1All over the world, cities are confronted with a multitude of (new) challenges. This paper explores two urban challenges, namely segregation and suburbanisation.
2In Northern America, people are increasingly choosing their residential environment on the basis of immaterial value orientations and lifestyle preferences. This clustering of like-minded Northern Americans has caused social communities of “kindred spirits” to arise – a process of political and lifestyle segregation that has recently been brought to the attention by Bill Bishop (Bishop, 2009). A recent research in the Netherlands (VROM-raad, 2009) confirms this trend of lifestyle segregation within a Northern European context. Regarding to suburbanisation, it is known that suburbanisation is one of the many causes of the increase in urban sprawl, and is still regarded as one of the major common challenges facing urban Europe today. The sprawling nature of Europe’s cities is critically important because of the major impacts that are evident in increased energy, land and soil consumption (European Environment Agency, 2006). In Belgium, despite a substantial urban funding program and a spatial policy in favour of the cities, the process of suburbanization did not come to an end. Moreover, since 2001 higher suburbanization percentages were reported (Moortgat and Vandekerckhove, 2007).
3These different surveys within a European and Northern American context state that segregation and suburbanisation indeed are important challenges. This paper explores the idea that these processes are related with residential choices (or preferences) made by residents with divergent lifestyles and value patterns.
4This can be applied to the specific situation in Belgium. Although considerable research has been devoted to Belgian political preferences in relation to value orientations as an aspect of lifestyles (Waege and Agneessens, 2001 ; De Maesschalck, 2010), less attention has been paid to relations between Belgian residential preferences and lifestyles (Pisman, 2007). Therefore, this paper explores the relation between lifestyles and residential choices or preferences of residents living in two urban and two suburban neighbourhoods in Ghent, a city in Flanders, the highly urbanized and Dutch speaking region in the North of Belgium. To start with, the paper examines the background of lifestyles within different academic fields, including the field of planning, paying specific attention to differences between urban and suburban lifestyles. Afterwards the research design, methodology and data are introduced. Finally, the results of the case study are presented, followed by the conclusions and discussions about lifestyles in relation to housing, suburbanisation and segregation and more specifically about lifestyles and residential preferences of urban and suburban residents in Belgium.
5The concept of lifestyle has different roots, and it has been given a wide range of meanings and uses (Cathelat, 1993 ; Aero, 2006 ; Bell and Hollows, 2006 ; Thorkild, 2006 ; Coppens and Oosterlynck, 2008 ; Heijs et al., 2009). In the last decades, it has gained importance in the relevant literature within the fields of Marketing, Social Psychology and Sociology (Holt, 1995 ; Holt, 1997 ; Kahle and Chiagouris, 1997 ; Crompton, 1998). In these fields, researchers use the concept of lifestyle to explore consumer diversity in relation to consumption (i.e. marketing), societal aspects (i.e. sociology) or individual value patterns and personality traits (i.e. social psychology). The meaning of the term is rather ambiguous, as there are a range of definitions, depending on the academic field. Lifestyles are indefinite with regard to the elements they cover, varying from behaviour (e.g. choice, acquisition, use, consumption), to behavioural domains (e.g. dwelling, work, transport) and factors that influence behaviour (e.g. intentions, preferences, values or structural variables) (Heijs et al., 2009). In this way, the lifestyle approach serves as an additional tool for traditional socio-demographic differentiation, because it includes aspects such as subjective patterns of values, intentions and preferences. Diverse societal changes, like individualization and cultural emancipation, are the agents behind the current success of the lifestyle concept (Bootsma et al., 1993 ; Crompton, 1998 ; Kipnis, 2004).
6In planning, the concept of lifestyle has only recently been used to describe residents’ diversity. In the past, planners often tried to characterize or understand residential migrations in relation to residents’ demographic or socio-economic profile (e.g. families with children, young families, the elderly, mid-income families or singles). This knowledge was used when designing specific houses for particular target groups (family houses, houses for the elderly, social houses or houses for starters). Recently, however, planners have devoted attention to the concept of lifestyle in different research contexts. Neighbourhood researchers now study the association between neighbourhood characteristics and individual outcomes (Friedrichs et al., 2003). Others use the concept of lifestyles to reveal new insights in the “driving factor” for the residential choices. “Residential self-selection”, the spatial division of lifestyle groups in different neighbourhoods as a result of the prior self-selection of residents into a built environment that is consistent with their predispositions, is considered to be one of the most important spatial impacts of lifestyles. In literature several quantitative researches dealing with this topic are found (Aero, 2006 ; Feijten et al., 2008 ; Mokhtarian and Cao, 2008 ; Van Acker, 2010). The conclusions of such researches must be approached critically in any case. The general pronouncement that lifestyles can (partly) explain residential choices must always be interpreted in consistency with the used definition of the lifestyle term and the elements they cover (Heijs et al., 2009).
7In the context of this article, the differences between urban and suburban residents, or rather urban and suburban lifestyles, warrant special attention.
- 1 VINEX refers to the “Vierde Nota Extra” (Supplement to the Fourth Report on Spatial Planning), a po (...)
8Several researchers have described or investigated the differences between urban and suburban lifestyles from a qualitative perspective. As early as 1938, Wirth already drew attention to urban lifestyles (LeGates and Stout, 1996). Walter Benjamin (1892-1940) taught us about the “etui-man” (Reijndorp, 2001) who fled the confusing and threatening urban life and resided in suburbia, while A. C. Spectorsky developed the term “urbanites” (Den Hartog, 2006) in 1955 to refer to people living outside the city that are still dependent on the city for work, relations or groceries. Bell and Gans focused on the urban-suburban divide in the USA in the 1960s and 1970s and found that suburbanization essentially constituted a quest for increased status and for like-minded neighbours (Aero, 2006). Within a more recent Northern European context, qualitative researchers like Reijndorp (Reijndorp et al., 1998 ; Reijndorp, 2004) have described the urban and suburban lifestyles of residents of Dutch city centres and VINEX1 neighbourhoods. According to Reijndorp, people who leave the cities and move to new housing developments on the outskirts have diverse motives. Amongst other things, they want to break away from older, busier and noisier urban districts, as well as immigrant neighbourhoods. He distinguishes six suburban lifestyles, referring to the economic and socio-cultural capital of the suburban residents, their housing history and their functional, social and emotional relationship with their residential environment. In Flanders an emerging suburban discourse focusing on safe and accessible cities for the suburban user was discovered within the Antwerp urban region (De Maesschalck, 2010). In his publication “La Distinction” in 1979, the French sociologist Bourdieu presented a quantitative analysis of different kinds of lifestyle-related topics, concluding that people differ from each other because they have different resources, particularly economic and cultural capital (Bourdieu, 1979). Bourdieu established an explicit link between lifestyles and residential choices in France in the seventies, showing that (most of the) French cultural elite was living in the city centres, while the economic elite preferred the more rural areas. In the Netherlands, a similar study by de Wijs-Mulkens (De Wijs-Mulkens, 1999) confirmed this thesis in the context of the nineties in and around Amsterdam and The Hague. Recently, statistical analyses of data from a survey on household activities and travel behaviour in Portland, Oregon (Walker and Li, 2007) have revealed three latent lifestyle segments : urban dwellers, suburban dwellers and transit riders (within a suburban setting). The authors found that socio-economic variables related to lifecycle had significant explanatory power, although the observable explanatory variables were still unable to explain several aspects of lifestyle preferences.
9It is widely accepted, amongst both qualitative and quantitative researchers, that urban and suburban lifestyles differ from each other. Although these researchers all use the concept of lifestyle in different perspectives – varying from lifestyles as result of behaviour, to lifestyles defined by intentions, preferences or values – there seems to be a growing consensus on the themes that are covered. Themes like economic and cultural status (Bourdieu, Bell and Gans, de Wijs-Mulkens), security (Benjamin, Reijndorp), and diversity (Wirth, Bell and Gans) are all used to describe and identify different lifestyles, in particular urban and suburban ones. These themes are equally important for Flemish inhabitants, as has been confirmed by recent research projects on the factors driving demand preferences for different housing environments in Flanders (Devogelaer, 2002 ; De Corte et al., 2003 ; Verhetsel et al., 2003 ; Devogelaer, 2004a ; Devogelaer, 2004b ; Pelfrene, 2004).
10In line with these insights, this study focuses on three dimensions : an economic, a cultural and a safety dimension to investigate lifestyles and housing preferences. However, one extra dimension was added, namely ecology, because of the current societal debates on ecology, sustainability, and sustainable building and planning in particular. Ecology is not often considered a study object in research on urban or suburban lifestyles, but we believe that this could change dramatically in future debates on housing preferences and lifestyles in particular.
11The following research questions will be addressed. Do the residents of the urban and suburban areas have different lifestyles ? Are like-minded Belgians tending to cluster together, just like Americans do ? Which are driving factors for urban and suburban residential choices ?
12In order to examine lifestyles, qualitative methods are often used. In this study, however, the findings are elaborated by means of quantitative methods, which entails some methodological challenges. Since no data on lifestyle aspects are available within Flanders, a structured questionnaire first needed to be developed on different aspects of lifestyle related to residential choice.
13The questionnaire was designed to provide information about the respondents’ current residential environment (revealed preference), their ideal residential environment (stated preference) and their lifestyles, supplemented with questions about their socio-economic and demographic background. The questions about lifestyles included involved the inhabitants’ behaviour, values and personality, with regard to the economic, the cultural, the safety and the ecological dimension.
- 2 The random walk method is appropriate in a situation where although the population is well defined, (...)
14This survey was conducted amongst residents over 18 years old in four case areas in Belgium, in the beginning of 2009. Respondents were selected by means of the random walk method2 (De Pelsmacker and Van Kenhove, 2002).
Table 1. Dimensions of the lifestyle concept and themes of the questionnaire.
Figure 1. Map of the four case areas.
- 3 The case areas were selected based on a cluster analysis of various Flemish databases at the level (...)
15The four residential neighbourhoods were carefully selected to cover a wide range of different areas3. The first urban neighbourhood is situated in the historical city centre of Ghent and is characterized by ethnic diversity, a lot of singles or smaller families and students who often rent their houses. The second urban neighbourhood is situated in the periphery of the historical city centre, although it does have some urban characteristics because of the mixed land use, and the variety of housing types and residents. The second urban neighbourhood is in fact a former village centre, which has been integrated in the urban morphological complex of the city. The suburban areas are both located in the urban fringe of Ghent. Suburban area 1 is located adjacent to urban area 2. Recently new houses were realised within this suburban area. Suburban area 2 is an older villa park, mainly realised in the period between 1945 and 1980. The suburban areas are both characterized by no other land uses besides the residential use and the dominance of homeowners and detached houses.
- 4 K-means cluster analysis or factor analysis can be applied to a wide range of data to define a limi (...)
16The lifestyle data were analyzed with different statistical methods. In studies aimed at quantifying lifestyles and their influence on behaviour, it is common to use a two-stage approach, in which the extraction of lifestyle is separated from the subsequent correlation of these lifestyle factors with residential location4 (Walker and Li, 2007). In this case, the collected data were first analyzed by means of factor analysis (with Varimax rotation) to reduce the different variables to a limited number of factors per dimension. Second, an Anova analysis was conducted to explore any meaningful differences between the scores in the four areas. The mean scores and standard deviations of the factors were analyzed to explore the values and spread of the data. Conclusions on trends related to the ideal housing situation were based on a descriptive analysis of the survey data.
Table 2. Characteristics of the four residential neighbourhoods.
- 5 Since no official demographic data on the residents living within the four neighbourhoods were ava (...)
17A total of 138 questionnaires (28-30-40-40 inhabitants per area) were collected within two urban and two suburban neighbourhoods in the Ghent Region, a city in Belgium. The sample was a result of a random walk through the neighbourhoods. Using this technique we can assume that the sample is a representative selection of the inhabitants5. Nevertheless, some results are striking or unexpected, and need further attention. The mean value of the age of the respondents in both the suburban areas and the urban 2 area is high (± 50 years old), but we were able to contact in each area younger as well as older people. The respondents in the urban area 1 are remarkably younger in average (32 years old). This is definitely related to the student population living within this neighbourhood during week days. These students are well represented in the sample since 15 out of 28 respondents within urban area 1, are students. The average age of the respondents is also reflected in the job participation of the respondents, for example 40 % of the respondents living in suburban area 2 are retired. This high percentage was not unexpected because this suburban neighbourhood was mainly realised in the sixties and seventies and a lot of the residents have “grown old” together with their homes.
18The interviews with a selection of inhabitants provided new data on the inhabitants’ behaviour, values and personality with regard to the residents’ lifestyles and the economic, cultural, ecological and safety dimensions.
19Further analysis focused on the differences between urban and suburban residents’ lifestyles.
20The factor analysis of the safety dimension revealed three factors (variance explained 65 %) : factor 1= unsafe feelings and expressions, factor 2= value orientation, factor 3= specific actions to secure private properties. The ANOVA analysis with the security factors as dependent variables and the case areas as independent variables affirms that factors 1 and 3 are significantly different among the four neighbourhoods. It is shown that the case areas are moderate related to feelings and expressions of safety (0,06 < η2 < 0,14), but are strongly related to the actions undertaken to secure the home against theft or burglary (η2 > 0,14).
Table 3. Socio-economic characteristics of the respondents.
21Inhabitants of urban neighbourhood 1, situated within the city centre, feel less safe in their neighbourhood and adjusted their individual behaviour in the public space by avoiding dangerous places. In contrast, suburban residents declare that they do not feel unsafe within their particular neighbourhood, but many of the respondents in suburban area 2 have nevertheless protected their private property by installing fences or anti burglary systems. The reasons for this suburban behaviour are not entirely clear since they are not related with feelings of unsafety. A probable explanation for this conduct is that the houses and gardens are simply more exclusive in these suburban areas and that the residents have more financial resources to pay for these security systems.
22The factor analysis of the economic dimension has led to three factors (variance explained 75 %) : factor 1 = education, job type and job participation as indicators for income, factor 2 = status behaviour, factor 3 = value orientation. For the factor related to status behaviour, the differences are significant but expected, with the highest score in suburban 2 and the lowest score in urban 1, the city centre. The differences between the value orientations of the respondents in the four case areas are rather limited, but still interesting for further analysis. The respondents of the suburban area 2 regard economic status as an important value, but in the city centre, people also score high on this item. These findings contradict the conclusions of Bourdieu and de Wijs-Mulkens, who stated that especially inhabitants of rural and suburban areas are more driven by economic capital. Unfortunately, the data at hand do not single out the reasons for this discrepancy. Since half of the central urban respondents are students, who are only staying in the city temporarily (during the week and study time), the scores of the urban residents are “coloured” since they might be urban (during weekdays) and suburban (during weekends).
Table 4. Results of the factor and ANOVA analysis.
23The ANOVA analysis, with the ecology factors as dependent variables and the four case areas as independent variables, reveals that only factors referring to behavioural aspects are meaningfully different for the case areas. With regard to global aspects of ecological behaviour (except the use of public transport) the most striking high score is that of the respondents in the second urban neighbourhood, the former village centre at the fringe. The respondents living in the city centre (urban area 1) use the public transport more often.
24Contrary to prior expectations, no meaningful differences were found for the cultural dimension. This data analysis could not confirm differences between the personalities of the urban and suburban respondents. Neither were differences confirmed between the social contacts or the use of daily services within their neighbourhood. In all four case areas about 25 % of the respondents claims to have most of the social contacts within their neighbourhoods, while about 40 % of the urban and the suburban residents makes use of local daily services. These findings seem to contradict with the idea that within the suburban areas some people are still dependent on the city for their relations or groceries, as Spectorsky has developed. However we must be careful with this conclusion since we do not have information on the locations of the contact places or daily services outside the neighbourhoods, they might be urban as well as suburban.
25A first analysis confirms the dissimilarities between urban and suburban lifestyle characteristics, but not for all themes. Especially for the economic and safety dimensions, meaningful differences were found. As expected, these differences were most clear for the urban area 1 in comparison with the suburban area 2. The findings suggest that the residents of the urban and suburban neighbourhoods have different lifestyles, which results in a social-spatial inequality and a polarization between the urban centre and the suburban fringe. This divergence was found to be more obvious for the residents’ behaviour and less for the factors behind their behaviour, such as value patterns and attitudes.
26This part of the paper focuses on the triggers for residential choices and migrations between types of residential environment in the future. In this perspective we collected data on the ideal housing of the respondents, asking several questions about their ideal house and housing environment. These findings are related to the lifestyles of the respondents and to the current residential situation of the residents, in order to explore the tensions between the current and the ideal housing.
27Recent research in the Netherlands (VROM-raad, 2009) has found that one of the dominant social cultural trends will be to live with peers, defined as people with a similar lifestyle, i.e. identical values and activity patterns. According to this research, living with like-minded people makes residents feel like they can control their environment, which gives them a feeling of safety and security. Yet, this trend is not supported by the quantitative inquiry within the four case areas in Flanders. When people were asked if residents in their ideal housing environment should be peers (in different aspects, such as age, ethnical background, household, income, or lifestyle), only a minority found this important. We can therefore assume that living with peers is not a driver for migration within Flanders.
Table 5. Results of analysis of the ideal house and housing environment.
28The aim for a more secure residential environment, however, is an important factor, since 94 % of the respondents consider traffic and social safety to be vital. This attitude definitely speaks in favour of the suburban areas, since respondents claimed to feel more secure there and people in general tend to perceive urban areas as centres of criminality (De Corte et al., 2003). The majority of the residents preferred to live in purely residential areas without any non-residential activities (79 %). However, we can perceive (expected) differences between 50 % of the respondents of the central urban area, who can appreciate the mixed land use, and the majority of the residents of the other areas, who preferred a more “quiet” residential neighbourhood. For 78 % of the respondents, detached single-family houses were the ideal housing type (yet, only 37 % of these respondents live in a detached house today). There is even a stronger consensus on the possession of a private garden. This type of dwelling is traditionally linked with affluence and status and with an upper-class way of living. These high scores confirm Bourdieu’s concept of hierarchised space, in which social distance is prevalent and the upper-class lifestyle is attractive. Within a Northern European context, these characteristics of the ideal housing type and environment are generally more associated with suburban or rural areas, and less with the central urban areas, as the latter are often characterized by a mix of activities, by criminality and by terraced houses or apartments.
29Focussing on differences between the current and the ideal housing situation, different patterns between the urban and suburban residents appear. The suburban respondents are generally more satisfied with their current house and housing environment, while only a minority of the urban respondents favours their own house, housing environment, and their urban community. More over, about 50 % of the urban respondents want to live in a more rural community.
30Primarily this paper aimed to assess the wider theoretical and academic discussion of lifestyles in relation to housing, migration and suburbanisation. While drawing on the theoretical work of several authors, we have developed a conceptualization of lifestyles which encompasses values as well as activity patterns and focuses on four dimensions that are relevant for housing preferences in Flanders, i.e. an economic, cultural, safety and ecological dimension. After all, especially the value aspect of the lifestyles was not very useful to make distinctions between the residents of the different case areas. The activity patterns of the respondents were found more interesting to analyse since some differences were found between the groups of respondents. However, these divergent urban and suburban activity patterns can’t be related in a one-way causal link with migration, since this research did not focus on causality. Probably the respondents carefully chose their urban or suburban neighbourhood in accordance with their activity pattern, or they have adapted their activities after moving to their current housing environment.
31Furthermore, the aim of the empirical part of the paper was to explore the relationship between lifestyles and residential choices for residents in two urban and two suburban neighbourhoods in Ghent and to focus on drivers for migration. Of course, this research is explorative and any conclusions need to be drawn with a lot of caution for numerous reasons. These conclusions are based on a case study (4 areas) and with a limited number of respondents (138). It would be interesting to repeat this research in other neighbourhoods, namely rural neighbourhoods or suburban areas situated at larger distance from the city centre. Moreover, differences between the residents were analysed in relation to the concept of lifestyle, but complementary research in relation to impact of age or impact of phase in the life course would be useful to explore in greater depth the impact of lifestyles versus the other aspects mentioned. Lastly, the results of the central urban area (28 respondents of which 15 students) need further research, to confirm possible differences between the residents who live in the city on a permanent basis and the students. Nevertheless, by analyzing the lifestyles and residential preferences of residents living in urban and suburban neighbourhoods in Ghent, we were able to draw a number of conclusions on lifestyles, segregation and suburbanization.
32Firstly, especially for the dimensions of security and economic status, the dissimilarities between urban and suburban lifestyles were confirmed. Urban and suburban residents act different in relation to security and economic status. Moreover we can confirm security as a driver for future suburban migrations since the aim for a more secure residential environment is shared amongst all residents and the suburban areas are generally perceived as more secure areas. Secondly, despite the societal discussions on diversity, polarization, segregation, etc., and contrary to expectations we were unable to draw any meaningful conclusions on the cultural dimension of the respondents’ lifestyles. Surprisingly, according to this survey, living with peers is not a driver for migration within Flanders. Thirdly, the data clearly illustrate the differences between the suburban residents, who are generally satisfied with their current house and housing environment, and the urban residents of whom only a minority favours their own house, housing environment, and their urban community. Furthermore, this research confirms the widespread ideal of the detached single-family house with private garden situated within a purely residential area.
33Our analysis does offer some insights in the potentials of lifestyle analysis in relation to residential choices. However, the relation between housing choices and lifestyle is limited, since we could only confirm differences between urban and suburban activity patterns. Especially the data on residential preferences and residential satisfaction were found to be interesting, since they allowed us to make some assumptions on urban or suburban drivers for migration. Referring to these data, suburbanisation remains one of the major common challenges facing Flanders, Belgium. As long as the demand for separate single-family homes with private gardens remains high, and suburbs provide the particular quiet and safe living environment which the inhabitants desired, land on the fringe of urbanised areas will be converted from open space to residential land use. This survey within a Flemish, Belgian context confirms that residents of urban and suburban zones have divergent activity patterns, but they agree on residential preferences and value patterns. This residential preferences might cause suburbanisation but do not need to lead to segregated social communities, since living with peers, does not seem to be a driver for migration in Flanders, Belgium.