Neighbourhood effects on health: Does it matter where you draw the boundaries?

“, 2001; Guest & Lee, 1984; Haney & Knowles, 1978; Lee & Campbell, 1997), or zone design techniques to explore the effects of boundary alterations on research results (Cockings & Martin, 2005; Flowerdrew et al., 2008; Haynes, Daras, Reading, & Jones, 2007; Martin, 1998).”

“At least the following criticisms have been put forward concerning such a procedure: a) Administrative units are not intended to capture health exposures meaningfully (Flowerdew et al., 2008; Wheeler et al., 2012); b) it is assumed that people in a neighborhood have similar exposures, independent of their daily mobili…”

“Specific to health research, Flowerdew, Manley, and Sabel (2008) demonstrate how results can vary dependent on how neighborhood boundaries are drawn.”

“…to neighbourhood delimitation, but more complex definitions of territorial neighbourhoods may consider built environment features, population characteristics, historical or social processes, and collective representations (Chappell et al., 2006; Flowerdew et al., 2008; Gauvin et al., 2007).”

“…and recent social science studies (especially environmental health and neighborhood effects research), this article further elaborates on the nature of the UGCoP and explores how recent advances in geographical information system (GIS) and geospatial technologies can help address the problem.”

“The modifiable areal unit problem is a persistent geographical-analysis issue whereby the boundaries (ie, municipalities or cycle zones) selected to aggregate spatial data can affect findings (Flowerdew et al, 2008; Horner and Murray, 2002; Jelinski and Wu, 1996).”

“It has been argued that although formal neighborhood units such as census tracts offer readily available data, they might not be at the right scale or zonation to accurately reflect or measure health-related processes and outcomes (Flowerdew, Manley, and Sabel 2008).”

“Moreover, internal homogeneity and resident identification are meaningful to draw neighborhood boundaries (Flowerdew et al., 2008).”

“…similar AZP-based algorithms for a range of purposes, including the development of standard geographies for the release of statistics and the creation of zoning systems for specific investigations (Cockings and Martin, 2005; Flowerdew et al, 2008; Grady and Enander, 2009; Haynes et al, 2007; 2008).”

“Scale effects is a dominant theme in spatial studies (Flowerdew et al. 2008; Lee et al. 2008; Villanueva et al. 2015), often described as the uncertain geographic context problem (UGCoP) (Kwan 2012).”

“For example, CaPS is situated in the South Wales valleys where there is high slope variability and an ongoing post-industrial transition of land use.”

“Further, there is some evidence of variations in the association of neighborhood characteristics with health outcomes when boundaries are changed (the modifiable area unit problem) (73).”

“This is when different associations are found for the same data when they are aggregated to units of a given shape that differ in size, or of a given size that differ in shape [28, 29].”

“It can be achieved discursively (manually) by grouping basic spatial units into larger ones [57-59], by combining social, statistical, and spatial analysis methods [60,61], and automatically through computationally intensive automated zoning software [9,15,37,62-65].”

“When it comes aggregating addresses into neighbourhoods, we face zoning and scaling problems (see Flowerdew et al. 2008 on the modifiable area unit problem).”

“However, other studies have shown that, when measuring direct area effects, “it does matter where [we] draw the boundary” (Flowerdew et al., 2008).”

“The influence of the well-established modifiable area unit problem (MAUP) has received much attention in the literature on health and place (Flowerdew, Manley, and Sabel 2008).”

“Nevertheless, we acknowledge the possibility that each of these definitions may mask real neighbourhood effects (Flowerdew et al. 2008) and may poorly reflect what ‘neighbourhood’ really means for individuals (Galster 2001).”

“…actively engaged in the debate, once again revisiting geographic theories such as the MAUP and applying them to health research (Haynes et al. 2007; Spielman and Yoo 2009) and examining the effect of spatial scale on neighborhood studies (Weiss et al. 2007; Flowerdew, Manley, and Sabel 2008).”

“the phenomenon of spatial autocorrelation of mortality, neighborhood problems and boundary issues (see, [26-33] or [34]).”

“This observation confirms that the way in which data are aggregated at the neighborhood level have a significant impact on the relationship between the neighborhood environment and health [62,63].”

“…boundaries and scales, are subject to resident perceptions and activities, which are individual specific and affected by the sociodemographic characteristics of the (sampled) individuals (Guest and Lee 1984; Beguin and Romero 1996; Coulton et al. 2001; Flowerdew et al. 2008; Chaix et al. 2009).”

“(Coulton et al. 2004; Flowerdew et al. 2008; Heywood et al. 1998).”

“However, they are subject to the modifiable areal unit problem (MAUP) where results can vary depending on the division of the study area, a zonation or aggregation effect [18], and the size of the units used, a scale effect [19].”

“Additionally, the methodology described in this study can be transferred to other regions to further evaluate MAUP implications in the context of deprivation studies.”

“Others concluded that modes of determining neighborhood boundaries that are sensitive to local context can improve inferences into neighborhood effects on health (Flowerdew et al., 2008; Krieger et al., 2002; Kwan, 2009; Riva et al., 2008; Riva et al., 2009).”

“For example, Flowerdew, Manley, and Sabel (2008) found that the spatial scale of the neighborhood significantly affected whether there was an association between place and long-term illness in England.”

“emerging from what has been described in the geographical literature the modifiable areal unit problem’s (MAUP’s) ‘scale’ and ‘zonation’ effects [69].”

“It also reinforces the need to think about the geography of “effective” units prior to analysis, and also to experiment with different scales and aggregations (see also Flowerdew, Manley, and Sabel 2008, Riva et al. 2008, Mobley, Kuo, and Andrews 2008, and Spielman and Yoo 2009).”

“19 20 Furthermore, the fixed effects approach removes the effects of unmeasured time-invariant confounders.”

“Deprivation is often studied at small area levels, whereas caution is also needed when drawing inferences at an individual level [32, 81].”

“We explore this by a variety of spatial scales (small areas and larger neighbourhoods or localities), as it has been noted that the extent to which area of residence may be important for health may depend on the spatial scale and neighbourhood boundaries used (Flowerdew et al., 2008).”

“However, the most practical way is to use census areas such as census blocks and census tracts (Flowerdew et al. 2008).”

“…effect - the number of areal units used (Kwan and Weber 2008; Schuurman et al. 2007; Smiley et al. 2010; Bell et al. 2013; Shah, Aspen, and Bell 2014), and 2) zonation effect - the choice of aggregation or boundaries (Flowerdew, Manley, and Sabel 2008; Stafford, Duke-Williams, and Shelton 2008).”

“53 Δ in visceral adipose tissue 711 (34) 729 (39) 694 (34) 604 (38) 0.”

“…tandem relationship can have significant influence on characterizing and modeling associations between the environment and health-related outcomes (Flowerdew et al., 2008; Haynes et al., 2007; Jackson, Davies, & Leyland, 2010; Reijneveld, Verheij, & de Bakker, 2000; Oliver & Hayes, 2007;…”

“Secondly, social behavior and interaction are not necessary restricted within one’s immediate neighborhood, especially when the neighborhood boundaries are defined in a way that does not coincide with one’s real-life experience such as census geography and postal code (Flowerdew et al. 2008; Guo and Bhat 2007; Riva et…”

“While using census boundaries and predetermined buffers may be a convenient means of approximating neighborhood, these proxies are inadequate in identifying the various spaces that individuals travel to and are exposed to and thereby the factors that may impact health [22-25].”

“Intra-unit homogeneity was defined as the level of homogeneity in exposure variables within the areas forming the geographical unit (Gauvin et al., 2007; Flowerdew et al., 2008; Grady and Enander, 2009).”

“…divisions and administrative zoning systems by municipalities are often preferred for data collection (e.g., surveys/censuses), they may not be consistent with residents' perceived neighbourhoods, their daily activities, and how they use the city itself (Chaix et al., 2009; Flowerdew et al., 2008).”

“Consequently, there is no one ideal way of defining a neighborhood and its spatial boundaries [47].”

“This issue is closely linked to the modifiable areal unit problem (MAUP), where the shape and size of an analytical unit affect both exposure estimates and their associations with human health 29–32 (figure 1).”

“…50 because of security requirements).1 The ECLS-B includes ZIP code data for the child’s residence at each wave, which are not as fine-grained as census tract “neighborhood” data but are an accepted proxy in the absence of other georeferenced data (Flowerdew et al. 2008; Krieger et al. 2002).”

“Local availability of health-promoting amenities and access to health and social services may influence health [45].”

“The modifiable areal unit problem has prompted many authors to think about the sociospatial context of people beyond standard administrative units at a single spatial scale (see, e.g., Flowerdew, Manley, and Sabel 2008; Kwan 2009).”

“Our buffer approach to delineate the geographic context circumvents methodological issues of administrative neighbourhood units (Flowerdew et al., 2008).”

“As Flowerdew, Manley, and Sabel (2008) showed, using different definitions of neighborhoods at nonconventional scales can change whether or not a so-called neighborhood effect is observable.”

“Importantly, the scale at which these dynamics operate and the appropriate spatial resolution at which to investigate their effects is unclear (Flowerdew et al., 2008); therefore, we conducted our analyses using two census geographies.”

“Also known as the modifiable areal unit problem, the type of geographic boundary used to aggregate data can affect variance, standard deviations, correlation, and regression analyses [38].”

“A limitation of this strategy may be the difficulty of obtaining a consensus view (Flowerdew et al., 2008).”

“Given the uncertainty around neighborhood boundaries it is perhaps reasonable to estimate neighborhood effects using multiple geographic units of analysis (as suggested by Flowerdew et al, 2008).”

“…are ‘quantitatively inclined’ – GIS are an important and increasingly necessary tool.28 Indeed, much contemporary research in spatial science would be virtually impracticable without the facility to collate large and complex data sets, exemplified in Flowerdew et al. (2008) and Sabel et al. (2013).”

“Due to concerns over heterogeneities in state and county size across the US and their potential effect on network structure (Flowerdew et al., 2008), we also approximated a state and county network using equally-sized geographic areas as nodes.”

“Flowerdew et al. (2008) set out to question whether it does in fact matter where zone boundaries are drawn, using 2001 census data on limiting long-term illness (LLTI) for Northamptonshire and Swindon, England.”

“55 We opted not to generate composite area-level measures of material deprivation 74 or social fragmentation.”

“Put together there is a background neighbourhood socioeconomic status (Stafford and Marmot, 2003) or ‘neighbourhood effect’, which can predispose an individual to be more at risk of disease and death (Stead et al., 2001; Flowerdew et al., 2008).”

“Noting the significant conceptual and methodological debates on defining meaningful neighbourhood boundaries (Flowerdew et al., 2008), in this study we adopted each participant’s self-defined conception of neighbourhood, an approach consistent with our interest in older adults’ perceptions and agency.”

“Interestingly, although a number of studies have investigated the neighbourhood, there is still no consensus definition on what ‘neighbourhood’ means (Young et al., 2004; McNeill et al., 2006; Flowerdew et al., 2008; Buffel et al., 2012; Alidoust et al., 2017).”

“The MAUP pertains to the variation in outcomes of analysis that arise from changes in the configuration and/or scaling of areal unit boundaries (Flowerdew, Manley, & Sabel, 2008; Jelinski & Wu, 1996), whereas the UGCoP pertains to errors in individual-level inferences resulting from the spatial…”