- Open Access
Lack of agreement between measured and self-reported distance from public green parks in Glasgow, Scotland
© Macintyre et al; licensee BioMed Central Ltd. 2008
- Received: 31 October 2007
- Accepted: 04 May 2008
- Published: 04 May 2008
Reviews have reported mixed findings for associations between physical activity and proximity to a range of environmental resources. Initially most studies used self reported proximity, but more are now using GIS techniques to measure proximity objectively. We know little about the extent of agreement between self reported and directly measured proximity of the same resource.
We used previously collected data in a community survey in Glasgow in which 658 respondents aged around 40 and 60 were asked whether they lived within half a mile of a public park. We compared their answers with GIS measures of whether there was a park within a half mile service area of their home (and whether their home was within a half mile crow fly buffer of a park).
Agreement was poor; percentage agreement between measured network distance and reported residence within 0.5 miles of a park was 62.0%, and the kappa value was 0.095. Agreement was no higher than poor in any socio-demographic subgroup, or when using crow fly buffers instead of service areas.
One should be cautious about assuming that respondents' self reports of proximity to a resource are a valid proxy for actual distance, or vice versa. Further research is needed to establish whether actual or self-reported proximity predict physical activity or other behaviours, and if so which is the strongest predictor. Further, qualitative study, also needs to examine the basis of people's judgements about the location of resources, and the possibility that these are shaped by their social and personal significance.
- Geographic Information System
- South West
- North West
- Public Park
- Occupational Social Class
In recent years there has been considerable interest in the potential health benefits of public parks, other green spaces, or greenery in general. Much of this work has focused on possible associations between proximity to parks and physical activity [1–6], but some has considered broader, public health and mental health aspects of proximity to parks [7–9]. Initially most of the investigations into associations between health or physical activity and proximity to parks was based on respondents' self-reported accounts of their access to parks [10, 11], but more recently Geographic Information Systems (GIS) (computerised systems which manage spatially referenced data ) or trained observers have been used to measure access to parks or the greenness of the immediate environment objectively [4–6, 13]. Reviews have found variability between studies in observed associations between physical activity and proximity to parks and other potential activity settings [11, 14–16].
One reason for this variability may be the use of different measures of proximity to settings in different studies, and in particular to the mixture of self-reported and GIS or observer-assessed distance. Sallis et al, for example, found that objective measures of access to exercise facilities were associated with physical activity while respondents' reports of access were not, and there was no significant correlation between the directly measured and the self reported . Tilt et al found that self-reported measured greenness in the neighbourhood was significantly associated with walking trips but objective greenness was not; objective accessibility of a number of resources was associated with walking trips but the authors do not report whether or not self-reported accessibility was. Objective accessibility of 17 destinations (banks, coffee shops, parks etc) within 0.4 miles (643.7 metres) was correlated with self-reported accessibility within 0.5 miles (804.7 metres) (R = 0.314, p < 0.0001). Controlling for socio-demographic factors, the regression of self-reported accessibility on objective accessibility produced R2 of 0.110 (i.e. 11% explained variance). Eighty-one percent of respondents to a community survey said they lived within 0.5 miles of a park, whereas 62% were measured by GIS as living within 0.4 miles of a park; kappa value for agreement between objective and self-reported accessibility to a park was 0.154 .
Kirtland et al compared directly measured and respondent assessed proximity to a range of environmental features in South Carolina, using two spatial scales (neighbourhood, within a 0.5 mile radius of the respondent's home, and community, within a ten mile radius). Levels of agreement between objective and self-reported access were fair to low for neighbourhood items (kappa = -0.02 to 0.37, with higher values, [0.19 to 0.37], for access to sidewalks, public recreation facilities, and streetlights), and low to slight for community items (kappa -0.07 to 0.25, with 0.25 being for the presence or absence of shopping malls). Kappa was 0.01 for whether there were any parks, playgrounds or sports fields in the community. Levels of agreement varied by physical activity levels; for example agreement on the presence of recreational facilities was higher among active respondents than insufficiently active or inactive respondents. The authors suggested that the relatively low level of agreement might relate to people's inability to judge distances, and their personal behaviours (for example transportation routes), beliefs and values .
In a study of two neighbourhoods in Adelaide, South Australia McCormack et al presented participants with a list of nine destinations including a park and asked them to report how long it would take to get from their home to the facility if they walked. Self-reported distance (based on an average walking speed of 4.5 km (2.8 miles)/hour) was calculated based on these self-reported walking times and then compared with the actual distance (using road network analysis). There was relatively poor agreement between the objective and self-reported measures of distance .
Sallis et al's study used self-reported convenience of 15 exercise facilities, and noted little correlation between that and measured density of all (aerobic) exercise facilities within 1 km of each resident's home. This is perhaps not surprising since self-reported convenience and local density may not be measuring the same thing . Tilt et al's study used different distance measures for objective (0.4 mile) and self-reported (0.5 mile) access . The Kirtland study used rather general measures, some of them qualitative rather than binary (e.g. standard of maintenance of sidewalk, safety, heaviness of traffic, quality of street lighting, whether unattended dogs were a problem), and the question on parks related to a ten mile radius rather than 0.5 mile radius . McCormack et al noted that a limitation of their study was that assumptions were made about the average walking speed of their study participants and this may have affected their results . Thus these studies were not examining self-reported and objectively measured versions of exactly the same variables.
In this paper we attempt to fill this gap by re-analysing an existing data set to address two questions: what is the extent of agreement between self-reported and objective measures of the same variable (whether or not the respondent lived within 0.5 miles of a public green open space)?; and does agreement between the self reported and directly assessed measures vary by socio-demographic and behavioural variables (given that previous literature suggests variation in perceptions of the local environment by such characteristics [21–25])?
We used data from the locality component of the 'The West of Scotland Twenty-07 Study: Health in the Community', the aim of which is to explore the social processes which produce or maintain differences in health by key social positions, in particular, by gender, age, social class, ethnicity, family composition and area of residence. The study is following up three cohorts (born in 1932, 1952 and 1972) using wide ranging home-based interviews and postal questionnaires. Respondents were aged 15, 35 and 55 years at first contact in 1987/88 .
At the 1992 sweep 690 respondents lived within the locality boundaries. Thirty two of them were excluded from this analysis; four because they did not answer the question on whether they lived within half a mile of a park, and 28 because their addresses in 1992 (16 from the NW and 12 from the SW) were not recognized when matched to the Ordnance Survey address point data because of inaccurate or out-of-date address data. This analysis therefore included 658 respondents; 352 within the NW locality and 306 in the SW.
Ordnance Survey maps (including addresses and roads, tracks, paths, buildings etc), were obtained, and ESRI ArcMap 9.1 was used to geocode the location of every respondent by their unit postcode (the smallest level of postal geography in the UK) and their point address (i.e. street number and name). All the twenty public green parks owned by Glasgow City Council within the boundaries of the localities, and half a mile outwith the boundaries, were also mapped. (Note that we did not include publicly maintained hard based spaces such as skate board or BMX parks since in the UK these would not normally be what people think of by the term 'parks'). According to the City Council, these comprised four City parks, four District parks, eleven Local parks and one wooded park; see Figures 1 and 2. Most of these were of nineteenth century origin and the one most recently taken over by the City was Pollok Park, which was gifted to the City for public use in 1966. Thus all the parks predated the question asked of respondents in 1992, and the geocoding of their homes, by at least 26 years. The ArcMap software was used to create service areas, with a network (i.e. road or path) distance of 1/2 mile, around each respondent's home. The presence or absence of a park within the 1/2 mile service area around each respondent home was determined (i.e. any part of the boundary of the park would need to intersect with the service area of the respondent to be regarded as present). We also measured whether each respondent lived within a half mile crow fly buffer (i.e. straight line Euclidean distance) from the edge of each park.
SPSS 14.0 for Windows was used to investigate the agreement between objective and self-reported proximity to public parks. Kappa values, and percentage agreement and disagreement, were calculated. We used the conventional cut offs for kappa of < 0.20 = poor, 0.21–0.40 = fair, 0.41–0.60 = moderate, 0.61–0.80 = good, and 0.81–1.00 = very good . Analysis was undertaken for all respondents together, and also by locality; sex; age; head of household occupational social class (using the UK Registrar General's Classification of Occupations  divided here into non-manual [professional and white collar], III manual [skilled blue collar] and IV/V manual [semi and unskilled blue collar]); deprivation of local small area (divided into quintiles using the Scottish Index of Multiple Deprivation ); whether they reported having taken a walk of more than 2 miles in the previous year (and whether that was in or outside their local area); and having a car or dog in the household. We have no data on the reliability or validity of the responses to the question on walking two miles in the last year. We were unable to control for or stratify by length of residence in the area, since although we had information on how long respondents had lived at their address when interviewed in 1992 (40 year old cohort mean 10.75, years, 60 year old cohort mean 18.5 years), we did not have information on their previous address so cannot distinguish between those who moved within and between areas.
Self reported living within half a mile of a park, by measured distance to a park.
Lives within half a mile of a park
Lives more than half a mile from a park
Believe they live within half a mile of a park
Believe they do not live within half a mile of a park
Agreement between self reported and directly measured distance of half a mile to a park, by sex, cohort, locality, class, area deprivation, car ownership, dog ownership, walking 2+ miles in previous year, and walking 2+ miles in previous year in area.
40 year olds
60 year olds
Walked 2+ miles (previous year)
Walked 2+ miles in area
Walker in area
Walker outside area
Agreement was slightly higher for households who did not own a car or have a dog (both groups who might be expected to walk about their local areas more) than for those who did not. There was however little difference between respondents who reported having taken a walk for more than two miles in the previous year and those who did not, nor was there much difference whether their walks were in or outside their local area (see Table 2). In none of the socio-demographic, asset ownership or behavioural subgroups was agreement any better than 'poor'.
There was relatively poor agreement between our GIS measure and respondents' perceptions of living near a public park. Nearly thirteen per cent of respondents whose homes we measured as being within half a mile of a park said they did not live near to a park; conversely, 79% of those we measured as living further than half a mile from a park said they did live within half a mile of a park.
Thus one should be cautious about assuming equivalence between such measures, i.e. that respondents' perceptions are a valid proxy for actual distance or vice versa. While lack of correlation between GIS or observer measured and respondent self-reported access to facilities and resources has been noted before, previous studies [17–20] have not compared observed and self-reported access to the same, single, resource using the same distance measures, and it might therefore have been thought that there would be more congruence when directly comparing like with like.
There were no noteworthy or consistent differences in agreement by car or dog ownership, or by walking, which we had hypothesised might be related to agreement (car owners being less, and dog owners and walkers more, accurate in estimates of access to local parks), nor by socio-demographic characteristics. Unfortunately we were not able to examine whether there were any differences by length of residence.
It might be objected that when thinking about proximity to a park respondents are not thinking of precise, road based network distances, but of more general proximity; however, our crow fly measures showed similar lack of agreement (63.4%, kappa 0.096). We only asked about a half mile distance so were unable to compare the agreement on this measure with that which might be observed with a different distance e.g. a mile. It is striking that several respondents, particularly in the SW (Figure 2) lived very close to a park but did not report doing so. Because our data were collected from a structured survey instrument it is not possible for us to analyse what respondents thought of as "a park", and the extent to which this corresponds with what local government and official maps define as parks in the UK.
It may be that what is important for people's perceptions is not actual or potential proximity but some sort of symbolic proximity; for example because a park is not somewhere they ever go, or which they feel is suitable for people like them, it may not be in a person's perceptual field. A recent qualitative study of the opening of a new supermarket in a deprived area of Glasgow found that some local residents did not see the new supermarket as being 'for them' even though it was very close and sold affordable food at low prices ; it is possible that similar processes may be at work in relation to parks. For example, two of the parks which had residents living in close proximity who said they were not within half a mile of a park, were relatively 'wild' rather than 'manicured' (Crookston Woods and Darnley Mill Country Park, see Figure 2); it is possible that what physical activity advocates or town planners think of as attractive, public, accessible, natural environments are not seen as such by all local residents, who may instead think of them as alien or frightening , or simply not anywhere they would go.
Thus researchers interested in associations between proximity to resources and use of them, or other behaviours (for example, in this case physical activity), should not assume that self-reported proximity is the same as actual proximity. Further work needs to be done to establish whether or not actual or self-reported proximity predict physical activity or other behaviours, and if so which is the strongest predictor. Some authors have noted that the match between actual and self-reported distance may be influenced by the perceived quality (e.g. attractiveness) of the end point of the destination [37, 38]. Further, qualitative study, therefore needs to examine the basis of people's judgements about the location of resources and the possibility that these are shaped by their social and personal significance.
All three authors are employed by the UK Medical Research Council, who also funded the West of Scotland Twenty-07 study on which this paper is based. This work is part of the Social and Spatial patterning of health programme (wbs U.1300.00.006). We are grateful to Geoff Der (MRC Social and Public Health Sciences Unit) for statistical advice and Stuart Clay (then also MRC Social and Public Health Sciences Unit) for GIS advice.
- Giles-Corti B, Broomhall MH, Kniuman M, Collins C, Douglas K, Ng K, Lange A, Donovan RJ: How Important is Distance To, Attractiveness, and Size of Public Open Space?. American Journal of Preventive Medicine. 2005, 28 (2S2): 169-176. 10.1016/j.amepre.2004.10.018.View ArticleGoogle Scholar
- Pierce JR, Denison AV, Arif AA, Rohrer JE: Living near a trail is associated with increased odds of walking among patients using community clinics. Journal of Community Health. 2006, 31 (4): 289-302. 10.1007/s10900-006-9014-8.View ArticleGoogle Scholar
- Hoehner CM, Brennan Ramirez LK, Elliott MB, Handy SL, Brownson RC: Perceived and Objective Environmental Measures and Physical Activity among Urban Adults. American Journal of Preventive Medicine. 2005, 28 (2S2): 105-116. 10.1016/j.amepre.2004.10.023.View ArticleGoogle Scholar
- Li F, Fisher J, Brownson RC, Bosworth M: Multilevel modelling of built environment characteristics related to neighbourhood walking activity in older adults. Journal of Epidemiology & Community Health. 2004, 59: 558-564. 10.1136/jech.2004.028399.View ArticleGoogle Scholar
- Ellaway A, Macintyre S, Bonnefoy X: Graffiti, greenery and obesity in adults. British Medical Journal. 2005, 331: 611-612. 10.1136/bmj.38575.664549.F7.View ArticleGoogle Scholar
- Duncan M, Mummery K: Psychosocial and environmental factors associated with physical activity among city dwellers in regional Queensland. Preventive Medicine. 2005, 40: 363-372. 10.1016/j.ypmed.2004.06.017.View ArticleGoogle Scholar
- Wen M, Hawkley L, Cacioppo J: Objective and perceived neighbourhood environment, individual SES and psychosocial factors and self rated health: an analysis of older adults in Cook County, Illinois. Social Science & Medicine. 2006, 63: 2575-2590. 10.1016/j.socscimed.2006.06.025.View ArticleGoogle Scholar
- Bedimo-Rung AL, Mowen AJ, Cohen DA: The Significance of Parks to Physical Activity and Public Health. A Conceptual Model. American Journal of Preventive Medicine. 2005, 28 (2S2): 159-168. 10.1016/j.amepre.2004.10.024.View ArticleGoogle Scholar
- Thrift J: Are better streets and parks a route to better mental health?. Journal of Public Mental Health. 2005, 4 (4): 6-9.View ArticleGoogle Scholar
- Ball K, Bauman A, Leslie E, Owen N: Perceived Environmental Aesthetics and Convenience and Company Are Associated with Walking for Exercise among Australian Adults. Preventive Medicine. 2001, 33: 434-440. 10.1006/pmed.2001.0912.View ArticleGoogle Scholar
- Humpel N, Owen N, Leslie E: Environmental factors associated with adults' participation in physical activity. A Review. American Journal of Preventive Medicine. 2002, 22 (3): 188-199. 10.1016/S0749-3797(01)00426-3.View ArticleGoogle Scholar
- Cromley EK, McLafferty SL: GIS and Public Health. 2002, New York , Guilford PressGoogle Scholar
- Lindsey G, Maraj M, Kuan S: Access, Equity, and Urban Greenways: An Exploratory Investigation. The Professional Geographer. 2001, 53 (3): 332-346. 10.1111/0033-0124.00288.View ArticleGoogle Scholar
- Bauman A, Bull F: Environmental correlates of physical activity and walking in adults and children: a review of reviews. 2007, National Insitute of Health and Clinical Excellence.Google Scholar
- Ferreira I, van der Horst K, Wendel-Vos W, Kremers S, van Lenthe F, Brug J: Environmental correlates of physical activity in youth, a review and update. Obesity Reviews. 2006, 8: 129-154. 10.1111/j.1467-789X.2006.00264.x.View ArticleGoogle Scholar
- Jones A, Bentham G, Foster C, Hillsdon M, Panter J: Obesogenic Environments Evidence review. 2007, Foresight Tackling Obesities; Future ChoicesGoogle Scholar
- Sallis J, Melbourne H, Hofstetter CR, Elder JP, Hackley M, Caspersen CJ, Powell KE: Distance between homes and exercise facilities related to frequency of exercise among San Diego residents. Public Health Reports. 1990, 105 (2): 179-185.Google Scholar
- Tilt JH, Unfried TM, Roca B: Using objective and subjective measures of neighborhood greenness and accessible destinations for understanding walking trips and BMI in Seattle, Washington. American Journal of Health Promotion. 2007, 21 (4): 371-379.View ArticleGoogle Scholar
- Kirtland KA, Porter DE, Addy CA, Neet MJ, Williams JE, Sharpe PA, Neff LJ, Kimsey CD, Ainsworth BE: Environmental Measures of Physical Activity Supports. Perception Versus Reality. American Journal of Preventive Medicine. 2003, 24 (4): 323-331. 10.1016/S0749-3797(03)00021-7.View ArticleGoogle Scholar
- McCormack G, Cerin E, Leslie E, DuToit L, Owen N: Objective versus perceived walking distance to destinations: correspondence and predictive validity. Environment and Behavior. 2008, 40 (3): 401-425. 10.1177/0013916507300560.View ArticleGoogle Scholar
- Ellaway A, Macintyre S, Kearns A: Perceptions of place and health in socially contrasting neighbourhoods. Urban Studies. 2001, 38 (12): 2299-2316. 10.1080/00420980120087171.View ArticleGoogle Scholar
- Humpel N, Marshall AL, Leslie E: Changes in neighbourhood walking are related to changes in perceptions of environmental attributes. Ann Behav Med. 2004, 27: 60-67. 10.1207/s15324796abm2701_8.View ArticleGoogle Scholar
- Pamalon R, Hamel D, DeKoninck M, Disant M: Perception of place and health: Differences between neighbourhoods in the Quebec City region. Social Science & Medicine. 2005, 61 (1): 95-111.Google Scholar
- Sooman A, Macintyre S: Health and perceptions of the local environment in socially contrasting neighbourhoods in Glasgow. Health and Place. 1995, 1: 15-26. 10.1016/1353-8292(95)00003-5.View ArticleGoogle Scholar
- Wood L, Giles-Corti B, Bulsara M: The pet connection: Pets as a conduit for social capital?. Social Science & Medicine. 2005, 61 (6): 1159-1173. 10.1016/j.socscimed.2005.01.017.View ArticleGoogle Scholar
- Macintyre S, Annandale E, Ecob R, Ford G, Hunt K, Jamieson B, Maciver S, West P, Wyke S: The West of Scotland Twenty-07 study; health in the community. Readings for a New Public Health. Edited by: Martin C, McQueen D. 1989, Edinburgh , Edinburgh University Press, 56-74.Google Scholar
- Macintyre S, Maciver S, Sooman A: Area, class and health; should we be focusing on places or people?. JSocPol. 1993, 22: 213-234.Google Scholar
- Maciver S, Macintyre S: West of Scotland Twenty-07 Study: selection of the study localities and the region. MRC Medical Sociology Unit Working Paper no 4. 1987, Glasgow, UK.Google Scholar
- De Bourdeaughuij I, Sallis J, Saelins B: Environmental Correlates of Physical Activity in a Sample of Belgian Adults. American Journal of Health Promotion. 2003, 18 (1): 83-92.View ArticleGoogle Scholar
- Saelens BE, Sallis JF, Black JB, Chen D: Neighborhood-Based Differences in Physical Activity: An Environment Scale Evaluation. Research and Practice. 2003, 93 (9): 1552-1558.Google Scholar
- Leslie E, Saelens B, Frank L, Owen N, Bauman A, Coffee N, Hugo G: Residents' perceptions of walkability attributes in objectively different neighbourhoods: a pilot study. Health & Place. 2005, 11: 227-236. 10.1016/j.healthplace.2004.05.005.View ArticleGoogle Scholar
- Altman D: Practical Statistics for Medical Research. 1993, London , Chapman and HallGoogle Scholar
- Office of Population Censuses and Surveys: Standard Occupational Classification. 1991, London , HMSOGoogle Scholar
- Scottish Executive: Scottish Executive: Scottish Index of Multiple Deprivation 2006. [http://www.scotland.gov.uk/Topics/Statistics/SIMD/simdbackgrounddata2simd06]
- Cummins S, Findlay A, Higgins C, Petticrew M, Sparks L, Thomson H: Reducing inequalities in health: findings from a study of the impact of a food retail development. Environment and Planning A. 2008, 40: 402-422. 10.1068/a38371.View ArticleGoogle Scholar
- Milligan C, Bingley A: Restorative places or scary spaces? The impact of woodland on the mental health of young adults. Health & Place. 2007, 13: 799-811. 10.1016/j.healthplace.2007.01.005.View ArticleGoogle Scholar
- Briggs R: Urban distance cognition. Image and Environment. Edited by: Downs RM, Stea D. 1976, Chicago , Aldine, 361-388.Google Scholar
- Walmsley DJ, Jenkins JM: Cognitive distance: A neglected issue in travel behavior. Journal of Travel Research. 1992, 24-29. 10.1177/004728759203100106.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.