You are on page 1of 14

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/49515869

The Measurement of Place Attachment: Personal, Community, and


Environmental Connections

Article  in  Journal of Environmental Psychology · December 2010


DOI: 10.1016/j.jenvp.2010.08.002 · Source: OAI

CITATIONS READS

650 20,025

3 authors:

Christopher M. Raymond Gregory Brown


University of Helsinki California Polytechnic State University, San Luis Obispo
111 PUBLICATIONS   10,883 CITATIONS    148 PUBLICATIONS   11,774 CITATIONS   

SEE PROFILE SEE PROFILE

Delene Weber
University of South Australia
42 PUBLICATIONS   2,088 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Assessing the co-benefits of nature-based solutions View project

International Place Attachment Network (IPAN) View project

All content following this page was uploaded by Christopher M. Raymond on 19 January 2019.

The user has requested enhancement of the downloaded file.


Journal of Environmental Psychology 30 (2010) 422e434

Contents lists available at ScienceDirect

Journal of Environmental Psychology


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

The measurement of place attachment: Personal, community, and environmental


connections
Christopher M. Raymond a, *, Gregory Brown b, Delene Weber c
a
Centre for Rural Health and Community Development, University of South Australia and Enviroconnect Pty Ltd, PO Box 190, Stirling SA, 5152, Australia
b
School of Geography Planning and Environmental Management, University of Queensland, St. Lucia Campus, Brisbane, QLD 4072, Australia
c
Barbara Hardy Institute, University of South Australia, Mawson Lakes Boulevard, Mawson Lakes, SA 5095, Australia

a b s t r a c t
Keywords: Place attachment has been researched extensively in the social and behavioural sciences over the past
Sense of place two decades. However, it is challenging for researchers to assimilate the mixed messages presented in
Place identity
the place attachment literature and to understand the multiple place attachment terms. In this study,
Community attachment
Scale development
a four-dimensional model of rural landholder attachments to their natural resource management region
Place attachment was conceptually and empirically developed with the aim of developing an integrated approach to the
Construct validity measurement of place attachment that clearly distinguishes between different elements of place
scholarship. A 29-item place attachment scale with the dimensions of place identity, place dependence,
nature bonding, and social bonding was tested on a random sample of rural landholders in the Adelaide
and Mount Lofty Ranges in South Australia (N ¼ 320). The majority of respondents were male (69.3%) and
the average age was 59 years. The scale was reduced to 20-items and then administered simultaneously
and in the same response format to rural landholders in two other areas of South Australia: the Northern
and Yorke region (N ¼ 664) and South Australian Murray-Darling Basin region (N ¼ 659). In both studies,
the majority of respondents were male (85%) and the average age was 55 years. Exploratory factor and
reliability analyses of Adelaide and Mount Lofty Ranges and Northern and Yorke datasets produced
a five-dimensional model of place attachment with high reliabilities. Social bonding divided into the
constructs of family bonding and friend bonding. The refined five-dimensional model was then exam-
ined for convergent validity, with moderate but significant correlations found between individual
attachment constructs and dependent variables expected to be related to the construct, such as place
identity and length of residence, and nature bonding and time currently spent in nature. We used
confirmatory factor analysis to test the goodness-of-fit of the South Australian Murray-Darling Basin
(SAMDB) dataset to the proposed five-dimensional model and then compared its fit to the traditional
two-dimensional model of place identity and place dependence. The five-dimensional model provided
moderate fit for the SAMDB data. We conclude with a discussion of the validity and reliability of the five-
dimensional model and its future role in place attachment research.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction Trentelman, 2009 for reviews). Trentelman (2009) provides an


eloquent review of the different strands of place scholarship, and
The construct of place attachment has received considerable highlights the tensions between the disciplines interested in: 1) the
attention in the environmental psychology and environmental socio-cultural dimensions of place, such as community attachment;
management literatures over the past two decades. However, it has 2) the biophysical dimensions of place, with emphasis on the
been challenging for researchers to assimilate the different place “setting or container”, and; 3) the integration of both socio-cultural
attachment concepts in these literatures and to understand its and natural setting dynamics within place attachment research.
jargon (see Patterson & Williams, 2005; Scannell & Gifford, 2010a; Researchers interested in the biophysical dimension of place have
placed considerable emphasis on a two-dimensional model of place
attachment comprising of place identity and place dependence (see
Williams, Patterson, Roggenbuck, & Watson, 1992). Place identity
* Corresponding author. Tel.: þ61 423 299 986; fax: þ61 8 8303 8590.
E-mail addresses: chris.raymond@enviroconnect.com.au (C.M. Raymond), greg. refers to those dimensions of self, such as the mixture of feelings
brown@uq.edu.au (G. Brown), delene.weber@unisa.edu.au (D. Weber). about specific physical settings and symbolic connections to place

0272-4944/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jenvp.2010.08.002
C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434 423

that define who we are (Proshansky, Fabian, & Kaminoff, 1983). Place in the Adelaide and Mount Lofty Ranges region of South Australia
dependence refers to the functional or goal-directed connections to using principal component and reliability analyses. The scale was
a setting; for example, it reflects the degree to which the physical refined and then included in a survey mailed to 1300 landholders in
setting provides conditions to support an intended use (Schreyer, each of the Northern and Yorke and South Australian Murray-
Jacob, & White, 1981). The validity and reliability of this two-dimen- Darling Basin (SAMDB) regions of South Australia. The construct
sional model is supported by several studies in the United States (e.g., validity and convergent validity of the refined scale was then tested
Bricker & Kerstetter, 2000; Jorgensen & Stedman, 2006; Kyle, Graefe, across all Northern and Yorke landholders. Finally, we tested the
& Manning, 2005; Williams & Vaske, 2003) and Australia (e.g., Brown goodness-of-fit of our conceptual model of place attachment against
& Raymond, 2007; Pretty, Chipuer, & Bramston, 2003). the SAMDB place attachment data. First, the theoretical basis for
However, this two-dimensional model overlooks important a three-pole conceptualization of place attachment is expounded.
connections to the natural and social environment which we Each of the proposed place attachment dimensions are defined and
propose are related constructs to the highly personalized attach- discussed.
ments associated with place identity and place dependence.
Gustafson (2001) provides a three-pole self-other-environment 1.1. Place attachment in personal context
theoretical framework to highlight these differences. The “self” pole
suggests some places are associated with highly personal meanings Place researchers have paid significant attention to the strength
related to life path, emotion, activity, and self-identification. The of individual or personal attachments to place. Most researchers
“other” pole suggests characteristics of the inhabitants influence have operationalised these personal place attachments using
place attachment. The “environment” pole, on the other hand, constructs of place identity and place dependence (Bricker &
reflects the role of the physical environment in developing place Kerstetter, 2000; Kyle, Absher, & Graefe, 2003; Moore & Graefe,
attachment. Recent empirical studies offer some support for this 1994; Williams et al., 1992). Although place identity and place
view. Techniques have been developed for measuring the extent of dependence are highly correlated, different relationships have been
individual’s connections to nature (the physical environment as it found between these constructs and dependent variables such as
exists without human beings), defined in terms of the extent of one’s recreation skill level (Bricker & Kerstetter, 2000), experience use
identity with nature (Clayton, 2003), affinity to nature (Kals, history (Hammitt, Backlund, & Bixler, 2004; Hammitt, Kyle, & Oh,
Schumacher, & Montada, 1999; Perkins, in press) or connectedness 2009), landscape values (Brown & Raymond, 2007) and visitors’
to nature (Schultz, 2001; Schultz, Shriver, Tabanico, & Khazian, 2004; spending preferences (Kyle, Graefe, Manning, & Bacon, 2004a,b;
Schultz & Tabanico, 2007). Researchers interested in social attach- Kyle et al., 2003). Researchers have also examined the relation-
ments have found that the social and geographic context of place ships between this two-dimensional model of place attachment
bonds requires greater consideration in place attachment research and pro-environmental behavior, with significant positive rela-
and have proposed constructs such as belongingness, and neigh- tionships found between place identity and willingness to engage
borhood attachment (Brown, Perkins, & Brown, 2003; Hay, 1998a; in pro-environmental behaviors (Stedman, 2002; Vaske & Kobrin,
Hidalgo & Hernandez, 2001; Kyle & Chick, 2007). 2001) and place-protective actions (Devine-Wright & Howes, in
Recent studies have developed psychometric scales for press).
measuring individuals’ attachments to place based upon their
interactions with both the community and the natural environment. 1.2. Place attachment in community context
Brehm, Eisenhauer, and Krannich (2006) examined the connection
between two dimensions of community attachment named social A second area of place scholarship highlights the importance of
and natural environment. Social attachment and natural environ- the role of the community in forming place attachments. This social
ment were distinct and separate dimensions of place attachment context has been operationalised in a variety of ways, including
and they had significant associations with environmental concern. community attachment, belongingness, rootedness, and familiarity.
Scannell and Gifford (2010b) developed and tested a place attach- The term ‘community’ is based on a systemic model of connection
ment scale comprising of natural and civic attachment dimensions. between residents and their communities (Kasarda & Janowitz,
The physical aspects were represented by natural environment 1974). This systemic model posits that community attachment is
attachment and the social aspects were represented by civic strongly related to individual connections to local social networks
attachment. Whilst both approaches clearly articulate the physical (bonds) and the interactions which occur with them. Kasarda and
and social dimensions of place attachment, they overlook how the Janowitz (1974) compared community attachments based on
physical and social dimensions of place interact with the highly a linear model of population size and density vs. the systemic model
personalized attachments related to place identity and place of social connection. They found that the social connectedness that
dependence (see Williams et al., 1992; Williams & Vaske, 2003). developed between people over the course of their residence in
Researchers are now calling for new integrated models which a given place was a more powerful predictor of community
consider the interactions between place as a natural and social attachment than population size or density of the community
setting and how the setting supports one’s self-identity or func- population. More recent studies build upon the systemic model in
tional goals (Davenport, Baker, Leahy, & Anderson, 2010; Sampson & different settings. Perkins and Long (2002) referred to these social
Goodrich, 2009). Linking nature-based and social attachments with connections in place as social bonding or the feelings of belonging-
more traditional measures of place attachment, such as place ness or membership to a group of people, as well as the emotional
identity and place dependence, may begin to address this need. connections based on shared history, interests or concerns.
In this study, we conceptualize and empirically examine a four- Researchers interested in the socio-cultural dimension of place
dimensional model of place attachment among three samples of argue that the two-dimensional model of place attachment is
rural landholders in regional South Australia for use in natural or inadequate in addressing place attachment and highlight the need
rural land-use contexts. The model includes place identity and place to consider the social context of place bonds, including the social
dependence (personal connections to place), nature bonding interaction through which place meanings are mediated (Hay,
(connections to the natural environment), and social bonding 1998a; Hidalgo & Hernandez, 2001; Kyle & Chick, 2007; Sampson
(connections to the community in place). The validity and reliability & Goodrich, 2009). Hay (1998b) found that ancestral and cultural
of the item-scales were examined among rural landholders who live connections are important to the development of a rooted
424 C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434

attachment to place. Fishwick and Vining (1992, p. 57) studied the individual and the natural world. Kals and colleagues showed that
experiences of college students in Illinois state parks and found that emotional affinity can be distinguished from its cognitive equivalent
students’ connections to these sites were an assortment of “setting, of ‘interest in nature’ and that emotional affinity is a powerful
landscape, ritual, routine, people, [and] personal experiences”. predictor of nature-protective behavior. More recently, researchers
Hidalgo and Hernandez (2001) observed that social attachments have developed a 7-point connectedness to nature scale (Dutcher,
were stronger than setting attachments within houses, neighbor- Finley, Luloff, & Johnson, 2007; Mayer & Frantz, 2004) with high
hoods, and cities. Kyle, Bricker, Graefe, and Wickham (2004) validity and reliability. Gosling and Williams (in press) used this
examined the relationships between involvement and place scale to examine the relationships between place attachment and
attachment for hikers along the Appalachian Trail. As recreationists’ connectedness to nature. Significant, but weak correlations were
social ties to the setting grew so too did their emotional attachment identified between these three constructs (r ¼ .25, p < .01). They also
(in terms of place identity) to the setting. Kyle and Chick (2007) found significant, but weak correlations between the desire to
highlighted the importance assigned to place experiences shared protect remnant vegetation and affinity to nature (r ¼ .27, p < .01).
with family and close friends on the development of place meanings A separate line of work has focused on the cognitive dimensions
within agricultural fairs. of environmental identity. Dutcher (2000) and Schultz (2001) used
Social attachments to places have been explained using a variety a modified version of the Inclusion of Other in Self scale (IOS) which
of constructs that in some instances have overlapping meanings. was developed by Aron and colleagues (Aron, Aron, & Smollan,
These constructs include: place belongingness where people feel 1992; Aron, Aron, Tudor, & Nelson, 1991) to measure the extent to
a ‘membership’ to an environment (Mesch & Manor, 1998; Milligan, which an individual includes nature within his or her cognitive
1998), place rootedness which refers to a very strong bond to home representation of self, in what is otherwise referred to as
(Hay, 1998b; Tuan, 1980), place familiarity defined as pleasant a connectedness to nature scale. Schultz and colleagues suggest that
memories, achievement memories, and environmental images individuals hold implicit cognitive associations between them-
associated with places (Roberts, 1996), and neighborhood attachment selves and the natural environment which influence their envi-
which captures one’s emotional connection to his or her immediate ronmental concerns (Schultz et al., 2004; Schultz & Tabanico, 2007).
surroundings (Brown et al., 2003). Hammitt and colleagues The recreational literature has also examined how the charac-
(Hammitt, Backlund, & Bixler, 2006) considered belongingness, teristics of the natural environment influence place attachment. A
rootedness, and familiarity (as well as place identity and place number of studies have examined relationships among place
dependence) in their design of a new place attachment scale. attachment, human use or experience of the leisure activity and the
Exploratory and reliability analyses indicated that these constructs characteristics of the natural environment. Hammitt, Backlund, and
were three separate and reliable dimensions of recreational place Bixler (2004, 2009) found a negative relationship between place
bonding. However, there were strong correlations between place dependence and experience use history. The authors suggest that
identity and place belongingness, and place identity explained most the availability of substitute natural settings moderate the degree of
of the variance in place attachment. Other researchers have inte- place dependence for any specific place. Kyle and colleagues
grated these three concepts under the term of social bonding. Kyle examined the relationships among place attachment, leisure
et al. (2005) developed a scale for measuring social bonding in activity involvement, and the characteristics of specific natural
terms of the social relationships which occur and are maintained in environments. In a study of hikers, boaters and anglers, the strength
recreational settings along an Appalachian Trail. Whilst their social of place attachment was related to specific motivations to be
bonding scale had low validity and reliability (i.e. low factor loadings involved in leisure activities and the type of natural setting in which
and t-values, low internal consistency), it explicitly addressed the this activity was undertaken (Kyle et al., 2004). When examining the
role of community as the intermediary between individuals and the relationships among place attachment, activity involvement, and
natural environment, as proposed by Gustafson (2001) and rein- setting density, place identity and place dependence were signifi-
forced by Sampson and Goodrich (2009). Kyle and colleagues also cant predictors of setting density, but to different degrees (Kyle
encourage the consideration of additional items of social bonding to et al., 2004a). Respondents scoring high on the place identity
improve the validity and reliability of this construct. Such refinement dimension were more inclined to feel crowded in the setting than
is one focus of this paper. those respondents with high place dependence. Similarly, respon-
dents scoring high on the place identity dimension were more likely
1.3. Place attachment in natural environment context to negatively appraise user impact on the natural environment
(Kyle, Graefe, Manning, & Bacon, 2004b).
Another area of place scholarship highlights the importance of A separate recreation and leisure literature has emphasized the
connections to the natural environment without human beings, role of both communities and physical settings in the formation of
which we refer to as nature bonding. Nature bonding has been place attachment. Stedman (2003, p. 673) encourages an “empirical
operationalised in a variety of ways, including environmental investigation between aspects of the natural environment, and its
identity, emotional affinity towards nature, and connectedness to meanings.” For example, he and colleagues explored the place
nature. Environmental identity theory (Clayton, 2003), which has its meanings and attachments of visitors to Jasper National Park,
origin in “deep ecology”, describes people’s experiences with nature Alberta, using photography. They observed that the meanings
as integral to one’s sense of self (Bragg, 1996; Fox, 1990; Nash, 1990; visitors associated with the park were driven by complex interac-
Zimmerman, Callicott, Sessions, Warren, & Clark, 1993). This theory tions between ecological and socio-cultural factors (Stedman,
also relates to the biophilia hypothesis (Kellert & Wilson, 1993; Beckley, Wallace, & Ambard, 2004). Sampson and Goodrich
Wilson, 1984) which claims that human’s connection to nature (2009) support Stedman’s (2003) view that there are interactions
and well-being is strongly influenced by their relationships with the between communities and the natural, physical setting. Commu-
surrounding natural world. Research on environmental identity, nities have particular characteristics which bind them to specific
otherwise termed emotional affinity towards nature (Kals et al., 1999; locales, but they are also bounded by particularities of the natural
Perkins, in press) focuses on individual emotional connections to environment. They also argue that communities provide the filter
nature, such as a ‘love of nature’. Unlike the definition of place through which individuals can develop identity with place.
identity presented in the leisure and recreational sciences, it has Coasters from the West Coast of the South Island, New Zealand,
a much greater emphasis on the connections between the shared strong anti-government sentiments about the loss of
C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434 425

forestry from their region which reinforced notions of identity and 1.4. An integrated model of place attachment e personal,
distinctiveness of place; however, they also described the capacity community, and natural environment connections
of the natural environment (e.g., amount of rainfall) to influence
the practices of people. Their study builds upon the concepts of The preceding literature review highlights that there are
place distinctiveness and continuity (Twigger-Ross & Uzzell, 1996). multiple, overlapping meanings of place attachment and its oper-
Distinctiveness theory suggests that individuals develop strong ational measures that have developed in different disciplines, such
identity with place because it facilitates “distinctiveness” from as social psychology, environmental psychology, and community
other place or affirms the uniqueness of the group. Continuity sociology. This section provides a model of place attachment that
theory suggests that the physical environment is conceptualized by attempts to integrate conceptually and empirically the many ways
individuals as a reference for past action and experience. Past which place attachment has been examined (Fig. 1). Researchers
action and experience in place plays an important role in main- interested in the biophysical dimension of place have either
taining individual and group identity. focused on: 1) the personal context of place attachment, specifically
We acknowledge that place attachment researchers have examining the highly individualized attachments of place identity
considered connections to the physical environment beyond and place dependence (Table 1), or; 2) the natural environment
natural settings. Manzo’s (2003) review of the place attachment context, examining the related constructs of environmental iden-
literature indicates that individual connections to places are tity, emotional affinity, and connectedness to nature, which we
dynamic and encompass a broad range of physical settings, such as propose can be considered using the overarching construct of
residential, recreational, and leisure settings. Kaltenborn (1997) nature bonding (Fig. 1) defined in Table 1. Nature bonding may not
investigated the place meanings of recreational homes. He be applicable for the measurement of place attachment in urban
defined two dimensions of place attachment: nature-culture settings, but it is relevant to natural and rural land-use contexts,
which relates to the place as both a natural environment and which is the focus of this paper. Researchers in the socio-cultural
a cultural landscape and family-social concerning family life at the dimension of place have largely focused on the community context
recreational home. Gustafson (2001) asked participants to identify of place attachment which has been examined using the constructs
attachments to their community/village, city, and country. Build- of neighborhood attachment, belongingness, and familiarity. We
ings and nature were coded under “physical environment.” Low propose these constructs form part of a larger construct of social
(2000) showed the importance of town plazas for Costa-Ricans. bonding (Fig. 1, Table 1).
Manzo (2005) identified the place experiences of residents of New Similarities and differences exist between the four hypothesized
York City. Individuals were attached to different types and scales of dimensions of place attachment presented in Fig. 1. Both place
physical settings. Some attachments related to beaches, parks, and identity and nature bonding include emotional connections about
lakes; others were churches, bars, Laundromats, and airports. physical settings. Following continuity theory, place identity is
Rollero and De Piccoli’s (2010) found that participants with high strongly related to residential history (see Hay, 1998a; Twigger-
place attachment described the city in a positive way, such as Ross & Uzzell, 1996); however, we argue that nature bonding is
beautiful and welcoming, whereas participants with low place unrelated to residential history within a given place. Rather, it is
attachment described the city negatively, such as the presence of related to experience or time spent in the natural environment. For
pollution. Whilst these studies show the importance of place example, a landholder who has lived in a region for a long period of
attachment to built environments, they do not provide scales for time may strongly identify with that region, but may have weak
measuring the intensity of these attachments across multiple nature bonds because he/she has not spent time visiting the natural
place contexts, including personal, community and natural envi- areas found in that region. We also recognize that there are inter-
ronment contexts. connections between nature bonding and social bonding. Nature,

Personal Context
Place identity
Place dependence
Rootedness

PLACE IDENTITY &


DEPENDENCE

Community Context Natural Environment Context


Neighborhood SOCIAL NATURE Connectedness to nature
attachment BONDING BONDING Environmental identity
Belongingness Affinity to nature
Familiarity

Fig. 1. Three-pole and four-dimensional conceptual model of place attachment which will be subject to empirical examination in this paper. Note: Place identity and place
dependence are included in the same pole (personal context) because they are related to highly personalized connections to place which are either symbolic (identity) or functional
(dependence) in nature. They may form through memories, experiences or events which are unrelated to the wider community or to the natural environment.
426 C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434

Table 1
Operational definitions of the four proposed dimensions of place attachment.

Pole Construct Definition Supporting literature


Personal Place identity Those dimensions of self, such as the mixture of feelings about Proshansky et al., 1983; Williams et al., 1992;
specific physical settings and symbolic connections to place, Williams & Vaske, 2003.
that define who we are.
Place dependence Functional connection based specifically on the individual physical Schreyer et al., 1981; Williams et al., 1992;
connection to a setting; for example, it reflects the degree to Williams & Vaske, 2003.
which the physical setting provides conditions to support an
intended use.
Community Social bonding Feelings of belongingness or membership to a group of people, Kasarda & Janowitz, 1974; Hay, 1998a; Perkins & Long, 2002;
such as friends and family, as well as the emotional connections Hidalgo & Hernandez, 2001; Stedman et al., 2004;
based on shared history, interests or concerns. Kyle & Chick, 2007; Sampson & Goodrich, 2009;
Trentelman, 2009.
Environment Nature bonding Implicit or explicit connection to some part of the non-human Kals et al., 1999; Clayton, 2003; Schultz, 2001;
natural environment, based on history, emotional response or Schultz et al., 2004.
cognitive representation (e.g., knowledge generation).

or the physical setting, provides the container for social experi- bonding will emerge as a separate dimension of place attachment
ences and the bonds which form through these experiences. with high construct validity and reliability. Another important
Our model is conceptually similar to the tripartite model of component of our model is that each individual is situated within
place attachment presented by Scannell and Gifford (2010a), but a place that has a personal, community, and natural environment
there are some important differences. Both models recognize that context. It is hypothesized that the strength and nature of our
connections to place can be based upon social or environmental attachments to place change in accordance with the personal,
attributes of place. Scannell and Gifford (2010b) separated the environmental, and community context in which landholders
social and environmental attributes of place into civic attachment operate. For example, the strength of social bonds that ‘landholder A’
and natural environment place attachment, respectively, and we has with ‘natural resource management region A’ may be weaker
divide these attributes into social bonding and nature bonding, than the social bonds ‘landholder B’ has with ‘natural resource
respectively. We explicitly separate attributes of the person into management region B’.
place identity and place dependence and the attributes of the Our approach involved three mail-based surveys across three
community into social bonding, but these object differences are distinct populations: The Adelaide and Mount Lofty Ranges, South
implicit within the natural environment place attachment and civic Australia, Northern and Yorke region, South Australia, and the
place attachment dimensions of the tripartite model. Our model South Australian Murray-Darling Basin (SAMDB). Semi-structured
does not consider how affect, cognition, and behavior are man- interviews conducted with landholders in the Eyre Peninsula
ifested in place attachment, whereas the tripartite model considers region informed the development of the original place attachment
the psychological process as a separate dimension of place scale (Raymond, 2009). This scale is first tested in the Adelaide and
attachment. We empirically test our model at the regional scale, Mount Lofty Ranges region (study 1). Results from this survey were
whereas Scannell and Gifford (2010b) test their model at the then used to inform and refine the survey instrument in the
community scale. Despite the differences, both models make an Northern and Yorke and SAMDB regions (study 2).
important contribution to the place attachment literature.
Our integrated model is based on a positivistic or psychometric 2. Study 1: dimensionality of the place attachment scale
approach to place scholarship which involves hypothesis testing
researcher-defined variables (see Trentelman, 2009). We develop The first study aimed to examine whether the place attachment
a series of scale items to measure four dimensions of place attach- instrument could reliably distinguish four place attachment
ment and then test hypotheses about the validity and reliability of dimensions (place identity, dependence, nature bonding, and social
the place attachment model. Whilst we adopt a positivistic method, bonding) in the Adelaide and Mount Lofty Ranges region. The
there are also qualitative approaches to place scholarship which have Adelaide and Mount Lofty Ranges covers an area of 5355 km2 and
an important role in place attachment research. Qualitative research includes Adelaide, the capital city of South Australia. The pop-
has focused on place-specific, unique meanings (e.g., Devine-Wright ulation of the region is approximately 1.1 million, with the majority
& Howes, in press; Gustafson, 2001; Manzo, 2003, 2005; Stokowski, of population found in the Adelaide metropolitan area. The region
2002) using social constructionist or applied natural resource social is known for multiple land uses (Adelaide and Mount Lofty Ranges
sciences approaches. Social constructionism is concerned with how Natural Resources Management Board, 2008).
place is socially constructed, by whom, and with what interests,
while the applied approach attempts to convince environmental 2.1. Method
managers to include meaning-oriented dimensions of “place” in
management decisions (see Trentelman, 2009). 2.1.1. Respondents and procedure
The remainder of this paper aims to empirically test our In 2009, 400 mail-based surveys were administered to a random
conceptual model. Following other studies (e.g., Williams & Vaske, sample of rural landholders in the Adelaide and Mount Lofty Ranges
2003), it is expected that place identity and dependence will region, South Australia. Surveys were administered using a modified
emerge as two separate dimensions of place attachment. Consistent Tailored Design Method (Dillman, 2007), which entailed sending
with the connectedness to nature theory (Schultz et al., 2004), we out a hand signed introductory letter (at time t), a first survey packet
hypothesize that nature bonding will emerge as a separate dimen- comprised of the survey and reply paid envelope (t þ one week),
sion of place attachment with high construct validity and reliability. reminder postcard (t þ three weeks), and second survey packet to
Further, in line with the community attachment theory presented by non-respondents to the first round of mailing (t þ five weeks). All
Kyle and colleagues (Kyle et al., 2005), we hypothesize that social survey packets included two postal stamps as an incentive to
C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434 427

participate. The survey consisted of three parts: 1) attachments to In the survey, scale items were preceded by the following
the Adelaide and Mount Lofty Ranges; 2) landholders’ personal directions: “Below are a set of statements about your attachment to
values, and; 3) socio-demographics. In total, 320 respondents the Adelaide and Mount Lofty Ranges. Please indicate your level of
completed the survey (77.8% valid response) of which 69.3% were agreement or disagreement with each statement”. Items were
male and 30.7% were female. The mean age was 59 years (SD ¼ 11.5). presented on a 5-point Likert scale where “1 ¼ Strongly Disagree”,
Respondents were older and had completed a higher level of “5 ¼ Strongly Agree”, and “3 ¼ Neither Agree or Disagree”.
education (42.9% with tertiary or postgraduate degree) than would
be expected based on comparable regional statistics collected by the 2.1.3. Analyses
Australian Bureau of Statistics (ABS, 2006). Further, proportionately We first created summated scales and analysed the item-to-total
more males than females completed the survey when compared to and inter-item correlations between the place attachment scale
regional ABS statistics. items. None of the item-to-total and inter-item correlations were
less than .5 and .3, the thresholds for deletion (Hair, Black, Babin,
2.1.2. Place attachment measures Anderson, & Tatham, 2006). The items’ ratings were then sub-
Part 1 of the survey instrument consisted of 29 place attachment jected to exploratory factor analysis (principal component analysis
statements, taken from studies previously demonstrating high with varimax rotation) to examine initial support for the place
internal consistency, in addition to items developed by the attachment items and proposed concept dimensions. The criteria
researchers. We measured place identity and place dependence used by Hammitt et al. (2006) was followed in forming factors, i.e.:
using 11 scale items developed and validated by Williams and Eigenvalues  1.0; factor loadings.40; item loadings on more than
colleagues (Williams et al., 1992; Williams & Roggenbuck, 1989; one factor had to differ by .10 in loading to be retained.
Williams & Vaske, 2003). Items concerning nature bonding were
developed according to descriptors of affiliation and connectedness 2.2. Results
with nature (Kals et al., 1999; Schultz, 2001). An additional item
was also included based on results from semi-structured interviews 2.2.1. Dimensionality and internal consistency of the place
with 30 rural landholders in the Eyre Peninsula region of South attachment scale
Australia (Raymond, 2009). Items concerning social bonding were Principal component and reliability analyses of responses to the
developed based on measures proposed by Kyle and colleagues 29-point place attachment scale in the Adelaide and Mount Lofty
(Kyle et al., 2005) in addition to connections to family and friends Ranges indicated five underlying dimensions with moderate reli-
which emerged during the thematic analysis of interviews con- abilities (Table 2, Cronbach alphas > .72). These five dimensions
ducted in the Eyre Peninsula region (Raymond, 2009). Both the comprised 19 items. Eight of the remaining 10 items were excluded
nature bonding and social bonding scale items were examined for because of high cross-loadings. A further two items loaded on a sixth
face validity by a panel of researchers and subsequently refined by dimension interpreted to be unrelated to place attachment (the
the authors prior to survey administration. importance of agriculture in the Adelaide and Mount Lofty Ranges).

Table 2
Principal component analysis of place attachment in the Adelaide and Mount Lofty Ranges region.

Item Factor Grand Item Standard Cronbach


loading mean mean deviation alpha
Place identitya (Eigenvalue ¼ 6.95; variance explained ¼ 36.6%) 3.95 .87
The Adelaide and Mount Lofty Ranges is very special to me .72 3.97 .73
The Adelaide and Mount Lofty Ranges mean a lot to me .75 4.10 .65
I am very attached to the Adelaide and Mount Lofty Ranges .76 4.19 .74
I identify strongly with the Adelaide and Mount Lofty Ranges .74 4.07 .77
Living in the Adelaide and Mount Lofty Ranges says a lot about who I am .57 3.70 .88
I feel the Adelaide and Mount Lofty Ranges is a part of me .58 3.67 .92
Nature bondingb (Eigenvalue ¼ 2.11; variance explained ¼ 11.1%) 3.83 .81
When I spend time in the natural environment in the Adelaide and Mount Lofty Ranges, .84 3.83 .87
I feel a deep feeling of oneness with the natural environmentc
I would feel less attached to the Adelaide and Mount Lofty Ranges if the native plants and .74 4.11 .90
animals that live here disappeared
I learn a lot about myself when spending time in the natural environment in the .73 3.50 .89
Adelaide and Mount Lofty Ranges
I am very attached to the natural environment in the Adelaide and Mount Lofty Ranges .69 3.89 .76
Place dependencea (Eigenvalue ¼ 1.60; variance explained ¼ 8.5%) 3.47 .81
No other place can compare to the Adelaide and Mount Lofty Ranges .78 3.21 1.08
I would not substitute any other area for the activities I do in the Adelaide and Mount Lofty Ranges .73 3.11 .96
Doing my activities in the Adelaide and Mount Lofty Ranges is more important to me .61 3.80 .93
than doing them in any other place
The Adelaide and Mount Lofty Ranges is the best place for the activities I like to do .60 3.77 .82
Family bondingb (Eigenvalue ¼ 1.33; variance explained ¼ 7.0%) 3.11 .72
I live in the Adelaide and Mount Lofty Ranges because my family is here .87 3.18 1.35
My relationships with family in the Adelaide and Mount Lofty Ranges are very special to me .82 3.63 1.15
Without my relationships with family in the Adelaide and Mount Lofty Ranges, I would probably move .65 2.54 .97
Friend bonding/belongingnessb (Eigenvalue ¼ 1.03; variance explained ¼ 5.4%) 3.36 .83
Belonging to volunteer groups in the Adelaide and Mount Lofty Ranges is very important to me .90 3.30 .94
The friendships developed by doing various community activities strongly connect .87 3.41 .92
me to the Adelaide and Mount Lofty Ranges

Place attachment items were measured on a scale ranging from 1 ¼ Strongly Disagree, 2 ¼ Disagree, 3 ¼ Neither Agree or Disagree, 4 ¼ Agree, 5 ¼ Strongly Agree.
a
all scale items which measure this dimension were developed and validated in previous studies (e.g., Williams et al., 1992; Williams & Vaske, 2003).
b
all scale items which measure this dimension are newly developed, unless otherwise specified.
c
scale item adapted from Kals et al. (1999).
428 C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434

The two previously validated dimensions of place identity and (1023 km2) agricultural production (South Australian Murray-
place dependence (e.g., Williams & Vaske, 2003; Brown & Darling Basin Natural Resources Management Board, 2009).
Raymond, 2007) were identified in this study. Six items previ-
ously found to be related to place identity loaded on a place identity 3.1. Method
dimension and a further four items found to be related to place
dependence loaded on a place dependence dimension with high 3.1.1. Respondents and survey procedure
reliability (Table 2, Cronbach alpha > .81). The scale items of “living In 2010, 1300 mail-based surveys were administered to
in the Adelaide and Mount Lofty Ranges says a lot about who I am” a random sample of rural landholders in each of the Northern and
and “I feel the Adelaide and Mount Lofty Ranges is a part of me” Yorke and SAMDB regions, using the same survey administration
loaded less strongly on the place identity dimension than other method as study 1. The two regional surveys were administered
measures of place identity; however, they were still significant and concurrently and contained the same questions. Research areas
strong contributors to overall place identity (factor loadings > .58). addressed in the survey included rural landholders’ place attach-
Nature bonding also emerged as a separate dimension of place ments, native vegetation conservation behavior, environmental
attachment with high construct validity and reliability. Four items concerns, human values, and attitudes toward government assis-
loaded strongly on this dimension (Table 2, Factor loadings > .69) tance programs for the management of natural resources. Part 3 of
with equal reliability to that of place dependence (Cronbach the survey instrument consisted of the refined 20 place attachment
alpha ¼ .81). Paired samples t-tests indicate that nature bonding items developed in study one. All items were randomly listed and
was significantly stronger for landholders in the Adelaide and presented in an identical response format as study one.
Mount Lofty Ranges than place dependence (t ¼ 7.88, p < .05). A total of 664 Northern and Yorke rural landholders returned
However, our hypothesis that social bonding would emerge as a completed survey (53.1% survey response after removing non-
a separate dimension of place attachment with high construct deliverable surveys). The majority of respondents were male (85.9%)
validity and reliability was not supported. The five scale items and the mean age was 55 years (SD ¼ 11.90). Over half of SAMDB
hypothesized to be related to social bonding divided into two rural landholders completed a survey (51.7% survey response)
dimensions (Table 2, factor loadings > .65) that relate to connec- providing a sample of 659. The majority of respondents were again
tions to place based upon family relationships (termed family male (86.9%) and the mean age was 55 years (SD ¼ 11.89). Response
bonding) and connections to place based upon other friendships rate from females was lower than expected given ABS (2006) data
(termed friend bonding). The item “Belonging to volunteer groups in indicates that females represent 18.2% of all farm owners or
the Adelaide and Mount Lofty Ranges is very important to me” managers in the Northern and Yorke region and 25.9% of all farm
appears related to the dimension of place belongingness discussed owners or managers in the South Australian Murray-Darling Basin.
by Hammitt et al. (2006). For this reason, we have termed the
dimension friend bonding/belongingness. 3.1.2. Analyses
3.1.2.1. Construct validity. We repeated principal component anal-
3. Study 2: refinement of the place attachment scale ysis with a varimax rotation to test the dimensionality of the
refined place attachment scale. Factor loading thresholds of >.50
In the second study, we tested the construct and convergent were used to discriminate between components. Reliability anal-
validity of a revised place attachment scale in the Northern and ysis was used to test the cohesiveness of the sub-scales. We also
Yorke region. We then conducted a confirmatory test of this revised compared the amount of variance in overall place attachment
scale in the South Australian Murray-Darling Basin (SAMDB) region. explained by two-dimensional place identity and place depen-
Hair et al. (2006) suggests that a construct should be reflected by dence model with our five-dimensional model. Similar or
a minimum of three items, preferably four. For this reason, we improved explanatory power will provide additional support for
included two extra friend bonding items in a mail-survey to land- our five-dimensional conceptualization.
holders in each of the Northern and Yorke and SAMDB regions of Another way to test the dimensionality of the place attachment
South Australia. The new items were: 1) “The friendships devel- scale is to compare the mean differences among place identity,
oped through volunteer activities in the Northern and Yorke region place dependence, nature bonding, family bonding, and friend
are very important to me”, and; 2) “The friendships developed bonding scores. To identify significant differences, paired samples
through sporting activities in the Northern and Yorke region are t-tests were run between the five dimensions. Significant difference
very important to me.” To further increase the reliability of the among mean response scores would support the hypothesis that
nature bonding scale, we included the additional item of “When I there are five rather than four dimensions of place attachment.
spend time in the natural environment in the Northern and Yorke Another measure of construct validity is the sensitivity of the
region, I feel at peace with myself.” place attachment scales for measuring attachments to different
The Northern and Yorke region covers an area of 34,500 km2 and places. Williams and Vaske (2003) suggest that place attachment
includes the townships of Hawker to the north, the Yorke Peninsula, should differentiate how one feels about various places. We used
the Northern Mount Lofty Ranges and the Southern Flinders Ranges. analysis of variance to examine the mean differences in place
The region supports a population of 89,000, with the major urban identity, dependence, nature bonding, family bonding, and friend
centres being the cities of Port Pirie, Port Augusta, Clare, and ‘The bonding scores among the three study regions of the Adelaide and
Copper Triangle’ towns of Kadina, Moonta, and Wallaroo. Dryland Mount Lofty Ranges, Northern and Yorke and SAMDB.
farming (crops and livestock) is the dominant land-use of the region
(Northern and Yorke Natural Resources Management Board, 2009). 3.1.2.2. Convergent validity. To test convergent validity, we exam-
The SAMDB region spans an area of just over 56,000 km2 and covers ined the expected correlations between the place attachment
the lower basin area of the Murray-Darling system. Around 81,000 dimensions and the variables of length of residence, length of
people reside in the region. The region includes the Coorong (a farming, and time currently spent in nature. Additionally, we ran
100 km long coastal wetland). The River Murray, a critical source of correlations between the place attachment dimensions and length
fresh water for South Australia, flows through the region entering of time the family has owned or operated the farm, generations the
the Ramsar listed lower lakes and Murray mouth. Land use in the family has been involved in farming, and longest period of time an
region is dominated by dryland (23,304 km2) and irrigated immediate family member has lived in the region. We also ran
C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434 429

a three-way full factorial model to determine potential interaction Yorke region (t > 3.10, p < .01), with the exception of place depen-
effects between these respondent variables and the five place dence and family bonding (p > .05). The largest mean differences
attachment dimensions. exist between place identity and place dependence (t ¼ 22.09,
p < .001) and place identity and nature bonding (t ¼ 13.97, p < .001).
3.1.2.3. Confirmatory test. We used confirmatory factor analysis to Significant, but weak differences exist between place dependence
determine the extent to which the SAMDB landholder place and family bonding (t ¼ 1.97, p < .05).
attachment data fit the five-dimensional model of place attach- We used principal component analysis to compare the explana-
ment. In the hypothesized model, identity, dependence, nature tory power of the two-dimensional, four-dimensional (place
bonding, family bonding, and friend bonding were latent variables identity, dependence, nature bonding, family bonding) and five-
and the corresponding scale items were the indicator measures. dimensional models of place attachment. The two-dimensional
Maximum likelihood estimation was used to estimate the model fit. model explained 66.6%, the four-dimensional model explained 66.4%
A variety of fit indices were used to test model fit, including the and the five-dimensional model explained 67.9% of the variance in
Comparative Fit Index (CFI), Normed Fit Index (NFI), and Root Mean overall landholder attachment to the Northern and Yorke region.
Square Error of Approximation (RMSEA). We then conducted post-
hoc modifications to improve the fit of the model using the 3.2.2. Convergent validity
modification indices provided in structural equation modelling Previous studies indicate that place attachment develops over
(SEM) with AMOS Version 17.0. To compare the fit of the 2- time, and is linked to length of residence, experience use history and
dimensional and 5-dimensional place attachment models, we used knowledge of a particular region (Altman & Low, 1992; Bricker &
the Akaike Information Criterion (AIC; Akaike, 1987) and the Kerstetter, 2000; Brown & Raymond, 2007; Hammitt et al., 2009;
parsimony measure named the Expected Cross-validation Index Moore & Graefe, 1994; Williams et al., 1992). Further, a study by
(ECVI). Kals et al. (1999) suggests significant positive relationships between
time spent in nature and connectedness with nature. To test the
3.2. Results convergent validity of our place attachment scale, we examined the
relationships between the five place attachment dimensions and the
3.2.1. Dimensionality and internal consistency of the refined place respondent variables of length of residence, perceived knowledge of
attachment scale the region, years involved in farming, and time spent in nature
Despite refining the social bonding scale items, principal (Table 5). We first ran a three-way full factorial model to determine
component analysis of the 20-item scale continued to show family potential interaction effects between these respondent variables
bonding and friend bonding loading on separate dimensions (Table and the five place attachment dimensions; however, no interactions
3, factor loadings > .64). Paired samples t-tests (Table 4) indicate were statistically significant. The length of residence, length of
significant differences between mean responses to family bonding farming, and time spent in nature variables explained 12.8% of the
and friend bonding across all Northern and Yorke landholders variance in place identity, 9.1% of the variance in place dependence,
(t ¼ 6.52, p < .001). Significant mean differences also exist between 16.8% of the variance in nature bonding, 12.3% of the variance in
the other four dimensions of place attachment in the Northern and family bonding, and 6.5% of the variance in friend bonding.

Table 3
Principal component analysis of place attachment in the Northern and Yorke region.

Item Factor Grand Item Standard Cronbach


loading mean mean deviation alpha
Place identitya (Eigenvalue ¼ 6.12; variance explained ¼ 41.4%) 3.80 .91
The Northern and Yorke region means a lot to me .75 3.79 .74
I am very attached to the Northern and Yorke region .72 3.76 .81
I have a lot of fond memories about the Northern and Yorke region .71 4.00 .68
The Nothern and Yorke region is very special to me .69 3.71 .77
I identify strongly with the Northern and Yorke region .70 3.77 .84
I feel the Northern and Yorke region is a part of me .66 3.77 .79
Nature bondingb (Eigenvalue ¼ 2.15; variance explained ¼ 9.8%) 3.55 .86
When I spend time in the natural environment in the Northern and Yorke region, .72 3.52 .88
I feel a deep feeling of oneness with the natural environmentc
I would feel less attached to the Northern and Yorke region if the native plants and animals that live here disappeared .77 3.65 .91
I learn a lot about myself when spending time in the natural environment in the Northern and Yorke region .68 3.25 .80
I am very attached to the natural environment in the Northern and Yorke region .68 3.55 .78
When I spend time in the natural environment in the Northern and Yorke region, I feel at peace with myself .66 3.79 .86
Place dependencea (Eigenvalue ¼ 1.33; variance explained ¼ 6.1%) 3.31 .85
I get more satisfaction out of living in the Northern and Yorke region than any other place .56 3.56 .92
No other place can compare to the Northern and Yorke region .82 2.96 .91
I would not substitute any other area for the activities I do in the Northern and Yorke region .79 3.07 .86
Doing my activities in the Northern and Yorke region is more important to me than doing them in any other place .55 3.45 .86
The Northern and Yorke region is the best place for the activities I like to do .64 3.50 .84
Family bondingb (Eigenvalue ¼ 1.12; variance explained ¼ 5.1%) 3.39 .70
I live in the Northern and Yorke region because my family is here .79 3.66 1.11
My relationships with family in the Northern and Yorke region are very special to me .64 3.79 .99
Friend bondingb (Eigenvalue ¼ 1.03; variance explained ¼ 4.7%) 3.63 .65
The friendships developed through volunteer activities in the Northern and Yorke region are very important to me .79 3.67 .85
The friendships developed through sporting activities in the Northern and Yorke region are very important to me .67 3.58 1.00

Place attachment items were measured on a scale ranging from 1 ¼ Strongly Disagree, 2 ¼ Disagree, 3 ¼ Neither Agree or Disagree, 4 ¼ Agree, 5 ¼ Strongly Agree.
a
all scale items which measure this dimension were developed and validated in previous studies (e.g., Williams & Vaske, 2003).
b
all scale items which measure this dimension are newly developed, unless otherwise specified.
c
scale item adapted from Kals et al. (1999).
430 C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434

Table 4
Mean differences between place attachment construct scores in the Northern and Yorke region.

Construct 1 Construct 2 Paired Paired mean Standard t


correlation difference deviation
Place identity Nature bonding .70 .30 .51 13.97 ***
Place identity Place dependence .69 .50 .52 22.09 ***
Place identity Family bonding .23 .42 .86 11.42 ***
Place identity Friend bonding .41 .20 .79 5.89 ***
Place dependence Nature bonding .64 .20 .58 7.99 ***
Place dependence Family bonding .19 .07 .91 1.97 ns
Place dependence Friend bonding .32 .32 .87 8.61 ***
Nature bonding Family bonding .12 .12 .93 3.10 **
Nature bonding Friend bonding .26 .12 .90 3.09 **
Family bonding Friend bonding .35 .24 .87 6.52 .***

Note: ***p < .001, **p < .01, *p < .05.


Place attachment items were measured on a scale ranging from 1 ¼ Strongly Disagree, 2 ¼ Disagree, 3 ¼ Neither Agree nor Disagree, 4 ¼ Agree, 5 ¼ Strongly Agree.

Bivariate correlations were then run between the respondent family bonding and the longest period of time lived by an immediate
variables and the place attachment dimensions (Table 5). As family member in the Northern and Yorke region (r ¼ .37, p < .01),
expected, we found significant positive relationships between the number of years the respondent’s family has owned the prop-
length of residence, place identity, and place dependence (r > .19, erty on which they currently live (r ¼ .25), and the number of
p < .01), as well as perceived knowledge of the Northern and Yorke generations the family has been involved in farming (r ¼ .22).
region and place identity and place dependence (r > .12, p < .01). The Significant, but weaker, relationships were identified between each
strongest positive relationship exists between time currently spent of these variables and friend bonding (r > .10, p < .05, Table 5).
in nature and nature bonding (r ¼ .40, p < .01), and length of resi- Rural landholders’ place attachments were compared across the
dence and family bonding (r ¼ .34, p < .01). Those respondents who three natural resource management regions (Table 6) in order to test
have been involved in farming or lived in the region for a longer the sensitivity of the place attachment scale for measuring attach-
period of time have stronger family bonds (r > .22, p < .05) ments to different places. Significant differences were found in the
compared to those who have been involved for shorter lengths of strength of place attachments across the three regions (F > 9.40,
times. Further, no significant relationship was found between p < .001). Rural landholders in the Adelaide and Mount Lofty Ranges
length of residence and nature bonding (r ¼ .07, p > .05). had significantly higher place identity, place dependence, nature
We examined the relationships between each place attachment bonding, and family bonding than Northern and Yorke and SAMDB
dimension and the number of years the farm has been owned or landholders. Northern and Yorke landholders had significantly
operated by the family, the number of generations the family has higher place identity than SAMDB landholders.
been involved in farming, and the longest period of time the
respondent has lived in the Northern and Yorke region (Table 5). The 3.2.3. Confirmatory factored dimensions
factorial model again indicated no interaction effects. The number of We used confirmatory factor analysis to determine the extent to
years the farm has been owned or operated by the family, the which the SAMDB landholder place attachment data fits the five-
number of generations the family has been involved in farming and dimensional model of place attachment. We then compared the five-
the longest period of time the respondent or their family lived in the dimensional model fit to the previously validated two-dimensional
Northern and Yorke region variables explained 2.6% of the variance model of place identity and dependence (Williams & Vaske, 2003). It
in place identity, 2.1% of the variance in place dependence, 1.5% of is important to note that the c2 statistic, the main measure of model
the variance in nature bonding, 12.4% of the variance in family fit, is sensitive to sample size. With a larger sample, the c2 value is
bonding, and 2.8% of the variance in friend bonding. Bivariate also likely to be bigger and the model more likely to be rejected.
correlations indicate significant, moderate relationships between Complex models with more parameters will also increase the c2

Table 5
Correlations between respondent variables and the five place attachment dimensions.

Variable Place Place Nature Family Friend


identity dependence bonding bonding bonding
Length of residence (0e100 yrs)
Pearson Correlation r .25** .19** .07 .34** .22**
Years involved in farming (0e100 yrs)
Pearson Correlation r .19** .16** .04 .22** .13**
Generations family involved in farming (0e6 generations)
Pearson Correlation r .01 .05 .08 .22** .10*
Years property on which currently live has been owned or operated by family (0e200 yrs)
Pearson Correlation r .08 .07 .08 .25** .13**

Nowadays, I spend a lot of time in the natural environment (1 ¼ strongly disagree to 5 ¼ strongly agree)
Pearson Correlation r .23** .21** .40** .03 .01
Longest period of time lived by an immediate family member in the Northern and Yorke region (0e100 yrs)
Pearson Correlation r .18** .11* .03 .37** .18**
Knowledge of Northern and Yorke region (1 ¼ Poor, 2 ¼ Fair, 3 ¼ Good, 4 ¼ Excellent)
Spearman Correlation r .25** .12** .21** .04 .09*

Note: **p < .01, *p < .05.


Place attachment items were measured on a scale ranging from 1 ¼ Strongly Disagree, 2 ¼ Disagree, 3 ¼ Neither Agree or Disagree, 4 ¼ Agree, 5 ¼ Strongly Agree.
C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434 431

Table 6 effort to integrate disparate place attachment literatures. Overall,


Comparison of place attachments across the three natural resource management our findings indicate that a five-dimensional model of place
regions.
attachment comprising of place identity, dependence, nature
Place Overall Adelaide Northern SA Murray- F bonding, family bonding, and friend bonding is a valid and reliable
attachment and Mt Lofty and Darling measure of rural landholder attachments to their natural resource
dimensions Ranges Yorke Basin
management region. Exploratory and confirmatory factor analyses
Place identity 3.79 3.94a 3.80b 3.69c 14.58 ***
revealed that our five-dimensional model explained a greater
Place dependence 3.32 3.47a 3.31b 3.25b 9.40 ***
Nature bonding 3.58 3.83a 3.50b 3.53b 26.25 *** amount of variance in overall place attachment than the traditional
Family bonding 3.30 3.12a 3.39b 3.30b 10.94 *** two-dimensional model of place identity and place dependence
Friend bonding N/A 3.62a 3.31b 10.94 *** in two of the three natural resource management regions, but the
Note: Note: ***p < .001. two-dimensional model provided a better fit for the SAMDB rural
Means were based on a scale ranging from 1 ¼ Strongly Disagree, 2 ¼ Disagree, landholder place attachment data.
3 ¼ Neither Agree or Disagree, 4 ¼ Agree, 5 ¼ Strongly Agree. Lettered superscripts A number of validity and reliability tests were undertaken as
indicate the mean is significantly different from the corresponding mean in the
part of this study. The first validity question tested was whether the
indicated cluster based on Bonferroni post-hoc testing (where p < .001).
N/A denotes not applicable. four hypothesized dimensions of place attachment constitute
distinct forms of place attachment. Exploratory factor analyses of
the Adelaide and Mount Lofty Ranges and Northern and Yorke
region revealed five statistically distinct factors comprising of place
value (Byrne, 2001); for example, Hair et al. (2006) state models identity, place dependence, nature bonding, family bonding, and
containing greater than 12 variables can be expected to have friend bonding. It was not expected that family bonding and friend
significant p-values for the chi-square test. The five-dimensional bonding would load on two separate dimensions. Place bonds
model was first tested using the SAMDB data, yielding only weak related to friends were generally stronger than those related to
support: chi-squared/degrees of freedom ratio (c2/d.f.) ¼ 4.58, family. One possible reason for this difference is that rural land-
comparative fit index (CFI) ¼ .91, root mean square error of holders, particularly commercial farmers, rely heavily upon their
approximation (RMSEA) ¼ .07. Post-hoc modifications were per- farm neighbors and friends as sources of information and support
formed in an attempt to develop a better fitting and more parsi- for on-farm practice. Northern and Yorke survey respondents most
monious model. According to the modification indices in AMOS, frequently cited farm neighbors or friends as their preferred source
model fit could be improved by adding co-variances between the of information about on-farm practices (>15.2% of responses,
indicators of family bonding and friend bonding and removing Raymond, 2010). The social bonding scale items were modified
a nature bonding indicator (“I learn a lot about myself when between the two studies with the goal of improving construct
spending time in the natural environment in X region”) which was validity and to develop a single social bonding measure, but they
strongly correlated with indicators of place identity. Rerunning of continued to load on separate dimensions. There are only marginal
the constrained model (Table 7) produced a substantially improved differences in explanatory power of the four and five-dimensional
fit ((c2/d.f.) ¼ 3.09, CFI ¼ .97, NFI ¼ .95, RMSEA ¼ .06). models, with the latter model explaining a slightly greater amount
We then compared the fit of the two-dimensional and five- of the variance in overall place attachment (67.9% vs. 66.4%).
dimensional model. AIC is a goodness-of-fit measure which adjusts Principal component analysis of the Adelaide and Mount Lofty
model chi-square to penalize for model complexity. Models with ranges data revealed that belonging to a volunteer group loaded on
smaller AIC values are favoured over models with larger AIC values the friend bonding factor. A number of studies have shown that
(Burnham & Anderson, 1998). A difference of greater than ten is feelings of belongingness or a communal bond with special friends
desired to provide compelling evidence that one model is superior is one of the significant sources of attachment to place (e.g., Guest &
to another (Akaike, 1987). The Expected Cross-Validation Index Lee, 1983; Hammitt, 2000). Volunteer groups in regional Australia
(ECVI) is also used to compare non-nested models and has the are an important outlet for the development of friendships. Those
advantage that it considers differences in parsimony. The smaller who feel they belong to volunteer groups may be more likely to
the ECVI value, the better the model. With this in mind, the two- have very strong friend bonds. The friendships developed through
dimensional model shows better fit overall to the SAMDB data volunteering may be one reason why Australia has a very high
(AIC ¼ 165.33 vs. 477.96, ECVI ¼ .26 vs. .75). volunteer rate. Australian Bureau of Statistics (2006) found that 5.2
million people (34%) of the Australian population aged 18 years and
4. Discussion over participate in voluntary work. The volunteering rate is higher
outside of cities (38%) and for people in the 45e54 age group (39%).
4.1. Validity and reliability of the place attachment model The internal consistency or reliability of the individual items
was determined. Place identity, place dependence, and nature
The objective of this study was to conceptualize and empirically bonding formed cohesive scales (Cronbach alphas > .80). However,
examine a four-dimensional model of place attachment among family bonding showed larger variance due to the negatively
three samples of rural landholders in regional South Australia in an worded statement of “without my relationships with family in the

Table 7
Confirmatory factor analysis of the five dimensions of place attachment across all South Australian Murray-Darling Basin landholders.

Dimensions c2 d.f. c2/d.f. 6c2 p CFI NFI RMSEA AIC ECVI


2 71.3 18 3.96 .98 .98 .07 165.33 .26
5 312.0 101 3.09 .97 .95 .06 487.96 .75
Model 2 vs. Model 5 240.7 <.001

Note: 2 dimensions ¼ place identity and place dependence; 5 dimensions ¼ place identity, place dependence, nature bonding, family bonding and friend bonding.
Acceptable ranges for the fit indices are c2/d.f. ratio of 2:1e5:1 (Marsh & Hocevar, 1985). CFI and NFI values in excess of .90 indicate an acceptable fit (Bentler, 1990; Bollen,
1989).
432 C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434

Adelaide and Mount Lofty Ranges, I would probably move.” This nature bonding without necessarily contributing to place identity
item was subsequently removed from further analyses. as an adult.
The spatial context of the population being measured appears to Our final test examined the fit of the revised five-dimensional
determine the relative strength and importance of the place model of place attachment against SAMDB landholder attachment
attachment dimensions. Analysis of variance and Bonferroni post- responses. The revised five-dimensional model of place attachment
hoc tests indicate that landholders in the Adelaide and Mount Lofty provided reasonable fit for the SAMDB dataset (chi-squared/
Ranges identify more strongly with place than landholders in the degrees of freedom ratio ((c2/d.f.) ¼ 3.09, CFI ¼ .95, NFI ¼ .95,
Northern and Yorke and SAMDB regions (F ¼ 14.58, p < .001). Further, RMSEA ¼ .06)). The model could be improved by increasing the
Adelaide and Mount Lofty Ranges have stronger place identity, place reliability of the family bonding and friend bonding constructs. We
dependence, and nature bonding, but weaker family bonding had to allow covariance between the errors on friend bonding and
attachments than Northern and Yorke and SAMDB landholders. One family bonding in the constrained model in order to improve model
reason for this difference is that the Adelaide and Mount Lofty Ranges fit, suggesting that family bonding and friend bonding may form
region has a higher proportion of hobby farmers than SAMDB and part of the one construct of social bonding in the SAMDB dataset.
Northern and Yorke regions. Results from the 2010 landholder Hence, it is still unclear whether family bonding and friend bonding
surveys indicate that hobby farmers have stronger nature bonds than are separate dimensions of place attachment or part of a larger
commercial farmers (Raymond, 2010). The social bonds of SAMDB construct named social bonding. One possible reason is that,
and Northern and Yorke residents may be stronger because of their despite our best efforts to develop cohesive sub-scales of family
remoteness from a major city. Landholders in remote areas have bonding and friend bonding, only two scale items loaded highly on
access to fewer services and are therefore more likely to require the each construct in the Northern and Yorke study. Future research
support of family and friends to achieve everyday farming tasks. We should increase the number of family bonding and friend bonding
also recognize that the results presented here are based on rural scale items prior to undertaking confirmatory analyses. Qualitative
landowners who live in three natural resource management regions. analyses of rural landholder attachments to place would have an
While they are geographically different regions, they are all similar important role in the development of these new scale items.
place types (rural and natural settings). Future research would We also need to be cautious about generalizing the place
benefit from the application of a consistent multi-dimensional model attachment results across regions. Confirmatory factor analysis
of place attachment across different places (a type of meta-analysis) revealed that the two-dimensional model provides a better fit for
to determine how the population location influences the relative the SAMDB rural landholder place attachment data. Place identity
importance of the various dimensions. and place dependence may be more robust measures of place
The identification of five dimensions of place attachment has attachment because they may be less dependent on the character-
some important implications for place attachment theory. Our istics of the social setting of a given study region than family bonding
results support previous studies (e.g., Hammitt et al., 2006; and friend bonding. Following the systemic model of connection
Williams & Vaske, 2003) who found that the personal meanings between residents and their communities (Kasarda & Janowitz,
of place identity and place dependence are separate dimensions of 1974), it would be expected that the region with the largest mean
place attachment. This personal context is missing from recent length of residence would have the strongest family and friend
studies (e.g., Brehm et al., 2006; Scannell & Gifford, 2010b) which bonds and this would be independent of the region’s population.
suggest that place attachment is related only to attributes of the Our results offer some support for this theory. Landholders in the
physical or social setting. Researchers are encouraged to examine Northern and Yorke region have a significantly longer mean length
how the community context of place attachment may be an of residence (p < .05) and their friend bonds are significantly
intermediary between the personal context and natural environ- stronger than the SAMDB region. However, the SAMDB region has
ment context. We also expand upon Hammitt and colleagues work a much larger population than the Northern and Yorke region. This
(Hammitt et al., 2006, 2009) by showing that attachments directly theory also holds for the differences in family bonds between the
related to family and friends as well as the natural environment Adelaide and Mount Lofty Ranges and SAMDB regions (Table 7).
need to be considered alongside the personal context of identity
and dependence in natural resource management research. There 4.2. Limitations
are important differences in the magnitude of family, friend, and
nature bonds which cannot simply be aggregated into place iden- The data collected in this study have several limitations. The first
tity, dependence or place belongingness measures. limitation is that the place attachment dimensions such as nature
The convergent validity analyses suggest that bonds to the bonding may be more relevant to rural than urban populations,
natural environment are largely independent of how long one has particularly those who own rural agricultural land or patches of
lived in a region or farming experience. When landholders native vegetation. We therefore cannot conclude that the place
acknowledge that they spend significant time in nature, rather than attachment findings here are generalisable to all land-use contexts.
a farming environment, bonds to nature increase. However, place Further research could expand the place attachment model pre-
identity increases with farming experience. These results are sented here to urban and peri-urban populations and contexts.
consistent with place continuity theory (see Hay, 1998a; Twigger- The scale items used to measure the friend bonding construct in
Ross & Uzzell, 1996), which suggests that past action and experi- the Adelaide and Mount Lofty Ranges and the Northern and Yorke
ence in place plays an important role in maintaining individual and region were different and therefore the structure of the construct is
group identity. Our results also support environmental identity not directly comparable. Another limitation is that significantly
theory and biophilia hypothesis which suggest that connection to more males than females completed the survey in all three study
nature and well-being is strongly influenced by a landholder’s regions. Researchers could compare and contrast the dimension-
experience and relationship with the surrounding natural world ality of place attachment based upon male and female responses,
(Clayton, 2003; Kellert & Wilson, 1993; Wilson, 1984). Future although to date, there is little empirical evidence that place
studies should further examine these proposed distinctions attachment is significantly related to gender.
between place identity and nature bonding by contrasting the We are concerned about the less than desired amount of vari-
experience in the region with other life experiences. For example, ance in place attachment accounted for by the different constructs.
childhood experiences with nature could contribute significantly to The five constructs accounted for over 68% of the variance in place
C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434 433

attachment; however, place identity accounted for the majority Akaike, H. (1987). Factor analysis and AIC. Psychometrika, 52, 317e332.
Altman, L. A., & Low, S. M. (1992). Place attachment. New York: Plenum.
(>39%) of it. Other studies (e.g., Hammitt et al., 2006, 2009) have
Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion of other in the self scale and the
reported similar findings. Whilst multiple conceptualizations of structure of interpersonal closeness. Journal of Personality and Social Psychology,
place were synthesized within the one model, other important 63(4), 596e612.
conceptualizations such as belongingness and familiarity have not Aron, A., Aron, E. N., Tudor, M., & Nelson, G. (1991). Close relationships as including
other in the self. Journal of Personality and Social Psychology, 60(2), 241e253.
been adequately considered. Further research could consider how Australian Bureau of Statistics. (2006). ABS census population and housing. Canberra:
these and other dimensions of place attachment interact with place Australian Bureau of Statistics. http://www.abs.gov.au Retrieved 15.05.10, from
identity, such as the negative attachments that people have with Australian Bureau of Statistics Web site:.
Bentler, P. M. (1990). Comparative fit indexes and structural models. Psychological
place (Manzo, 2003). This is particularly important in rural South Bulletin, 107, 238e246.
Australia where recent drought times have led to a high suicide rate Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley.
and people leaving their farms. Bragg, E. A. (1996). Towards ecological self: Deep ecology meets constructionist
self-theory. Journal of Environmental Psychology, 16(2), 93e108.
Brehm, J. M., Eisenhauer, B. W., & Krannich, R. S. (2006). Community attachments as
5. Conclusions predictors of local environmental concern: The case for multiple dimensions of
attachment. American Behavioral Scientist, 50(2), 142e165.
Bricker, K. S., & Kerstetter, D. L. (2000). Level of specialization and place attachment:
The purpose of this study was to develop and test a model of An exploratory study of whitewater recreationists. Leisure Sciences, 22,
place attachment which integrates multiple approaches to place 233e257.
Brown, B., Perkins, D. D., & Brown, G. (2003). Place attachment in a revitalizing
scholarship, including personal, community, and natural environ-
neighborhood: Individual and block levels of analysis. Journal of Environmental
ment contexts. Whilst not conclusive, our findings indicate that Psychology, 23(3), 259e271.
a five-dimensional model of place attachment comprising of place Brown, G., & Raymond, C. M. (2007). The relationship between place attachment
and landscape values: Toward mapping place attachment. Applied Geography,
identity, dependence, nature bonding, family bonding, and friend
27(2), 89e111.
bonding is a valid and reliable measure of rural landholder Burnham, K. P., & Anderson, D. R. (1998). Model selection and inference: A practical
attachments to their natural resource management region. In information-theoretic approach. New York: Springer-Verlag.
support of our integrated conceptual model of place attachment, Byrne, B. (2001). Structural equation modeling with AMOS: Basic concepts, applica-
tions, and programming. New Jersey: Lawrence Erlbaum.
the results suggest that rural landholders develop strong attach- Clayton, S. (2003). Environmental identity: A conceptual and an operational defi-
ments to both the physical and social characteristics of place. These nition. In S.Clayton., & S.Opotow. (Eds.), Identity and the natural environment (pp.
characteristics also have an important role in supporting place 45e66). Massachusetts: Massachusetts Institute of Technology.
Davenport, M. E., Baker, M. L., Leahy, J. E., & Anderson, D. H. (2010). Exploring
identity and place dependence. The attributes of the physical and multiple place meanings at an Illinois state park. Journal of Park and Recreation
social setting therefore cannot be viewed in isolation of the highly Administration, 28(1), 52e69.
personalized emotions formed in these settings which we refer to Devine-Wright, P., & Howes, Y. Disruption to place attachment and the protection of
restorative environments: A wind energy case study. Journal of Environmental
as the personal context to place attachment. Convergent validity Psychology, in press, corrected proof.
tests also revealed some conceptual and empirical distinctions Dillman, D. (2007). Mail and internet surveys: The tailored design method (2nd ed.).
between place identity and nature bonding, highlighting the New Jersey: John Wiley and Sons.
Dutcher, D. (2000). Landowner perceptions of protecting and establishing riparian
importance of considering the personal context of place attachment
forests in central Pennsylvania. Unpublished PhD Dissertation. Pennsylvania
separately to the environmental context. The interactions between State University e Pennsylvania.
the personal, social, and physical context of place have implications Dutcher, D. D., Finley, J. C., Luloff, A. E., & Johnson, J. B. (2007). Connectivity with
nature as a measure of environmental values. Environment and Behavior, 39(4),
for the systemic model of social interaction, place continuity theory,
474e493.
and environmental identity theory. Individual connection with Fishwick, L., & Vining, J. (1992). Toward a phenomenology of recreation place.
place is not just a function of experience with nature or social Journal of Environmental Psychology, 12(1), 57e63.
interaction with friends and family in setting, but also how indi- Fox, W. (1990). Transpersonal Ecology: “Psychologising” ecophilosophy. Journal of
Transpersonal Psychology, 22(1), 59e96.
viduals construct their own identity through their residential and Gosling, E., & Williams, K. J. H. Connectedness to nature, place attachment and
farming histories. It is possible that the spatial location and context conservation behaviour: Testing connectedness theory among farmers. Journal
of the population being measured also influences the strength and of Environmental Psychology, in press, corrected proof.
Guest, A. M., & Lee, B. A. (1983). Sentiment and evaluation as ecological variables.
relative importance of the place attachment dimensions. Future Sociological Perspectives, 26(2), 159e184.
research could expand the place attachment model presented here Gustafson, P. E. R. (2001). Meanings of place: Everyday experience and theoretical
by comparing urban and regional attachments, or across different conceptualisations. Journal of Environmental Psychology, 21(1), 5e16.
Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multi-
sub-regions of residence. The model presented does however variate data analysis (6th ed.). New Jersey: Pearson Prentice Hall.
provide researchers with a valid and inclusive method to measure Hammitt, W. E. (2000). The relation between being away and privacy in urban
place attachment in rural and natural settings. forest recreation environments. Environment and Behavior, 32(4), 521e540.
Hammitt, W. E., Backlund, E. A., & Bixler, R. D. (2004). Experience use history, place
bonding and resource substitution of trout anglers during recreation engage-
Acknowledgements ments. Journal of Leisure Research, 36(3), 356e378.
Hammitt, W. E., Backlund, E. A., & Bixler, R. D. (2006). Place bonding for recreation
places: Conceptual and empirical development. Leisure Studies, 25(1), 17e41.
Christopher Raymond acknowledges the generous financial Hammitt, W. E., Kyle, G. T., & Oh, C. O. (2009). Comparison of place bonding models
support provided by Jean Pearce. We thank the three anonymous in recreation resource management. Journal of Leisure Research, 41(1), 57e72.
reviewers and Prof. Guy Robinson for their thorough and Hay, R. (1998a). Sense of place in developmental context. Journal of Environmental
Psychology, 18(1), 5e29.
constructive feedback on this publication. Funding for this project Hay, R. (1998b). A rooted sense of place in cross-cultural perspective. Canadian
was provided by the Centre for Rural Health and Community Geographer, 42, 245e266.
Development, University of South Australia, South Australian Hidalgo, M. C., & Hernandez, B. (2001). Place attachment: Conceptual and empirical
questions. Journal of Environmental Psychology, 21(3), 273e281.
Murray-Darling Basin Natural Resources Management Board and Jorgensen, B. S., & Stedman, R. C. (2006). A comparative analysis of predictors of sense
the Northern and Yorke Natural Resources Management Board. of place dimensions: Attachment to, dependence on, and identification with
lakeshore properties. Journal of Environmental Management, 79(3), 316e327.
Kals, E., Schumacher, D., & Montada, L. (1999). Emotional affinity toward nature as
References a motivational basis to protect nature. Environment and Behavior, 31, 178e202.
Kaltenborn, B. P. (1997). Nature of place attachment: A study among recreation
Adelaide and Mount Lofty Ranges Natural Resources Management Board. (2008). homeowners in Southern Norway. Leisure Sciences, 19(3), 175e189.
Creating a sustainable future. In State of the region report, Vol. A. Adelaide: Kasarda, J., & Janowitz, M. (1974). Community attachment in mass society. American
Adelaide and Mount Lofty Ranges Natural Resources Management Board. Sociological Review, 39, 328e339.
434 C.M. Raymond et al. / Journal of Environmental Psychology 30 (2010) 422e434

Kellert, S. R., & Wilson, E. O. (1993). The biophilia hypothesis. Washington, D.C.: Roberts, E. (1996). Place and spirit in public land management. In B. L. Driver,
Island Press. D. Dustin, T. Baltic, G. Elsner, & G. Peterson (Eds.), Nature and the human spirit
Kyle, G., Bricker, K., Graefe, A., & Wickham, T. (2004). An examination of recreationists’ (pp. 61e78). State College, PA: Venture Publishers.
relationships with activities and settings. Leisure Sciences, 26(2), 123e142. Rollero, C., & De Piccoli, N. (2010). Place attachment, identification and environment
Kyle, G., & Chick, G. (2007). The social construction of a sense of place. Leisure perception: An empirical study. Journal of Environmental Psychology, 30(2),
Sciences, 29(3), 209e226. 198e205.
Kyle, G., Graefe, A., & Manning, R. (2005). Testing the dimensionality of place Sampson, K. A., & Goodrich, C. G. (2009). Making place: Identity construction and
attachment in recreational settings. Environment and Behavior, 37(2), 153e177. community formation through “sense of place” in Westland, New Zealand.
Kyle, G., Graefe, A., Manning, R., & Bacon, J. (2004a). Effect of activity involvement Society & Natural Resources, 22(10), 901e915.
and place attachment on recreationists’ perceptions of setting density. Journal Scannell, L., & Gifford, R. (2010a). Defining place attachment: A tripartite organizing
of Leisure Research, 36(2), 209e231. framework. Journal of Environmental Psychology, 30, 1e10.
Kyle, G., Graefe, A., Manning, R., & Bacon, J. (2004b). Effects of place attachment on Scannell, L., & Gifford, R (2010). The relations between natural and civic place
users’ perceptions of social and environmental conditions in a natural setting. attachment and pro-environmental behavior. Journal of Environmental
Journal of Environmental Psychology, 24(2), 213e225. Psychology, 30(3), 289e297. doi:10.1016/j.jenvp.2010.01.010.
Kyle, G. T., Absher, J. D., & Graefe, A. R. (2003). The moderating role of place Schreyer, R., Jacob, G., & White, R. (1981). Environmental meaning as a determinant
attachment on the relationship between attitudes toward fees and spending of spatial behavior in recreation. In R. Schreyer, G. Jacob, & R. White (Eds.),
preferences. Leisure Sciences, 25(1), 33e50. Proceedings of the applied geography conferences, Vol. 4 (pp. 294e300). Bing-
Low, S. (2000). On the Plaza: The politics of public space and culture. Austin: hampton, NY: SUNY Binghampton.
University of Texas Press. Schultz, P. W. (2001). The structure of environmental concern: Concern for self, other
Manzo, L. C. (2003). Beyond house and haven: Toward a revisioning of emotional people, and the biosphere. Journal of Environmental Psychology, 21(4), 327e339.
relationships with places. Journal of Environmental Psychology, 23(1), 47e61. Schultz, P. W., Shriver, C., Tabanico, J. J., & Khazian, A. M. (2004). Implicit connec-
Manzo, L. C. (2005). For better or worse: Exploring multiple dimensions of place tions with nature. Journal of Environmental Psychology, 24(1), 31e42.
meaning. Journal of Environmental Psychology, 25(1), 67e86. Schultz, P. W., & Tabanico, J. (2007). Self, identity, and the natural environment:
Marsh, H. W., & Hocevar, D. (1985). Application of confirmatory factor analysis to Exploring implicit connections with nature. Journal of Applied Social Psychology,
the study of self-concept: First and higher order factor models and their 37(6), 1219e1247.
invariance across groups. Psychological Bulletin, 97, 562e582. South Australian Murray-Darling Basin Natural Resources Management Board.
Mayer, F. S., & Frantz, C. M. (2004). The connectedness to nature scale: A measure of (2009). South Australian Murray-Darling Basin natural resources management
individuals’ feeling in community with nature. Journal of Environmental plan. In State of the region report, Vol. B. Adelaide: South Australian Murray-
Psychology, 24(4), 503e515. Darling Basin Natural Resources Management Board.
Mesch, G. S., & Manor, G. (1998). Social ties, environmental perception, and local Stedman, R. (2002). Toward a social psychology of place: Predicting behavior from
attachment. Environment and Behavior, 30(4), 504e519. place-based cognitions, attitude, and identity. Environment and Behavior, 34(5),
Milligan, M. J. (1998). Interactional past and potential: The social construction of 561e581.
place attachment. Symbolic Interactionism, 21, 1e33. Stedman, R. (2003). Is it really just a social construction? The contribution of the
Moore, R. L., & Graefe, A. R. (1994). Attachments to recreation settings: The case of physical environment to sense of place. Society and Natural Resources, 16,
rail-trail users. Leisure Sciences, 16(1), 17e31. 671e685.
Nash, R. (1990). American environmentalism: Readings in conservation history (3rd Stedman, R., Beckley, T., Wallace, S., & Ambard, M. (2004). A picture and 1000
ed.). New York: McGraw-Hill. words: Using resident-employed photography to understand attachment to
Northern and Yorke Natural Resources Management Board. (2009). Northern and high amenity places. Journal of Leisure Research, 36(4), 580e606.
Yorke natural resources management plan. In State of the region report, Vol. A. Stokowski, P. A. (2002). Languages of place and discourses of power: Constructing
Adelaide: Northern and Yorke Natural Resources Management Board. new senses of place. Journal of Leisure Research, 34(4), 368.
Patterson, M. E., & Williams, D. R. (2005). Maintaining research traditions on place: Trentelman, C. K. (2009). Place attachment and community attachment: A primer
Diversity of thought and scientific progress. Journal of Environmental Psychology, grounded in the lived experience of a community sociologist. Society & Natural
25(4), 361e380. Resources, 22(3), 191e210.
Perkins, D. D., & Long, A. D. (2002). Neighborhood sense of community and social Tuan, Y. F. (1980). Rootedness versus sense of place. Landscape, 24, 3e8.
capital: A multi-level analysis. In A. Fisher, C. Sonn, & B. Bishop (Eds.), Psycho- Twigger-Ross, C. L., & Uzzell, D. L. (1996). Place and identity processes. Journal of
logical sense of community: Research, applications and implications (pp. Environmental Psychology, 16(3), 205e220.
291e318). New York: Plenum Press. Vaske, J. J., & Kobrin, K. C. (2001). Place attachment and environmentally respon-
Perkins, H. E. Measuring love and care for nature. Journal of Environmental sible behavior. Journal of Environmental Education, 32, 16e21.
Psychology, in press. Williams, D. R., Patterson, M. E., Roggenbuck, J. W., & Watson, A. E. (1992). Beyond
Pretty, G. H., Chipuer, H. M., & Bramston, P. (2003). Sense of place amongst the commodity metaphor: Examining emotional and symbolic attachment to
adolescents and adults in two rural Australian towns: The discriminating place. Leisure Sciences, 14, 29e46.
features of place attachment, sense of community and place dependence in Williams, D. R., & Vaske, J. J. (2003). The measurement of place attachment: Validity
relation to place identity. Journal of Environmental Psychology, 23(3), 273e287. and generalizability of a psychometric approach. Forest Science, 49(6),
Proshansky, H. M., Fabian, A. K., & Kaminoff, R. (1983). Place-identity: Physical 830e840.
world socialization of the self. Journal of Environmental Psychology, 3(1), 57e83. Williams, D. R., & Roggenbuck, J. W. (1989). Measuring place attachment: Some
Raymond, C. M. (2009). Assessment of rural landholder and NRM staff attitudes, goals preliminary results. In Abstracts of the 1989 Leisure Research Symposium.
and place values in the Eyre Peninsula region, South Australia. Adelaide: Envir- October 20–24, San Antonio, TX.
oconnect Pty Ltd. Wilson, E. O. (1984). Biophilia. Cambridge: Harvard University Press.
Raymond, C. M. (2010). Results of the 2010 Northern and Yorke rural landholder Zimmerman, M., Callicott, J. B., Sessions, G., Warren, K. J., & Clark, J. (1993). Envi-
survey. Adelaide: Centre for Rural Health and Community Development, ronmental philosophy: From animal rights to radical ecology (3rd ed.). UK:
University of South Australia. Prentice-Hall.

View publication stats

You might also like