Building on existing literatures that examined residents’ perceptions of neighbourhood characteristics that impact their MWB using concept mapping [14, 17], our work is the first to examine these relationships while considering the influences of age and depression status. We identified a clear set of neighbourhood attributes that our participants perceived to be influencing MWB. In the 5 participating neighbourhoods, most respondents emphasized that aesthetic attributes, social and physical environment, accessibility, and availability of a range of services within the neighbourhood were all seen as important to their good MWB. Conversely crime and negative environmental attributes were seen as contributors to poor MWB. Attributes that cause a neighbourhood to be aesthetically pleasing, were of interest, because it was both the most subjective and least well defined concept to arise from the mapping sessions.
Heterogeneity within neighbourhoods was emphasized by residents, not uniformity. Those features contributing to poor MWB indicated that although residents wanted a diverse community, there was a point at which certain features could become problematic (e.g. while close proximity to neighbours was often seen as positive, it could give way to overcrowded housing and excessive levels of traffic and noise). While having a visible marginalized population (e.g. homeless) was seen as a negative feature, residents also emphasized the need for services to support the homeless and marginally housed within their communities.
We found considerable agreement among participants regarding features that were perceived to contribute to poor MWB. Crime, negative neighbourhood environment (including overcrowded housing, noise, litter, heavy traffic) and social concerns (e.g. safety, isolation, class discrimination, gentrification) were consistently cited as the three most important clusters influencing poor MWB, consistent with data from the United Kingdom . This held true regardless of gender, household income, and age. Using successive waves of British crime data, Markowitz and colleagues (2006) suggest that there is a feedback loop acting in progression between several of these negative neighbourhood characteristics. Their model proposes that decreased neighbourhood cohesion leads to increased crime, promoting increased levels of fear which in turn decreases neighbourhood cohesion . It is possible that similar feedback loops may be operating in the Toronto context.
There was less agreement among participants as to which features are most important to good MWB. As did Burke et al., we identified differences in how clusters were rated based on income status . Low-income participants were more likely to indicate that aesthetic attributes (e.g. physical beauty, parks), availability of beneficial community services (e.g. public recreational facilities, libraries, health services, access to places of worship), and accessible transportation were influential to good MWB. Nielsen and Hansen suggest that access to a green area from an individual’s residence is associated with less stress, although the frequency of visits to a green area, was not associated with any health benefits . Perhaps ‘just knowing’ the green area and/or other community services are available is important for good MWB among low-income participants. The importance of ‘knowing’ neighbours to visit, with whom to borrow or exchange favours has been shown to protect MWB in areas of deprivation in a multilevel population analysis in Wales . This aspect of trust in other members of the community to explain health inequalities is supported by other Canadian data [17, 27]. By contrast those experiencing higher family incomes and are younger than the age of 30 years old cited crime, negative neighbourhood environment, and social concerns as having the greatest influence on good MWB, whereas these were ranked very low among the older sample (Figure 4). Figure 3 demonstrates the contrast between the income groups; the contrasts are much less pronounced between men and women and between depressed/non-depressed participants.
The reasons for the income-related differences remain unclear. Perhaps higher-income individuals are more concerned with private property (i.e. more likely to possess it and have more to lose as victims of property crime). These concerns may in turn manifest themselves in anxiety and fear. A review of the literature of neighbourhood characteristics and maternal and child health reported that over one-third of the articles (n = 31) included a measure of residents’ perceptions regarding issues of social resources within their neighbourhoods .
Among the non-low-income participants, age appeared to have an effect on cluster ratings related to good MWB. The pattern matches in Figure 4 show that those participants under age 30 prioritized crime, negative neighbourhood environment and social concerns as highly important to good MWB – as opposed to their over-age-30 counterparts, who emphasized neighbourhood aesthetics, the presence of essential services, and a positive community environment. The reasons for such age-related differences remain unclear. These findings may reflect individual experiences rather than neighbourhood-specific ones. It will require further study to determine if younger respondents are more influenced than older residents by experiences of their peers (rather than their experiences of neighbourhood per se). Interestingly, for low-income groups, the age effect disappeared, showing considerable agreement on cluster ranking for good MWB.
While our findings varied somewhat based on participants’ demographic characteristics, they sketched complex linkages between neighbourhood features and residents’ MWB. Despite the small sample size, these data demonstrate that the contributors to good MWB are not simply a corollary of factors contributing to poor MWB. As such, this study represents a unique contribution to the existing literature on neighbourhoods and health.
A number of interesting questions were raised in the concept mapping sessions. We had assumed at the beginning of our study that built environment (as reflected by the exterior of buildings) was what constituted one of the primary physical features of a neighbourhood. This was challenged by participants during the brainstorming sessions. Those living in high-rises classified spaces within buildings (and between individual home units) as important features of neighbourhood. The importance of ‘internal spaces’ to positive MWB is further corroborated in the literature .
It is unlikely that employing a single methodology would be sufficient to capture the complex linkages between neighbourhood features and the MWB of residents. In order to fully understand the pathways by which neighbourhood features influence MWB, multiple methods are required . In addition, perceptions of factors that are related to MWB may be different in Canadian neighbourhoods as compared to similarly sized American neighbourhoods. In a related study, our research team evaluated a composite systematic social observation (SSO) tool in these same neighbourhoods. This instrument is premised on the assumption that physical and social disorder negatively impact residents’ health that was derived from blockface observation tools pioneered primarily in the U.S.A. This SSO investigation contributed to our understanding of ‘neighbourhood disorder’ and led us to question whether this is the appropriate construct for the Toronto context . This failure could be due to the nature of the tool, or differences between US and Canadian cities, or both. Therefore, the theoretical framework developed from concept mapping enhances our understanding of these relationships by revealing linkages between neighbourhood features (parks, tree lined streets, well maintained streets and sidewalks, mixed use of neighbourhood structures) and residents’ perceptions regarding their contribution to MWB. Our study questions the homogeneity with which area level factors have been judged to impinge on MWB. It suggests that it is important to acknowledge the differential impact of neighbourhood factors by demographic variables such as gender and income, newcomer status, and age.
This study has a number of limitations. We were only able to conduct interactive mapping sessions in 3 of the 5 neighbourhoods. The relatively low sample size resulted from difficulties in recruitment and retention of participants for subsequent group sessions. The sample was characterized by a skewed distribution in terms of age and gender – and by a preponderance of low-income participants and those not in the work force. Relatively few home owners and parents of young children were represented. Also, the high proportion of participants with relatively high depression ratings on the CES-D may also suggest a selection bias. The cross-sectional nature of this study failed to adequately capture residents’ experiences with neighbourhoods previously inhabited and how this might have shaped their perceptions of their current neighbourhoods.
The study has implications for other areas of research involving neighbourhoods, such as those using SSO methods. These data suggest that we need to redefine traditional constructs of neighbourhoods, for example looking at spaces within multi-residence buildings and residents’ perceptions of neighbourhood boundaries. Our data indicated that the relationship of neighbourhoods to MWB is one characterized by complexity – for example, residents spoke eloquently to both positive and negative aspects of gentrification, a process they saw as both potentially beneficial and harmful; in the end it was placed in the cluster of social concerns, a cluster more likely to contribute to poor MWB than to good.