Page 1
Health & Place 11 (2005) 157–171
ARTICLE IN PRESS
*Correspond
03-5454-7767.
E-mail addr
1353-8292/$ - se
doi:10.1016/j.he
Changing places: the impact of rural restructuring on mentalhealth in Australia
Cait Frasera,*, Henry Jacksonb, Fiona Judda,c, Angela Komitia,d, Garry Robinsb,Greg Murraye, John Humphreysf, Pip Pattisonb, Gene Hodginsa,b
a Centre for Rural Mental Health, Bendigo Health Care Group, PO Box 126, Bendigo, Vic. 3552, Australiab Department of Psychology, The University of Melbourne, Melbourne, Vic., Australia
c School of Psychology, Psychiatry, and Psychological Medicine, Monash University, Melbourne, Vic., Australiad Department of Psychiatry, The University of Melbourne, Melbourne, Vic., Australia
e School of Social and Behavioural Sciences, Swinburne University of Technology, Melbourne, Vic., Australiaf School of Rural Health, Monash University, Bendigo, Australia
Accepted 16 March 2004
Abstract
Significant demographic, social and economic change has come to characterise much of rural Australia, with some
authors arguing there are now two sharply differentiated zones, one of growth and one of decline. This restructuring
process, which has been similar to other western nations, has had a profound impact upon rural places—socially,
economically and physically. Findings from research investigating the relationship between health, place and income
inequality suggest that rural ‘desertification’, which is characterised by decline of the agricultural sector, net population
loss and the deterioration of demographic structures, may negatively influence mental health outcomes in these areas.
By contrast, the growth in rural areas, which is associated with expanding employment opportunities and the movement
of capital and people, may confer positive benefits to mental health. The aim of this study was to investigate differences
in mental health and well-being between rural communities experiencing growth and decline as measured by net
population change. Utilising a survey methodology, questionnaires were distributed to 20,000 people randomly
sampled from the electoral role in rural Australia. We selected four sub-regions from the sample area that were
characteristic of areas experiencing population growth and decline in Australia and analysed the results of respondents
from these four regions (n ¼ 1334). The analysis provided support for our hypothesis that living in a declining area is
associated with poorer mental health status; however, the factors that underpin growth and decline may also be
important in influencing mental health. Discussed are the mechanisms by which demographic and social change
influence mental health. The findings of this study highlight the diversity of health outcomes in rural areas and suggest
that aspects of place in declining rural areas may present risk factors for mental health.
r 2004 Elsevier Ltd. All rights reserved.
Keywords: Rural; Mental health; Place
Introduction
There is an increasing body of empirical and
theoretical work examining the link between health
ing author. Tel.: +61-03-5454-7753; fax: +61-
ess: [email protected] (C. Fraser).
e front matter r 2004 Elsevier Ltd. All rights reserve
althplace.2004.03.003
and place. Whilst this work has yielded significant
insights into the mechanisms through which place
influences health (Curtis and Rees Jones, 1998; Macin-
tyre et al., 2002; Sampson et al., 2002), there have been
few studies that have investigated place and health in
rural contexts (Kobetz et al., 2003). Compared with
urban neighbourhoods, which have tended to be the
focus of urban health and place studies, there are
d.
Page 2
ARTICLE IN PRESSC. Fraser et al. / Health & Place 11 (2005) 157–171158
well-documented differences in rural towns, such as
geographical isolation, and lower rates of health service
per capita, that may be important to health outcomes. In
addition, characteristics of communities that are known
to be important to poor health, such as high levels of
socio-economic disadvantage, poor quality housing and
limited access to labour markets, may have differential
impacts across urban and rural settings as they are
mediated by distinct sets of cultural and social condi-
tions.
The aim of this study is to investigate the relationship
of one feature of rural places, population growth and
decline, to mental health. Population loss and popula-
tion growth are salient features of rural areas in
Australia and many other developed nations
(McKenzie, 1994a b; Stockdale et al., 2000). Examining
population growth is relevant to health as it is both an
indicator of, and contributor to, changes to broader
social and economic conditions of an area. Rural decline
or ‘desertification’ is not only characterised by net
population loss but also the undermining of demo-
graphic structures, the withdrawal and closure of
services and declining employment opportunities (Caw-
ley, 1994). Similarly, population growth or ‘counter-
urbanisation’ is associated not only with the movement
of people to an area but also capital and employment
(Curry et al., 2001; Stockdale et al., 2000). Population
growth, and the processes that underpin it, therefore has
the potential to result in significant changes to the nature
of places and potentially impact upon health outcomes.
This paper will draw upon research investigating
health, place and income inequality, which provides a
basis for understanding how living in areas of growth
and decline can influence health. These studies have
identified a range of features of local areas that can
influence health outcomes including the ‘neo-material
environment’ (Lynch et al., 2000), access to labour
markets, increased exposure to environmental stressors
(Macintyre et al., 1993) and the psychological impact of
living in a deprived area (Wilkinson, 1997).
Background: population change in rural areas
In Australia rural population change has generally
reflected international trends. Like many other devel-
oped nations the movement of people from rural to
urban areas has been occurring in Australia for much of
this century (Productivity Commission, 1999). Typically,
this type of decline led to an aging and relatively
impoverished rural population, with access to limited
services (Stockdale et al., 2000). However, by the 1970s a
process of net migration gain to rural areas—‘counter-
urbanisation’—was found to be taking place in a
number of western countries including the US (Beale,
1977) Australia (Hugo, 1994), Canada (Dahms and
McComb, 1999) and the UK (Champion, 1989; Stock-
dale et al., 2000). By the 1980s, however, many authors
were asserting that the ‘rural renaissance’ as it was
described, had ended, and that rural areas would
continue to decline as metropolitan areas grew (Dahms
and McComb, 1999). Data from Australia suggest a
more complex pattern of population change in rural
areas, with two sharply differentiated zones of growth
and decline emerging in rural Australia (Hugo, 2001).
Whilst some of the strongest growth rates in the country
are being recorded in the attractive coastal and
mountain environments, and urban fringe areas (Hugo,
1994), widespread decline is continuing in inland
agricultural regions, and to a lesser extent mining towns
(McKenzie, 1994a b).
Drivers of growth and decline in rural Australia
Negative and positive population growth in rural
Australia is essentially a symptom of more widespread
political, economic and social change. The factors that
underpin rural decline and growth are often exogenous
to rural areas themselves and are tied to national and
international economic and political dynamics (Cloke,
1996). This has clearly been true in the Australian
setting. Rural decline has been strongly tied to the
diminishing terms of trade for primary produce,
particularly agriculture, within global markets
(McKenzie, 1994b). The 1980s and 1990s proved to be
one of the most difficult periods for farmers in Australia
with the collapse of the wheat, wool and barley markets
(Budge, 1996), a dramatic fall in land values and
simultaneous increase in interest rates (Smailes, 2000)
and widespread drought (Budge, 1996). During this time
the process of agro-industrialisation was occurring,
leading to a centralisation of agricultural production
(Burch and Rickson, 2001), and a drive towards
economically competitive production (McMichael and
Lawrence, 2001). The result has been farm amalgama-
tions, fewer farming families (a decline of 22% since
1986) (Australian Bureau of Statistics, 2003a, b), and
reduced employment of paid labour (Lawrence and
Williams, 1990). The declining number of farms and the
wider sourcing of farm inputs (Lawrence, 1996) also had
a detrimental effect upon the economy and population
of the towns servicing farming areas (Budge, 1996). As
Lawrence and Williams (1990, p. 40) have noted: ‘‘more
productive agriculture is coming to mean less productive
and viable rural communities’’.
Mining areas have also been vulnerable to substantial
population decline due to the unstable nature of
many mineral markets and their economic depen-
dence on a single, exhaustible resource (McKenzie,
1994b). As the majority of mines operate in remote
Page 3
ARTICLE IN PRESSC. Fraser et al. / Health & Place 11 (2005) 157–171 159
areas and these towns tend to have an undiversified
economy, mine closure can result in rapid population
decline.
As Walmsley and Sorensen (1990) have outlined, a
significant fall in population can lead to a ‘vicious cycle’
of decline. Falling population leads to a reduction in the
demand for goods, a subsequent decline in the produc-
tion of goods and services, decreased employment
opportunities and further out-migration in rural areas
(Fincher, 1999; McKenzie, 1994a). In Australia this
cycle was compounded by a trend towards the centra-
lisation of public and private sector services (particularly
health, education and financial services) taking jobs,
capital and people out of rural areas to large regional or
metropolitan cities (Lawrence, 1999; Raiston and Beal,
1997). The withdrawal of public sector employment,
which accounts for up to one-third of employment in
small country towns (Jones and Tonts, 1995), also
removed the regular ‘drought-proof’ income from
communities, which can be heavily affected by the
seasonal variations in the agricultural economy
(Smailes, 1997). Whilst this process of decline leads to
out-migration, it can also lead to ‘entrapment’; as
property values fall and there are few buyers, house-
holds may be unable to sell their property at prices that
allows them to relocate to more prosperous areas
(Budge, 1996).
A number of studies have shown that the loss of
amenities such as schools, hospitals and banks can also
have a social and psychological impact upon the town’s
remaining residents (Argent and Rolley, 2000; Raiston
and Beal, 1997) and reduce the sense of cohesion and
participation of community members (Witten et al.,
2001). Smailes (2000) argues that the removal of services,
such as banks and shops, to larger centres, undermines
community cohesion, as the mutually reinforcing nexus
between social and commercial activity is removed.
Another social dimension to depopulation is the
characteristics of those who stay and those who leave
declining rural areas. The majority of out-migrants are
between 15 and 35 years of age (Black et al., 2000;
Productivity Commission, 1999). The loss of this age
group often leads to a shortage of individuals who have
the capacity for maintaining and participating in
sporting clubs, recreational and volunteer organisations
which are often a strong component of a town’s sense of
identity and community (Productivity Commission,
1999). This group also represents the ‘child bearing’
section of the community, pointing to further popula-
tion decline (Hugo, 1994). Some studies from the US
and Australia have found that people who leave
declining rural areas are more likely to be better
educated and have greater job prospects (Fitchen,
1995), whilst those who move into these areas are
overwhelmingly from low-income groups (Fincher and
Wulff, 2001). Another feature of declining areas is the
migration of low-income families attracted by depressed
rural property prices and lower costs of living (Fincher
and Wulff, 1997). In some instances this has led to a high
concentration of low-income families in areas creating
an ‘us’ and ‘them’ mentality between newer and older
residents and inadequate levels of service provision for
new households (Fincher and Wulff, 1997).
The rural renaissance
Whilst rural decline is widespread, it is far from
uniform across all rural areas. The contrasting picture in
rural Australia, and many developed counties, is of
significant population growth and migration to pictur-
esque and coastal rural towns in close proximity to
major urban centres (known as exurban areas). No
single theory has been developed to explain rural
repopulation (Champion, 1989)—however, a number
of reasons have been suggested including the high value
placed on the amenity of rural areas, the desire for the
‘rural lifestyle’, the role of planning policies, regional
restructuring, the flexibility of labour and the flexibility
of capital in post-Fordist economies (Dahms and
McComb, 1999). In Australia, research suggests that
changing employment patterns have underpinned coun-
terurbanisation trends—with a range of industries
including transport, utilities and tertiary sectors all
experiencing growth in rural areas. Those employed in
these occupations, the 25–39 age group with young
families, dominated internal migration flows to rural
areas (Hugo, 1989). The movement of retirees and
others to coastal communities for lifestyle reasons and
the reduced cost of living has also been a driver of
growth in these regions (Curry et al., 2001). Other areas
of population growth are large inland regional centres
that are not dependent upon agriculture and mining
(Eves, 1998).
The reported impacts of counterurbanisation have
been mixed. The migration of sections of the urban
population to rural areas has been identified as bringing
a range of opportunities to local communities as there is
increased demand for new functions and service
(Paquette and Domon, 2003). These changes can lead
to increased rural income levels, employment opportu-
nities and net investment in the rural housing stock
(Stockdale et al., 2000). However, counterurbanisation
has been described as a double-edged sword (Fielding,
1990) creating tensions between ‘old’ and ‘new’ residents
in some communities (Curry et al., 2001; Jones and
Tonts, 2003), increased demand for limited housing
stock and reduced affordability for ‘locals’ (Jones and
Tonts, 2003), the impact of changing social structures on
local decision-making processes (Greive and Tonts,
1996), and the environmental impact of new housing
Page 4
ARTICLE IN PRESSC. Fraser et al. / Health & Place 11 (2005) 157–171160
and agricultural pursuits (Black et al., 2000). US studies
that have examined rapid population growth have found
that it has been associated with a range of negative
social impacts such as rising crime rates, domestic
violence and increased use of alcohol and illicit drugs
(Broadway, 2000; Gouveia and Stull, 1995). This
population growth, however, has been the result of high
demand for (predominantly male) semi-skilled labour in
industries such as meatpacking (Broadway, 2000).
Population growth in rural Australia is not typically
driven by the same factors as the areas examined in US
research and there are no data to indicate rising crime
and social disorder in growing rural areas.
Mental health and population change
Mental health in rural areas has been poorly
investigated (Fraser et al., 2002). Most of the research
has focused on determining whether disorders are more
or less prevalent in rural or urban settings (Judd et al.,
2002). Little attention has been directed to investigating
how rural place and its economic, social, environmental
and socio-cultural components, might influence mental
health. We have proposed population growth and
decline as one index of place, and as described above,
it is of particular relevance to rural places. Population
change in rural areas is associated with two significant
factors that may influence mental health outcomes—
changes to population composition and changes to local
environments. We have previously argued (Judd et al.,
2002) that any investigation of the mental health of rural
residents needs to examine variables that relate to place
as well as to individual differences between people. Most
obviously, population change alters the composition of
a community, typically increasing the concentration of
low-income families in declining rural areas (Dear and
Wolch, 1987; Dembling et al., 2002; Fitchen, 1995;
Leighton et al., 1963). Numerous studies have shown
rates of almost all disorders decline monotonically with
increased income and education (Bruce et al., 1991;
Kessler et al., 1994), suggesting there may be a higher
prevalence of mental health problems in declining rural
areas, indicating that those who move to, or remain in,
declining rural areas are at greater risk of mental health
problems.
Secondly, population growth or decline is associated
with changes to the social, economic and physical
characteristics of a town (McKenzie, 1994a). Research
investigating the relationship between health, place and
income inequality provides a theoretical (Macintyre
et al., 2002) and empirical basis for understanding how
area interacts with individual-level variables to influence
health outcomes, with a range of studies demonstrating
the effect of local environments on mortality (Kaplan,
1996; Waitzman and Smith, 1998), chronic illness
(Robert, 1998), mental illness (Halpern, 1995) and
health behaviours such as smoking and alcohol use
(Curry et al., 1993; Reijneveld, 1998; Stead et al., 2001).
Whilst the relationship between health and place is
complex, a range of hypotheses has been proposed to
explain the mechanisms through which area influences
health. We draw upon two of these theories in
examining health and place in declining rural areas.
Lynch et al. (2001) have described the importance of
the ‘neo-material’ environment to health. This concept
refers to the ‘neo-material matrix of contemporary
life’—housing, health services, education, access to
healthy foodstuffs and recreational opportunities. The
authors argue that in disadvantaged communities the
neo-material environment is diminished due to the
limited resources of individuals and the systematic
underinvestment in community infrastructure by gov-
ernment (Lynch, 2000). A number of studies have found
that community infrastructures, such as transport,
schools and health centres, are required not only to
deliver health care or education but also as venues for
community interaction (Warin et al., 2000; Witten et al.,
2001). Such interaction can generate social capital, social
support and social inclusion (Warin et al., 2000), which
have a well-recognised health benefit (Kawachi and
Berkman, 2003). The importance of the neo-material
environment to health is also supported by the work of
Macintyre et al. (1993, 2002), who have argued that
living in particular areas can impact upon health
outcomes through access to resources important for
health and exposure to environmental stressors. In
declining and deprived areas there are fewer of the
resources required for residents to live a healthy life,
such as job opportunities, access to health care services,
adequate transport and good quality housing stock
(Black et al., 2000; Jones and Tonts, 2003; McKenzie,
1994a). Environmental stressors may be in the form of
crime and violence (which tend to be higher in deprived
areas) (Stead et al., 2001) or from air and water
pollutants (Halpern, 1995). However, there are no data
to indicate that these characteristics are more prevalent
in declining rural areas.
The present study examined how population growth
and decline is related to self-reported mental health in
rural settings. Using a case study approach, four local
government areas (LGAs) were selected which are
typical of the types of communities experiencing
population change (declining agricultural regions, de-
clining mining communities, growing exurban areas and
growing regional cities). LGAs refer to the catchment
area of local councils. These LGAs also represent the
extremes of population growth and decline in our
sample. We tested three hypotheses regarding popula-
tion change and mental health: (1) that those in
declining areas would be older, less educated and more
likely to be unemployed; (2) that those in declining areas
Page 5
ARTICLE IN PRESS
Fig. 1. Case study areas.
C. Fraser et al. / Health & Place 11 (2005) 157–171 161
would score higher (i.e., poorer) on measures of mental
ill-health and higher on measures of substance use; and
(3) that those in declining areas would score lower on
measures of mental well-being. We believed that the
demographic characteristics of declining towns and the
characteristics of the places themselves would explain
the poorer mental health status of residents in these
areas.
We selected measures of anxiety, depression and
substance misuse (smoking, drinking and drug use) as
measures of mental ill-health, as they are the most
common disorders in the western societies, including the
USA (Kessler et al., 1994), UK (Paykel et al., 2000),
Australia (Andrews et al., 1999) and New Zealand
(Oakley-Browne et al., 1989). We extended previous
work reviewed by Judd et al. (2002) by including
measures of disability and well-being alongside measures
of mental health symptoms. This allowed us to take a
broader perspective looking beyond morbidity alone
by examining measures of functioning and satisfaction
with life.
Method
This study was part of a larger project investigating
the mental health of rural Australians in which 20,000
residents were randomly sampled from the electoral roll
in Victoria and New South Wales (NSW)—Australia’s
two most populous states. The geographic area of the
sampling frame covered approximately 540,000 km2,
beginning in southern Victoria and extending to north-
ern NSW (or the NSW/Queensland border). The total
population of the area was approximately 400,000
people. The population was predominantly Australian-
born and included relatively large indigenous commu-
nities in central and northern NSW and northern
Victoria. The average unemployment rate for the region
was approximately 10%; however, this varies across the
region. Agriculture, forestry and fishing and government
services are the region’s major employers. Significant
proportions of the workforce are also employed in retail,
manufacturing, education and health and community
services sectors.
From this sample four LGAs were selected. The areas
were selected because they are characteristic of commu-
nities experiencing population growth and decline across
rural Australia. Therefore, we selected a declining
agricultural region, a declining mining region, a growing
regional city and a picturesque growing exurban area.
The communities
The communities selected were the shire of Broken
Hill, located in the far west of NSW, the shire of Dubbo
located in central NSW, the shire of Buloke in central
Victoria and the shire of Macedon Ranges in southern
Victoria and are shown in Fig. 1. Demographic details
of the shires are listed in Table 1.
The demographic characteristics of the four regions
vary considerably, with the declining communities
having a greater proportion of older residents—a typical
difference between declining and growing communities.
The history and economic base of the areas also vary
considerably.
The city of Broken Hill is located in outback NSW
and covers an area of approximately 70 km2. This city
has a prominent place in Australian history due to its
mining wealth and strong history of union activity. In
1985, four mines were operating in Broken Hill,
employing approximately 4000 people, but by 1995
three mines had closed due to the high cost of extraction
(New South Wales Department of Mineral Resources,
2002). The remaining working mine in the town is
scheduled to close in less than 15 years. Without the
discovery of further mineral deposits in the area, the
future for Broken Hill appears bleak (New South Wales
Department of Mineral Resources, 2002).
The Shire of Buloke is located in North Western
Victoria and encompasses five major towns and five
smaller townships. Agriculture accounts for 65% of
businesses in the shire; however, it only accounts for
10% of employment. Government sector services
(health, community services and education), retail trade
and manufacturing are the largest employers in the
region (Department of Education and Training, 2003).
Since the 1980s, the shift to a post-industrial economy
has led to changes in the region including a decrease in
the number of people employed in the agriculture,
forestry and fishing, public administration and
defence sectors (Department of Education and Training,
2003).
Page 6
ARTICLE IN PRESS
Table 1
Australian bureau of statistic regional statistics data—Victoria and New South Wales 1999
Variables Growing communities Declining communities
Macedon Ranges Dubbo Buloke Broken Hill
Estimated resident population 35,844 37,396 7518 20,934
Annual population growth 1991–2001 1.9% 1.35% �1.85% �1.45%
Percentage persons aged 0–14 years 25.00% 25.50% 19.50% 20.60%
Percentage persons aged 65+ 8.80% 10.50% 19.50% 17.20%
Unemployment rate 5.90% 4.80% 4.70% 9.30%
Centrelink customersa 2003 7819 1863 7425
Centrelink customers as a % of population (15 years+) 20.62% 28.10% 30.81% 44.70%
Average annual taxable income $35,071 $32,370 $24,532 $33,581
a Centrelink is the government agency responsible for the distribution of social security payments including unemployment and
disability and sickness benefits and pensions.
C. Fraser et al. / Health & Place 11 (2005) 157–171162
Macedon Ranges Shire is located in central Victoria
and is within close proximity to the state’s capital city.
The region is picturesque and one of the fastest growing
shires in the state (Australian Bureau of Statistics,
2003a, b). Growth in the region is principally due to the
migration from the state’s capital Melbourne, and 51%
of the Shire’s working population commute to Mel-
bourne for work (an increase from 14% since 1991)
(Department of Human Services, 1999). There have
been changes in employment in the region, with an
increase in employment in the service industries and a
decline in the number of people employed in the primary
industries (Department of Human Services, 1999).
The city of Dubbo is located in central NSW and is a
service centre for the agricultural towns in the region
(Eves, 1998). Much of the area surrounding Dubbo is
used to farm cattle, sheep and wheat. Light industry and
the service sector also make up a significant proportion
of economic activity within the city itself. The popula-
tion growth that has been experienced in the city is
strongly associated with the decline of smaller towns and
the population shift to Dubbo (Eves, 1998).
Measures
Socio-demographic variables
Nine socio-demographic variables were measured
essentially for descriptive purposes: age, gender, rela-
tionship status, highest level of education (a surrogate
measure of SES, as previously employed by, for
example, Blazer et al., 1985), employment, nationality,
indigenous heritage, duration of local residence and
household composition. However, as specified pre-
viously, for three of the variables, namely age, education
and employment, we hypothesised differences between
growing and declining areas.
Mental health measures
Three scales were used to measure anxiety and
depressive symptoms: the positive and negative affect
scale (PANAS) (Watson et al., 1988), the K10 (Kessler
and Mroczec, 1994; Kessler and Mroczec, 1992) and the
SPHERE-12 (Hickie et al., 2001).
The 20-item PANAS was used to measure mood
characterised in terms of two quasi-orthogonal mea-
sures: positive affect (PA) and negative affect (NA).
Extensive research suggests that PANAS is a reliable
and valid measure of the two fundamental constructs
(Watson et al., 1988). Given that the authors of the
PANAS did not provide cut-points, based on the
distributions of the current data set, a score of 26 or
more was accepted as indicative of high scores on PA,
whereas a score of 21 or more was accepted as indicative
of high scores on NA (Murray et al., in press).
The K10 is a 10-question screening scale of psycho-
logical distress, containing items about the experience of
anxiety and depressive symptoms over the preceding 4
weeks (Kessler and Mroczec, 1994). High scores are
associated with the presence of diagnosed anxiety and
mood disorders as well as consultations for a mental
problem (Andrews and Slade, 2001). Internal reliability
for the K10 has been reported as 0.93, and the
instrument strongly discriminates between community
cases and non-cases of DSM-IV disorders (Kessler et al.,
2002). Given the distribution of the data in the current
study, a score of 21 and above was accepted as indicative
of high levels of distress (see Murray et al., in press).
SPHERE-12 is a short (12-item) form of the SPHERE
questionnaire, designed to measure psychological and
somatic symptoms in a primary care population (Hickie
et al., 2001). Two scores are derived from the scale.
‘‘Psych6’’ refers to psychological symptoms (both
anxiety and depressive) and ‘‘Soma6’’ refers to somatic
symptoms. The internal reliability for Psych6 and
Page 7
ARTICLE IN PRESSC. Fraser et al. / Health & Place 11 (2005) 157–171 163
Soma6 has been reported as 0.90 and 0.80, respectively
(Hickie et al., 2001). The authors of the scale stipulated a
cut-point of three or more on the Soma6 as indicative of
caseness for somatic symptoms, and two or more on the
Psych6 as indicative of caseness for psychological
symptoms (Hickie et al., 2001).
For substance use, participants were assessed on their
use of alcohol, cigarettes, marijuana and amphetamines
by single-item questions. Participants were asked to
choose the option that best described their typical use of
substances. For the purposes of this report, we
dichotomised the responses as follows: for alcohol, the
options were: alcohol 1=never or rarely drinks versus
2=drinks 1or 2 or more standard drinks per day. For
cigarette use: 1=non-smoker (plus ex-smoker) versus
2=smokes. For marijuana use: 1=never used versus
2=used, whilst for amphetamine use: 1=never used
versus 2=used.
Disability
The SF-12 is an abbreviated (12-item) version of the
Medical Outcomes Study 36-Item Short Form Health
Survey (Ware et al., 1996)—a multipurpose short-form
measure of disability. Two subscale scores (physical
component summary (PCS) and mental component
summary (MCS)) are calculated using weights designed
to maximise separation between the two factors. Given
the distribution of the scores in the current study, scores
of 41 or more for the PCS, and 46 or more for the MCS,
respectively, were indicative of good functioning (see
Murray et al., in press).
Well-being
Satisfaction with life was measured on the five-item
satisfaction with life scale (Diener et al., 1985), for which
a range of validating data exists (Pavot and Diener,
1993). Internal consistency has been reported as
0.87, and 2-month test–retest reliability at 0.82
(Diener et al., 1985). Again, given the absence of a
recognised cut-point, based on the distribution of the
current data set, a score of 21 or more was accepted as
indicative of greater satisfaction with life (see Murray
et al., in press).
Results response rate
Of the initial 20,000 individuals contacted, 3106 were
identified as ineligible (2823 incorrect addresses,1 155
deceased, 128 incapable). Of these, 1922 could be
1 The high number of incorrect addresses was due to the age
and accuracy of the electoral roles used for sampling. Whilst the
most up-to-date version of the roles were obtained for the study
they were still approximately 3 years old.
replaced in the sampling frame during the surveying
period, leaving 1184 individuals from the original frame
who were ineligible and not replaced. Response rate was
calculated as (n returned)/(n mailed�n ineligible) (de
Vaus, 1991). Using this formula, the response rate was
40.5% (N ¼ 7615).
From the sample 1334 respondents were identified as
living in the four LGAs selected. Using the formula
mentioned above, response rates were calculated for
each area with the Shire of Buloke having a significantly
higher response rate (54.4%) compared with Macedon
Ranges (32.4%), Dubbo (31.7%) and Broken Hill
(31.3%).
Socio-demographic variables
Inspection of Table 2 shows that the overall sig-
nificance value for all eight variables achieved statistical
significance.
The mean age of Macedon Ranges residents in our
sample was significantly lower than those of Broken
Hill and Buloke as confirmed by direct contrasts,
although the comparison between Macedon and
Dubbo showed a trend towards significance in the
same direction (p ¼ 0:058). Broken Hill citizens in
our sample were significantly younger than those
living in Buloke (p ¼ 0:03). Dubbo residents in our
sample were also found to be younger than residents
living in Buloke (p ¼ 0:004). There were no differences
between Broken Hill and Dubbo on this variable
(p ¼ 1:00).
As regards gender, a greater percentage of respon-
dents in Dubbo was female than was the case in
the other three LGAs. For marital status, inspection
of the adjusted residuals found that the proportion
of Broken Hill and Buloke residents who were
married was significantly less than the proportion
of residents of Dubbo who were married. Macedon
Ranges had the smallest proportion of people living
alone, followed by Dubbo (see Table 2). The adjusted
residuals found that the proportion of people alone in
Macedon Ranges was smaller than expected by chance,
whereas the percentage living alone in Buloke was
greater than expected by chance. There were no other
significant differences between the four LGAs for
this variable.
Adjusted residuals indicated that individuals living in
Macedon Ranges and Dubbo were more likely to be
employed compared with individuals living in Buloke.
The data in Table 2 also show differences in the
percentages of people with post-secondary qualifications
training. Inspection of adjusted residuals shows that
Macedon Ranges’ residents were more likely to have
post-secondary qualifications compared with residents
in Buloke. There were no other differences between the
four LGAs.
Page 8
ARTICLE IN PRESS
Table 3
Mental health, disability and satisfaction with life variables
Variable Growing Declining
Macedon Ranges n ¼ 210 Dubbo n ¼ 540 Buloke n ¼ 196 Broken Hill n ¼ 388
No. % No. % No. % No. %
SWL 169 82.0 379 74.0 128 68.1 258 70.7
MCS 153 73.2 401 74.5 136 69.7 281 72.4
PCS 176 84.2 393 73.0 134 68.7 266 68.6
Positive affect 157 77.3 383 75.1 133 72.3 252 70.2
Negative affect 154 75.9 403 78.9 141 77.0 258 72.3
Psych6 151 72.2 371 70.7 128 68.1 229 60.6
Soma6 138 66.7 346 65.4 117 61.6 196 52.0
K-10 155 73.8 395 73.6 131 67.2 265 68.5
Note: All numbers and percentages shown represent people who are better functioning and are non-cases.
Table 2
Socio-demographic variables for the four local governmental areas
Variable Growing Declining
Macedon
Ranges
n ¼ 210
Dubbo
n ¼ 540
Buloke
n ¼ 196
Broken Hill
n ¼ 388
Chi-Square or
ANOVA
Significance
value
Age (age in years) means (SDs) 47.64 (15.43) 51.07 (16.18) 55.73 (17.04) 51.76 (16.12) F ð3Þ ¼ 8:51 0.000
Gender (no. and % of females) 109 (52%) 333 (62%) 106 (55%) 207 (55%) X 2ð3Þ ¼ 8:58 0.04
Married (no. and %) 156 (75%) 410 (77%) 125 (64%) 262 (67%) X 2ð3Þ ¼ 14:18 0.003
Living alone (no. and % alone) 19 (9%) 77 (14%) 51 (26%) 70 (18%) X 2ð3Þ ¼ 25:01 0.000
Employment (no. and
% employed)
141 (67%) 321 (60%) 95 (49%) 179 (46%) X 2ð3Þ ¼ 33:10 0.000
Education (no. and
% tertiary educated)
121 (57%) 227 (42%) 56 (29%) 153 (40%) X 2ð3Þ ¼ 34:33 0.000
Nationality (no. and
% Australian born)
176 (84%) 511(94%) 186 (96%) 368 (95%) X 2ð3Þ ¼ 36:30 0.000
Aboriginal or Torres Street
Islander (no. and %)
2 (1.0%) 14 (2.6%) 1 (0.5%) 8 (2.1%) X 2ð3Þ ¼ 4:49 0.21
Duration of residence in the
area (no. and % with less than 5
years in area)
22 (10.5%) 38 (7.1%) 19 (4.9%) 7 (3.6%) X 2ð3Þ ¼ 10:06 0.02
C. Fraser et al. / Health & Place 11 (2005) 157–171164
Table 2 indicates that Macedon Ranges had the
lowest proportion of Australian-born citizens. The
adjusted residuals data confirmed that this was indeed
the only significant difference for the four LGAs.
Differences in the proportions of people of Aboriginal
or Torres Street Islander descent were not significantly
different across the four LGAs.
Finally, as regards the duration of time a person was
resident in the area (coded more than 5 years in the area
versus 5 years or less in the area), adjusted residuals
showed that Macedon Ranges contained a greater
proportion of new residents than would be expected by
chance.
Mental health, disability and satisfaction with life
variables
Table 3 shows the numbers and percentages of people
within each of the four LGAs who have high levels of
satisfaction with life and do not have mental (MCS) or
physical (PCS) disability. For all three variables, both
growing areas obtained higher percentages than the two
declining areas and for two of those three variables,
Macedon Ranges contained higher percentages of
‘‘better functioning’’ people than Dubbo.
Table 3 indicates that for the two measures of
affect—positive and negative, both growing areas
Page 9
ARTICLE IN PRESS
Table 4
Logistic loglinear models for psychological variables and local government area: contrasts for growth/decline
Variable Comparison of growth/decline model
with saturated model: difference in
likelihood ratio statistics (df=2)
Comparison of growth/decline model
with main effects model: difference in
likelihood ratio statistics (df=1)
Odds ratios for growth/decline
contrast (with 95% confidence
intervals)
SWL 5.82 (p ¼ 0:054) 6.78 (po0:05) 1.18 (1.04–1.34)
MCS 0.59 (p > 0:05) 1.15 (p > 0:05) 1.07 (0.95–1.20)
PCS 10.97 (po0:05) 9.41 (po0:05) 1.21 (1.07–1.37)
PA 0.66 (p > 0:05) 3.70 (p > 005) 1.13 (0.998–1.28)
NA 2.21 (p > 0:05) 2.87 (p > 005) 1.12 (0.98–1.28)
Psych6 3.25 (p > 0:05) 9.41 (po0:05) 1.20 (1.07–1.35)
Soma6 4.84 (p > 0:05) 15.0 (po0:05) 1.25 (1.12–1.40)
K10 0.11 (p > 0:05) 4.96 (po0:05) 1.15 (1.02–1.29)
C. Fraser et al. / Health & Place 11 (2005) 157–171 165
contained a higher percentage of people who scored
over the cut-score in terms of PA and under the cut-
point for NA.
Table 3 shows that the two growing areas have higher
proportions of people who are non-cases on the two
SPHERE subscales and the K-10 than the two declining
areas. For the Somatic subscale of the SPHERE, Broken
Hill contained a comparatively low number of non-cases.
Table 4 provides results for logistic regression models
for each of the dichotomised psychological variables
depicted in Table 3, using a growth/decline contrast. The
saturated model includes the interaction between the
psychological variable and the categorical variable
specifying the four LGAs. The growth/decline model
includes the interaction between the psychological
variable and a categorical variable pertaining to
growth/decline (in fact, a contrast on the four-category
LGA variable). The main effects model includes only the
main effect for LGA. If there is a significant difference
between the growth/decline model and the saturated
model (column 2 in the table), then the growth/decline
contrast does not explain all of the distribution of the
psychological variable across the four LGAs. If there is
a significant difference between the growth/decline
model and the main effects model, however, (column
3) the growth/decline contrast explains some of the
distribution of the psychological variable. The fourth
column provides 95% confidence intervals for the odds
ratio for the growth/decline contrast in the growth/
decline model. As usual, a confidence interval for the
odds ratio that does not contain 1 represents a
significant result. The coding of the variables is such
that a confidence interval above 1 indicates a better
outcome in growing LGAs.
Accordingly, a pattern with po0:05 in column 3 and
p > 0:05 in column 2 provides evidence that variation in
the psychological measures across the four LGAs is
significant and that this variation can be more parsimo-
niously explained by growth/decline rather than by the
four-fold categorisation. We see this result for Psych6,
for Soma6 and for K10, and marginally for SWL. A
pattern with po0:05 in column 3 and po0:05 in column
2 provides evidence that variation in the psychological
across the four LGAs is significant, that this variation
can be partly explained by growth/decline, but there is
still some variation attributable to differences among the
individual LGAs. We see this pattern for PCS. Finally, a
pattern with p > 0:05 in both columns 2 and 3 suggests
that the psychological variable is independent of LGA
and of growth/decline. We see this for MCS, PA and
NA.
For the variables Psych6, Soma6, K10 and SWL, we
investigated a logistic regression model controlling for
sex, age category (less than 30, 30–50, 50–70, over 70),
education (tertiary or not), nationality (country of
origin), duration of time domiciled in the area and
household arrangements (living alone or not). In each of
these models controlling for demographic characteris-
tics, the odds ratio for the growth contrast remained
significant (for K10, 1.14, 95% CIs=1.00–1.29; for
Psych6, 1.19, 95% CIs=1.05–1.34; for Soma6, 1.22,
CIs=1.08–1.37; for SWL, 1.15, CIs=1.01–1.31).
Employment status was not included in the model as it
may be either an outcome or a compositional factor
(part of the growth/decline construct).
Substance use
Substance use across the four LGAs is shown in
Table 5 whilst Table 6 contains the logistic regressions
for substance abuse variables and the LGA contrasts for
growth/decline.
The two tables for dichotomous substance abuse
variables show no evidence that growth/decline relates
to substance use. The coding of the variables in Table 4
is such that a confidence interval above 1 indicates
proportionately lower substance use in growing areas.
The 95% confidence interval for alcohol suggests the
possibility of a trend for proportionately greater alcohol
Page 10
ARTICLE IN PRESS
Table 6
Logistic loglinear models for substance abuse variables and local government area: contrasts for growth/decline
Variable Comparison of growth/decline
model with saturated model:
difference in likelihood ratio
statistics (df=2)
Comparison of growth/decline
model with main effects model:
difference in likelihood ratio
statistics (df=1)
Generalised odds ratios for
growth/decline contrast (with
95% confidence intervals)
Alcohol 4.56 (p > 0:05) 2.41 (p > 0:05) 0.92 (0.82–1.02)
Smoking 5.65 (p ¼ 0:059) 0.11 (p > 0:05) 0.98 (0.86–1.12)
Marijuana 0.45 (p > 0:05) 0.67 (p > 0:05) 0.88 (0.64–1.20)
Note: Because of the very small number of amphetamine users within each of the four local government areas, loglinear modelling was
not conducted for amphetamine use.
Table 5
Numbers (and percentages) of people engaging in substance use across the four local government areas
Variable Growing Declining
Macedon Ranges
n ¼ 210
Dubbo
n ¼ 540
Buloke
n ¼ 196
Broken Hill
N ¼ 388
Smoking (no. and % of smokers) 42 (20%) 111 (21%) 28 (14%) 87 (23%)
Alcohol use (no. and % who drink alcohol regularly) 96 (46%) 202 (37%) 70 (36%) 137 (35%)
Amphetamine use (no. and % of amphetamine users) 2 (1%) 4 (0.7%) 0 (0%) 3 (0.8%)
Marijuana use (no. and % of marijuana users) 8 (3.8%) 20 (3.7%) 7 (3.6%) 10 (2.6%)
C. Fraser et al. / Health & Place 11 (2005) 157–171166
use in growing areas (in fact, this trend is largely being
driven by greater alcohol usage in Macedon shown in
Table 3).
We also investigated logistic regressions for a three-
level categorisation of alcohol and smoking behaviours
(no use, occasional or light use, heavy use). Again, there
was no evidence for the importance of a growth/decline
contrast for alcohol use. For cigarette use, an interaction
effect between smoking and growth suggested evidence
for proportionately more heavy, rather than light,
smoking in declining areas, with the odds-ratio esti-
mated at 1.33, with 95% confidence intervals from 1.04
to 1.71.
Discussion
Few studies have examined the impact of local
environments outside urban settings and as Kobetz
et al. (2003) have argued, there are limited theoretical
perspectives to assist in understanding how rural settings
influence health. We have drawn upon the health and
place literature to understand how one aspect of place in
rural environments, population growth and decline, can
influence mental health. In addition, we have also
examined migration as a factor that may influence the
prevalence of mental health problems in rural areas.
There have been numerous studies into migration and
mental illness (Dear and Wolch, 1987; Dembling et al.,
2002; Gleeson et al., 1998). These studies have tended to
focus on the migration and concentration of people
experiencing mental illness in inner-city ‘service depen-
dent ghettos’. In contrast to these studies we focused on
the movement of people to, and within, rural areas and
its impact upon the prevalence of mental health
problems in these communities.
This study found clear demographic differences
between growing and declining areas that support our
initial hypothesis that residents in growing areas would
be more likely to be younger, employed and have higher
levels of education; however, the pattern was more
complex than initially anticipated. The disparity be-
tween growing and declining regions was driven
primarily by differences between the Shires of Macedon
Ranges and Buloke. On a number of variables (e.g., age,
education, living alone and marital status) there were no
significant differences between the LGAs of Dubbo and
Broken Hill. There were also differences within growing
and declining areas. There was a higher proportion of
people who had tertiary education and were employed in
Macedon Ranges, compared with Dubbo, and the
average age of Dubbo residents was higher. In the
declining regions a lower proportion of those living in
Buloke had tertiary qualifications, whilst the average age
of Buloke residents was also higher compared with
Broken Hill. These findings indicate that population
growth and decline did not lead to uniform demographic
changes, rather it produced different outcomes in
Page 11
ARTICLE IN PRESSC. Fraser et al. / Health & Place 11 (2005) 157–171 167
different areas. In addition to examining growth or
decline it may also be necessary to understand the
drivers of these changes. In this study, growth in
Macedon Ranges is predominantly due to exurban
migration, bringing younger, more affluent families into
the region (Department of Human Services, 1999). The
growth of Dubbo, by contrast, has been attributed to
the decline of smaller rural towns in the region and the
subsequent migration of residents from these declining
towns to Dubbo (Eves, 1998). Decline in Buloke is due
to the downturn in agriculture (Department of Educa-
tion and Training, 2003) whilst the decline in Broken
Hill is primarily due to the closure of the major mines in
the area (New South Wales Department of Mineral
Resources, 2002).
In addition to examining compositional differences
between growing and declining regions, we also looked
at contextual differences that were likely to occur in
these areas. Population decline has clear impacts upon
place in terms of both the built environment and the
socio-cultural aspects of the environment (Black et al.,
2000; Jones and Tonts, 2003; McKenzie, 1994a). Lynch
et al.’s (Lynch, 2000) concept of neo-material environ-
ment is particularly useful in understanding how these
environmental characteristics can influence health out-
comes. This concept has particular resonance for
declining rural towns, where communities have experi-
enced deterioration in their material living condition
through growing unemployment and the concurrent
withdrawal of community infrastructure such as schools,
transport and hospitals. Again, however, our findings
only provide some support for these arguments. On the
measures of mental health and mental well-being there
were a number of small but significant differences (with
odds ratios indicating about 15–25% difference) be-
tween growing and declining areas and across the four
regions. In general, the data tended to indicate that
those living in declining areas were less happy (lower
PA), less satisfied with life, and had a greater number of
psychological and physical symptoms. However, there
were a number of measures on which there were no
differences between growing and declining communities.
Drug and alcohol use does not vary across the regions.
There was a trend, although not significant, for people
who lived in growing areas being more likely to report
alcohol use. This trend may in part be due to the
younger average age of people living in Macedon
Ranges, as in Australia, older age groups (50 years
and over) are less likely to be smokers and drinkers
(AIHW, 2002; Fleming, 2003).
In general, the findings of this study of population
growth and decline have some relevance to health that
may be due to the spatial redistribution of those at risk
of mental illness, or due to changes in the neo-material
environment, or social comparisons. Future research
investigating the spatial variations in mental health and
well-being across rural communities requires a more
fine-grained, longitudinal approach that examines the
physical and social properties of rural communities and
the processes of change in these areas.
Limitations
First, the research was not conducted longitudinally
and as such a causal relationship cannot be drawn from
the associations identified between declining rural
communities and the socio-demographic characteristics
and psychological measures. The second limitation
concerns the representativeness of the sample. Although
potential participants were randomly selected from the
electoral roll in each area, the design cannot guarantee
that respondents were representative of these commu-
nities and the communities selected for the study may
not reflect the experiences of all declining and growing
rural areas. It is also possible that a bias was produced
by the reliance upon self-report questionnaires, as those
with severe mental health problems may be less likely to
participate.
Another limitation of the study was the geographic
areas sampled and the geographic units of analysis used.
The study did not include a growing coastal town.
Coastal towns are some of the fastest growing rural
regions in Australia. Unlike the growing exurban area
that was sampled in this study, which predominantly
attracts high-income families, studies of coastal towns
have found that they can attract both high- and low-
income households (Beer et al., 1994). This factor limits
the applicability of findings of the study to growing
coastal regions. The study also used LGAs. LGAs are
useful in that data are routinely collected at the LGA
level. However, these districts can be both spatially and
compositionally diverse. Thus, one town’s experiences
may be different from those of another town even
though they ‘‘placed’’ within the same LGA. This
suggests the importance of the concept of ‘‘place’’ and
use of smaller geographic units of analysis.
Conclusion
The current study is important because to the best of
our knowledge this is the first study to empirically
examine mental health, substance use, well-being and
disability across four areas indicative of either popula-
tion growth or decline. The study used measures with
good psychometric properties to index the various
constructs of interest. A large population was surveyed
via a postal survey and we obtained a good response rate
for such a survey.
This study found that there is some relationship
between mental health and living in a community with a
Page 12
ARTICLE IN PRESSC. Fraser et al. / Health & Place 11 (2005) 157–171168
declining population. Importantly, our significant
results for three of the mental heath measures
(i.e., Psych6, Soma6, K10) and the SWL—a measure
of well-being, held up even when six demographic
variables were entered into the regression models as
covariates. This suggests, that at least for the demo-
graphic variables we investigated, it is simply not an
effect of composition of these areas that accounts for
the results. There is evidence that the very fact of
living in declining areas might lead to some vulnerability
in terms of mental health and reduced well-being.
However, the study also raised questions about the
causal mechanisms between population change and
mental health. The findings suggest that drivers
behind population growth and decline could play a
role in influencing the level of mental health problems
in a community; nonetheless, further longitudinal
studies are required to develop a better understanding
of this.
Whilst this study has focused on the Australian
context, the findings have broader relevance. The
experiences of rural desertification and counterurbanisa-
tion are not unique to Australia, but continue to occur
in many developed nations including the US (Beale,
1977), UK (Cawely, 1984; Stockdale et al., 2000), parts
of Europe including France (Dean, 1988), Germany
(Kontuly and Vogelsang, 1988), Italy (Dematteis, 1986)
and Norway (Illeris, 1984). Based on this study,
generalisations about the specific impacts of mental
health on these changes cannot be made. However,
data strongly suggest that changes to place that are
important to health have occurred in many communities
that have experienced population growth or decline.
Impacts that have been reported in studies from the US,
the UK and Europe include increased fear of crime
(Hunter et al., 2002) increased rates of crime and
other social problems (Broadway, 2000), hostility
between ‘locals’ and ‘newcomers’ (Murdoch and Day,
1998), rising property prices (Gilligan, 1987), changes
in local employment patterns and the abandonment
of formerly productive land (Cawley, 1994). As popula-
tion growth is a symptom of broader dynamics,
the drivers behind growth and decline in these commu-
nities will also influence how these changes are played
out at the local level and the impact they may have on
health.
What is clearly demonstrated in the study is that there
are different patterns of mental health among rural
communities. This finding is important in terms of rural
mental health policy and research, which tends to have
an unsophisticated conceptualisation of what is ‘rural’
and treats all areas beyond the metropolitan fringe as
homogeneous. The study also suggests that population
growth and decline is an important aspect of place that
may warrant further investigation in terms of its impact
on health outcomes.
Acknowledgements
The authors gratefully acknowledge the financial
support provided by beyondblue: The National Depres-
sion Initiative.
References
Andrews, G., Slade, T., 2001. Interpreting scores on the Kessler
psychological distress scale (K10). Australian and New
Zealand Journal of Public Health 25, 494–497.
Andrews, G., Hall, W., Teeson, M., Henderson, S., 1999. The
National Survey of Mental Health and Wellbeing: the
Mental Health of Australians, Mental Health Branch.
Commonwealth Department of Health and Aged Care,
Canberra.
Argent, N., Rolley, F., 2000. Financial exclusion in rural and
remote New South Wales, Australia: a geography of bank
branch rationalisation, 1981–1998. Australian Geographical
Studies 38, 182–203.
Australian Bureau of Statistics, 2003a. Australian Social
Trends. Family and Community—Living Arrangements:
Farming Families. Australian Bureau of Statistics,
Canberra.
Australian Bureau of Statistics, 2003b. Estimated Resident
Population. Local Government Areas, Victoria (Table 1b)
[Cat. No. 3218.0.55.001].
Australian Institute of Health and Welfare (AIHW), 2002. 2001
National Drug Strategy Household Survey First
Results. Australian Institute of Health and Welfare,
Canberra, PHE 35.
Beale, C., 1977. The recent shift of United States population to
non-metropolitan areas. International Regional Science
Review 2, 113–122.
Beer, A., Bolam, A., Maude, A., 1994. Beyond the Capitals:
Urban Growth in Regional Australia. Australian Govern-
ment Publishing Service, Canberra.
Black, A., Duff, J., Saggers, S., Bains, P., Jennings, A., Bowen,
P., 2000. Rural Communities and Rural Social Issues:
Priorities for Research. Rural Industries Research and
Development Corporation, Canberra.
Blazer, D., George, L.K., Landerman, R., Pennybacker, M.,
Melville, M.L., Woodbury, M., Manton, K.G., Jordan, K.,
Locke, B., 1985. Psychiatric disorders. A rural/urban
comparison. Archives of General Psychiatry 42, 651–656.
Broadway, M.J., 2000. Planning for change in small towns or
trying to avoid the slaughterhouse blues. Journal of Rural
Studies 16, 37–46.
Bruce, M.L., Takeuchi, D.T., Leaf, P.J., 1991. Poverty and
psychiatric status: longitudinal evidence from the new haven
epidemiologic catchment area study. Archives of General
Psychiatry 48, 470–474.
Budge, T., 1996. Population decline in Victoria and Tasmania.
In: Newton, P., Bell, M. (Eds.), Population Decline in
Victoria and Tasmania. Australian Government Publishing,
Canberra, pp. 192–204.
Burch, D., Rickson, R., 2001. Industrialised agriculture:
agribusiness, input-dependency and vertical integration.
In: Lockie, S., Bourke, L. (Eds.), Rurality Bites: the Social
Page 13
ARTICLE IN PRESSC. Fraser et al. / Health & Place 11 (2005) 157–171 169
and Environmental Transformation of Rural Australia.
Pluto, Annandale, pp. 165–178.
Cawley, M.E., 1994. Desertification: measuring population
decline in rural Ireland. Journal of Rural Studies 10,
395–407.
Champion, A., 1989. Counterurbanization. Edward Arnold,
London.
Cloke, P., 1996. Rural life-styles: material opportunity, cultural
experience, and how theory can undermine policy. Econom-
ic Geography 72, 433–449.
Curry, G., Koczberski, G., Selwood, J., 2001. Cashing out,
cashing in: rural change on the south coast of Western
Australia. Australian Geographer 32, 109–124.
Curry, S.J., Wagner, E.H., Cheadle, A., Diehr, P., Koepsell, T.,
Psaty, B., McBride, C., 1993. Assessment of community-
level influences on individuals’ attitudes about cigarette
smoking, alcohol use, and consumption of dietary fat.
American Journal of Preventive Medicine 9, 78–84.
Curtis, S., Rees Jones, I., 1998. Is there a place for geography in
the analysis of health inequality. In: Bentley, M., Blanc, D.,
Dovey Smith, G. (Eds.), The Sociology of Health Inequal-
ities. Blackwell Publishers, Oxford, pp. 85–111.
Dahms, F., McComb, H., 1999. Counterurbanisation, interac-
tion and functional change in a rural amenity area—a
Canadian example. Journal of Rural Studies 15, 129–146.
Dean, K.G., 1988. Inter-regional flows of economically active
persons in France 1975–1982. Demography 25, 81–98.
Dear, M., Wolch, J., 1987. Landscapes of Despair: From
Deinstitutionalization to Homelessness. Polity, Cambridge.
Dematteis, G., 1986. Urbanisation and counterurbanisation in
Italy. Ekistics 316/317, 26–33.
Dembling, B.P., Rovnyak, V., Mackey, S., Blank, M., 2002.
Effect of geographic migration on SMI prevalence esti-
mates. Mental Health Services Research 4, 7–12.
Department of Education and Training, 2003. Local learning
and employment network statistics resource. Department of
Education and Training, Victoria.
Department of Human Services, 1999. A Profile of the Shire
of Macedon Ranges. Department of Human Services,
Bendigo.
Diener, E., Emmons, R.A., Larsen, R.J., Griffin, S., 1985. The
satisfaction with life scale. Journal of Personality Assess-
ment 49, 71–75.
de Vaus, D.A., 1991. Surveys in Social Research. Allen &
Unwin, St Leonards.
Eves, C., 1998. A comparison of agricultural, residential and
commercial property in rural regions. Australian Property
Journal, August, 273–278.
Fielding, A., 1990. Counterurbanisation: threat or blessing? In:
Pinter, D. (Ed.), Western Europe: Challenge and Change.
Belhaven, London, pp. 226–239.
Fincher, R., 1999. New geographies of disadvantage: implica-
tions for service delivery. Australian Journal of Public
Administration 58, 55–60.
Fincher, R., Wulff, M., 1997. Locations of poverty and
disadvantage. In: Fincher, R., Nieuwenhuysen, J. (Eds.),
Australian Poverty Now and Then. Melbourne University
Press, Melbourne, pp. 144–164.
Fincher, R., Wulff, M., 2001. Moving in and out of
disadvantage: population mobility and Australian places.
In: Fincher, R., Saunders, P. (Eds.), Creating Unequal
Futures?: Rethinking Poverty, Inequality and Disadvan-
tage. Allen & Unwin, Crows Nest, pp. 158–193.
Fitchen, J.M., 1995. Spatial redistribution of poverty through
migration of poor people to depressed rural communities.
Rural Sociology 60, 181–202.
Fleming, J., 2003. The epidemiology of alcohol use in
Australian women: findings from a national survey of
women’s drinking. Addiction 91, 1324–1334.
Fraser, C., Judd, F., Jackosn, H., Murray, G., Humphrey, J.,
Hodgins, G., 2002. Does one size really fit all? Why the
mental health of rural Australians requires further research.
Australian Journal of Rural Health 10, 288–296.
Gilligan, J., 1987. Visitors, tourists, and outsiders in a Cornish
town. In: Bouquet, M., Winter, M. (Eds.), Who from their
Labours Rest? Conflict and Practice in Rural Tourism.
Aldershot, Avebury, pp. 65–82.
Gleeson, B., Hay, C., Law, R., 1998. The geography of mental
health in Dunedin, New Zealand. Health and Place 4, 1–14.
Gouveia, L., Stull, D.D., 1995. Dances with cows: beefpack-
ing’s impact on garden city, Kansas and Lexington,
Nebraska. In: Stull, D.D., Broadway, M.J., Griffith, D.
(Eds.), Any Way You Cut It: Meat Processing and
Smalltown America. University Press of Kansas, Lawrence,
pp. 85–107.
Greive, S., Tonts, M., 1996. Regulation, land development
and the contested countryside: reflections of Bridge-
town, western Australia. New Zealand Geographer 52,
18–23.
Halpern, D., 1995. Mental Health and the Built Environment:
More than Bricks and Mortar? Taylor & Francis, London.
Hickie, I.B., Davenport, T.A., Hadzi-Pavlovic, D., Koschera,
A., Naismith, S.L., Scott, E.M., Wilhelm, K.A., 2001.
Development of a simple screening tool for common mental
disorders in general practice. Medical Journal of Australia
175, S17.
Hugo, G., 1989. Australia: the spatial concentration of the
turnaround. In: Champion, A. (Ed.), Counterurbanization:
the Changing Pace and Nature of Population Deconcentra-
tion. Edward Arnold, London, pp. 62–82.
Hugo, G., 1994. The turnaround in Australia: some first
observations from the 1994 census. Australian Geographer
25, 1–17.
Hugo, G., 2001. What is really happening in regional and rural
populations. In: Rogers, M.F., Collins, Y.MJ. (Eds.), The
Future of Australia’s Country Towns. La Trobe University,
Bendigo, pp. 57–71.
Hunter, L., Krannich, R.S., Smith, M.D., 2002. Rural
migration, rapid growth, and fear of crime. Rural Sociology
67, 71–89.
Illeris, S., 1984. Danish regional development during economic
crisis. Geografiske Tidsskrift 84, 53–62.
Jones, R., Tonts, M., 1995. Rural restructuring and social
sustainability: some reflections on the western Australian
Wheatbelt. Australian Geographer 26, 133–140.
Jones, R., Tonts, M., 2003. Transition and diversity in
rural housing provision. Australian Geographer 34,
47–59.
Judd, F.K., Jackson, H.J., Komiti, A., Murray, G., Hodgins,
G., Fraser, C., 2002. High prevalence disorders in urban and
rural communities. Australian and New Zealand of
Psychiatry 36, 104–113.
Page 14
ARTICLE IN PRESSC. Fraser et al. / Health & Place 11 (2005) 157–171170
Kaplan, G.A., 1996. People and places: contrasting perspectives
on the association between social class and health. Interna-
tional Journal of Health Services 26, 507–519.
Kawachi, I., Berkman, L., 2003. Neighborhoods and Health.
Oxford University Press, Oxford pp. 20–40.
Kessler, R., Mroczec, D., 1992. An Update of the Development
of Mental Health Screening Scales for the US National
Health Interview Survey. Ann Arbor Science, Ann
Arbor, MI.
Kessler, R., Mroczec, D., 1994. Final Versions of Our Non-
Specific Psychological Distress Scale. Ann Arbor Science,
Ann Arbor, MI.
Kessler, R.C., McGonagle, K.A., Zhao, S., Nelson, C.B.,
Hughes, M., Eshleman, S., Witchen, H., Kendler, K.S.,
1994. Lifetime and 12 month prevalence of DSM-III R
psychiatric disorders in the United States. Results from the
national comorbidity survey. Archives of General Psychia-
try 51, 8–19.
Kessler, R.C., Andrews, G., Colpe, L.J., Hiripi, E., Mroczek,
D.K., Normand, S.-L.T., Walters, E.E., Zaslavsky, A.M.,
2002. Short screening scales to monitor population pre-
valences and trends in non-specific psychological distress.
Psychological Medicine 32, 959–976.
Kobetz, E., Daniel, M., Earp, J.A., 2003. Neighbourhood
poverty and self-reported health among low-income, rural
women, 50 years and older. Health and Place 9, 263–271.
Kontuly, T., Vogelsang, R., 1988. Explanations for the
intensification of counterurbanisation in the Federal Re-
public of Germany. Professional Geographer 40, 42–54.
Lawrence, G., 1996. The future of agriculture and rural society
in central Queensland. In: Cryle, D., Griffin, G., Stehlik, D.
(Eds.), Future for Central Queensland, Rockhampton:
Rural and Social Economic Research Centre. University
of Queensland, Rockhampton, pp. 53–58.
Lawrence, G., 1999. Agri-food restructuring: a synthesis of
recent Australian research. Rural Sociology 64, 186–202.
Lawrence, G., Williams, C., 1990. The dynamics of decline:
implications for social welfare delivery in rural Australia.
In: Cullen, T., Dunn, P., Lawrence, G. (Eds.), Wagga
Wagga: Centre for Rural Welfare Research. Charles Sturt
University, Wagga Wagga, pp. 38–59.
Leighton, D.C., Harding, J.S., Macklin, D.B., MacMillan,
A.M., Leighton, A.H., 1963. The Character of Danger:
Psychiatric Symptoms in Selected Communities. Basic
Books, New York.
Lynch, J., 2000. Income inequality and health: expanding the
debate. Social Science & Medicine 51, 1001–1005.
Lynch, J., Due, P., Muntaner, C., Smith, G.D., 2000. Social
capital—is it a good investment strategy for public health?’.
Journal of Epidemiology and Community Health 54, 404–
408 (comment).
Lynch, J., Smith, G.D., Hillemeier, M., Shaw, M., Raghu-
nathan, T., Kaplan, G., 2001. Income inequality, the
psychosocial environment, and health: comparisons of
wealthy nations. Lancet 358, 194–200.
Macintyre, S., Maciver, S., Sooman, A., 1993. Area, class and
health: should we be studying people or places? Journal of
Social Policy 22, 213–234.
Macintyre, S., Ellaway, A., Cummins, S., 2002. Place effects on
health: how can we conceptualise, operationalise and
measure them. Social Science & Medicine 55, 125–139.
McKenzie, F., 1994a. Population decline in non-metropolitan
Australia: impacts and policy implications. Urban Policy
and Research 12, 253–263.
McKenzie, F., 1994b. Regional Population Decline in Aus-
tralia: Impacts and Policy Implications. Australian Govern-
ment Publishing Service, Canberra.
McMichael, P., Lawrence, G., 2001. Globalising agriculture:
structures of constraint for rural farmers. In: Lockie, S.,
Bourke, L. (Eds.), Rurality Bites: the Social and Environ-
mental Transformation of Rural Australia. Pluto, Annan-
dale, pp. 153–165.
Murdoch, J., Day, G., 1998. Middle class mobility, rural
communities and the politics of exclusion. In: Boyle, P.,
Halfacree, K. (Eds.), Migration into Rural Areas. Wiley,
Chichester, pp. 186–199.
Murray, G., Judd, F., Jackson, H., Fraser, C., Komiti, A.,
Hodgins, G., Pattison, P., Humpheys, J., Robins, G., in
press. A community survey of mental health and well-being
in rural Australia: the role of remoteness. Australian and
New Zealand Journal of Psychiatry.
New South Wales Department of Mineral Resources, 2002.
Broken Hill–Silver City. New South Wales Department of
Mineral Resources. (http://www.minerals.nsw.gov.au/mine/
silver.htm accessed 24/06/2003).
Oakley-Browne, M.A., Joyce, P.R., Wells, J.E., Bushnell, J.A.,
Hornblow, A.R., 1989. Christchurch psychiatric epidemiol-
ogy study, Part 2: six month and other period prevalences of
specific psychiatric disorders. Australian and New Zealand
Journal of Psychiatry 23, 327–340.
Paquette, S., Domon, G., 2003. Changing ruralities, changing
landscapes: exploring social recomposition using a multi-
scale approach. Journal of Rural Studies 19, 425–444.
Pavot, W., Diener, E., 1993. Review of the satisfaction with life
scale. Psychological Assessment 5, 164–172.
Paykel, E.S., Abbott, R., Jenkins, R., Brugha, T.S., Meltzer,
H., 2000. Urban–rural mental health differences in Great
Britain: findings from the national morbidity survey.
Psychological Medicine 30, 269–280.
Productivity Commission, 1999. Impact of Competition Policy
Reforms on Rural and Regional Australia. Australian
Government Publishing, Canberra.
Raiston, D., Beal, D., 1997. Bank branch closures in rural
areas. Australian Banker 3, 126–129.
Reijneveld, S.A., 1998. The impact of individual and area
characteristics on urban socioeconomic differences in health
and smoking. International Journal of Epidemiology 27, 33–40.
Robert, S.A., 1998. Community-level socioeconomic status
effects on adult health. Journal of Health and Social
Behavior 39, 18–37.
Sampson, R.J., Morenoff, J.D., Gannon-Rowley, T., 2002.
Assessing ‘‘neighborhood effects’’ social processes and new
directions in research. Annual Review of Sociology 28, 443–478.
Smailes, P., 1997. Socioeconomic change and rural morale in
South Australian 1982–1993. Journal of Rural Studies 13,
19–42.
Smailes, P., 2000. The diverging geographies of social and
business interaction patterns: a case study of rural south
Australia. Australian Geographical Studies 38, 158–181.
Stockdale, A., Findlay, A., Short, D., 2000. The repopulation
of rural Scotland: opportunity and threat. Journal of Rural
Studies 16, 243–257.
Page 15
ARTICLE IN PRESSC. Fraser et al. / Health & Place 11 (2005) 157–171 171
Stead, M., MacAskill, S., Mackintosh, A., Reece, J., Eadie, D.,
2001. It’s as if your’re locked in: qualitative explanations for
area effects on smoking in disadvantaged communities.
Health and Place 7, 333–343.
Waitzman, N.J., Smith, K.R., 1998. Phantom of the
area: poverty-area residence and mortality in the
United States. American Journal of Public Health 88,
973–976.
Walmsley, D.J., Sorensen, A.D., 1990. Contemporary Austra-
lia: Explorations in Economy, Society and Geography.
Longman Cheshire, Melbourne.
Ware, J.E.J., Kosinski, M., Keller, S., 1996. A 12-item
short-form health survey: construction of scales and
preliminary tests of reliability and validity. Medical Care
34, 220–233.
Warin, M., Baum, F., Kalucy, E., Murray, C., Veale, B., 2000.
The power of place: space and time in women’s and
community health centres in south Australia. Social Science
& Medicine 50, 1863–1875.
Watson, D., Clark, L.A., Tellegen, A., 1988. Development and
validation of brief measures of positive affect and negative
affect: the Panas scales. Journal of Personality and Social
Psychology 54, 1063–1070.
Wilkinson, R.G., 1997. Socioeconomic determinants of health:
health inequalities: relative or absolute material standards?
British Medical Journal 314, 591–595.
Witten, K., McCreanor, T., Kearns, R., Ramasubramanian, L.,
2001. The impacts of a school closure on neighbourhood
social cohesion: narratives from Invercargill, New Zealand.
Health and Place 7, 307–317.