Residential mobility and neighbourhood attachment in Guangzhou,China

Abstract

Wholesale redevelopment, suburbanization and increased population mobility in recent decades have brought significant social and spatial changes to urban neighbourhoods in Chinese cities, not least the subjective feelings of residents about their neighbourhoods. While there is a substantial literature on urban restructuring and migration at different geographical scales, relatively little is known about how feelings such as neighbourhood attachment are conditioned upon residential mobility and neighbourhood change in Chinese cities. To address this deficiency in the literature, multi-level models are employed to explore the extent to which residential mobility affects three different dimensions of neighbourhood attachment based on a large-scale household survey conducted in Guangzhou in 2012. The findings show that mobility experience and neighbourhood-related factors exert discernible influences on the attitudes towards the neighbourhood. Specifically, while people staying in reform/work-unit housing compounds tend to have better knowledge of their neighbours, those moving from reform/work-unit compounds to commodity housing estates exhibits greater involvements in the affairs of the new neighbourhood. The built environment, population size and frequency of population turnovers of the neighbourhood underpin residents’ attachment to it.

Keywords

Residential mobility neighbourhood attachment neighbourhood involvement neighbour acquaintances Guangzhou

Introduction

Residential mobility or intra-urban migration is a major factor underlying changes in urban spatial structure. It can be seen as a spatial adjustment process on the part of an individual and his/her family to accommodate different housing and location needs in different stages of the family life cycle or personal life course (Brown and Moore, 1970; Clark, 2013). Alternatively, it can be seen as an attempt by the household to respond to changes in the neighbourhood environment, as in the case of ‘white flight’ from the central city in the United States (Wright et al., 2015), or to major restructuring of the urban economy underscoring deepening economic globalization and de-industrialization. In comparison with studies on the decision to move and the resultant spatial sorting of population in an urban area, relatively few studies have delved into the impacts of the residential relocation process on the subjective feelings of individuals and households towards their neighbourhoods. Yet, the latter has become a critical issue within the discourse and practice of public policy (Thomas et al., 2016).

Empirical research has demonstrated that an individual’s socioeconomic status, life cycle stage, duration of residence in the dwelling and/or neighbourhood and neighbourhood environment contribute to his/her attachment to the neighbourhood (Hummon, 1992; Kasarda and Janowitz, 1974; Theodori and Luloff, 2000). It is clear, however, that these variables explain only part of the variations in neighbourhood attachment. One relatively unexplored set of variables concerns residential mobility. As Sharp (2013) reveals, for newcomers in a neighbourhood, where they come from and the reasons for moving to the current residence likely affect how well they adjust to the new environment, and the extent to which they develop sentimental attachment to it. The spatial features of the move also matter. In this light, Bolan (1997) demonstrates that households who move shorter distances are significantly more attached to the destination neighbourhoods. Surely, neighbourhood attachment is better understood by taking account of residential mobility, especially in a society undergoing dramatic economic, social and spatial changes.

China has been experiencing unprecedented urbanization since the mid-1990s. Hundreds of millions of peasant-migrants have flocked to the cities, and the abundant and cheap labour they provide underlies much of the country’s phenomenal economic and urban growth. Yet most rural migrants are precluded from transferring their place of household registration or hukou from the home town or village to the city of current residence and, as such, are confronted with immense discrimination in the labour and housing markets (Chan, 2014). Inside the city, new neighbourhoods in the form of large housing estates mushroom in the suburbs, while high-rise condominium apartments are rapidly replacing old neighbourhoods that pre-dated the founding of the People’s Republic in 1949 and work-unit compounds that used to characterize the urban development scene in the pre-reform and early reform times. Concomitantly, the remaining old neighbourhoods and work-unit compounds are undergoing rapid transformation in respect to building usage, housing tenure mix and population composition (Li et al., 2012). All these changes are likely to impinge heavily on social relationships at the neighbourhood level. Despite the scale and pace of change, and despite the efforts of the state to foster community building by enhancing the functions of residents’ committees and other grassroots state organizations (Bray, 2006; Fu et al., 2015; He, 2015), little has been written on how urban growth and transformation and the associated population reshuffling impinge on the daily routines of individuals and families in the neighbourhood, neighbourly relations and attachment to the place of residence.

Forrest and Yip (2007), among the few exceptions, argue that social ties at the neighbourhood level are being eroded as the Chinese economy and society become more commodified. Nonetheless, Zhu et al. (2012) demonstrate that recent in-movers in newly built commodity housing estates exhibit strong neighbourhood attachment. Likewise, Li and Li (2014) document the extensive use of online–offline neighbourhood forums by young professionals in Guangzhou to foster community spirit, who moved into newly built suburban housing estates in search of a better living environment. Not surprisingly, studies on villages-in-the-city, that is migrant enclaves on former urban–rural fringes constructed by local villagers to generate rental incomes, find that migrants’ attachment to the neighbourhood is weak. The migrant-tenants see themselves as perennial outsiders, even though many have lived in the enclave for years (Du and Li, 2010; Du et al., 2017; Wu and Logan, 2016).

The present study builds upon the above-cited works and explores further the extent to which residential mobility affects neighbourhood attachment, using data from a large-scale survey conducted in Guangzhou in 2012. The survey provides a rich set of information on neighbourhood social networks and on residents’ participation in neighbourhood organizations such as residents’ committees and homeowners’ associations. The rest of the paper is organized as follows. The next section discusses in greater detail extant works on neighbourhood attachment in general and in Chinese cities in particular. Following that, we describe the dataset and outline the modelling strategy before reporting the findings. The final section concludes.

Literature review

Perspectives on neighbourhood attachment

The terms sense of place and place attachment refer to bonds with a place developed through daily spatial experiences. The former can be defined as the collection of meanings, beliefs, symbols, values and feelings that human beings associate with a place (Paradis, 2000; Williams et al., 2010), while the latter involves a set of feelings towards the place (Bailey et al., 2012; Lewicka, 2011). Often these two terms are used interchangeably (Patterson and Williams, 2005). Attachment to neighbourhood constitutes a significant aspect of a person’s place attachment that underscores the nature of social life in an urban area (Hummon, 1992; Kasarda and Janowitz, 1974; Matarrita-Cascante and Luloff, 2010). In addition to attitudinal indices including measures of sentiment towards the neighbourhood, behavioural dimensions including neighbourly relations and involvement in neighbourhood affairs are often invoked when measuring neighbourhood attachment (Bolan, 1997; Liu et al., 1998; Stinner et al., 1990).

Regarding the determinants of neighbourhood attachment, two major perspectives can be identified. First, the linear model posits that neighbourhood attachment is inversely related to neighbourhood size and population density. Second, the systemic model examines the relationships between neighbourhood attachment and three sets of personal factors, namely length of residence in the neighbourhood, socioeconomic position and life cycle stage. In particular, it is postulated that more time spent in the neighbourhood, higher socioeconomic status and increase in age will generate stronger emotional sentiments and a sense of belonging towards it (Theodori and Luloff, 2000). It has been argued that personal factors are better predictors of community sentiment than ecological or community characteristics such as size and density, because sentiments are products of personal choice (Goudy, 1990; Kasarda and Janowitz, 1974). Kasarda and Janowitz (1974) reveal that length of residence has the most powerful and consistent effects on local bonds. Moreover, past research has demonstrated that factors such as homeownership and presence of children in the family influence an individual’s feelings towards the neighbourhood as well as his/her involvements in it (Stinner et al., 1990). Recent works have incorporated interactions between personal attributes, on the one hand, and the built environment and communal space, on the other, into multivariate models of attachment (Brehm et al., 2004, 2006; Zhu and Fu, 2017).

Variables such as length of residence and neighbourhood demographics to a significant extent are also measures of population turnover and residential change. Research on the long-term effects of residential moves on an individual’s health and well-being indicates that a move can create persistent changes in the individual’s life situation (Stokols and Shumaker, 1982; Tucker et al., 1998). Also, Bolan (1997) finds that mobility experiences are closely associated with attachment to the new residence. Clark and Lisowski (2017) reveal that people who do not intend to move home experience higher levels of satisfaction with housing and community. Migration history, motivation for moving and the distance of the move are found to impact on how well an individual adapts to the new environment, and hence shape his or her subsequent attitudes towards the destination neighbourhood. Sharp (2013) argues that, for people who have resided in a neighbourhood for a long time, attachment to it would persist along multiple dimensions, irrespective of profound changes in personal life over time. Moreover, neighbourhood composition and dynamics have an important role in the development of social capital and neighbourhood attachment (Thomas et al., 2016). On one hand, neighbourhood attachment is significantly lower in more deprived neighbourhoods, primarily because of weaker social cohesion (Bailey et al., 2012). For ethnic minorities, strong neighbourhood belonging is associated with high co-ethnic density (Finney and Jivraj, 2013). On the other hand, high levels of population turnover or immigration undermine community cohesion and are associated with lower levels of neighbourhood belonging (Finney and Jivraj, 2013).

Neighbourhood attachment and residential mobility flows in Chinese cities

Studies on processes of transformation in urban neighbourhoods are of major significance in Chinese cities, which are undergoing unprecedented growth and restructuring (Forrest and Yip, 2007). The 40 years of market-oriented reform, particularly the ending of the welfare allocation of housing in 1998, has fundamentally changed the housing provision scene (Li and Yi, 2007). In the city core, wholesale redevelopments of urban neighbourhoods have given rise to landmark monuments, glossy office towers and shopping arcades, as well as luxurious condominiums. In the industrial belt of the socialist era, state enterprises, enticed by rising land values, are eager to relocate productive activities to the development zones on the city outskirts and join forces with such property development giants as Wanke and Country Garden to redevelop their former compounds for commodity housing projects. In the ever-expanding suburbs, enormous housing estates consisting of tens of thousands of multi-storey apartments, townhouses and detached dwellings proliferate, catering to the country’s rising middle class and nouveaux riches (Giroir, 2014; Li and Li, 2014). Often located side-by-side with new commodity housing developments in the inner and outer suburbs are urbanized villages where hundreds of thousands of migrant workers and other socially deprived people congregate. Characterizing the contemporary urban sociospatial fabrics are increasingly variegated built environments, co-presence of marketized and socialist party-state governance at the community or neighbourhood level, complex housing tenure mixes and juxtapositions of population groups of diversified socioeconomic and migratory status. Under such circumstances, neighbourhood life in Chinese cities has become all the more complex.

Forrest and Yip (2007) find that accompanying the transformation of work-unit compounds to commodity housing estates is a general decline in social interactions. Xu (2010) provides evidence in support of the ‘community lost’ thesis under urbanization in post-reform China, but Li and Li’s (2014) study of in-movers to commodity housing estates in Panyu in suburban Guangzhou suggests otherwise. Li et al. (2012), based on a survey conducted in Guangzhou, find that the patterns of neighbouring vary across different types of neighbourhoods; furthermore, despite the presence of higher neighbourhood attachment and satisfaction, commodity housing estates generally manifest weak local social networks. Zhu et al. (2012) similarly report strong neighbourhood attachment but weak neighbourly interactions in commodity housing estates in Guangzhou.

Variously, estate management firms and homeowners’ associations have assumed increasing pre-eminence in managing neighbourhood affairs such as landscaping, traffic control, security, recreation facilities provision and organization of social activities, thus posing severe challenges to established grassroots state organizations such as residents’ committees and street offices in neighbourhood governance (Fu, 2014; He, 2015; Sun and Yip, 2014). In general, despite diminishing neighbourly relations (Forrest and Yip, 2007), neighbourhood attachment continues to exist but perhaps in different forms; more specifically, attachment to the physical or built environment of the neighbourhood seems to have strengthened upon moving to gated commodity housing communities.

Underscoring the immense changes in neighbourhood demographics and neighbourly relations are heightened population mobility at various geographical scales: rural–urban, inter-regional, and intra-urban (Chan, 2014; Li and Siu, 2001). Compared with the immense literature on rural–urban and inter-regional migration, works on intra-urban migration or residential mobility in Chinese cities are much scarcer. Earlier studies adopting the life cycle or life course approach reveal increases in residential mobility in early reform times (Li, 2004), but more recent research points towards greater residential stability, reflecting China’s sudden conversion to a nation of homeowners with the implementation of the 1998 housing reform, which called for the cessation of the welfare allocation of housing and promotion of homeownership (Li and Mao, 2016; Li and Zhu, 2014).

Regarding how residential moves are related to neighbourhood affairs, Li and Li (2014) document the extensive use of online–offline neighbourhood forums to foster friendship ties and community activities by young professionals in Guangzhou, many of whom are of non-local origin, who have moved into suburban gated commodity housing estates in search of better residential environments. Understandably, scholarly attention has focused on forced relocation under massive redevelopment programmes initiated by municipal governments. Song et al. (2012) reveal that many who were relocated to remote suburbs from central Shanghai would like to see imminent transformation of their current remote and ‘rural’ places of residence to new urban centres. Also, irrespective of the inaccessibility and cessation of long-established social ties, in general the displaced residents were not too dissatisfied with their new dwellings, which offer much-improved living conditions (Li and Song, 2009).

The above highlights two relatively thin strands of literature on Chinese urban studies: the study of residential location and relocation decisions; and neighbourhood restructuring and diversity in metropolitan development, including the issues of community building and neighbourhood attachment. Obviously, the processes involved in structuring residential distributions and neighbourhood dynamics are closely intertwined. Based on the two strands of the literature reviewed, we attempt to empirically analyse the following: (1) the relationship between personal sociodemographic attributes, including life course variables and neighbourhood attachment; (2) the extent to which residential experiences, particularly residential relocation, contribute to the development and experience of attachment towards the neighbourhood; (3) whether the findings of research conducted in the United States and elsewhere, which shows that neighbourhood instability, large neighbourhood size and low levels of satisfaction of the built environment would curtail neighbourhood attachment, and that socioeconomic composition of the neighbourhood may have an influence on attachment, apply also to the Chinese case in rather different political and economic contexts.

Data and research method

Data

A research team comprising academics from Hong Kong Baptist, Sun Yat-sen and Duke University undertook a large-scale household survey on neighbourhoods with homeowners’ associations in Guangzhou over the period October 2012 to January 2013. The survey is particularly rich in information about neighbourly relations and neighbourhood social networks, participation in community affairs including neighbourhood online forums and offline social activities, homeowners’ associations and residents’ committee, perceptions of the neighbourhood physical and social environment, and objective measures of dwelling and environmental attributes. Also available are aspects of respondents’ residential histories. The present work is based on the data generated from the survey.

To maximize the spatial representation of the sample, a three-stage sampling strategy was adopted. First, 30 streets (jiedaos) or sub-districts were selected from a total of 139 streets within the outer ring road of the city. Each street was assigned to one of the three distance zones: inner core, inner suburb or outer suburb. The boundaries of distance zones were demarcated with reference to the city’s history of urban development, land-use patterns and residential density. Among the 139 streets, 52 belonged to the inner core, 45 to the inner suburbs and 42 to the outer suburbs. In proportion to the population size of each street, 12 streets were chosen from the inner core, 11 from the inner suburbs, and 7 from the outer suburbs. Second, in each street, one neighbourhood was randomly chosen, using the list of neighbourhoods with functioning homeowners’ associations provided by the South China Centre for Harmonious Community Development, a non-government organization seeking to promote the interests of homeowners’ associations. An additional neighbourhood would be added to the street if the interviewers failed to get access to one-third of the target households in the original neighbourhood due to such difficulties as refusal of interview after several attempts by the interviewers. The resulting sample comprised 17 neighbourhoods from the inner core, 14 from the inner suburbs and 8 from the outer suburbs. Figure 1 plots the location of the surveyed neighbourhoods. Third, within a chosen neighbourhood, a number of households or residents were recruited using an interval sampling strategy based on home address and neighbourhood population size. The neighbour next door would be approached if a chosen resident refused to participate in the survey. A total of 1804 valid questionnaires from 39 neighbourhoods was achieved. After excluding cases with missing values, the subsequent analyses are performed on the remaining 1694 respondents.

Figure 1.

Distribution of surveyed neighbourhoods.

Formation of homeowners’ associations under the supervision of local residents’ committees and street offices has been encouraged by the Guangzhou Municipal Government (Guangzhou Land Resource and Housing Management Bureau, 2009). The sample covers neighbourhoods in various settings: commodity housing estates, work-unit and reform housing (former work-unit housing) compounds, public housing and relocation housing estates, and inner-city neighbourhoods pre-dating the socialist era. It is unlikely that homeowners’ associations are found in villages-in-the-city or self-built housing communities where migrants from rural areas without the local Guangzhou hukou concentrate. Because of this, the survey is likely to under-sample rural migrants. We acknowledge that the absence of migrant enclaves and under-sampling of non-hukou rural migrants is a limitation of the present study. However, the different types of neighbourhood under examination in a large part constitute the city’s formal housing market and house the majority of its population. With growing maturity of the housing market, especially with the development of a thriving rental market, quite a few rural migrants without the local hukou have infiltrated inner-city neighbourhoods and work-unit/reform housing compounds. In fact, some have made their way to commodity housing estates. In our sample, 16% of respondents used to live in villages-in-the-city, self-built settlements or dormitories. For this, non-hukou migrants are not totally unrepresented in the sample. Comparison with data from the 2010 Population Census of China reveals broad agreements in the age and gender distributions between the sample and the population of Guangzhou.¹

Dependent variable: measures of neighbourhood attachment

Neighbourhood attachment is a multi-dimensional concept. Based on the literature survey, three dimensions of neighbourhood attachment can be identified: (1) emotional attachment; (2) neighbourhood involvement; and (3) neighbour acquaintances (Bolan, 1997; Flaherty and Brown, 2010; Stinner et al., 1990). The first dimension is mainly concerned with attitudes towards the neighbourhood, while the latter two are aspects of residents’ neighbourly behaviours. The three dimensions capture different facets of neighbourhood life. Concerning the attitudinal dimension, in the survey five 5-point Likert scales were employed to gauge various aspects of respondents’ feelings about the neighbourhood. Table 1 gives the mean scores of the scales. The majority of respondents (71.66%) feel attached to the neighbourhood. Between 60% and 80% indicate that they like the neighbourhood and feel at home with it. About half would feel sorry if they had to leave the neighbourhood. However, just less than half agree that friendships and personal associations within the neighbourhood mean a lot to them. In order to provide a fuller picture of respondents’ subjective feelings, we make use of principle component analysis (PCA) on the five scales to construct a composite index for the attitudinal dimension.² The result confirms that one single dimension that explains 53% of total variance is adequate to represent the five scales with high reliability (Cronbach’s alpha = 0.71). The component score is used in the subsequent analyses.

Table 1.

Principal components analysis of emotional attachment to the neighbourhood.

Statements	% ‘strongly agree’	Mean	Component
(a) Overall, I am very attached to this neighbourhood	71.66	3.78	0.839
(b) As a living place, I like the neighbourhood	69.85	3.75	0.834
(c) I feel at home to live in this neighbourhood	72.93	3.71	0.376
(d) I feel sorry to leave the neighbourhood	52.78	3.35	0.738
(e) The friendships and associations that I have with other people in this neighbourhood mean a lot to me	48.61	3.35	0.628

Neighbourhood involvement is generally gauged by respondents’ participation in activities associated with community life (Dekker, 2007). In the survey, a total of 10 activities were identified, based on pilot interviews with neighbourhood activists and experts (see Table 2). Overall, 79.59% of the respondents indicate that they have been involved in at least one activity over the past year. Voting for homeowners’ associations has the highest percentage (43.86%), whereas participation in activities organized by the homeowners’ association has the lowest (37.25%). Again, PCA reveals uni-dimensionality of these activities (Cronbach’s alpha = 0.80). The component score is thus employed to represent participation in neighbourhood affairs.³

Table 2.

Principal component analysis of neighbourhood involvement.

Neighbourhood activity	Yes	%	Component
Voting for RC members	564	31.26	0.628
Giving advice to RC	420	23.28	0.693
Voting for HOA members	784	43.46	0.680
Giving advice to HOA	598	33.15	0.737
Attending homeowners’ assembly	440	24.39	0.695
Joining HOA activities	169	9.37	0.600
Attending hearings	214	11.86	0.606
Complaints about neighbourhood problems	388	21.51	0.438
Participation in activities self-organized by homeowners	672	37.25	0.427
Participation in activities organized by PMC	666	36.92	0.463

HOA, homeowners associations; PMC, property management company; RC, residents’ committee.

Finally, the third dimension of neighbourhood attachment, neighbour acquaintances, is measured by the number of adults in the neighbourhood known by name. To contain the effect of extreme values on the regression models, answers with more than 500 acquaintances are set at 500. The resultant mean value is 12.

Bivariate correlations between each pair of dimensions are given in Table 3. The three dimensions are positively correlated, but the correlations are weak (all below 0.22), confirming the presence of noticeable differences between the dimensions (Guest and Lee, 1983).

Table 3.

Correlations among the three dimensions of neighbourhood attachment.

	Emotional attachment	Neighbourhood involvement	Neighbour acquaintances
Emotional attachment	1	0.215	0.215
Neighbourhood involvement		1	0.210
Neighbour acquaintances			1

Independent variables

To incorporate both micro- or individual-level and ecological or neighbourhood variables in a single framework, a two-level regression model is attempted. Level 1 variables include personal attributes and aspects of mobility experience. Level 2 variables pertain to aspects of variations across neighbourhoods. The list of independent variables together with summary statistics (percentages for categorical variables and means for continuous variables) is given in Table 4. The extent to which the effects of given level 1 variables depend on neighbourhood conditions is analysed by estimating the interaction terms.

Table 4.

Descriptive statistics of predictor variables.

	N	%
Gender
Male	745	44.0
Female	949	56.0
Presence of children
Yes	891	52.6
No	803	47.4
Hukou status
Local	1213	71.6
Migrant	481	28.4
Education
Junior secondary or below	422	24.9
Senior secondary	430	25.4
Associate degree	345	20.4
University and above	497	29.3
Occupation
Manual/services worker	211	12.5
Clerical/technical worker	225	13.3
Administration/professional	372	22.0
Others (self-employed, retired, unemployed, etc.)	886	52.3
Homeownership
Owner	1352	79.8
Renter	342	20.2
Self-perceived socioeconomic status
Lower or lower-middle class	979	57.8
Middle class or above	715	42.2
Current housing type
Commodity housing	1395	82.4
Work-unit/reform housing	134	7.9
Others (public housing, relocation housing, etc.)	165	9.7
Previous housing type
Commodity housing	924	54.6
Work-unit/reform housing	378	22.3
Others (public housing, relocation housing, urban village, dormitory, etc.)	392	23.1
Migrant concentration
20% or below	629	37.1
20–50%	896	52.9
50% or above	169	10.0
Neighbourhood location
Inner core	577	34.1
Inner suburb	824	48.6
Outer suburb	293	17.3
	Mean	SD
Age	44.9	15.0
Length of residence	7.3	5.3
Neighbourhood stability	0.6	0.2
Neighbourhood size	1968.6	3015.2
Neighbourhood homeownership rate	0.8	0.1
Neighbourhood evaluation	2.8	0.6

SD, standard deviation.

The personal attributes employed include: (1) age, gender and presence of children aged 12 or younger – these demographic variables are very much tied to a person’s stage in the life course or family life cycle; (2) hukou status (1 = local and 0 = non-local) – this rather unique demarcation of a person’s identity significantly influences his or her position in the labour and housing market as well as the intention to stay in the city of current domicile or to return to the town or village of origin (Du and Li, 2012); and (3) educational attainment, occupational rank and homeownership, which provide objective measures of socioeconomic status. Homeownership is also related to the costs involved in a house move. In addition, we add perceived socioeconomic status (1 = lower and lower-middle class; 2 = middle class or above) to gauge the effects of subjective assessment of one’s status on attachment to the neighbourhood. Other than hukou status, most variables are common to those employed in prior studies (Bolan, 1997; Goudy, 1990; Kasarda and Janowitz, 1974; Sharp, 2013). Among the 1694 respondents, the average age is 45 years; 44.0% are males; 52.6% have at least one child in the household; 71.6% hold the Guangzhou hukou; 49.7% have completed post-secondary or higher education; 54.3% are employed while 28.6% are retired, and among those employed, 40.6% are administrative or professional workers; 79.8% own their homes; and 42.2% consider themselves belonging to the middle or above class.

Mobility experience, the main focus of the present work, is measured by three aspects of residential history: length of residence in the present dwelling (mean = 7.3 years); previous housing type (public housing, relocation housing, urban villages and dormitories, 23.1%; work-unit/reform housing, 22.3%; commodity housing, 54.6%); and current housing type (public housing and relocation housing, 9.7%; work-unit/reform housing, 7.9%; and commodity housing, 82.4%).⁴

Level 2 or neighbourhood variables, which can be seen as the aggregated outcomes of mobility behaviours, include the following:

Neighbourhood stability (inversely, rate of population turnover). It is given by the percentage of residents having lived in the same dwelling for more than five years in the neighbourhood. In the sample, an average of 57% of residents per neighbourhood have lived in their present residence for more than five years.

Neighbourhood size or the total number of households residing in it. Previous works show that larger neighbourhoods dampen neighbourhood attachment, because it is more difficult for residents to know one another in such a setting (Kasarda and Janowitz, 1974). Among the 39 neighbourhoods surveyed, neighbourhood size varies from 112 to 11,322 households, with an average of 1967 households per neighbourhood.

Evaluation of the neighbourhood built environment. Trained surveyors were asked to independently assess the overall environment of each neighbourhood on a five-point scale from 1 ‘low quality’ to 5 ‘high quality’. The number of surveyors for each neighbourhood ranges from 6 to 33, with an average of 14. The variable is defined as the mean of surveyors’ ratings for each neighbourhood.

Neighbourhood demographic composition. This is given by the relative share of homeowners (mean = 79.4%), and also that of migrants without the Guangzhou hukou in the neighbourhood (mean = 28.3%).

Location of the present neighbourhood in the distance zone, the spatial distribution of which was spelled out in the sampling frame outlined above.

Dimensions of neighbourhood attachment

The result of the two-level regression analysis (random intercept models) is given in Table 5. Included in the parentheses are the standard errors of the corresponding regression coefficient estimates. Collinearity tests do not show significant multicollinearity. Analysis of cross-level interactions indicates that adding interaction terms only significantly improve model fit for Model 2 – that is, the model on neighbourhood involvement – but not for the other two models. Below we first present the fixed-effect results for both level 1 variables, which include personal attributes and residential mobility experiences, and for level 2 or neighbourhood variables. We then discuss the significant cross-level interaction effects.

Table 5.

Impacts of personal and neighbourhood attributes on neighbourhood attachment.

	Emotional attachment		Neighbourhood involvement		Neighbour acquaintances
Fixed effects
Constant	–1.298**	(0.418)	–1.002	(0.581)	–19.25*	(9.689)
Age	–0.009	(0.010)	0.010	(0.010)	0.927***	(0.240)
Age/10 squared	0.014	(0.010)	–0.022*	(0.010)	–0.909***	(0.243)
Gender (1 = male)	–0.011	(0.049)	0.015	(0.145)	–0.331	(1.164)
Presence of children (1 = yes)	0.111*	(0.049)	–0.046	(0.046)	1.310	(1.164)
Hukou status (1 = local)	–0.099	(0.060)	0.389*	(0.165)	0.839	(1.440)
Education (reference = junior secondary or below)
Senior secondary	–0.203**	(0.068)	0.104	(0.065)	–0.773	(1.622)
Associate degree	–0.226**	(0.079)	0.124	(0.075)	–3.379	(1.877)
University and above	–0.198*	(0.082)	0.020	(0.077)	–5.673**	(1.940)
Job rank (reference = manual/services worker)
Clerical/technical worker	–0.145	(0.094)	–0.031	(0.089)	–4.385*	(2.227)
Administration/professional	–0.157	(0.087)	–0.014	(0.082)	–1.086	(2.064)
Others (self-employed, retired, unemployed, etc.)	–0.083	(0.081)	–0.110	(0.076)	–3.041	(1.921)
Homeownership (1 = owner)	–0.102	(0.074)	0.290***	(0.070)	–0.979	(1.757)
Self-perceived socioeconomic status (1 = Middle class or above)	0.134**	(0.050)	–0.187	(0.149)	1.881	(1.194)
Length of residence	0.031***	(0.005)	0.034***	(0.005)	0.604***	(0.128)
Current housing type (reference = commodity housing)
Work-unit/reform housing	0.088	(0.097)	–0.082	(0.092)	4.354	(2.290)
Others (public housing, relocation housing, etc.)	–0.041	(0.089)	0.055	(0.084)	–3.981	(2.123)
Previous housing type (reference = commodity housing)
Work-unit/reform housing	0.080	(0.065)	0.150*	(0.061)	0.530	(1.535)
Others (public housing, relocation housing, urban village, dormitory, etc.)	0.052	(0.061)	0.021	(0.058)	0.886	(1.457)
Neighbourhood stability	0.518**	(0.199)	–1.471**	(0.539)	9.552*	(4.582)
Neighbourhood size/1000	–0.005	(0.021)	–0.013	(0.030)	–1.227*	(0.483)
Neighbourhood homeownership rate	0.491	(0.379)	0.341	(0.516)	7.989	(8.684)
Migrant concentration (reference = 20% or below)
20–50%	–0.070	(0.078)	–0.123	(0.108)	1.844	(1.785)
50% or above	0.043	(0.201)	0.099	(0.280)	19.22***	(4.610)
Evaluation of neighbourhood built environment	0.261***	(0.068)	0.083	(0.093)	–0.409	(1.552)
Neighbourhood location (reference= inner suburb)
Inner core	0.038	(0.081)	0.117	(0.114)	–6.573***	(1.854)
Outer suburb	0.361***	(0.099)	0.419**	(0.137)	–1.261	(2.250)
Age × neighbourhood stability			0.025**	(0.008)
Gender × neighbourhood stability			0.165	(0.243)
Hukou × neighbourhood stability			–0.321	(0.286)
Self-perceived socioeconomic status × neighbourhood stability			0.480	(0.249)
Random effects
Level 2 neighbourhood: intercept variance	0.132	(0.039)	0.057	(0.018)	8.031	(5.042)
Level 1 individual: variance (residual)	0.941	(0.016)	0.792	(0.028)	506.895	(17.642)
–2 log likelihood	4600		4459		15,378

*p < 0.05, **p < 0.01, ***p < 0.001

Level 1 variables: personal attributes

The results show that age is not significant in Model 1 (emotional attachment). However, age squared is significant in Model 2 (neighbourhood involvement) and both age and age squared are significant in Model 3 (neighbour acquaintances). Older people would have more time to get involved in neighbourhood affairs and be acquainted with neighbours, but such a relationship is non-linear. We will further explore the age effect when discussing the interaction between level 1 and level 2 variables below.

It is generally believed that females, who traditionally play the role of homemakers and spend more time at home, tend to be more attached to the neighbourhood, particularly in respect to neighbour acquaintances; however, gender is insignificant in all three models, although it carries a negative sign in Model 1 (emotional attachment) and Model 3 (neighbour acquaintances).

Events such as walking and strolling, playing with children in communal space and asking neighbours to help look after children would generate social bonding and deepen the emotional attachment to the neighbourhood and increase neighbour acquaintances. The results suggest that presence of children in the household enhances emotional attachment; however, the variable is not significant in Model 3 (neighbour acquaintances) although the coefficient estimate carries a positive sign.

Unlike migrants in villages-in-the-city who would consider themselves as outsiders even after several years of living in the neighbourhood (Du et al., 2017), migrants without the local hukou in the sample do not show significantly lower levels of neighbourhood sentiment than do local hukou holders. Prior studies suggest that migrants tend to depend on neighbours who are kinsmen and tongxiangs – that is, people from the same home village or province – for information on job and housing opportunities, as well as on material assistance when encountering financial difficulties (Wu and Logan, 2016). Manual and service workers particularly rely on such help. However, such reliance on neighbours is likely to be restricted among kinsmen and tongxiangs. Migrants’ relationships with local people and others are tenuous if not outright precarious. Expectedly, the regression models show that local hukou holders’ participation in community affairs is significantly higher than that of non-hukou migrants.

Regarding the measures on socioeconomic status, educational attainment shows a significant and negative effect on emotional attachment to the neighbourhood as well as the number of adult neighbours known by name. This is understandable. Highly educated people tend to have more extensive social networks and are less tied to the neighbourhood (Guest and Wierzbicki, 1999). Moreover, developments in science and technology and improvements in transport connections, including the increasing prevalence of car ownership, make it easier for the highly educated to maintain social networks as well as to rely on resources available outside the neighbourhood. Probably because of the close connection between educational attainment and occupational rank, with the exception of ‘clerical and technical worker’, none of the occupational rank dummies is significant.

Attainment of homeownership is often considered a major personal accomplishment. The model results show that homeowners tend to participate more in neighbourhood activities, which is consistent with findings in Western cities (Stinner et al., 1990). In the case of China, evolving and still ill-defined property rights under repeated rounds of reform quite naturally underpin contestations between the suddenly emerged class of homeowners, on the one side, and various parties with property development interests, including property developers, estate management firms and local government agencies, on the other, leading to spontaneous formation of homeowners’ associations in Beijing and other cities (Fu et al., 2015; Read, 2003). Active participation by selected homeowners in neighbourhood affairs is expected and confirmed by the regression results. However, the estimated models suggest that being a homeowner does not necessarily heighten emotional attachment to the neighbourhood; nor does it necessarily increase acquaintance with neighbours.

We tried models with and without the inclusion of self-evaluation on class belonging; invariably inclusion of this variable led to significant improvements in the likelihood function Chi-squared estimate. The results show that people who consider themselves of higher social class tend to exhibit stronger emotional attachment to the neighbourhood, echoing Woolever (1992) and Dekker (2007). A plausible explanation is that such individuals are likely to be those who are able to move up the housing and neighbourhood ladder, and hence experience higher neighbourhood satisfaction and form emotional bonding to it.

Level 1 variables: residential experiences

Among the different aspects of residential experience examined, in line with the findings of previous studies (e.g. Kasarda and Janowitz, 1974), length of residence is found to have the strongest effect on neighbourhood attachment: it is significant and positively related to all three dimensions, of emotional attachment, involvement in neighbourhood activities and acquaintance of neighbours. Although it is difficult to infer causal relationships in cross-sectional analysis (Davies and Pickles, 1985), the results strongly suggest that length of residence is an essential ingredient for the development of neighbourhood-based emotional ties and social capital accumulation.

The results for previous and current housing type dummies also appear reasonable. Continuing residence in a work-unit/reform housing compound tends to be associated with more neighbour acquaintances, although the relationship is only of marginal significance. On average, reform and work-unit housing residents know 19 neighbours by name. In comparison, residents of commodity housing know 11, and residents of other types of housing know 10. However, it seems that such acquaintance may not lead to deeper emotional attachment to the neighbourhood.

Moving from a deteriorating reform housing or work-unit compound to a new commodity housing estate with all sorts of amenities including landscaped gardens and club house facilities not only represents significant improvements in the housing lot, but also a major personal achievement on the part of the individual. Even though the move may be associated with severance of ties in the former neighbourhood, the sense of pride arising from upward residential mobility would foster emotional attachment to and hence willingness to participate in the activities of the new neighbourhood (Li et al., 2012; Wu and Logan, 2016; Zhu et al., 2012). The results reiterate previous findings.

Level 2 variables: neighbourhood attributes

Again, the results on neighbourhood attributes are generally in line with those reported in the literature. Neighbourhood stability exerts significant positive influence on emotional attachment to the neighbourhood as well as acquaintance of neighbours. This echoes Flaherty and Brown (2010) and Thomas et al. (2016), who suggest that neighbourhood stability adds to the opportunity and potential for residents to have more neighbourhood-based social interactions and thus is conducive to the formation of stronger social ties. However, emotional attachment and acquaintance of neighbours may not be translated to more involvement in neighbourhood affairs, as the coefficient estimate for Model 2 (involvement in neighbourhood affairs) is of the ‘wrong’ sign.

Regarding the effect of neighbourhood size, affirming Kasarda and Janowitz (1974), residence in larger neighbourhoods is associated with less acquainted neighbours. However, the effect of neighbourhood size on the other two dimensions of neighbourhood attachment is not significant.

The coefficient estimates of neighbourhood population composition are in line with expectations, too. Higher share of homeowners in a community appears to be associated with heightening emotional attachment to, and participation in/with, the neighbourhood and neighbour acquaintances, although none of the regression coefficients is significant. A second measure of neighbourhood composition employed is migrants’ share of the population, which is found to be positively associated with acquaintance of neighbours. As pointed out above, for migrants neighbourhood-based reciprocal help among kinsmen and people from the same village or province of origin constitute an important source of social support (Du and Li, 2010; Wang et al., 2017; Wu and Logan, 2016).

The results pertaining to the evaluation of neighbourhood built environment are also reasonable. Reiterating Zhu et al. (2012) and Forrest and Yip (2007), better neighbourhood built environment, underscored by cleanliness, security and abundance of amenities such as community centres and sports arenas, significantly enhances respondents’ emotional attachment to it. However, evaluation of neighbourhood built environment seems to have little to do with either involvement in neighbourhood activities or acquaintance of neighbours.

Residential location is significantly related to the behavioural dimensions of attachment. People residing in the outer suburbs tend to have higher emotional attachment as well as higher levels of participation in neighbourhood affairs, compared with those in the inner suburbs. It may be pointed out that 52.2% of the sampled respondents in the outer suburbs are young homeowners under 40 years old. Many choose to live there to bring up their family; as such, they tend to spend more time with children in the neighbourhood communal space. At the same time, residents in new-build suburban housing are often confronted with such problems as faulty internal fittings, lack of quality schools, unreliable and inadequate public transport and problems related to estate management. The recent in-movers, under such circumstances, are more inclined to join forces with their neighbours, who are often of similar socioeconomic status and migration histories, to tackle their shared problems (Li and Li, 2014); also, such collaborative efforts are likely to contribute to higher emotional attachment to the neighbourhood. Perhaps as a corollary, residents living in the inner-city core tend to have less acquaintance with neighbours. On average, inner-core residents only know 7 neighbours by name; in comparison, inner-suburb residents know 15 and outer-suburban residents know 13. Correspondingly, 30% of inner-core residents do not know any neighbour by name; the respective figures for inner- and outer-suburb residents are 21% and 26%.

Interaction between level 1 and 2 variables

The results show that for the dimension ‘involvement in neighbourhood activities’, the respondent’s age interacts with neighbourhood stability. More specifically, older people tend to get more involved in stable neighbourhoods. Unlike neighbourhoods with higher population turnover rates, where residents tend to be younger (Forrest and Yip, 2007), stable neighbourhoods have higher percentages of long-term and older residents. Yip et al. (2013) find that the latter inhabitants, especially retirees, play a more active part in residents’ committee and homeowners’ association than do other groups. Availability of time is perhaps a major reason. Moreover, as also evidenced by Li et al. (2012), there exists a negative relationship between age and neighbourhood attachment in new commodity housing estates, but not in work-unit and reform housing compounds. Apparently, senior citizens tend to have greater difficulty adjusting to the new environment.

Conclusions

Wholesale redevelopment, phenomenal suburbanization and massive in-migration flows have characterized China’s urban development scene in recent decades. There is a rapidly growing literature on China’s unprecedented urban transformation. Much has been written, for example, on the hukou system, which aggravates urban–rural disparities, pushing ever more peasants to urban areas, and institutionally erecting invisible walls within major cities that separate rural migrants from de jure urbanites (Chan, 2014). The hukou system is closely tied to the dual land tenure system, under which urban land belongs to the state while rural land belongs to rural collectives. Such a structural duality underpins the emergence of villages-in-the-city, which are almost without exception migrant enclaves on former urban–rural fringes, another major topic of contemporary research (see, for example, Du and Li, 2012; Wu and Logan, 2016).

However, to date there has not been much written on residential relocation within an urban area. In particular, virtually nothing is known about how residential moves and the resulting population turnovers and changes in sociophysical environments in urban neighbourhoods affect an individual’s attachment to the neighbourhood of his or her residence in the Chinese context. To help fill the vacuum in the literature, in this study we estimated several two-level regression models to explore three different dimensions of neighbourhood attachment – emotional attachment, involvement in neighbourhood activities and neighbour acquaintances – and analysed the extent to which these dimensions are affected by personal residential histories as well as the type and characteristics of the current and previous residence, which are closely related to the time of development and hence distance to the city centre.

The results confirm that attachment to the neighbourhood is related to aspects of residential history. Length of residence, which is related to a person’s position in the life course and homeownership, in particular, has the strongest effect: it is statistically significant for all three dimensions. Senior citizens, especially retirees, would likely to have more time to participate in neighbourhood affairs. However, the results show that the effect of age depends on neighbourhood stability; in particular, older people play a more active part in neighbourhood affairs only in those neighbourhoods with lower population turnover. Also, people living in old work-unit/reform housing compounds for many years have developed stronger relationships with their neighbours than those residing in newly constructed commodity housing estates. However, confirming the findings of Zhu et al. (2012), the results also show that somewhat higher emotional attachment to the new neighbourhood is revealed for those who have moved out of the work-unit/reform housing compound, which is almost invariably inside the urban built-up area in pre-reform and early reform times, to a commodity housing estate, which is likely to be located further into the suburbs.

Population mobility, be it rural–urban, inter-urban or intra-urban, impinges heavily on demographic dynamics at the neighbourhood level. Neighbourhood stability – that is, lack of population turnovers – is positively related to respondents’ emotional attachment to the neighbourhood as well as acquaintance of neighbours. Population composition of the neighbourhood also matters. A higher percentage of homeowners is positively associated with all three dimensions of neighbourhood attachment, although none of the regression coefficients is significant. Migrant-rich neighbourhoods, on the other hand, are associated with more neighbour acquaintances. Zhu and Fu (2017) contend that in Chinese cities today the sense of home is based more on neighbourhood physical attributes than on social relations. Results of the two-level regression models lend further credence to their observation. Finally, living in different parts of the city would lead to different behavioural inclinations with respect to neighbourhood attachment. Residents in the inner core are less likely to develop friendships within the neighbourhood, while people residing in the outer suburbs tend to participate more actively in neighbourhood affairs.

We acknowledge the absence of village-in-the-city in the sampling frame, which precluded us from examining how emotional attachment, engagement in neighbourhood affairs and development of personal networks by both migrant-tenants and indigenous villagers manifest in such migrant enclaves. We also acknowledge the usefulness of constructing ethnographies to complement the present work in unravelling the meaning of neighbourhood and neighbourhood attachment to individuals and families. However, we believe that the above analysis has enriched the literature on urban transformation and community life in Chinese cities by taking into account the multi-dimensional nature of neighbourhood attachment, incorporating both micro- and ecological variables, and highlighting the significance of residential location and relocation in our analysis of neighbourhood attachment. The results demonstrate that an individual’s or household’s attachment to the neighbourhood can be better understood by taking into consideration mobility-related factors, which include the residential experiences of individual households, as well as the aggregate impacts of residential relocations on given neighbourhoods. Findings of the study will help inform policy-makers in China who are eager to rebuild community vitality amid the bewildering changes under China’s Great Urban Transformation, to quote Hsing (2010).

Footnotes

Acknowledgement

Thanks are due to Professor Shenjing He of the University of Hong Kong for helping to implement the 2012 Guangzhou questionnaire survey, upon which the present study is based.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Hong Kong Research Grant Council General Research Fund (Grant No. HKBU 245511) and Hong Kong Baptist University Faculty Research Grant (Grant No. FRG216-17006).

Notes

References

Bailey

Kearns

Livingston

(2012) Place attachment in deprived neighbourhoods: The impacts of population turnover and social mix. Housing Studies 27(2): 208–231.

Bolan

(1997) The mobility experience and neighborhood attachment. Demography 34(2): 225–237.

Bray

(2006) Building ‘community’: New strategies of governance in urban China. Economy and Society 35(4): 530–549.

Brehm

Eisenhauer

Krannich

(2004) Dimensions of community attachment and their relationship to well-being in the amenity-rich rural west. Rural Sociology 69(3): 405–429.

Brehm

Eisenhauer

Krannich

(2006) Community attachments as predictors of local environmental concern: The case for multiple dimensions of attachment. American Behavioral Scientist 50(2): 142–165.

Brown

Moore

(1970) The intra-urban migration process: A perspective. Geografiska Annaler. Series B, Human Geography 52(1): 1–13.

Chan

(2014) China’s urbanization 2020: A new blueprint and direction. Eurasian Geography and Economics 55(1): 1–9.

Clark

WAV

(2013) Life course events and residential change: Unpacking the age effects on the probability of moving. Journal of Population Research 30(4): 319–339.

Clark

WAV

Lisowski

(2017) Decisions to move and decisions to stay: Life course events and mobility outcomes. Housing Studies 32(5): 547–565.

10.

Davies

, and Pickles

(1985) Longitudinal versus cross-sectional methods for behavioural research: A first-round knockout. Environment and Planning A 17(10): 1315–1329.

11.

Dekker

(2007) Social capital, neighbourhood attachment and participation in distressed urban areas: A case study in The Hague and Utrecht, the Netherlands. Housing Studies 22(3), 355–379.

12.

(2010) Migrants, urban villages, and community sentiments: A case study of Guangzhou, China. Asian Geographer 27(1–2): 93–108.

13.

(2012) Is it really just a rational choice? The contribution of emotional attachment to temporary migrants’ intention to stay in the host city in Guangzhou. China Review 12(1): 73–94.

14.

Hao

(2017) ‘Anyway, you are an outsider’: Temporary migrants in urban China. Urban Studies. Epub ahead of print 14 February 2017. DOI: 10.1177/0042098017691464.

15.

Finney

Jivraj

(2013) Ethnic group population change and neighbourhood belonging. Urban Studies 50(16): 3323–3341.

16.

Flaherty

Brown

(2010) A multilevel systemic model of community attachment: Assessing the relative importance of the community and individual levels. American Journal of Sociology 116(2): 503–542.

17.

Forrest

Yip

(2007) Neighbourhood and neighbouring in contemporary Guangzhou. Journal of Contemporary China 16(50): 47–64.

18.

(2014) The contentious democracy. In: Huang

(eds) Housing Inequality in Chinese Cities. New York: Routledge, pp.201–216.

19.

Zhu

et al . (2015) Toward a relational account of neighborhood governance: Territory-based networks and residential outcomes in urban China. American Behavioral Scientist, 59(8): 992–1006.

20.

Giroir

(2014) The gated communities of chateaux in China: Back to feudalism? In: Huang

(eds) Housing Inequality in Chinese Cities. New York: Routledge, pp.143–160.

21.

Goudy

(1990) Community attachment in a rural region. Rural Sociology 55(2):178–198.

22.

Guangzhou Land Resource and Housing Management Bureau (2009) Draft Regulation for Homeowners’ Assembly and Homeowners’ Associations in Guangzhou for public comments. Available at: http://www.gzlpc.gov.cn/gtzy/qtlgg/2009-09/04/content_163197.shtml (accessed 20 August 2017).

23.

Guest

Lee

(1983) Sentiment and evaluation as ecological variables. Sociological Perspectives 26(2), 159–184.

24.

Guest

Wierzbicki

(1999) Social ties at the neighborhood level: Two decades of GSS evidence. Urban Affairs Review 35(1): 92–111.

25.

(2015) Homeowners associations and neighborhood governance in Guangzhou, China. Eurasian Geography and Economics 56(3): 260–284.

26.

Hsing

Y-T

(2010) The Great Urban Transformation: Politics of Land and Property in China. Oxford: Oxford University Press.

27.

Hummon

(1992) Community attachment: Local sentiment and sense of place. In Altman

Low

(eds) Place Attachment. New York: Plenum Press, pp.253–278.

28.

Kasarda

Janowitz

(1974) Community attachment in mass society. American Sociological Review 39(3): 328–339.

29.

Lewicka

(2011) Place attachment: How far have we come in the last 40 years? Journal of Environmental Psychology 31(3): 207–230.

30.

(2014) Living the networked life in the commodity housing estates: Everyday use of online neighborhood forums and community participation in urban China. In: Huang

(eds) Housing Inequality in Chinese Cities. New York: Routledge, pp.181–198.

31.

(2004) Life course and residential mobility in Beijing, China. Environment and Planning A 36(1): 27–43.

32.

Mao

(2016) Exploring residential mobility in Chinese cities: An empirical analysis of Guangzhou. Urban Studies. Epub ahead of print 6 December 2016. DOI: 10.1177/0042098016679774.

33.

Siu

(2001) Residential mobility and urban restructuring under market transition: A study of Guangzhou, China. Professional Geographer 53(2):219–229.

34.

Song

(2009) Redevelopment, displacement, housing conditions, and residential satisfaction: Study of Shanghai. Environment and Planning A 41(5):1090–1108.

35.

(2007) The road to homeownership under market transition Beijing 1980–2001. Urban Affairs Review 42(3): 342–368.

36.

Zhu

(2014) Residential mobility within Guangzhou city, China, 1990–2010: Local residents versus migrants. Eurasian Geography and Economics 55(4): 313–332.

37.

Zhu

, and Li

(2012) Neighborhood type, gatedness, and residential experiences in Chinese cities: A study of Guangzhou. Urban Geography 33(2): 237–255.

38.

Liu

Ryan

Aurbach

et al . (1998) The influence of local Church participation on rural community attachment. Rural Sociology 63(3): 432–450.

39.

Matarrita-Cascante

DRS

Luloff

(2010) Permanent and seasonal residents’ community attachment in natural amenity-rich areas: Exploring the contribution of landscape-related factors. Environment and Behavior 42(2): 197–220.

40.

Paradis

(2000) Main street transformed: Community sense of place for nonmetropolitan tourism business districts. Urban Geography 21(7): 609–639.

41.

Patterson

Williams

(2005) Maintaining research traditions on place: Diversity of thought and scientific progress. Journal of Environmental Psychology 25(4): 361–380.

42.

Read

(2003) Democratizing the neighborhood? New private housing and home-owner self-organization in urban China. The China Journal 49: 806–820.

43.

Sharp

(2013) Neighborhood attachment in dynamic perspective: Changing people and changing contexts. Thesis, Pennsylvania State University, USA.

44.

Song

Chiang

(2012) The place attachment of residents displaced by urban redevelopment projects in Shanghai. Issues & Studies 48 (3): 43–73.

45.

Stinner

Loon

Chung

et al . (1990) Community size, individual social position, and community attachment. Rural Sociology 55(4): 494–521.

46.

Stokols

Shumaker

(1982) The psychological context of residential mobility and well-being. Journal of Social Issues 38(3): 149–171.

47.

Sun

Yip

(2014) Managing the nouveaux riches: Neighborhood governance in upmarket residential developments in Shanghai. In: Huang

(eds) Housing Inequality in Chinese Cities. New York: Routledge, pp.217–233.

48.

Theodori

Luloff

(2000) Urbanization and community attachment in rural areas. Society & Natural Resources 13(5): 399–420.

49.

Thomas

Stillwell

Gould

(2016) Modelling the duration of residence and plans for future residential relocation: A multilevel analysis. Transactions of the Institute of British Geographers 41(3): 297–312.

50.

Tucker

Marx

Long

(1998) ‘Moving on’: Residential mobility and children’s school lives. Sociology of Education 71(2): 111–129.

51.

Wang

Zhang

(2017) Neighbourhood cohesion under the influx of migrants in Shanghai. Environment and Planning A 49(2): 407–425.

52.

Williams

Kitchen

DeMiglio

et al . (2010) Sense of place in Hamilton, Ontario: Empirical results of a neighborhood-based survey. Urban Geography 31(7): 905–931.

53.

Woolever

(1992) A contextual approach to neighbourhood attachment. Urban Studies 29(1): 99–116.

54.

Wright

Ellis

Holloway

(2015) Neighbourhood racial diversity and white residential segregation in the United States. In: Lloyd

Shuttleworth

Wong

(eds) Social-Spatial Segregation: Concepts, Processes and Outcomes. Bristol: Policy Press, pp. 111–134.

55.

Logan

(2016) Do rural migrants ‘float’ in urban China? Neighbouring and neighbourhood sentiment in Beijing. Urban Studies 53(14): 2973–2990.

56.

(2010) The ‘community question’ in the state-led urbanization process of post-reform China: A case study of Shanghai. Thesis, Hong Kong Polytechnic University, Hong Kong.

57.

Yip

Leung

Huang

(2013) Impact of community on personal well-being in urban China. Journal of Social Service Research 39(5): 675–689.

58.

Zhu

(2017) Deciphering the civic virtue of communal space: Neighborhood attachment, social capital, and neighborhood participation in urban China. Environment and Behavior 49(2): 161–191.

59.

Zhu

Breitung

(2012) The changing meaning of neighbourhood attachment in Chinese commodity housing estates: Evidence from Guangzhou. Urban Studies 49 (11): 2439–2457.