Face-to-face interviews lasting 20-30 minutes were conducted with adults (aged 20-65 years) from an existing cohort randomly selected from 10 urban areas in Stoke-on-Trent, England [7 (link)]. For each of the 10 Lower Level Super Output Areas in the city (LSOA; mean population 1500) [21 ] GIS maps were created using 1:10,000 scale Ordnance Survey Street View mapping. Maps displayed an area of 1 mile around the boundary of each LSOA (approximately centred on the participants' residence) and included road names and places of local interest such as green space, shops, services, schools and physical activity facilities.
Each participant was shown the printed GIS map for their local area. The interviewer helped to orientate them by pointing out the location of their residence, main roads, and local landmarks. Using maps for reference or simply from recall, participants were asked to:
(i) Recall all recent (last seven days) and usual walking destinations from their home. In the event that participants were unable to identify any (more) walking destinations, interviewer prompts were used to ask them to recall any places they had walked from their home without using the map, or used the map to identify possible destinations (e.g., local shopping areas, pubs/bars, family/friends).
(ii) Draw their 'neighbourhood area' on the map. Participants were advised that it could be any size or shape, and that there was no right or wrong answer.
Annotated maps from all participants were scanned back into a GIS for analysis. All recalled destination points and 'neighbourhood area' boundaries were digitised and the annotation from the maps recorded as feature attributes. GIS analysis was used to create a number of Euclidean and network distance buffers around the address location of each participant. A Euclidean buffer is a straight line circular radius around an address, whereas the network buffers were calculated by measuring a defined distance along the pedestrian street network (i.e., roads and pathways used by pedestrians) in all possible directions away from a participant's address. The end points of these routes were joined together to form an enclosed area representing the total area within a defined walking distance of the address.
For each participant we produced five different neighbourhood areas (Figure1 ): 1 mile (or 1.6 km) around the individual's home, network and Euclidean buffers; 1 km around the individual's home, network and Euclidean buffers; participant's perceived neighbourhood area. The areas of the perceived neighbourhoods were calculated as a proportion of the different network and Euclidean buffers to explore discrepancies in area size [i.e., area within perceived neighbourhood boundary (m2)/area within network or Euclidean neighbourhood boundary (m2) × 100].
Reported walking destinations were placed into eight categories used in the recently developed European ALPHA questionnaire [14 (link)]: retail (e.g., shops, supermarkets, grocers); local services (e.g., banks, libraries); eating and drinking (e.g., pubs, cafes, restaurants); family and friends; work/school; bus stops; green space (e.g., parks and common areas); and physical activities facilities (e.g., leisure centres, private gyms, swimming pools). Each geocoded destination was defined as lying within or outside each of the five different neighbourhood areas. These were examined for the sample as a proportion of total walking destinations.
The study was approved by the Staffordshire University ethics committee.
Each participant was shown the printed GIS map for their local area. The interviewer helped to orientate them by pointing out the location of their residence, main roads, and local landmarks. Using maps for reference or simply from recall, participants were asked to:
(i) Recall all recent (last seven days) and usual walking destinations from their home. In the event that participants were unable to identify any (more) walking destinations, interviewer prompts were used to ask them to recall any places they had walked from their home without using the map, or used the map to identify possible destinations (e.g., local shopping areas, pubs/bars, family/friends).
(ii) Draw their 'neighbourhood area' on the map. Participants were advised that it could be any size or shape, and that there was no right or wrong answer.
Annotated maps from all participants were scanned back into a GIS for analysis. All recalled destination points and 'neighbourhood area' boundaries were digitised and the annotation from the maps recorded as feature attributes. GIS analysis was used to create a number of Euclidean and network distance buffers around the address location of each participant. A Euclidean buffer is a straight line circular radius around an address, whereas the network buffers were calculated by measuring a defined distance along the pedestrian street network (i.e., roads and pathways used by pedestrians) in all possible directions away from a participant's address. The end points of these routes were joined together to form an enclosed area representing the total area within a defined walking distance of the address.
For each participant we produced five different neighbourhood areas (Figure
Reported walking destinations were placed into eight categories used in the recently developed European ALPHA questionnaire [14 (link)]: retail (e.g., shops, supermarkets, grocers); local services (e.g., banks, libraries); eating and drinking (e.g., pubs, cafes, restaurants); family and friends; work/school; bus stops; green space (e.g., parks and common areas); and physical activities facilities (e.g., leisure centres, private gyms, swimming pools). Each geocoded destination was defined as lying within or outside each of the five different neighbourhood areas. These were examined for the sample as a proportion of total walking destinations.
The study was approved by the Staffordshire University ethics committee.
Free full text: Click here
