- Email: [email protected]
An investigation into the synergistic wellbeing benefits of greenspace and physical activity: Moving beyond the mean
Urban Forestry & Urban Greening 19 (2016) 7–12
Contents lists available at ScienceDirect
Urban Forestry & Urban Greening journal homepage: www.elsevier.com/locate/ufug
Short communication
An investigation into the synergistic wellbeing benefits of greenspace and physical activity: Moving beyond the mean Christopher L. Ambrey Cities Research Centre, Gold Coast Campus, Griffith University, Queensland 4222, Australia
a r t i c l e
i n f o
Article history: Received 3 June 2015 Received in revised form 17 March 2016 Accepted 21 June 2016 Available online 4 July 2016 Keywords: Geographic information systems (GIS) Greenspace Household Income and labour dynamics in Australia (HILDA) Life satisfaction Mental health Physical activity Psychological distress Wellbeing
a b s t r a c t The purpose of this study is to shed light on: (1) how greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of wellbeing; and also (2) how the potential synergies between greenspace and physical activity might have heterogeneous impacts across the distribution of wellbeing. Using data from the Household, Income and Labour Dynamics in Australia survey and data from Geographic Information Systems this study finds, for the case of major Australian cities, that greenspace and physical activity are independently positively associated with life satisfaction, mental health and negatively associated with psychological distress. A finding which is stronger for physical activity than it is for greenspace. Across measures of life satisfaction, mental health and psychological distress, the results lend support to the hypothesis that physical activity may be relatively more effective at mitigating the likelihood of experiencing a serious dearth of wellbeing, compared to promoting higher levels of wellbeing. Unexpectedly, the results do not provide support for the hypothesised greenspace-physical activity synergy. A result found to be common across the wellbeing distribution. While further research is required to draw a definitive conclusion, this result may indicate that physical activity is simply not conducive to the realisation of the restorative benefits of exposure to nature, and the other co-benefits of ‘green exercise’. In all, the findings presented in this study add to the existing stock of knowledge from a socialecological perspective and also raise new questions for future research. The results presented in this study may also prove useful to policy makers wrestling with the challenges of maintaining or improving residents’ wellbeing and reducing residents’ ill-being in the face of continuing population growth and declining per capita greenspace. © 2016 Elsevier GmbH. All rights reserved.
1. Introduction Throughout the world urbanisation is occurring at an unprecedented rate (United Nations Economic and Social Affairs, 2014a,b; United Nations Population Fund, 2015). Australia, already one of the most urbanised countries in the world (The World Bank, 2015) is sharing in this global trend. However, urbanisation and land use change are irrevocably altering the environment in which many residents live (Sachs, 2008), impinging on recreational spaces like parks and bushland (Lavelle, 2006) and leading to the congestion of those green spaces that remain (Arnberger, 2012) with implications for the health and wellbeing of a city’s local residents (Barton, 2009). Acknowledging this relationship between greenspace and wellbeing a number of studies (cf. Bodin and Hartig, 2003; Hug et al., 2009; Mitchell, 2013; Pretty et al., 2007; Thompson Coon et al., 2011) have hypothesised that physical activity in natural environ-
E-mail address: c.ambrey@griffith.edu.au http://dx.doi.org/10.1016/j.ufug.2016.06.020 1618-8667/© 2016 Elsevier GmbH. All rights reserved.
ments might produce greater mental health benefits than physical activity elsewhere (Mitchell, 2013). This hypothesised synergistic link builds on the well-established physiological and psychological benefits of physical activity (cf. US Department of Health, 1996). It also combines: (1) the restorative effects of contact with a natural environment (cf. Kaplan, 1995); and (2) the co-benefits of ‘green exercise’ for example, feelings of connectedness to nature and an increased appreciation of nature, which may bolster the longevity of one’s engagement in physical activity (Thompson Coon et al., 2011). Despite the best efforts of earlier investigators many caveats continue to surround the triadic relationship between the greenspace, physical activity and wellbeing (Mitchell, 2013). A recent systematic review of the literature by Thompson Coon et al. (2011) demonstrates a real paucity of high quality evidence on which to base recommendations. Of the studies reviewed, the samples tend to contain of between 8 and 269 respondents. Further, most of these respondents are young university students, in the US, who are already physically active. The use of single shortterm walking or running experiments also provides little indication
8
C.L. Ambrey / Urban Forestry & Urban Greening 19 (2016) 7–12
of the efficacy of interventions over the long-term. Moreover, the lack of easily transferrable outcome measures and the presence of concurrent interventions obscures precisely what conclusions may be drawn from some earlier evidence. Distinct from this body of literature, there is evidence emerging from the empirical economic and psychological literature which suggests that positive and negative wellbeing are more than merely opposite ends of the same phenomenon (Boes and Winkelmann, 2010). It may be the case for example, that individuals experiencing mild depression gain larger benefits from physical activity (Thompson Coon et al., 2011). In this regard, the purpose of this study is to go beyond earlier research efforts by revealing both: (1) how greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of wellbeing; and also (2) how the potential synergies between greenspace and physical activity might have heterogeneous impacts across the distribution of wellbeing. In particular, this study investigates the following hypotheses: H1: Whether or not the greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of life satisfaction. H2: Whether or not any greenspace and physical activity synergy differs across the distribution of life satisfaction. H3: Whether or not the greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of mental health. H4: Whether or not the greenspace and physical activity synergy differs across the distribution of mental health. H5: Whether or not the greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of psychological distress. H6: Whether or not the greenspace and physical activity synergy differs across the distribution of psychological distress. In doing so, for the case of major Australian cities this study contributes to the stock of knowledge regarding the interplay between greenspace, physical activity and wellbeing. The findings presented in this study may prove useful to policy makers wrestling with the challenges of maintaining or improving residents’ wellbeing and reducing residents’ ill-being in the face of continuing population growth and declining per capita greenspace. In what follows, Section 2 reports the data and method employed. Section 3 provides an account of the results and Section 4 discusses the findings and Section 5 concludes. 2. Materials and method In terms of the socioeconomic data on the 6082 residents this is obtained from wave 13 (2013) of the Household, Income and Labour Dynamics in Australia (HILDA) Survey, subset to the major capital cities of Australia.1 The sampling design of the survey involves the selection of households into the sample by a multi-stage process. In wave 1 (2001) of the HILDA Survey, a random sample of 488 Census Collection Districts (CDs) based on the 1996 census boundaries was selected from across Australia, stratified by State, and within the five largest States in terms of population, by metropolitan and non-metropolitan regions, each CD consisting of approximately 200–250 households. The CDs were sampled with probability proportional to their size as measured by the number of dwellings (unoccupied and occupied) recorded in the 1996 Census with some adjustments for population growth since the Census. Within each of these CDs, all dwellings were fully enumerated and a sample of 22–34 dwellings randomly sampled based on the
expected response and occupancy rates within each area (Watson and Wooden, 2002). The life satisfaction dependent variable is obtained from residents’ responses to the question: ‘All things considered, how satisfied are you with your life?’ The life satisfaction variable is an ordinal variable, the resident choosing a number between 0 (totally dissatisfied with life) and 10 (totally satisfied with life). The mental health dependent variable is obtained using data from the Short-Form General Health Survey (SF-36) instrument (collected within the HILDA Survey), an internationally recognised diagnostic tool for assessing functional health status and wellbeing. The Mental Component Summary (MCS) used in this study is derived from 14 items on four scales; vitality, social functioning, role-emotional and mental health, transformed to a 0–100 index using 1995 Australian Bureau of Statistics population norms (Australian Bureau of Statistics, 1995; Ware et al., 2000). A higher mental health score indicates better mental health while a lower mental health score indicates the reverse. The psychological distress dependent variable is measured by the Kessler Psychological Distress Scale (K10) also collected in the HILDA Survey. The ten questions and their selection are described at length in Kessler et al. (2002), as explained by Wooden (2009) the K10 score was derived by scoring responses on each of the items using a simple linear scale running from 5 (all of the time) to 1 (none of the time), and summing across all items. The overall score thus ranges from 10 to 50, where a higher score indicates greater psychological distress and a lower score indicates lower psychological distress. Apart from the different dependent variables employed, the key measure of physical activity is derived from dichotomising the total physical activity Metabolic Equivalent of Task (MET) minutes per week (International Physical Activity Questionnaire). The variable is defined as ‘Exercising as recommended (MET)’ (1) or not (0), where exercising as recommended is defined as MET minutes per week greater than 840 and less than 10,000. That is, the equivalent of 30 or more min × week × 4 MET. To avoid measurement error due to over-reporting, those reporting energy expenditure of 10,000 MET (min/week) or more were excluded (Giles-Corti and Donovan, 2002). Exercising as recommended was defined as the accumulation of the equivalent of 30 min or more of moderate physical activity on most days of the week (US Department of Health, 1996). Data from the HILDA survey are linked to Geographic Information Systems (GIS) data on greenspace through the resident’s Census Collection District (CD). Using GIS CDs are overlayed with greenspace measured from the PSMA Australia Limited Transport and Topography dataset (the 2010 release). Greenspace includes for instance, public parks, community gardens, cemeteries, sports fields, national parks and wilderness areas (cf. Bell et al., 2008). The variable is the number of hectares of greenspace per resident in the CD. A detailed description and summary statistics of the key variables are provided Table 1.2 Hypotheses 1–6 are investigated through the estimation of four conditional logistic regression models, one for each quartile of the dependent variable within a seemingly unrelated regression (SUR) system of equations.3 The dependent variable WBq1...4 r,k represents a resident’s life satisfaction, or mental health or psychological distress each of which have been disaggregated into q1...4 quartiles. The different quartiles for each wellbeing measure allow heterogeneity across the distribution to be revealed.
2
Full summary statistics and results are available as supplementary material. The models estimated for each quartile are based on the same data, they are not independent from each other and hence their residuals are likely correlated. As such the regression results for the separate estimations, obtained using Stata/SE 13.1, are combined using Stata’s ‘suest’ postestimation command. 3
1 Major capital cities in Australia include: Adelaide, Brisbane, Canberra, Darwin, Melbourne, Perth and Sydney.
C.L. Ambrey / Urban Forestry & Urban Greening 19 (2016) 7–12
9
Table 1 Key model variables. Variable name
Definition
Dependent variables Life satisfaction
Mean (std. dev.)
Resident’s life satisfaction (0–10) Quartile Min. Max.
Mental health
7.9 (1.4) Q1 0 7
Resident’s SF-36 Mental Component Summary (MCS) (0–100) Quartile Min. Max.
Psychological distress
Greenspace (ha) per capita
Population density (residents per ha)
Q2 8 8
Q3 9 9
Q4 10 10 48.8 (10.3)
Q1 3.0 44.1
Q2 44.1 51.5
Q3 51.5 56.0
Q4 56.0 71.7 15.7 (6.2)
Resident’s Kessler Psychological Distress Scale (K10) score (10–50) Quartile Min. Max.
Independent variables Exercise as recommended (MET)
%
Q1 10 10
Resident has Metabolic Equivalent of Task (MET) minutes per week greater than 840 and less than 10,000. That is, the equivalent of 30 or more min × week × 4 MET The amount of greenspace in a resident’s CD per resident in the CD. Greenspace in this instance, includes for instance, public parks, community gardens, cemeteries, sports fields, national parks and wilderness areasa Residents in the CD per hectare
Q2 11 13
Q3 14 17
Q4 18 50 63.6%
20.1 (27.6)
61.0 (62.9)
a
For reasons of confidentiality, the CD is lowest level of spatial reference for residents in the HILDA survey. While this means that the precise residential address of respondents is unknown; this spatial reference also provides an upper bound for the degree of measurement error. To be specific, if the CD polygons are assumed to take approximately the shape a circle, the resident is at most, on average (or median), 614 (309) metres from the captured in the greenspace variable. Major factors influencing the use of greenspace are size and proximity (Schipperijn et al., 2010); the use of the CD permits the convenient synthesis of these two factors into a single variable. Despite the precise location of where the resident engages in physical activity and the usability of the greenspace remaining unknown; it is reasonable to expect that residents are more likely to use or experience the greenspace that is nearby (cf. Coombes et al., 2010; Larkin, 2003).
Eq. (1) illustrates the general form of the quartile-specific conditional logistic regression models, for resident r in location k, within the seemingly unrelated regression (SUR) system: WBq1...4 r,k = 0 +
m
j=1
j xjr,k + ␥xr,k zr,k + k + r,k
(1)
Where, xjr,k denotes variables j = 1. . .m. These variables include for example, age, gender, ethnicity and importantly the measure of physical activity and the measure of greenspace. xr,k zr,k represents the two-way interaction term of greenspace × physical activity. k represents the Local Government Area (LGA)-specific fixed effects. Finally, r,k is the error term. It is worth noting that the terms greenspace and physical activity that are not interacted in each quartile-specific conditional logistic regression provide evidence on hypotheses 1, 3 and 5. In contrast, the two-way interaction terms (greenspace × physical activity) in each quartile-specific conditional logistic regression yield evidence on hypotheses 2, 4 and 6. 3. Results The key results, for Eq. (1) are reported in Tables 2–4. With regards to hypothesis 1, Table 2 columns 1–4 suggest that, independently and distinct from any synergistic effect, both greenspace (odds of reporting a life satisfaction score in the first quartile = 0.9452, [0.9023,0.9902], p-value < 0.05) and physical activity (odds of reporting a life satisfaction score in the first quartile = 0.8128, [0.7114,0.9288], p-value < 0.01), reduce the likelihood of reporting a life satisfaction score in the first quartile. In terms of hypothesis 2, while the estimates for greenspace and physical activity in column 4 are stronger and more precise than those of the other quartiles, the results provide no evidence of the
hypothesised greenspace-physical activity synergy at conventional levels of statistical significance. This lack of statistical significance appears to be a homogenous result across the life satisfaction distribution. Table 3 columns 1–4 suggest, pertinent to hypothesis 3, that greenspace is only statistically significant for the third quartile (odds of reporting a mental health score in the third quartile = 1.0997, [1.0213,1.1841], p-value < 0.05), net of any synergistic effects. Furthermore, physical activity is found to have an independently greater estimated effect on reducing the likelihood of reporting a mental health score in the first quartile (odds of reporting a mental health score in the first quartile = 0.7828, [0.6995,0.8760], p-value < 0.01) compared to the other quartiles, distinct from any hypothesised synergistic psychological benefits. Table 3 columns 1–4 yield no evidence of synergistic psychological benefits across the mental health distribution (hypothesis 4). For the case of hypothesis 5 which relates to psychological distress4 Table 4 columns 1–4 reveal that only physical activity, net of hypothesised synergistic psychological benefits is generally associated with lower levels of psychological distress. Greenspace is not found to be statistically significantly linked to psychological distress holding constant any synergistic effect. However, physical activity (odds of reporting a psychological distress score in the first quartile = 1.3025, [1.1020,1.5396], p < 0.01) is linked to a higher estimated likelihood of reporting lower levels of psychological distress. Corroborating this general finding, although marginally less compelling, the results for the third (odds of reporting a psycho-
4 In contrast to the preceding life satisfaction and mental health results (for which higher quartiles relate to higher levels of wellbeing); higher psychological distress quartiles relate to lower levels of wellbeing (see Section 2).
10
C.L. Ambrey / Urban Forestry & Urban Greening 19 (2016) 7–12
Table 2 Key life satisfaction model resultsa . (1)
(2)
(3)
(4)
First quartile
Second quartile
Third quartile
Fourth quartile
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
0.9452** (0.0224) [0.9023,0.9902]
1.2037 (0.2054) [0.8616,1.6816]
0.8673 (0.2378) [0.5067,1.4845]
0.6022 (0.1861) [0.3286,1.1035]
Exercise as recommended (MET)
0.8128*** (0.0553) [0.7114,0.9288]
1.0464 (0.0608) [0.9337,1.1726]
1.0776 (0.0803) [0.9312,1.2471]
1.1550 (0.1247) [0.9347,1.4271]
Greenspace (ha) per capita × Exercise as recommended (MET)
0.9908 (0.0766) [0.8514,1.1530]
0.7333 (0.1477) [0.4942,1.0882]
1.3784 (0.3850) [0.7972,2.3831]
1.6605 (0.5187) [0.9002,3.0630]
Life satisfaction Greenspace (ha) per capita
– Summary statistics Observations Groups
6082 131
**
p < 0.05. p < 0.01 Standard errors in parentheses adjusted for clustering at the LGA level. Other controls included: Age, gender, ethnicity, marital status, parenting, health, educational attainment, employment status, manual work, income, social desirability bias, free time, social interaction, household members engaged in physical activity, personality traits, years at current address, years interviewed, proximity to lake, proximity to river, proximity to coastline and the SEIFA 2011 Index. a Note, variance inflation factors of a base model with no interaction terms shows no sign of worrisome multicollinearity. ***
Table 3 Key mental health model resultsa . (1)
(2)
(3)
(4)
First quartile
Second quartile
Third quartile
Fourth quartile
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
0.9735 (0.0323) [0.9122,1.0389]
0.8105 (0.1318) [0.5892,1.1148]
1.0997** (0.0415) [1.0213,1.1841]
0.9526 (0.0396) [0.8781,1.0335]
Exercise as recommended (MET)
0.7828*** (0.0449) [0.6995,0.8760]
0.9096 (0.0630) [0.7942,1.0418]
1.1704** (0.0809) [1.0221,1.3403]
1.1384 (0.0828) [0.9871,1.3128]
Greenspace (ha) per capita × Exercise as recommended (MET)
0.9964 (0.0823) [0.8474,1.1716]
1.1895 (0.2174) [0.8314,1.7020]
0.9933 (0.0493) [0.9013,1.0947]
1.0032 (0.0593) [0.8934,1.1265]
Mental health Greenspace (ha) per capita
– Summary statistics Observations Groups
6082 131
**
p < 0.05. p < 0.01 Standard errors in parentheses adjusted for clustering at the LGA level. Other controls included: Age, gender, ethnicity, marital status, parenting, health, educational attainment, employment status, manual work, income, social desirability bias, free time, social interaction, household members engaged in physical activity, personality traits, years at current address, years interviewed, proximity to lake, proximity to river, proximity to coastline and the SEIFA 2011 Index. a Note, variance inflation factors of a base model with no interaction terms shows no sign of worrisome multicollinearity. ***
logical distress score in the third quartile = 0.8515, [0.7461,0.9718], p < 0.05) and fourth (odds of reporting a psychological distress score in the fourth quartile = 0.8622, [0.7536,0.9863], p < 0.05) quartiles point to a lower chance of reporting higher degrees of psychological distress for those who engage in physical activity. Similar to the case of life satisfaction (hypothesis 2) and mental health (hypothesis 4) Table 4 columns 1–4 also fall short of providing evidence of synergistic psychological benefits (hypothesis 6). 4. Discussion This study set out to go beyond earlier research efforts by revealing both: (1) how greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of wellbeing; and also (2) how the potential synergies
between greenspace and physical activity might have heterogeneous impacts across the distribution of wellbeing. To begin with, the results provide some evidence that greenspace and physical activity are independently (net of any hypothesised synergy) positively associated with life satisfaction, mental health and negatively associated with psychological distress (hypotheses 1, 3 and 5). A finding robust to a battery of controls and replicable using easily transferable outcome measures (Thompson Coon et al., 2011). This general finding is found to be stronger for physical activity than for greenspace. Delving more deeply into the results relating to hypotheses 1, 3 and 5 it is possible to glean further insights. The results provide some evidence to support the hypothesis that physical activity, independent of any synergy, is heterogeneously linked across the distribution of wellbeing. Specifically, physical activity (although
C.L. Ambrey / Urban Forestry & Urban Greening 19 (2016) 7–12
11
Table 4 Key psychological distress model resultsa . (1)
(2)
(3)
(4)
First quartile
Second quartile
Third quartile
Fourth quartile
Odds ratio (standard error)
Odds ratio (standard error)
Odds ratio (standard error)
Odds ratio (standard error)
1.2593 (0.5505) [0.5346,2.9666]
0.8110 (0.1669) [0.5418,1.2140]
0.9566 (0.0406) [0.8802,1.0397]
0.9510 (0.0472) [0.8629,1.0481]
Exercise as recommended (MET)
1.3025*** (0.1111) [1.1020,1.5396]
1.0587 (0.0623) [0.9435,1.1880]
0.8515** (0.0574) [0.7461,0.9718]
0.8622** (0.0592) [0.7536,0.9863]
Greenspace (ha) per capita × Exercise as recommended (MET)
0.7885 (0.3463) [0.3334,1.8650]
1.3319 (0.2681) [0.8977,1.9761]
1.0113 (0.0621) [0.8967,1.1405]
0.9422 (0.1233) [0.7291,1.2176]
Psychological distress Greenspace (ha) per capita
– Summary statistics Observations Groups
6076 130
**
p < 0.05. p < 0.01 Standard errors in parentheses adjusted for clustering at the LGA level. Other controls included: Age, gender, ethnicity, marital status, parenting, health, educational attainment, employment status, manual work, income, social desirability bias, free time, social interaction, household members engaged in physical activity, personality traits, years at current address, years interviewed, proximity to lake, proximity to river, proximity to coastline and the SEIFA 2011 Index. a Note, variance inflation factors of a base model with no interaction terms shows no sign of worrisome multicollinearity. ***
not greenspace), distinct from any hypothesised synergy, may actually be more relatively effective at mitigating the likelihood of experiencing a serious dearth of wellbeing, rather than cultivating higher levels of subjectively measured wellbeing. This result for is a convincing finding given that estimated effects for poor wellbeing are likely to be bias downwards (cf. Mitchell, 2013; 133). This finding implies that physical activity may be a more effective remedy to poor wellbeing, the minimisation of which is also an idea that policy makers may be more comfortable with, rather than maximising positive wellbeing or the more nebulous concept of happiness per se (Kahneman and Krueger, 2006). Moreover, the results relating to hypotheses 2, 4 and 6 provide a lack of evidence of the hypothesised synergistic wellbeing benefits of greenspace and physical activity at conventional levels of statistical significance. Furthermore, these same results indicate that this lack of statistical significance does not differ over the distribution of wellbeing. This lack of evidence however, does not constitute evidence that no such relationship exists, rather that this study has merely been unable to find such a relationship. The absence of a link between greenspace and physical activity while not commonly hypothesised is not unheard of, Hoehner et al. (2005) and Hillsdon et al. (2006) fail to find a link between greenspace and hours of recreational physical activity. Further, Mytton et al. (2012), despite finding a positive link between greenspace and physical activity also conclude that it may not necessarily be attributable to greenspace being used for physical activity in the ways we might expect, e.g. for running, cycling, walking, or football/rugby. For these reasons, the triadic relationship between greenspace, physical activity and wellbeing may in fact be more complicated than has tended to be hypothesised. For instance, the realisation of hypothesised synergistic benefits may be contingent on particular individual, social environmental and physical environmental conditions such as fear of crime in the neighbourhood (cf. Sreetheran and van den Bosch, 2014) or different dimensions of a resident’s social environment (cf. McNeill et al., 2006). It may also be the case that physical activity does not allow the restorative benefits of exposure to nature, and the co-benefits of ‘green exercise’ to be realised. Residents may walk, run or cycle briskly through nature. In doing so, residents may find that physical activity in natural environments is as visually stimulating or more visually stimulating than in other environments. For this rea-
son ‘green exercise’ may not be restorative. Also, by engaging in physical activity residents may not find that they have the opportunity to connect with or appreciate nature. This may also explain why Thompson Coon et al. (2011), in a recent systematic review of the literature, observed a tendency for lower (although not always statistically significantly so) reported feelings of calmness and tranquillity for outdoor as opposed to indoor walking (cf. Focht, 2009; Plante et al., 2003; Plante et al., 2006; Plante et al., 2007). Further research, operationalising these concepts is required to draw a more definitive conclusion. It should also be acknowledged that not unlike earlier research efforts, the results presented here have their own limitations. To begin with, the measure of greenspace relates to a resident’s CD. However, resident or household level measures greenspace would likely have lower levels of measurement error. Ceteris paribus, this would increase the likelihood of rejecting the null hypothesis. Also, this study relies on cross-sectional data and despite a gamut of controls falls short of establishing causal links. This remains a perennial challenge for future research. In this regard, experimental and quasi-experimental research designs offer great potential.
5. Conclusions In all, this study provides new insights on the confluence of greenspace, physical activity and wellbeing. Specifically, this study provides some evidence that greenspace and physical activity are independently (net of any hypothesised synergy) positively associated with wellbeing. Furthermore, this link is found to be more pronounced for physical activity than for greenspace. The results of this study also highlight differences and at times, the lack thereof, across the wellbeing distribution and beyond the mean. Notably, the finding that physical activity, separate from any hypothesised synergy may reduce the risk of experiencing a serious dearth of wellbeing, rather than fostering greater wellbeing. This is not found to be the case for greenspace. By drawing attention to these differences and similarities this study builds on the existing literature and uniquely gives a voice to those residents who fall further away from the mean. The findings of this investigation, beyond adding to existing knowledge, may also prove useful to decision makers confronted with the challenge of maintaining or improving the
12
C.L. Ambrey / Urban Forestry & Urban Greening 19 (2016) 7–12
health and wellbeing of residents in the face of continued population growth in urban centres. Acknowledgements The author is grateful for the anonymous reviewers’ constructive comments and feedback. Also, this paper was presented at the Regional Studies Association Inaugural Australasian Conference at which it was awarded the Best Early Career Conference Paper for 2015. The author is grateful for the comments and feedback of conference participants. This paper uses unit record data from the Household, Income and Labour Dynamics in Australia (HILDA) survey. The HILDA Project was initiated and is funded by the Australian Government Department of Social Services (DSS) and is managed by the Melbourne Institute of Applied Economic and Social Research (Melbourne Institute). The findings and views reported in this paper, however, are those of the author and should not be attributed to either DSS or the Melbourne Institute. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ufug.2016.06. 020. References Arnberger, A., 2012. Urban densification and recreational quality of public urban green spaces—a Viennese case study. Sustainability 4, 703. Australian Bureau of Statistics, 1995. 4399.0 ? National Health Survey: SF36 Population Norms. Australian Bureau of Statistics, Canberra. Barton, H., 2009. Land use planning and health and well-being. Land Use Policy 26S, S115–S123. Bell, S., Hamilton, V., Montarzino, A., Rothnie, H., Travlou, P., Alves, S., 2008. Greenspace Scotland Research Report: Greenspace and Quality of Life: A Critical Literature Review. In: Scotland, G. (Ed.), Stirling. Bodin, M., Hartig, T., 2003. Does the outdoor environment matter for psychological restoration gained through running? Psychol. Sport Exercise 4, 141–153. Boes, S., Winkelmann, R., 2010. The effect of income on general life satisfaction and dissatisfaction. Social Indic. Res. 95, 111–128. Coombes, E., Jones, A., Hillsdon, M., 2010. The relationship of physical activity and overweight to objectively measured green space accessibility and use. Social Sci. Med. 70, 816–822. Focht, B.C., 2009. Brief walks in outdoor and laboratory environments. Res. Q. Exercise Sport 80, 611–620. Giles-Corti, B., Donovan, R., 2002. The relative influence of individual, social and physical environment determinants of physical activity. Social Sci. Med. 54. Hillsdon, M., Panter, J., Foster, C., Jones, A., 2006. The relationship between access and quality of urban green space with population physical activity. Public Health 120, 1127–1132. Hoehner, C.M., Brennan Ramirez, L.K., Elliott, M.B., Handy, S.L., Brownson, R.C., 2005. Perceived and objective environmental measures and physical activity among urban adults. Am. J. Prev. Med. 28, 105–116. Hug, S.-M., Hartig, T., Hansmann, R., Seeland, K., Hornung, R., 2009. Restorative qualities of indoor and outdoor exercise settings as predictors of exercise frequency. Health Place 15, 971–980.
Kahneman, D., Krueger, A., 2006. Developments in the measurement of subjective well-being. J. Econ. Perspect. 20, 3–24. Kaplan, S., 1995. The restorative benefits of nature: toward an integrative framework. J. Environ. Psychol. 15, 169–182. Kessler, R., nbsp, C., Andrews, G., Colpe, L., et al., 2002. Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol. Med. 32, 959–976. Larkin, M., 2003. Can cities be designed to fight obesity? Lancet 362, 1046–1047. Lavelle, P., 2006. Are Our Cities Killing Us? ABC Health & Wellbeing: The Pulse: Australian Broadcasting Corporation. McNeill, L., Kreuter, M., Subramanian, S., 2006. Social environment and physical activity: a review of concepts and evidence. Social Sci. Med. 63, 1011–1022. Mitchell, R., 2013. Is physical activity in natural environments better for mental health than physical activity in other environments? Social Sci. Med. 91, 130–134. Mytton, O.T., Townsend, N., Rutter, H., Foster, C., 2012. Green space and physical activity: an observational study using health survey for England data. Health Place 18, 1034–1041. Plante, T.G., Aldridge, A., Su, D., Bogdan, R., Belo, M., Kahn, K., 2003. Does virtual reality enhance the management of stress when paired with exercise?: An exploratory study. Int. J. Stress Manage. 10, 203–216. Plante, T.G., Cage, C., Clements, S., Stover, A., 2006. Psychological benefits of exercise paired with virtual reality: outdoor exercise energizes whereas indoor virtual exercise relaxes. Int. J. Stress Manage. 13, 108–117. Plante, T.G., Gores, C., Brecht, C., Carrow, J., Imbs, A., Willemsen, E., 2007. Does exercise environment enhance the psychological benefits of exercise for women? Int. J. Stress Manage. 14, 88–98. Pretty, J., Peacock, J., Hine, R., Sellens, M., South, N., Griffin, M., 2007. Green exercise in the UK countryside: effects on health and psychological well-being: and implications for policy and planning. J. Environ. Plann. Manage. 50, 211–231. Sachs, J., 2008. Common Wealth: Economics for a Crowded Planet. Penguin Books, London. Schipperijn, J., Stigsdotter, U., Randrup, T., Troelsen, J., 2010. Influences on the use of urban green space − A case study in Odense, Denmark. Urban For. Urban Greening 9, 25–32. Sreetheran, M., van den Bosch, C.C.K., 2014. A socio-ecological exploration of fear of crime in urban green spaces −A systematic review. Urban For. Urban Greening 13, 1–18. The World Bank, 2015. Urban Population (% of Total). The World Bank. Thompson Coon, J., Boddy, K., Stein, K., Whear, R., Barton, J., Depledge, M.H., 2011. Does participating in physical activity in outdoor natural environments have a greater effect on physical and mental wellbeing than physical activity indoors? A systematic review. Environ. Sci. Technol. 45, 1761–1772. US Department of Health, H.S, 1996. Physical Activity and Health. A Report of the Surgeon General. US Department of Health and Human Services, Centres for Disease Control and Prevention National Center for Chronic Disease Prevention and Health Promotion, Atlanta, Georgia. United Nations Economic & Social Affairs, 2014a. Concise Report on the World Population Situation in 2014. ST/ESA/SER.A/354. United Nations, New York. United Nations Economic & Social Affairs, 2014b. World Urbanization Prospects: The 2014 Revision. ST/ESA/SER.A/352. United Nation, New York. United Nations Population Fund, 2015. Urbanization. United Nations, New York. Ware, J., Snow, K., Kosinski, M., 2000. SF-36 Health Survey: Manual and Interpretation Guide. Quality Metric Incorporated, Lincoln. RI. Watson, N., Wooden, M., 2002. The Household, Income and Labour Dynamics in Australia (HILDA) Survey: Wave 1 Survey Methodology. Project Technical Paper Series No. 1/02. Melbourne Institute of Applied Economic and Social Research, The University of Melbourne, Melbourne. Wooden, M., 2009. Use of the Kessler Psychological Distress Scale in the HILDA Survey HILDA Project Discussion Paper Series, No. 2/09. Melbourne Institute, Melbourne.
Contents lists available at ScienceDirect
Urban Forestry & Urban Greening journal homepage: www.elsevier.com/locate/ufug
Short communication
An investigation into the synergistic wellbeing benefits of greenspace and physical activity: Moving beyond the mean Christopher L. Ambrey Cities Research Centre, Gold Coast Campus, Griffith University, Queensland 4222, Australia
a r t i c l e
i n f o
Article history: Received 3 June 2015 Received in revised form 17 March 2016 Accepted 21 June 2016 Available online 4 July 2016 Keywords: Geographic information systems (GIS) Greenspace Household Income and labour dynamics in Australia (HILDA) Life satisfaction Mental health Physical activity Psychological distress Wellbeing
a b s t r a c t The purpose of this study is to shed light on: (1) how greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of wellbeing; and also (2) how the potential synergies between greenspace and physical activity might have heterogeneous impacts across the distribution of wellbeing. Using data from the Household, Income and Labour Dynamics in Australia survey and data from Geographic Information Systems this study finds, for the case of major Australian cities, that greenspace and physical activity are independently positively associated with life satisfaction, mental health and negatively associated with psychological distress. A finding which is stronger for physical activity than it is for greenspace. Across measures of life satisfaction, mental health and psychological distress, the results lend support to the hypothesis that physical activity may be relatively more effective at mitigating the likelihood of experiencing a serious dearth of wellbeing, compared to promoting higher levels of wellbeing. Unexpectedly, the results do not provide support for the hypothesised greenspace-physical activity synergy. A result found to be common across the wellbeing distribution. While further research is required to draw a definitive conclusion, this result may indicate that physical activity is simply not conducive to the realisation of the restorative benefits of exposure to nature, and the other co-benefits of ‘green exercise’. In all, the findings presented in this study add to the existing stock of knowledge from a socialecological perspective and also raise new questions for future research. The results presented in this study may also prove useful to policy makers wrestling with the challenges of maintaining or improving residents’ wellbeing and reducing residents’ ill-being in the face of continuing population growth and declining per capita greenspace. © 2016 Elsevier GmbH. All rights reserved.
1. Introduction Throughout the world urbanisation is occurring at an unprecedented rate (United Nations Economic and Social Affairs, 2014a,b; United Nations Population Fund, 2015). Australia, already one of the most urbanised countries in the world (The World Bank, 2015) is sharing in this global trend. However, urbanisation and land use change are irrevocably altering the environment in which many residents live (Sachs, 2008), impinging on recreational spaces like parks and bushland (Lavelle, 2006) and leading to the congestion of those green spaces that remain (Arnberger, 2012) with implications for the health and wellbeing of a city’s local residents (Barton, 2009). Acknowledging this relationship between greenspace and wellbeing a number of studies (cf. Bodin and Hartig, 2003; Hug et al., 2009; Mitchell, 2013; Pretty et al., 2007; Thompson Coon et al., 2011) have hypothesised that physical activity in natural environ-
E-mail address: c.ambrey@griffith.edu.au http://dx.doi.org/10.1016/j.ufug.2016.06.020 1618-8667/© 2016 Elsevier GmbH. All rights reserved.
ments might produce greater mental health benefits than physical activity elsewhere (Mitchell, 2013). This hypothesised synergistic link builds on the well-established physiological and psychological benefits of physical activity (cf. US Department of Health, 1996). It also combines: (1) the restorative effects of contact with a natural environment (cf. Kaplan, 1995); and (2) the co-benefits of ‘green exercise’ for example, feelings of connectedness to nature and an increased appreciation of nature, which may bolster the longevity of one’s engagement in physical activity (Thompson Coon et al., 2011). Despite the best efforts of earlier investigators many caveats continue to surround the triadic relationship between the greenspace, physical activity and wellbeing (Mitchell, 2013). A recent systematic review of the literature by Thompson Coon et al. (2011) demonstrates a real paucity of high quality evidence on which to base recommendations. Of the studies reviewed, the samples tend to contain of between 8 and 269 respondents. Further, most of these respondents are young university students, in the US, who are already physically active. The use of single shortterm walking or running experiments also provides little indication
8
C.L. Ambrey / Urban Forestry & Urban Greening 19 (2016) 7–12
of the efficacy of interventions over the long-term. Moreover, the lack of easily transferrable outcome measures and the presence of concurrent interventions obscures precisely what conclusions may be drawn from some earlier evidence. Distinct from this body of literature, there is evidence emerging from the empirical economic and psychological literature which suggests that positive and negative wellbeing are more than merely opposite ends of the same phenomenon (Boes and Winkelmann, 2010). It may be the case for example, that individuals experiencing mild depression gain larger benefits from physical activity (Thompson Coon et al., 2011). In this regard, the purpose of this study is to go beyond earlier research efforts by revealing both: (1) how greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of wellbeing; and also (2) how the potential synergies between greenspace and physical activity might have heterogeneous impacts across the distribution of wellbeing. In particular, this study investigates the following hypotheses: H1: Whether or not the greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of life satisfaction. H2: Whether or not any greenspace and physical activity synergy differs across the distribution of life satisfaction. H3: Whether or not the greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of mental health. H4: Whether or not the greenspace and physical activity synergy differs across the distribution of mental health. H5: Whether or not the greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of psychological distress. H6: Whether or not the greenspace and physical activity synergy differs across the distribution of psychological distress. In doing so, for the case of major Australian cities this study contributes to the stock of knowledge regarding the interplay between greenspace, physical activity and wellbeing. The findings presented in this study may prove useful to policy makers wrestling with the challenges of maintaining or improving residents’ wellbeing and reducing residents’ ill-being in the face of continuing population growth and declining per capita greenspace. In what follows, Section 2 reports the data and method employed. Section 3 provides an account of the results and Section 4 discusses the findings and Section 5 concludes. 2. Materials and method In terms of the socioeconomic data on the 6082 residents this is obtained from wave 13 (2013) of the Household, Income and Labour Dynamics in Australia (HILDA) Survey, subset to the major capital cities of Australia.1 The sampling design of the survey involves the selection of households into the sample by a multi-stage process. In wave 1 (2001) of the HILDA Survey, a random sample of 488 Census Collection Districts (CDs) based on the 1996 census boundaries was selected from across Australia, stratified by State, and within the five largest States in terms of population, by metropolitan and non-metropolitan regions, each CD consisting of approximately 200–250 households. The CDs were sampled with probability proportional to their size as measured by the number of dwellings (unoccupied and occupied) recorded in the 1996 Census with some adjustments for population growth since the Census. Within each of these CDs, all dwellings were fully enumerated and a sample of 22–34 dwellings randomly sampled based on the
expected response and occupancy rates within each area (Watson and Wooden, 2002). The life satisfaction dependent variable is obtained from residents’ responses to the question: ‘All things considered, how satisfied are you with your life?’ The life satisfaction variable is an ordinal variable, the resident choosing a number between 0 (totally dissatisfied with life) and 10 (totally satisfied with life). The mental health dependent variable is obtained using data from the Short-Form General Health Survey (SF-36) instrument (collected within the HILDA Survey), an internationally recognised diagnostic tool for assessing functional health status and wellbeing. The Mental Component Summary (MCS) used in this study is derived from 14 items on four scales; vitality, social functioning, role-emotional and mental health, transformed to a 0–100 index using 1995 Australian Bureau of Statistics population norms (Australian Bureau of Statistics, 1995; Ware et al., 2000). A higher mental health score indicates better mental health while a lower mental health score indicates the reverse. The psychological distress dependent variable is measured by the Kessler Psychological Distress Scale (K10) also collected in the HILDA Survey. The ten questions and their selection are described at length in Kessler et al. (2002), as explained by Wooden (2009) the K10 score was derived by scoring responses on each of the items using a simple linear scale running from 5 (all of the time) to 1 (none of the time), and summing across all items. The overall score thus ranges from 10 to 50, where a higher score indicates greater psychological distress and a lower score indicates lower psychological distress. Apart from the different dependent variables employed, the key measure of physical activity is derived from dichotomising the total physical activity Metabolic Equivalent of Task (MET) minutes per week (International Physical Activity Questionnaire). The variable is defined as ‘Exercising as recommended (MET)’ (1) or not (0), where exercising as recommended is defined as MET minutes per week greater than 840 and less than 10,000. That is, the equivalent of 30 or more min × week × 4 MET. To avoid measurement error due to over-reporting, those reporting energy expenditure of 10,000 MET (min/week) or more were excluded (Giles-Corti and Donovan, 2002). Exercising as recommended was defined as the accumulation of the equivalent of 30 min or more of moderate physical activity on most days of the week (US Department of Health, 1996). Data from the HILDA survey are linked to Geographic Information Systems (GIS) data on greenspace through the resident’s Census Collection District (CD). Using GIS CDs are overlayed with greenspace measured from the PSMA Australia Limited Transport and Topography dataset (the 2010 release). Greenspace includes for instance, public parks, community gardens, cemeteries, sports fields, national parks and wilderness areas (cf. Bell et al., 2008). The variable is the number of hectares of greenspace per resident in the CD. A detailed description and summary statistics of the key variables are provided Table 1.2 Hypotheses 1–6 are investigated through the estimation of four conditional logistic regression models, one for each quartile of the dependent variable within a seemingly unrelated regression (SUR) system of equations.3 The dependent variable WBq1...4 r,k represents a resident’s life satisfaction, or mental health or psychological distress each of which have been disaggregated into q1...4 quartiles. The different quartiles for each wellbeing measure allow heterogeneity across the distribution to be revealed.
2
Full summary statistics and results are available as supplementary material. The models estimated for each quartile are based on the same data, they are not independent from each other and hence their residuals are likely correlated. As such the regression results for the separate estimations, obtained using Stata/SE 13.1, are combined using Stata’s ‘suest’ postestimation command. 3
1 Major capital cities in Australia include: Adelaide, Brisbane, Canberra, Darwin, Melbourne, Perth and Sydney.
C.L. Ambrey / Urban Forestry & Urban Greening 19 (2016) 7–12
9
Table 1 Key model variables. Variable name
Definition
Dependent variables Life satisfaction
Mean (std. dev.)
Resident’s life satisfaction (0–10) Quartile Min. Max.
Mental health
7.9 (1.4) Q1 0 7
Resident’s SF-36 Mental Component Summary (MCS) (0–100) Quartile Min. Max.
Psychological distress
Greenspace (ha) per capita
Population density (residents per ha)
Q2 8 8
Q3 9 9
Q4 10 10 48.8 (10.3)
Q1 3.0 44.1
Q2 44.1 51.5
Q3 51.5 56.0
Q4 56.0 71.7 15.7 (6.2)
Resident’s Kessler Psychological Distress Scale (K10) score (10–50) Quartile Min. Max.
Independent variables Exercise as recommended (MET)
%
Q1 10 10
Resident has Metabolic Equivalent of Task (MET) minutes per week greater than 840 and less than 10,000. That is, the equivalent of 30 or more min × week × 4 MET The amount of greenspace in a resident’s CD per resident in the CD. Greenspace in this instance, includes for instance, public parks, community gardens, cemeteries, sports fields, national parks and wilderness areasa Residents in the CD per hectare
Q2 11 13
Q3 14 17
Q4 18 50 63.6%
20.1 (27.6)
61.0 (62.9)
a
For reasons of confidentiality, the CD is lowest level of spatial reference for residents in the HILDA survey. While this means that the precise residential address of respondents is unknown; this spatial reference also provides an upper bound for the degree of measurement error. To be specific, if the CD polygons are assumed to take approximately the shape a circle, the resident is at most, on average (or median), 614 (309) metres from the captured in the greenspace variable. Major factors influencing the use of greenspace are size and proximity (Schipperijn et al., 2010); the use of the CD permits the convenient synthesis of these two factors into a single variable. Despite the precise location of where the resident engages in physical activity and the usability of the greenspace remaining unknown; it is reasonable to expect that residents are more likely to use or experience the greenspace that is nearby (cf. Coombes et al., 2010; Larkin, 2003).
Eq. (1) illustrates the general form of the quartile-specific conditional logistic regression models, for resident r in location k, within the seemingly unrelated regression (SUR) system: WBq1...4 r,k = 0 +
m
j=1
j xjr,k + ␥xr,k zr,k + k + r,k
(1)
Where, xjr,k denotes variables j = 1. . .m. These variables include for example, age, gender, ethnicity and importantly the measure of physical activity and the measure of greenspace. xr,k zr,k represents the two-way interaction term of greenspace × physical activity. k represents the Local Government Area (LGA)-specific fixed effects. Finally, r,k is the error term. It is worth noting that the terms greenspace and physical activity that are not interacted in each quartile-specific conditional logistic regression provide evidence on hypotheses 1, 3 and 5. In contrast, the two-way interaction terms (greenspace × physical activity) in each quartile-specific conditional logistic regression yield evidence on hypotheses 2, 4 and 6. 3. Results The key results, for Eq. (1) are reported in Tables 2–4. With regards to hypothesis 1, Table 2 columns 1–4 suggest that, independently and distinct from any synergistic effect, both greenspace (odds of reporting a life satisfaction score in the first quartile = 0.9452, [0.9023,0.9902], p-value < 0.05) and physical activity (odds of reporting a life satisfaction score in the first quartile = 0.8128, [0.7114,0.9288], p-value < 0.01), reduce the likelihood of reporting a life satisfaction score in the first quartile. In terms of hypothesis 2, while the estimates for greenspace and physical activity in column 4 are stronger and more precise than those of the other quartiles, the results provide no evidence of the
hypothesised greenspace-physical activity synergy at conventional levels of statistical significance. This lack of statistical significance appears to be a homogenous result across the life satisfaction distribution. Table 3 columns 1–4 suggest, pertinent to hypothesis 3, that greenspace is only statistically significant for the third quartile (odds of reporting a mental health score in the third quartile = 1.0997, [1.0213,1.1841], p-value < 0.05), net of any synergistic effects. Furthermore, physical activity is found to have an independently greater estimated effect on reducing the likelihood of reporting a mental health score in the first quartile (odds of reporting a mental health score in the first quartile = 0.7828, [0.6995,0.8760], p-value < 0.01) compared to the other quartiles, distinct from any hypothesised synergistic psychological benefits. Table 3 columns 1–4 yield no evidence of synergistic psychological benefits across the mental health distribution (hypothesis 4). For the case of hypothesis 5 which relates to psychological distress4 Table 4 columns 1–4 reveal that only physical activity, net of hypothesised synergistic psychological benefits is generally associated with lower levels of psychological distress. Greenspace is not found to be statistically significantly linked to psychological distress holding constant any synergistic effect. However, physical activity (odds of reporting a psychological distress score in the first quartile = 1.3025, [1.1020,1.5396], p < 0.01) is linked to a higher estimated likelihood of reporting lower levels of psychological distress. Corroborating this general finding, although marginally less compelling, the results for the third (odds of reporting a psycho-
4 In contrast to the preceding life satisfaction and mental health results (for which higher quartiles relate to higher levels of wellbeing); higher psychological distress quartiles relate to lower levels of wellbeing (see Section 2).
10
C.L. Ambrey / Urban Forestry & Urban Greening 19 (2016) 7–12
Table 2 Key life satisfaction model resultsa . (1)
(2)
(3)
(4)
First quartile
Second quartile
Third quartile
Fourth quartile
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
0.9452** (0.0224) [0.9023,0.9902]
1.2037 (0.2054) [0.8616,1.6816]
0.8673 (0.2378) [0.5067,1.4845]
0.6022 (0.1861) [0.3286,1.1035]
Exercise as recommended (MET)
0.8128*** (0.0553) [0.7114,0.9288]
1.0464 (0.0608) [0.9337,1.1726]
1.0776 (0.0803) [0.9312,1.2471]
1.1550 (0.1247) [0.9347,1.4271]
Greenspace (ha) per capita × Exercise as recommended (MET)
0.9908 (0.0766) [0.8514,1.1530]
0.7333 (0.1477) [0.4942,1.0882]
1.3784 (0.3850) [0.7972,2.3831]
1.6605 (0.5187) [0.9002,3.0630]
Life satisfaction Greenspace (ha) per capita
– Summary statistics Observations Groups
6082 131
**
p < 0.05. p < 0.01 Standard errors in parentheses adjusted for clustering at the LGA level. Other controls included: Age, gender, ethnicity, marital status, parenting, health, educational attainment, employment status, manual work, income, social desirability bias, free time, social interaction, household members engaged in physical activity, personality traits, years at current address, years interviewed, proximity to lake, proximity to river, proximity to coastline and the SEIFA 2011 Index. a Note, variance inflation factors of a base model with no interaction terms shows no sign of worrisome multicollinearity. ***
Table 3 Key mental health model resultsa . (1)
(2)
(3)
(4)
First quartile
Second quartile
Third quartile
Fourth quartile
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
Odds ratio (standard error) [95% CI]
0.9735 (0.0323) [0.9122,1.0389]
0.8105 (0.1318) [0.5892,1.1148]
1.0997** (0.0415) [1.0213,1.1841]
0.9526 (0.0396) [0.8781,1.0335]
Exercise as recommended (MET)
0.7828*** (0.0449) [0.6995,0.8760]
0.9096 (0.0630) [0.7942,1.0418]
1.1704** (0.0809) [1.0221,1.3403]
1.1384 (0.0828) [0.9871,1.3128]
Greenspace (ha) per capita × Exercise as recommended (MET)
0.9964 (0.0823) [0.8474,1.1716]
1.1895 (0.2174) [0.8314,1.7020]
0.9933 (0.0493) [0.9013,1.0947]
1.0032 (0.0593) [0.8934,1.1265]
Mental health Greenspace (ha) per capita
– Summary statistics Observations Groups
6082 131
**
p < 0.05. p < 0.01 Standard errors in parentheses adjusted for clustering at the LGA level. Other controls included: Age, gender, ethnicity, marital status, parenting, health, educational attainment, employment status, manual work, income, social desirability bias, free time, social interaction, household members engaged in physical activity, personality traits, years at current address, years interviewed, proximity to lake, proximity to river, proximity to coastline and the SEIFA 2011 Index. a Note, variance inflation factors of a base model with no interaction terms shows no sign of worrisome multicollinearity. ***
logical distress score in the third quartile = 0.8515, [0.7461,0.9718], p < 0.05) and fourth (odds of reporting a psychological distress score in the fourth quartile = 0.8622, [0.7536,0.9863], p < 0.05) quartiles point to a lower chance of reporting higher degrees of psychological distress for those who engage in physical activity. Similar to the case of life satisfaction (hypothesis 2) and mental health (hypothesis 4) Table 4 columns 1–4 also fall short of providing evidence of synergistic psychological benefits (hypothesis 6). 4. Discussion This study set out to go beyond earlier research efforts by revealing both: (1) how greenspace and physical activity, independent of any synergy, are heterogeneously linked across the distribution of wellbeing; and also (2) how the potential synergies
between greenspace and physical activity might have heterogeneous impacts across the distribution of wellbeing. To begin with, the results provide some evidence that greenspace and physical activity are independently (net of any hypothesised synergy) positively associated with life satisfaction, mental health and negatively associated with psychological distress (hypotheses 1, 3 and 5). A finding robust to a battery of controls and replicable using easily transferable outcome measures (Thompson Coon et al., 2011). This general finding is found to be stronger for physical activity than for greenspace. Delving more deeply into the results relating to hypotheses 1, 3 and 5 it is possible to glean further insights. The results provide some evidence to support the hypothesis that physical activity, independent of any synergy, is heterogeneously linked across the distribution of wellbeing. Specifically, physical activity (although
C.L. Ambrey / Urban Forestry & Urban Greening 19 (2016) 7–12
11
Table 4 Key psychological distress model resultsa . (1)
(2)
(3)
(4)
First quartile
Second quartile
Third quartile
Fourth quartile
Odds ratio (standard error)
Odds ratio (standard error)
Odds ratio (standard error)
Odds ratio (standard error)
1.2593 (0.5505) [0.5346,2.9666]
0.8110 (0.1669) [0.5418,1.2140]
0.9566 (0.0406) [0.8802,1.0397]
0.9510 (0.0472) [0.8629,1.0481]
Exercise as recommended (MET)
1.3025*** (0.1111) [1.1020,1.5396]
1.0587 (0.0623) [0.9435,1.1880]
0.8515** (0.0574) [0.7461,0.9718]
0.8622** (0.0592) [0.7536,0.9863]
Greenspace (ha) per capita × Exercise as recommended (MET)
0.7885 (0.3463) [0.3334,1.8650]
1.3319 (0.2681) [0.8977,1.9761]
1.0113 (0.0621) [0.8967,1.1405]
0.9422 (0.1233) [0.7291,1.2176]
Psychological distress Greenspace (ha) per capita
– Summary statistics Observations Groups
6076 130
**
p < 0.05. p < 0.01 Standard errors in parentheses adjusted for clustering at the LGA level. Other controls included: Age, gender, ethnicity, marital status, parenting, health, educational attainment, employment status, manual work, income, social desirability bias, free time, social interaction, household members engaged in physical activity, personality traits, years at current address, years interviewed, proximity to lake, proximity to river, proximity to coastline and the SEIFA 2011 Index. a Note, variance inflation factors of a base model with no interaction terms shows no sign of worrisome multicollinearity. ***
not greenspace), distinct from any hypothesised synergy, may actually be more relatively effective at mitigating the likelihood of experiencing a serious dearth of wellbeing, rather than cultivating higher levels of subjectively measured wellbeing. This result for is a convincing finding given that estimated effects for poor wellbeing are likely to be bias downwards (cf. Mitchell, 2013; 133). This finding implies that physical activity may be a more effective remedy to poor wellbeing, the minimisation of which is also an idea that policy makers may be more comfortable with, rather than maximising positive wellbeing or the more nebulous concept of happiness per se (Kahneman and Krueger, 2006). Moreover, the results relating to hypotheses 2, 4 and 6 provide a lack of evidence of the hypothesised synergistic wellbeing benefits of greenspace and physical activity at conventional levels of statistical significance. Furthermore, these same results indicate that this lack of statistical significance does not differ over the distribution of wellbeing. This lack of evidence however, does not constitute evidence that no such relationship exists, rather that this study has merely been unable to find such a relationship. The absence of a link between greenspace and physical activity while not commonly hypothesised is not unheard of, Hoehner et al. (2005) and Hillsdon et al. (2006) fail to find a link between greenspace and hours of recreational physical activity. Further, Mytton et al. (2012), despite finding a positive link between greenspace and physical activity also conclude that it may not necessarily be attributable to greenspace being used for physical activity in the ways we might expect, e.g. for running, cycling, walking, or football/rugby. For these reasons, the triadic relationship between greenspace, physical activity and wellbeing may in fact be more complicated than has tended to be hypothesised. For instance, the realisation of hypothesised synergistic benefits may be contingent on particular individual, social environmental and physical environmental conditions such as fear of crime in the neighbourhood (cf. Sreetheran and van den Bosch, 2014) or different dimensions of a resident’s social environment (cf. McNeill et al., 2006). It may also be the case that physical activity does not allow the restorative benefits of exposure to nature, and the co-benefits of ‘green exercise’ to be realised. Residents may walk, run or cycle briskly through nature. In doing so, residents may find that physical activity in natural environments is as visually stimulating or more visually stimulating than in other environments. For this rea-
son ‘green exercise’ may not be restorative. Also, by engaging in physical activity residents may not find that they have the opportunity to connect with or appreciate nature. This may also explain why Thompson Coon et al. (2011), in a recent systematic review of the literature, observed a tendency for lower (although not always statistically significantly so) reported feelings of calmness and tranquillity for outdoor as opposed to indoor walking (cf. Focht, 2009; Plante et al., 2003; Plante et al., 2006; Plante et al., 2007). Further research, operationalising these concepts is required to draw a more definitive conclusion. It should also be acknowledged that not unlike earlier research efforts, the results presented here have their own limitations. To begin with, the measure of greenspace relates to a resident’s CD. However, resident or household level measures greenspace would likely have lower levels of measurement error. Ceteris paribus, this would increase the likelihood of rejecting the null hypothesis. Also, this study relies on cross-sectional data and despite a gamut of controls falls short of establishing causal links. This remains a perennial challenge for future research. In this regard, experimental and quasi-experimental research designs offer great potential.
5. Conclusions In all, this study provides new insights on the confluence of greenspace, physical activity and wellbeing. Specifically, this study provides some evidence that greenspace and physical activity are independently (net of any hypothesised synergy) positively associated with wellbeing. Furthermore, this link is found to be more pronounced for physical activity than for greenspace. The results of this study also highlight differences and at times, the lack thereof, across the wellbeing distribution and beyond the mean. Notably, the finding that physical activity, separate from any hypothesised synergy may reduce the risk of experiencing a serious dearth of wellbeing, rather than fostering greater wellbeing. This is not found to be the case for greenspace. By drawing attention to these differences and similarities this study builds on the existing literature and uniquely gives a voice to those residents who fall further away from the mean. The findings of this investigation, beyond adding to existing knowledge, may also prove useful to decision makers confronted with the challenge of maintaining or improving the
12
C.L. Ambrey / Urban Forestry & Urban Greening 19 (2016) 7–12
health and wellbeing of residents in the face of continued population growth in urban centres. Acknowledgements The author is grateful for the anonymous reviewers’ constructive comments and feedback. Also, this paper was presented at the Regional Studies Association Inaugural Australasian Conference at which it was awarded the Best Early Career Conference Paper for 2015. The author is grateful for the comments and feedback of conference participants. This paper uses unit record data from the Household, Income and Labour Dynamics in Australia (HILDA) survey. The HILDA Project was initiated and is funded by the Australian Government Department of Social Services (DSS) and is managed by the Melbourne Institute of Applied Economic and Social Research (Melbourne Institute). The findings and views reported in this paper, however, are those of the author and should not be attributed to either DSS or the Melbourne Institute. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ufug.2016.06. 020. References Arnberger, A., 2012. Urban densification and recreational quality of public urban green spaces—a Viennese case study. Sustainability 4, 703. Australian Bureau of Statistics, 1995. 4399.0 ? National Health Survey: SF36 Population Norms. Australian Bureau of Statistics, Canberra. Barton, H., 2009. Land use planning and health and well-being. Land Use Policy 26S, S115–S123. Bell, S., Hamilton, V., Montarzino, A., Rothnie, H., Travlou, P., Alves, S., 2008. Greenspace Scotland Research Report: Greenspace and Quality of Life: A Critical Literature Review. In: Scotland, G. (Ed.), Stirling. Bodin, M., Hartig, T., 2003. Does the outdoor environment matter for psychological restoration gained through running? Psychol. Sport Exercise 4, 141–153. Boes, S., Winkelmann, R., 2010. The effect of income on general life satisfaction and dissatisfaction. Social Indic. Res. 95, 111–128. Coombes, E., Jones, A., Hillsdon, M., 2010. The relationship of physical activity and overweight to objectively measured green space accessibility and use. Social Sci. Med. 70, 816–822. Focht, B.C., 2009. Brief walks in outdoor and laboratory environments. Res. Q. Exercise Sport 80, 611–620. Giles-Corti, B., Donovan, R., 2002. The relative influence of individual, social and physical environment determinants of physical activity. Social Sci. Med. 54. Hillsdon, M., Panter, J., Foster, C., Jones, A., 2006. The relationship between access and quality of urban green space with population physical activity. Public Health 120, 1127–1132. Hoehner, C.M., Brennan Ramirez, L.K., Elliott, M.B., Handy, S.L., Brownson, R.C., 2005. Perceived and objective environmental measures and physical activity among urban adults. Am. J. Prev. Med. 28, 105–116. Hug, S.-M., Hartig, T., Hansmann, R., Seeland, K., Hornung, R., 2009. Restorative qualities of indoor and outdoor exercise settings as predictors of exercise frequency. Health Place 15, 971–980.
Kahneman, D., Krueger, A., 2006. Developments in the measurement of subjective well-being. J. Econ. Perspect. 20, 3–24. Kaplan, S., 1995. The restorative benefits of nature: toward an integrative framework. J. Environ. Psychol. 15, 169–182. Kessler, R., nbsp, C., Andrews, G., Colpe, L., et al., 2002. Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol. Med. 32, 959–976. Larkin, M., 2003. Can cities be designed to fight obesity? Lancet 362, 1046–1047. Lavelle, P., 2006. Are Our Cities Killing Us? ABC Health & Wellbeing: The Pulse: Australian Broadcasting Corporation. McNeill, L., Kreuter, M., Subramanian, S., 2006. Social environment and physical activity: a review of concepts and evidence. Social Sci. Med. 63, 1011–1022. Mitchell, R., 2013. Is physical activity in natural environments better for mental health than physical activity in other environments? Social Sci. Med. 91, 130–134. Mytton, O.T., Townsend, N., Rutter, H., Foster, C., 2012. Green space and physical activity: an observational study using health survey for England data. Health Place 18, 1034–1041. Plante, T.G., Aldridge, A., Su, D., Bogdan, R., Belo, M., Kahn, K., 2003. Does virtual reality enhance the management of stress when paired with exercise?: An exploratory study. Int. J. Stress Manage. 10, 203–216. Plante, T.G., Cage, C., Clements, S., Stover, A., 2006. Psychological benefits of exercise paired with virtual reality: outdoor exercise energizes whereas indoor virtual exercise relaxes. Int. J. Stress Manage. 13, 108–117. Plante, T.G., Gores, C., Brecht, C., Carrow, J., Imbs, A., Willemsen, E., 2007. Does exercise environment enhance the psychological benefits of exercise for women? Int. J. Stress Manage. 14, 88–98. Pretty, J., Peacock, J., Hine, R., Sellens, M., South, N., Griffin, M., 2007. Green exercise in the UK countryside: effects on health and psychological well-being: and implications for policy and planning. J. Environ. Plann. Manage. 50, 211–231. Sachs, J., 2008. Common Wealth: Economics for a Crowded Planet. Penguin Books, London. Schipperijn, J., Stigsdotter, U., Randrup, T., Troelsen, J., 2010. Influences on the use of urban green space − A case study in Odense, Denmark. Urban For. Urban Greening 9, 25–32. Sreetheran, M., van den Bosch, C.C.K., 2014. A socio-ecological exploration of fear of crime in urban green spaces −A systematic review. Urban For. Urban Greening 13, 1–18. The World Bank, 2015. Urban Population (% of Total). The World Bank. Thompson Coon, J., Boddy, K., Stein, K., Whear, R., Barton, J., Depledge, M.H., 2011. Does participating in physical activity in outdoor natural environments have a greater effect on physical and mental wellbeing than physical activity indoors? A systematic review. Environ. Sci. Technol. 45, 1761–1772. US Department of Health, H.S, 1996. Physical Activity and Health. A Report of the Surgeon General. US Department of Health and Human Services, Centres for Disease Control and Prevention National Center for Chronic Disease Prevention and Health Promotion, Atlanta, Georgia. United Nations Economic & Social Affairs, 2014a. Concise Report on the World Population Situation in 2014. ST/ESA/SER.A/354. United Nations, New York. United Nations Economic & Social Affairs, 2014b. World Urbanization Prospects: The 2014 Revision. ST/ESA/SER.A/352. United Nation, New York. United Nations Population Fund, 2015. Urbanization. United Nations, New York. Ware, J., Snow, K., Kosinski, M., 2000. SF-36 Health Survey: Manual and Interpretation Guide. Quality Metric Incorporated, Lincoln. RI. Watson, N., Wooden, M., 2002. The Household, Income and Labour Dynamics in Australia (HILDA) Survey: Wave 1 Survey Methodology. Project Technical Paper Series No. 1/02. Melbourne Institute of Applied Economic and Social Research, The University of Melbourne, Melbourne. Wooden, M., 2009. Use of the Kessler Psychological Distress Scale in the HILDA Survey HILDA Project Discussion Paper Series, No. 2/09. Melbourne Institute, Melbourne.