sustainability Science, 14, 607–623.
Peng, K., Nisbett, R. E., & Wong, N. Y. (1997). Validity problems comparing values across
cultures and possible solutions. Psychological Methods, 2(4), 329–344.
Quinn, T., Bousquet, F., & Guerbois, C. (2019). Changing places: The role of sense of
place in perceptions of social, environmental and overdevelopment risks. Global
Environmental Change, 57, Article 101930.
Ramm, T. D., Graham, S., White, C. J., & Watson, C. S. (2017). Advancing values-based
approaches to climate change adaptation: A case study from Australia. Environmental
Science & Policy, 76, 113–123.
Raymond, C. M., & Brown, G. (2011). Assessing spatial associations between perceptions
of landscape value and climate change risk for use in climate change planning.
Climatic Change, 104(3–4), 653–678.
Reid, K., Beilin, R., & McLennan, J. (2020). Communities and responsibility: Narratives
of place-identity in Australian bushre landscapes. Geoforum, 109, 35–43.
Ruxton, G. D., & Beauchamp, G. (2008). Time for some a priori thinking about post hoc
testing. Behavioural Ecology, 19(3), 690–693.
Ryeld, F., Cabana, D., Brannigan, J., & Crowe, T. (2019). Conceptualizing ‘sense of
place’in cultural ecosystem services: A framework for interdisciplinary research.
Ecosystem Services, 36, Article 100907.
Scannell, L., & Gifford, R. (2010). The relations between natural and civic place
attachment and pro-environmental behaviour. Journal of Environmental Psychology,
30, 289–297.
Tschakert, P., Barnett, J., Ellis, N., Lawrence, C., Tuana, N., New, M., Elrick-Barr, C.,
Pandit, R., & Pannell, D. (2017). Climate change and loss, as if people mattered:
Values, places, and experiences. Wiley Interdisciplinary Reviews: Climate Change, 8(5).
Tschakert, P., Ellis, N. R., Anderson, C., Kelly, A., & Obeng, J. (2019). One thousand ways
to experience loss: A systematic analysis of climate-related intangible harm from
around the world. Global Environmental Change, 55, 58–72.
Tuan, Y. F. (1977). Space and place: The perspective of experience. U of Minnesota Press.
Wagenmakers, E. J., & Farrell, S. (2004). AIC model selection using Akaike weights.
Psychonomic Bulletin & Review, 11(1), 192–196.
Wartmann, F. M., & Purves, R. S. (2018). Investigating sense of place as a cultural
ecosystem service in different landscapes through the lens of language. Landscape
and Urban Planning, 175, 169–183.
Williams, D. R., & Vaske, J. J. (2003). The measurement of place attachment: Validity
and generalizability of a psychometric approach. Forest Science, 49(6), 830–840.
Wooltorton, S., Collard, L., & Horwitz, P. (2017). The land still Speaks: Ni, Katitj. PAN:
Philosophy Activism Nature, 1(13), 57–67.
A. Kelly et al.
Applied Geography 172 (2024) 103424
11