Does “Where We Live” Reflect “What We Do”? An Overview of Empirical Evidence on Place-People Correspondence with Regards to Sustainability Performance

Despite effort to uncover the link between people’s behaviors and places regarding their sustainability performance, the findings remain ambiguous. This paper presents an overview of the empirical evidence on this issue and provides a framework to help design of studies on the performance of residential choice considering the three pillars of sustainability. We identified the papers through a search in multidisciplinary databases from 1994–2017 and experts’ recommendations. We analyzed them considering these questions: 1) what aspects of places and people’s behaviors are evaluated? 2) To which pillar of sustainability do they refer? 3) How are places and people related; and, 4) why are they linked the way they are? The conclusion underlines the need to embrace the complexity of residential choice and the associated daily mobility as well as for policymakers to define actions to be taken for unsustainable places to be improved and more sustainable behaviors to be adopted.


Search Strategy
A literature search using keywords Residential choice AND Lifestyle OR Behavior AND Residential neighborhood OR Built environment OR Urban forms AND Sustainability, was performed on databases (e.g. Web of science and SpringerLink). In addition to the web search, 22 references were provided from two experts. A total of 140 papers published after 1992 1 were thus first identified on the basis of their titles. Since this review aims at understanding the mechanisms linking people's behaviors to their places of residence with regard to sustainability, we considered the two following exclusion criteria after reading the abstracts: 1) abstracts focusing exclusively on either places or behaviors were excluded; 2) papers focusing on scenario situations instead of real life situations were also excluded. Of the 140 initial papers, we were left with 40 articles which were read entirely; after which nearly half of them were further excluded because of the similarity between their contents.

Corpus Description
Of the twenty-two articles retained for the critical overview (Table 1), six were written by geographers, five by civil engineers, one by environmental psychologist; one by economist; one by sociologist; one by computer scientist. The seven others were led by interdisciplinary teams among which transport and urban planners and only one including architects. The majority of the research was conducted in Western countries, with about 40 percent, in European countries (3 in the UK, 2 in Belgium, 1 in Austria, 1 in Denmark, 1 in Finland and 1 in Germany). The rest were conducted in North America (4 in the US, 1 in Canada); Australia (3); China (3); Korea (1), and finally, Iran (1). Two main types of strategies are used for evaluating the sustainability of residential choice and the level of congruity between people's place and behaviors. The first type favors the objective measurement of the built environment and socio-economic factors (fifteen studies focus strictly on these dimensions). The second favors subjective variables associated with psychological factors (eight studies target solely these dimensions). Finally, ten studies mix both approaches, incorporating objective and subjective variables. Most studies used cross-sectional survey designs (14 quantitative, 1 qualitative and 2 mixed), the others proceeded through cohort observation with quantitative research design: quasi-longitudinal (n=2), prospective (n=1) and retrospective (n=2). The sample sizes are highly variable (from n=26 to n=75331) among studies. The primary criterion for sample selection is to have respondents residing in different types of places (e.g. suburban vs. urban neighborhoods; transit-oriented development vs. "regular" neighborhood). For cohort studies, another considered criterion that respondents have experienced relocation. ** Urbanization: refers to the traditional suburb/urban approach whereas more urbanized structure refers to inner city areas.
***New inner city development refers to new residential neighborhoods within the existing built structure, which has a lower density compared to the inner city structures (Jarass & Heinrichs, 2014).

Operationalizing the Sustainability Performance of People-Place Interaction
Among our corpus of 22 articles, we identified a diversity of indicators of different nature used to evaluate the sustainability performance of people-place interaction (Table.1). To evaluate the sustainability of places, the indicators used to describe the residential location in terms of its built environment or socioeconomic characteristics (e.g. Boussauw & Witlox, 2011;Delmelle, Haslauer, & Prinz, 2013;Etminani-Ghasrodashti & Ardeshiri, 2016;Farber & Li, 2013;Figueroa, Nielsen, & Siren, 2014). With regard to evaluating the sustainability of the built environment, most indicators provide quantifiable measures. Boussauw & Witlox (2011), Fan et al.(2011 and Sung & Lee, (2015) considered the contribution of accessibility, population density, and residential density. In other studies, it is a typology of urban developments that is considered, e.g. transit-oriented development (TOD) versus non-TOD (Kamruzzaman et al., 2013) or inner city, inner suburb and outer suburb (e.g. Valkila & Saari, 2013). Concerning the definition of the territorial unit of measurement, the most common approach is to define a buffer around each respondent's home. Fan et al. (2011), Kamruzzaman et al.(2013) and Yu et al. (2012) respectively uses radiuses of 0.4 km, 0.6 km, and 1.2 km; Boussauw & Witlox, (2011), radiuses of 1 km, 4 km, and 8 km. A variation is to define the buffer zone around a rail or rapid bus transit (Nahlik & Chester, 2014). Another avenue is to use a homogeneous dwelling density, e.g. precincts defined as 30 or more dwellings/hectare (Buys & Miller, 2011) or existing administrative boundaries, e.g. census Lower Super Output Area (LSOA) (Aditjandra et al., 2012). The incorporation of subjective indicators to evaluate the sustainability performance of place is also necessary. Indeed, even a homogeneous socioeconomic group living in the same location may differ in terms of individual behaviors (De Vos, Derudder, Van Acker, & Witlox, 2012).
To evaluate the sustainability of people's behaviors, indicators of different natures were used. All but five studies used objective spatiotemporal indicators related to daily mobility. They measure traveled space and time, trip frequency, or specify travel purpose or transport modes. Researchers use these variables in combination e.g. travel mode and travel purpose (Barr & Prillwitz, 2012;Buys & Miller, 2011), travel mode and travel distance (Figueroa et al., 2014), or travel length and travel mode (Valkila & Saari, 2013). Some also manipulate variables through mathematical formulations to carve out their area of interest. For instance, Fan et al. (2011) combine a set of variables to define the indicators required travel, and excess travel. Required travel is "a function of the relative distances among daily activity locations", while excess travel is a function of "the relative distances between the actual residential location and the daily activity locations" (p. 1242). A lower level of required travel means to have smaller geographical areas and less dispersed activity locations; a lower level of excess travel means to have a better coordination between home and activity locations and vice versa, which may encourage or promote sustainability. Buys & Miller (2011) incorporate four variables to qualify daily mobility, bringing forward the concept of convenience. It is defined as the intersection of utilitarian and psycho-social elements and is identified as a determinant factor of transportation choice. It is the outcome of four objective measures: time-efficiency, seamless journey 2 , distance to the destination and purpose of the journey.
The social imperatives of sustainability were considered in six studies. The subjective indicators measured alternately the satisfaction with the social composition of the neighborhood or the perception of its social cohesion, the sense place or commitment with the neighborhood, the social interaction with the neighbors, and the satisfaction with current domicile and residential preferences. Some studies refer to two useful concepts. The first one is the residential dissonance 3 which refers to the mismatch between actual and preferred residential neighborhoods (De Vos et al., 2012;Kamruzzaman, Baker, Washington, & Turrell, 2013). The second is residential self-selection which refers to the "an individual's inclination to choose a particular neighbourhood according to their travel abilities, needs, and preferences" (Aditjandra et al., 2012;Yu et al., 2012;Kamruzzaman et al., 2013).
Four studies evaluated the sustainability of people's behaviors from an environmental standpoint from three indicators: energy use, carbon footprint and vehicle occupancy (Figueroa et al., 2014;Jarass & Heinrichs, 2014;Valkila & Saari, 2013;Yu et al., 2012). Finally, the economic aspect of behaviors were only considered in one study (Yu et al., 2012). Table 2 summarizes the variables used to measure people's behavior identified in the literature.

Which Imperative of Sustainability Is Evaluated in the Studies?
It is well-known and accepted that sustainability encompasses at least three imperatives of environmental, social and economic natures. Environmental sustainability deals with the impact of the development process on biodiversity of habitats and the utilization of natural resources (Deakin, Curwell, & Lombardi, 2001). Social sustainability refers to the strong of sociocultural life, social involvement, access to services, safety and security and overall human well-being both mentally and physically (Bacon, Douglas, Woodcraft, & Brown, 2012;Deakin et al., 2001;Woodcraft, 2012). Regarding economic sustainability, we endorse the definition of Markandya & Pearce (1988) who define it as "the use of resources today should not reduce real incomes in the future" (p. 5) and agree with Moldan, Janoušková, & Hák (2012) that in view of recent economic and financial crises, economic issues should be addressed "on their own merits, with no apparent connection to the environmental aspects" (p. 4).
A first group of variables are unidimensional in that they measure only one dimension of sustainability. For instance, energy use, carbon footprints, or vehicle occupancy relate to environmental sustainability. Social satisfaction, sense of place or neighborhood commitment rather pertains to social sustainability. A second group of variables could be labeled multidimensional in that they relate to more than one dimension of sustainability. For instance, with regard to transport modes, the use of public transit may contribute to environmental sustainability, but have a negative impact on social sustainability by increasing travel time which may reduce time for social interactions with family or neighbors. Similarly, walking may contribute to environmental, social, and economic sustainability by diminishing traffic, pollution, improving health and social relations, and diminishing car-associated monetary expenditures. Finally, a last group of variables allows for a better jsd.ccsenet. understand sustainabil

How C
The review place perfo places and unsustaina sustainable significant   Vol. 11, No. 2;. This being said, even when walking was their primary transport mode, it was affected by the perceived safety (Buys & Miller, 2011). According to Fan et al. (2011), the presence of children in households increased daily mobility. Longer travel distances were related to families with children. Indeed, school quality and location were strong influences on residential choice, and reduce households' opportunities to concentrate their daily trips on smaller geographical areas. The number of children had no significant effect on excess travel (Fan et al., 2011) although larger households and families with children were found to rely more on car (Aditjandra et al., 2012;Susilo, Williams, Lindsay, & Dair, 2012), which may be used to conciliate parental, familial, or professional obligations. This being said, larger households were also found to use public transportation because of limited access to cars for all members. It is not clear whether income influences travel mode choice through increasing car ownership and whether it has a direct effect on mobility (Aditjandra et al., 2012;Fan et al., 2011;Figueroa et al., 2014;Yang, Fan, & Zheng, 2016). As for the impact of education and employment, highly educated professional workers tend to depend less on car and walk more compared to other socioeconomic groups (Kamruzzaman et al., 2013;Yu et al., 2012). Bike ownership incites biking (Barr & Prillwitz, 2012;Jarass & Heinrichs, 2014) and car ownership encourages driving (Aditjandra et al., 2012;Barr & Prillwitz, 2012;Delmelle et al., 2013;Etminani-Ghasrodashti & Ardeshiri, 2016;Kamruzzaman et al., 2013;Yang et al., 2016). This latter is also the only significant factor pertaining to life situation that influences social satisfaction (Delmelle et al., 2013). Interestingly, the same factor has a negative impact on environmental sustainability.

Lifestyle
Variables related to lifestyle refer to preferences and attitudes (Aditjandra et al., 2012). A total of 10 papers out of 22 investigated the influence of lifestyle factors on the congruity between the level of sustainability of people's behaviors and their places of residence. Barr & Prillwitz (2012) identified four behavioral profiles with regard to transport: "Addicted Car Users", who used the car most frequently and lived in low-density places did not show pro-environmental attitudes. "Aspiring Green Travellers", who still relied on the car, but used other transport modes, especially active transport, and had strong environmental attitudes. "Reluctant Public Transport Users" used public transport as their primary transport mode, but had relatively negative environmental attitudes. Finally, "Committed Green Travellers", whose attitudes matched their behavior. These individuals relied on walking and had strong pro-environmental attitudes. With no surprise people favoring public and active transport drove less, and those who preferred easy access to shopping facilities drove more (Aditjandra et al., 2012;Yang et al., 2016).
Residential consonance or dissonance refers to the match or mismatch between actual and preferred residential neighborhood. It was studied by De Vos et al. (2012) and Kamruzzaman et al. (2013). Both studies found that the residential consonance or dissonance had a significant influence on travel mode choice, especially on public and active transport. On the other hand, built environment had a stronger impact on car use compared to the other modes of transport. Namely, urban consonants and rural dissonant were more likely to use AT and PT and urban dissonant and rural consonants were more likely to use car, with a lower degree of association compared to the previous group because car use is more influenced by built environment. Consideration for School quality considerations were positively associated with required travel and negatively related to excess travel (Fan et al., 2011). People who considered schools quality in their residential choice had fewer opportunities to concentrate their other daily activities (Fan et al., 2011). According to (Fan et al., 2011), smaller excess travel may result from the fact that this group of people faced more temporal constraints and tried to create more spatially coordination between residences and daily destinations. Consideration for neighborhood security was negatively associated with excess travel and had no influence on required travel, which may be explained by the fact that people with security concerns had limited housing options which may keep them from residing in the neighborhoods that can minimize their daily travel.
As we have observed above, even though the lifestyle factors were important in the determination of sustainable travel behavior, they were not always reflected in people's behaviors. People did not or could not always act on their attitudes and preferences because of their life situation, the built environment where they live, or simply for convenience.

Convenience
Convenience corresponds to the intersection between utilitarian and psychosocial dimensions. It is considered as a determinant factor of transportation choice and developed used by Buys & Miller, (2011). They define it through three key elements: time-efficiency, seamless journey, as well as distance to and purpose of the journey.

Conclusions, Limitations and Future Research
This article provides interesting insights into the complexity of measuring people-places performances in terms of sustainability, as well as interpreting the congruity of performances. As our results show, "where we live" might not necessarily reflect "what we do" and vice-versa. This being said, this critical overview does not purport to represent all disciplinary fields, neither distinguish approaches with regards to place performance or describe the involved dimensions in sufficient detail because of space limitations. Nevertheless, this paper provides a conceptual umbrella embracing elements of the utmost importance, which can be used to orient future studies and feed further discussions. It could be expanded with additional dimensions (e.g. desirability and upstream impact) and developed into a theoretical model to explain degrees of congruity between places and behaviors. This paper also suggests that for the development and implementation of solutions to move forward to a more sustainable society, individual behaviors must be understood and considered as part of the solution. There is also much work to be done to go beyond mobility-associated behaviors and this calls for interdisciplinary approaches to be developed.
A limitation of the overview derives from the fact that this paper focuses only on real life situations and excludes scenario situations, which constitutes a considerable portion of the body of knowledge on place-people correspondence with regards to sustainability performance. Other limitations concern the nature of the paper, which follows a holistic approach to provide a conceptual framework. These limitations were grave, but did not stop us, as was seen, from extracting significant indicators from the body of literature and developing a conceptual framework to help design of studies on the performance of residential choice considering the three pillars of sustainability.
The limitations in this paper suggest particular needs for future research. First, while analysis of real situations provides an understanding of sustainability of an existing system, future research is needed to study scenario situations to explore some other aspects of the use of indicators to gauge the complex systems of place-people interactions. Although a holistic approach is helpful to develop an overall framework for reference, future research is needed to use an analytic approach and to delve into the complex determinants of place and people performance considering the environmental, social and economic (ESE) pillars of sustainability.