Social Relations and Place Identity of Development-Induced Migrants: A Case Study of Rural Migrants Relocated from the Three Gorges Dam, China

Abstract

The resettlement of development-induced migrants is a complex socioeconomic and cultural process. The levels of place identity among migrants may profoundly affect their long-term stability and sustainable development in host communities. For long-distance displaced migrants, their social relations undergo drastic changes, and the extent of social-relations reconstruction determines their place identity in the resettlement area. Previous studies mainly concentrated on migrant compensation schemes, housing, and land allocation, and livelihood restoration. However, insufficient attention has been paid to the mechanism by which social-relations reconstruction shapes place identity. Drawing on sample survey data collected during 2022–2023 from migrants displaced from the Three Gorges Dam (TGD) in China to various other provinces, this study conceptualizes place identity in three dimensions: group identity, permanent settlement intention, and expectations for children. A structural equation model (SEM) was employed to investigate how social relations, categorized as geographical, home-tied, and carried-over relations, mediate the influence of multiple factors on place identity. The findings are that: (1) among development-induced migrants, social relations exerted significant positive effects on place identity, along with personal characteristics, socioeconomic characteristics, land, and housing factors. Moreover, family features, socioeconomic characteristics, and land and housing factors all had a significantly positive impact on social relations; (2) social relations acted as a full mediator between family features and place identity, and a partial mediator between socioeconomic characteristics/land and housing factors and place identity; (3) among geographical, home-tied, and carried-over social relations, geographical social relations had the largest effect on place identity, followed by carried-over relations. However, continuous dependence on home-tied social relations negatively affected the migrants’ development of place identity in new resettlement areas. This study elucidates the role of social relations in the socioeconomic and cultural reconstruction during migrant resettlement, offering insight for improving resettlement policies and promoting sustainable community integration.

1. Introduction

Development-induced migration refers to the forced relocation of people caused by large-scale development projects, such as dam construction, mineral resource exploitation, or urban renewal [1,2]. These involuntary displacements are vast in scale worldwide. During the twentieth century, dam construction alone forced an estimated 40–80 million predominantly rural residents to relocate worldwide, and development-induced infrastructure projects (including dams, mines, and urban expansion) continue to uproot roughly 10–15 million people each year [3]. Since migrants have little personal choice in such projects, they often face grave socioeconomic challenges: disrupted livelihoods, fragmented communities, and the loss of emotional ties to homelands. If resettlement measures are inadequate, migrants may remain impoverished or be at risk of further instability [4,5,6,7]. Empirical evidence suggests that promoting the long-term integration and stability of large-scale development-induced rural migrants is a persistent policy challenge. For instance, in India’s Sardar Sarovar Dam project, where tens of thousands of tribal households were relocated to host villages, discriminatory practices and deficient impact assessments severed traditional kinship bonds, leaving many people socially isolated and excluded from community participation (Flood, 1997) [8]. Ghana’s Akosombo Dam development, in which about 80,000 people were relocated to settlements lacking adequate social-service infrastructure or alignment with local customs, resulted in the erosion of communal solidarity, loss of shared burial grounds, and a depopulation of more than half of some resettlement villages within two decades. These disruptions, as extensively recorded in subsequent studies of downstream communities, starkly highlighted the disintegration of traditional support systems and persistent social marginalization [9,10]. After the Danjiangkou Reservoir project in China in the 1950s, many migrants faced social maladjustment and psychosocial stress, as evidenced by weak participation in local governance, fragmented support networks, and collective return migration due to integration difficulties. Between 1959 and 1981, five major waves of return migration were recorded, during which 108,000 migrants, constituting 26.21% of the total displaced population, returned to their original hometowns [11]. Some critics argue that dam projects are a form of “technological tyranny” that “tears communities apart, undermines social cohesion, and causes hardship on a vast scale” [12]. Given the state-led and involuntary nature of these relocations, the physical act of moving people is relatively straightforward. However, achieving true long-term integration and well-being in the host location remains highly challenging. This gap between “moving” and “staying settled” lies at the very core of the development-induced migration problem. Addressing this gap is crucial not only for the well-being of the migrants themselves but also for maintaining social stability and promoting sustainable development in both the origin and host regions [13].

The Three Gorges Dam in China provides a critical case study through which to examine these issues. The Three Gorges Project in China constitutes the world’s largest dam-related relocation initiative, having displaced about 1.29 million people since 1993. Around 153,000 of those rural migrants were resettled outside the reservoir area to 254 counties in 12 provinces or municipalities, with some traveling as far as 1500 km from their home regions. Studies have shown that cultural disparities and environmental mismatches are principal reasons for the repeated re-migration of these long-distance migrants [14,15]. Many of the Three Gorges Dam migrants have entered locales drastically different in terms of natural conditions, social customs, and even language. The substantial diversity across their host communities creates unprecedented challenges to migrants’ adaptation. Under these circumstances, it is vital to support migrants in establishing new lives and developing a sense of belonging, often referred to as “staying settled”. In this context, place identity becomes a decisive factor in the success or failure of resettlement efforts. Place identity refers to the emotional attachment and sense of membership formed through ongoing interactions with a specific location [16,17,18]. For migrants, forming a place identity involves making the psychological transition from outsider to community member. A strong sense of place identity fosters stronger social connections and deeper civic engagement [19,20]. When displaced people regard a new location as “home”, they are more inclined to devote resources to improving local living conditions and strengthening household and community activities [21]. As such, the establishment of place identity not only indicates the extent of social integration for development-induced migrants but also shapes the long-term stability and sustainability of migrant communities.

However, place identity cannot be cultivated overnight, especially for rural migrants compelled to relocate. They face a complex, progressive journey of breaking and rebuilding social connections. With relocation, their primary challenge is to reconstitute a supportive social network: existing kinship ties and neighborly relationships from their original village are often fractured, disrupting informal networks and depleting their social capital [22,23,24]. As researchers argued regarding social disarticulation theory, forced migration unravels established community structures by scattering former communities and cutting off well-established social bonds. In this process, migrants lose mutual support networks and social support mechanisms essential to their livelihoods [25,26]. These disruptions are even more pronounced among Chinese rural migrants, who previously lived in so-called “acquaintance societies”, as described by [27]. Within rural communities, reciprocal trust and routine cooperation are fundamental to everyday security and interpersonal belonging [28]. Once displaced from this familiar environment and placed into a tightly integrated receiving community, migrants are often subject to deeper social isolation and adaptation pressures, much more so than voluntary migrants. Consequently, whether migrants can swiftly rebuild new social networks will largely determine whether their construction of place identity succeeds or fails [29]. Studying how migrants reconstruct social bonds to foster a sense of place identity thus has significant theoretical and practical significance.

Although the broader literature on place identity has explored its conceptual foundations, perceptual dimensions, and personal or environmental determinants, very few studies have interrogated the distinctive trajectory by which development-induced rural migrants reconstruct place identity after involuntary relocation [30]. Existing work on dam resettlement has focused mainly on economic compensation and infrastructure, with limited attention to how ruptured and subsequently reassembled social networks condition migrants’ sense of belonging [31]. Research framed around urban-renewal displacement (e.g., gentrification) does examine identity shifts, but it involves mostly urban residents with stronger formal safety nets and differs markedly from state-organized rural-to-rural or rural-to-peri-urban dam resettlement [32]. Consequently, the mediating role of social relations in shaping place identity among development-induced rural migrants remains underexplored.

Against this background, the present study focused on rural migrants from the Three Gorges Dam, undertaking an in-depth inquiry into how social relations influence place identity. Drawing on field surveys and questionnaire data, we quantitatively evaluated migrants’ place identity attributes and assessed how different forms of social interaction affected their group identity, permanent settlement intention, and expectations for children in newly established communities. We then employed structural equation modeling (SEM) to examine the mediating role of social relations in linking various factors to place identity. Our ultimate goal was twofold: from a theoretical perspective, we aimed to broaden place-identity research on development-induced migrants; from an applied standpoint, we seek to offer recommendations for improving resettlement policies, enhancing community governance, and ensuring migrants’ long-term well-being. By exploring how social relations and place identity intersect among these Three Gorges Dam migrants, this work endeavored to illuminate the critical mechanisms allowing this group to reconstruct their sense of belonging in a new homeland.

2. Theoretical Framework

2.1. Conceptual Foundations: Place Identity and Forced Migration

Conventional place identity theories in environmental psychology depict place identity as an interplay of perceptions, emotions, values, and behavioral tendencies developed, whether consciously or subconsciously, through interactions with one’s environment [21,33]. In human geography, place identity further emphasizes the “sense of place”, understood as an affective attachment and sense of belonging fostered by engagement with a physical setting and its cultural connotations [34,35,36].

On the one hand, voluntary migrants generally exhibit a high degree of agency in migration decisions, often evaluating the environmental conditions and social opportunities of the destination beforehand. Equipped with certain expectations and preparatory strategies, they are more likely to retain parts of their original social networks or quickly establish new ones upon arrival. As a result, the reconstruction of place identity among voluntary movers typically follows a trajectory of active integration and gradual affiliation [37].

On the other hand, involuntary migrants frequently experience what has been termed the “Ulysses syndrome”, a prolonged sense of uprootedness accompanied by cultural bereavement, social isolation, and psychological maladaptation, all of which severely hinder the re-establishment of place identity [38,39]. According to the theory of cultural bereavement, when forced migrants are unable to sustain their original cultural practices or social structures, they are particularly vulnerable to identity disruption and belonging crises [40,41]. Moreover, forced displacement often results in a dual rupture in the human–place relationship, characterized by both disconnection and the difficult process of reintegration. Research shows that the longer the migration distance, the more extensive the breakdown of social networks, and the slower the recovery of place attachment [42]. These findings underscore that the traumatic rupture caused by forced migration often requires additional social and psychological interventions to support identity reconstruction, yet such support is frequently lacking for involuntary migrants [43].

In sum, while the fundamental mechanisms of place identity operate in both voluntary and involuntary migration contexts, the former typically unfolds through incremental adaptation and self-directed relocation, whereas the latter entails a more complex emotional and social rebuilding process due to the non-volitional nature of migration and the severing of social relations. For development-induced migrant populations, the systematic reconstruction of social networks, cultural continuity, and community participation is essential to restoring place identity and achieving long-term settlement stability.

2.2. Special Features in Rural China and Evaluation Dimensions

However, the way development-induced migrants construct place identity differs from how voluntary migrants do so, in two notable ways, especially in rural China. First, in rural Chinese “acquaintance societies”, social life is organized according to Fei Xiaotong’s differential mode of association (chaxu geju), whereby interpersonal ties form concentric rings emanating from the self, with each successive ring representing weaker—and more generalized—social obligations and support [27,44]. These concentric, highly localized networks underpin both personal self-concepts and place belonging, as villagers rely on stable kinship and neighborhood ties for everyday assistance and emotional security [45]. When forced displacement severs these inner circles, rural migrants experience acute social fragmentation: longitudinal studies of China’s poverty-alleviation relocations document dramatic declines in neighborly interaction and intergroup support within two years of resettlement [46]. Second, because development-induced relocations often span long distances across provinces, migrants confront significant cultural dissonance and institutional unfamiliarity. Empirical evidence from China’s internal-migrant surveys shows that greater migration distance correlates with lower levels of social integration and subjective well-being among rural migrants, with social integration mediating the negative impact of distance on happiness [47]. Moreover, the involuntary nature of such displacement compounds emotional distress and nostalgia: forced movers cannot exercise agency over their destination, leading to heightened psychosocial stress and a stronger pull toward return migration, as documented in studies of dam-affected communities and broader forced-displacement reviews [48,49].

Based on the above discussion, development-induced migrants not only endure the trauma of being forced to leave their familiar environment but also face the challenge of adapting to and reshaping their identities in a new setting. We define place identity in this context as the socio-psychological process through which individuals or groups perceive themselves as part of a specific place by continuously interacting with it [21]. This concept incorporates both how migrants adapt to tangible facets of their new environment (e.g., housing, employment, or family life) and how they strive to reconstruct emotional bonds after losing their original social networks.

In this study, we break down the place identity of development-induced migrants into three dimensions: group identity, permanent settlement intention, and expectations for children, which are based on a temporal sequence of migrant integration. Identity development frameworks propose that individuals first internalize new social and cultural norms, then commit behaviorally, and finally project their sense of self into the future [50,51]. Group identity aligns with the cognitive–affective internalization stage, where migrants begin to see themselves as community members, consistent with place identity as an aspect of self shaped by environmental interactions [52]. Permanent settlement intention reflects the exploration–commitment transition in identity-status theory: after initial internalization, individuals form concrete intentions to stay, signaling behavioral commitment to the locale [53,54]. Empirical studies confirm that identity formation is inherently temporal and multilayered: research on place identity underscores distinct personal and social dimensions that evolve across cognitive, behavioral, and affective stages, and lifespan identity work highlights the need to capture early internalization, mid-life commitment, and later reflective projection [55]. Thus, these three indicators map directly onto recognized phases of identity construction—short-term cognitive integration, medium-term behavioral commitment, and long-term intergenerational projection—providing a theoretically robust way to measure place identity formation in development-induced migrants.

2.3. Influencing Factors for Place Identity

Researchers generally categorize factors affecting place identity into personal traits, the physical environment, and the social environment [56,57]. Nevertheless, the forced nature of development-induced migration and the abrupt dissolution and reconstruction of social networks require tailored factor selection.

Personal characteristics. Factors such as age, gender, educational background, occupation, and motives for migration are widely recognized as determinants of how fast and effectively migrants can adapt and integrate [58,59,60]. Because development-induced migrants are required to relocate, many may lack the skills or education to thrive in their new homes, further complicating adaptation.

Family features. The attitudes and mutual support of family members play a decisive role in whether migrants accept and integrate into the local community [16,61]. Development-induced migration frequently involves entire families; whether they collectively participate in community networks or provide each other emotional support can critically reinforce or inhibit their place identity.

Socioeconomic characteristics. Local economic vitality, infrastructure, and job prospects are crucial for migrant integration [62,63,64,65]. Because involuntary migrants cannot freely choose their destinations, a sharp mismatch between their original and new environments can deepen alienation and hinder place-identity formation.

Land and housing factors. In rural China, this factor often determines whether an individual can truly “take root” in a new community [51]. Development-induced migrants generally rely on government-allocated housing and farmland to rebuild their livelihoods. If these fail to meet their daily production and living needs, it becomes challenging to establish a sense of security and belonging [66].

Rupture and reconstruction of social relations. This feature is particularly prominent for development-induced migrants. If they cannot promptly rebuild supportive social networks in their relocation sites, lacking emotional and social capital, their path to place identity may be undermined [67].

Accordingly, the five key factors examined in this study are personal characteristics, family features, socioeconomic characteristics, land and housing factors, and social relations.

2.4. The Bridging Role of Social Relations

Social relations serve as a critical bridge in the formation of place identity for development-induced migrants, more so than for voluntary movers, because their original networks are involuntarily severed. In this forced context, social ties not only directly shape migrants’ access to emotional support and community resources, but also strongly mediate how personal, family, socioeconomic, and land–housing factors are converted into new place identity [68,69].

First, the rupture and subsequent rebuilding of social networks is far more acute for forced migrants. When kinship and local ties are disrupted without the agency to choose alternative support groups, establishing new social bonds in the host community becomes indispensable for emotional security and resource sharing [70,71]. In voluntary migration scenarios, by contrast, migrants often retain old contacts or pre-plan new ones; forced migrants must rebuild entirely from scratch, making social relations their primary survival mechanism [50,70].

Second, while personal and family characteristics shape integration for all migrants, for development-induced migrants their direct influence on place identity is largely channeled through social relations. Factors such as age, skill level, or household support may predict network-building ability, but unlike in voluntary moves where higher education or younger age can directly translate into faster integration, here they only matter to the extent that they facilitate or hinder the reconstruction of local ties [72]. A family’s negative outlook, for instance, undermines collective efforts to forge new connections, blocking the emotional anchoring that underpins identity formation.

Third, socioeconomic conditions and land–housing elements likewise exert their effects primarily by enabling or constraining the scope and depth of social interactions in the new locale. Although favorable economic environments and secure housing can expedite integration for all migrants, involuntary movers cannot leverage these assets as freely: they depend on social relations to access job networks, navigate local institutions, and translate material resources into community standing [73,74].

Given the forced nature of development-induced migration and the inherent challenges in reconstructing social networks, social relations stand not only as a direct factor but also as a central mediator that converts other influences into place identity for these migrants. As shown in Figure 1, well-established social relations foster a sense of belonging, whereas social isolation impedes it, even if migrants hold favorable personal characteristics or secure land–housing resources.

3. Materials and Methods

3.1. Study Region and Data

The Three Gorges Project is so far the world’s largest hydropower-related resettlement initiative. Since 1993, it has displaced approximately 1.29 million individuals, of whom 153,000 were relocated outside the reservoir area to 12 provinces (or municipalities) across China, involving 254 counties at distances of 100–1500 km from the dam site. The diverse natural environments, languages, and resettlement approaches across these host locations present valuable opportunities for this study.

The sampling frame was drawn from the Ministry of Water Resources’ Monitoring Sample Database for Three Gorges Dam migrants, which randomly selected two to three county-level units in each of 11 provinces (or municipalities): Anhui, Fujian, Guangdong, Hubei, Hunan, Jiangsu, Jiangxi, Shanghai, Sichuan, Zhejiang, and Chongqing. Within each of the resulting 22 sample counties (districts), two migrant villages were randomly selected. Using local registration lists of relocated households remaining in each village, the number of respondents to survey was determined in proportion to the total relocated population, and individual migrants were then chosen via simple random sampling. Face-to-face, door-to-door interviews were conducted with each selected respondent between June 2022 and September 2023.

A total of 89 villages were surveyed and 353 valid responses obtained, with each province contributing 20–50 interviews. Given that these 22 counties (districts) host roughly 3300 relocated households, the sample of 353 achieved a 95% confidence level with a margin of error below 6%. Demographic characteristics of the respondents are summarized in

Table 1. Sample descriptive statistics.

Indicator	Category	Number	Proportion
Gender	Male	268	75.92%
	Female	85	24.08%
Age	0–14	0 *	0.00%
	15–64	306	86.69%
	65 and above	47	13.31%
Education	Primary school or below	152	43.06%
	Junior high school	160	45.33%
	High school	28	7.93%
	College or above	13	3.68%
Occupation	Worker	257	72.80%
	Business owner	17	4.81%
	Other	10	2.83%
	Self-employed farmer	69	19.54%
Resettlement Year	1999	17	4.82%
	2000	101	28.61%
	2001	104	29.46%
	2002	75	21.25%
	2003	5	1.42%
Land Status	No land	54	15.30%
	Possessing 0–4 mu	185	52.41%
	Possessing 4–10 mu	91	25.78%
	Possessing over 10 mu	23	6.51%

* As this study focused exclusively on first-generation migrants and did not include second-generation individuals, the number of respondents aged 0–14 was zero.

.

3.2. Methods

We employed statistical analysis, factor analysis, and structural equation modeling (SEM). Statistical analysis was used for basic sample characteristics. Factor analysis identified common factors from numerous social interaction variables, enabling us to classify different types of social relations. Finally, SEM was used to model multi-factor, multi-effect relationships, constructing a theoretical framework linking personal characteristics, family features, socioeconomic conditions, land and housing factors, social relations, and place identity.

3.3. Variables

3.3.1. Exogenous Latent Variables

In the SEM, the main exogenous latent variables for place identity included personal characteristics, family features, socioeconomic characteristics, and land and housing factors. Personal characteristics typically cover a migrant’s gender, age, education, occupation, and a variable distinguishing first-generation from second-generation migrants. Due to the fact that all migrants in the sample completed their resettlement between 2001 and 2002, the variation in relocation timing was negligible; consequently, “length of residence” was not included as an additional control variable. For family features, prior research indicates that family’s emotional bonds and support help migrants find a sense of belonging in a new place [75]. Hence, a high residential continuity and a larger number of household members in the new location presumably correlate with stronger local integration. Socio-economic characteristics integrate micro-level indicators (e.g., personal and household disposable income) with macro-level measures of host-city economic performance (notably regional GDP per capita). Recent studies have shown that while individual income directly affects migrants’ capacity to participate in local economies, city-level economic context (e.g., GDP size, growth rate) significantly shapes overall socioeconomic integration and place-identity formation [76,77]. Land and housing factors capture the degree to which rural households have “taken root” in the new community. In the Chinese agrarian context, land-use rights and rural homestead rights constitute core assets for farmers’ livelihood security; for instance, collective land ownership of homestead tenure has been shown to significantly enhance farmers’ sense of security and community participation, thereby strengthening their attachment to the resettlement area [78,79]. In this study, “housing status” and “per-capita housing area” are employed as representative indicators to reflect migrants’ material settlement status and the extent of their linkage to local society.

Table 2. Variables of the SEM model.

Latent Variable	Measurement Variable	Coding Scheme
Personal Characteristics	Gender	Female: 1; Male: 2
	Age/year	Below 40: 1; 40–49: 2; 50–59: 3; 60–69: 4; sbove 70: 5
	Relation to migrants	Migrant: 1; descendant of migrant: 2
	Education	primary school or below: 1; junior high school: 2; high school: 3; college or above: 4
	Occupation	self-employed farmer: 1; worker: 2; business owner: 3; other: 4
Family Features	Household size/person	3 or fewer: 1; 4–5: 2; 6–7: 3; 8 or more: 4
	Residential continuity	Refer to Table 3
	Proportion of employed family members	Below 25%: 1; 26–50%: 2; 51–75%: 3; 76–100%: 4
Socio-Economic Characteristics	Urban economic gap (based on GDP of affiliated prefecture—level cities in 2022)/RMB trillion	Below 2.5: 1; 2.6–5.9: 2; 6.0–14.9: 3; above 15.0: 4
	per capita disposable income of migrant household/RMB	Below 10,000: 1; 10,001–19,999: 2; 20,000–29,999: 3; 30,000–49,999: 4; above 50,000: 5
Land and Housing Factors	Housing status	Allocated as the only housing: 1; with extra self-built or purchased housing: 2
	Per capita housing area/m2	Below 40: 1; 40–60: 2; 60–80: 3; above 80: 4
	Land dependence	Landless: 1; owned but uncultivated: 2; cultivating less than 30% of allocated land: 3; cultivating 31–99% of allocated land: 4; cultivating all allocated land: 5; renting additional land for agricultural activities: 6

summarizes the measurement variables for the five latent factors.

Because “residential continuity” cannot be directly coded, we derived a weighted index based on the main residence locations and travel durations of each household member.

Table 3. Weight coefficient of residential continuity index.

	Absence Time	No Out-Migration	Short-Term Out (Less than 3 Months per Year)	Intermediate Out (3–6 Months per Year)	Long-Term Out (6–12 Months per Year)	Permanent Out
Residence
Village/Community		1.0	0.8	0.6	0.4	0.2
County		0.9	0.7	0.5	0.3	0.1
Province		0.8	0.6	0.4	0.2	0
Across Provinces		0.7	0.5	0.3	0.1	0

presents the weighting coefficients.

3.3.2. Mediating Variable

In line with the multidimensional measurement of social integration in migration studies, a social interaction intensity index was devised, consisting of three types of social relations: geographical, home-tied, and carried-over. These represent connections newly formed in the host community, those maintained with non-relocated individuals in the place of origin, and those carried forward from co-relocated kin or fellow villagers, respectively. Considering the characteristics of social relations, the following three indices are selected for description.

Number of contacts (s): Measures the size of available social capital, resonating with “bridging” and “bonding” social capital in network theory [80].

Interaction frequency (q): Indicates the regularity of daily interactions, representing the maturity of trust and communication channels.

Interaction depth (d): Ranges from routine daily exchanges to consultation, emotional support, and resource sharing, capturing the strength of relational support.

The overall intensity for each relation type j is calculated as:

SC_j = s_j × q_j × d_j ……

(1)

The specific assignment of the indicators is detailed in

Table 4. Design of social interaction indicators.

Measured Indicator	Coding Specification
Number of Contacts (s)	1: 1–2 persons; 2: 3–5 persons; 3: 6–10 persons; 4: 11–20 persons; 5: >20 persons
Interaction Frequency (q)	1/1: Nearly every day; 1/2: approximately once per week; 1/3: approximately once per month; 1/4: once every several months; 0: none
Interaction Depth (d)	0: None; 1: daily-life exchange, leisure, work-related contact; 2: consultation, emotional support; 3: borrowing money, handling affairs

. This multiplicative formulation accounts for the multiplicative effect among dimensions (i.e., if any single dimension approaches zero, the overall intensity is constrained), and aligns with social network research practices that combine “relationship quality × relationship strength” into a composite measure [81].

3.3.3. Endogenous Latent Variable

This study adopted place identity as the endogenous latent variable, assessed from three perspectives: group identity, permanent settlement intention, and expectations for children. Specifically, we focused on these three questions: “Do you consider yourself already a local resident?”, “Do you plan to return to your original home?”, and “Do you hope the next generation stays here for the long term?”. We used a five-point Likert scale, with 1 to 5 representing strongly disagree, disagree, neutral, agree, and strongly agree, respectively. Note that the data for “Do you plan to return to your original home?” are reverse-coded. These questions collectively captured whether development-induced migrants were willing to put down roots in the host area, thereby evaluating the extent to which place identity was formed.

4. Empirical Analysis

4.1. Overview of Migrants’ Basic Characteristics and Place Identity

We first conducted a descriptive statistical analysis of the different dimensions of place identity among Three Gorges Dam migrants (

Table 5. Scores comparison of different components in place identity.

Score	Share (%)
	Group Identity	Permanent Settlement Intention	Expectations for Children
1	4.53%	1.42%	1.42%
2	11.90%	2.27%	3.68%
3	22.38%	30.88%	29.46%
4	40.79%	46.46%	42.49%
5	20.40%	18.98%	22.95%
Mean	3.61	3.79	3.82

). Overall, the mean scores for group identity, permanent settlement intention, and expectations for children each exceeded the midpoint of 3, indicating that these rural migrants generally displayed moderate to high levels of acceptance and integration into their host communities. About 40% of respondents scored 4 on these dimensions, and more than 20% scored 5, suggesting that most held either positive or very positive attitudes toward their new environment, with only a small fraction reporting negative or neutral views.

Among the three dimensions, group identity presented the lowest mean score and likewise a relatively smaller proportion of 5-point responses. This suggests that, although many migrants achieved basic living stability, they had not fully come to see themselves as local residents and may still have felt somewhat distant from local culture and norms. Additionally, while the proportion scoring 1 or 2 was not large, it remained higher for group identity than for the other indicators, suggesting that changes in identity require more time for adaptation and transformation. On one hand, the involuntary nature of long-distance relocation left these migrants with less emotional attachment to their new surroundings; on the other hand, deep-rooted bonds to their original hometown, including customs, dialects, and social networks, continued to influence how their identity evolves.

Meanwhile, permanent settlement intention ranked in the middle but boasted a mean score near or over 3.7, indicating that migrants generally took a favorable stance on remaining in their new location. When focusing on those who marked 4 or 5, more than 60% of respondents deemed the local social and economic conditions conducive to long-term settlement. Contributing factors included government support, stable employment, and improved community facilities, all of which appeared to reinforce migrants’ determination to stay.

Lastly, expectations for children obtained the highest mean. Even if some migrants remained uncertain about their own sense of belonging, they tended to be more willing for the next generation to settle in the new community. This pattern is common in intergenerational migration contexts, given that children and young people often adapt more smoothly to local customs and norms, establishing stronger local networks and thus enhancing the family’s overall integration.

In sum, despite some lingering caution in group identity, migrants showed an active orientation in permanent settlement intention and especially in expectations for children, underscoring the key role of intergenerational perspectives in strengthening place identity. Simultaneously, this pattern reflects their recognition of the host area’s material and social environment, as well as the significance of generational continuity in fostering a sense of place.

4.2. Analysis of Factors Influencing Place Identity Among Migrants

4.2.1. SEM Validation

As shown in

Table 6. Estimation of the variables in the SEM model.

Factor	Variable	Unstandardized Loading	Standardized Loading	z	S.E.	p
Place Identity	Group identity	1.000	0.528	-	-	-
	Permanent settlement intention	0.456	0.330	2.667	0.171	0.008 ***
	Expectations for children	1.211	0.777	4.531	0.267	0.000 ***
Personal Characteristics	Gender	1	0.129	-	-	-
	Age	−0.772	−0.114	−0.825	0.937	0.41
	Education	16.503	0.932	1.102	14.977	0.271
	Occupation	−7.859	−0.556	−1.192	6.593	0.233
	Relation to migrants	2.233	0.468	2.143	1.042	0.032 **
Family Features	Household size	1.000	0.723	-	-	-
	Residential continuity	0.049	0.471	4.084	0.012	0.000 ***
	Proportion of employed family Members	0.874	0.689	5.693	0.153	0.000 ***
Socio-Economic Characteristics	Urban economic gap	1.000	0.909	-	-	-
	Per capita disposable income	0.100	0.766	0.530	0.188	0.006 ***
Land and Housing Factors	Housing status	1.000	1.000	-	-	-
	Per capita housing area	2.816	0.874	7.076	0.398	0.000 ***
	Land dependence	0.368	0.277	2.297	0.16	0.022 **
Social Relations	Number of total contacts	1.000	0.735	-	-	-
	Home-tied social relations	−0.060	−0.181	−1.678	0.036	0.093 *
	Geographical social relations	1.215	0.747	5.958	0.204	0.000 ***
	Carried-over social relations	0.115	0.463	4.297	0.027	0.000 ***

Note: *, **, and *** denote significance at the 0.1, 0.05, and 0.01 levels, respectively.

, Cronbach’s alpha coefficients for all six latent variables surpassed the accepted threshold of 0.70, indicating strong internal consistency. All latent constructs exhibited average variance extracted (AVE) above 0.50, demonstrating sound convergent validity, and the square root of each construct’s AVE was larger than the absolute value of its correlation with any other construct, thereby confirming discriminant validity. Applying maximum likelihood estimation (MLE) in Amos 28 to assess global model fit, we note that χ²/df = 1.821 (<3.0) and RMSEA = 0.053 (<0.06), GFI = 0.980 (>0.95), AGFI = 0.960 (>0.90), and CFI = 0.932, IFI = 0.932, NFI = 0.928 (all > 0.90) [82]. These results indicate that the model demonstrates excellent fit and adequately interprets the structural relationships among latent variables.

4.2.2. Measurement Model Path Analysis

Given the favorable model fit, we examined the standardized factor loadings and significance tests (Table 6) to determine which observed variables meaningfully indicated their underlying constructs. Figure 2 displays the measurement model results, where significant paths are shown in bold and nonsignificant ones are dashed.

Among personal characteristics, only gender and migrant generation emerged as significant, with migrant generation (coefficient 0.468) being more influential. This highlights the importance of generational differences in shaping place identity among development-induced migrants: first-generation migrants, with deeper emotional ties to their original homes, struggled more to adapt, whereas second- or later-generation migrants were more likely to internalize local culture and norms. Turning to family features, both the number of registered household members and residential continuity had significantly positive loadings (0.723 and 0.471), implying that larger co-residing households provided stronger internal support and reduced social isolation. In socioeconomic characteristics, the “urban economic gap” index stood out (0.909), reflecting how economic disparities between origin and destination either compounded or reduced barriers. Regarding land and housing factors, “housing status” (1.000) and “per capita housing area” (0.874) most directly affected migrants’ daily living standards. In social relations, geographical social relations (0.747) were the strongest, underscoring that newly formed local connections were vital for assimilation. In contrast, the negative factor loading (−0.181) for home-tied relations indicates that excessive dependence on origin-based networks may have hindered building new ties. Although “number of total contacts” loaded moderately onto the latent factor (0.735), the findings suggest that simply having more contacts did not necessarily lead to stronger place identity. The distinct types of social relations demonstrated more pronounced and direct effects. Finally, in place identity, expectations for children (0.777) explained the most variance, surpassing both group identity (0.528) and permanent settlement intention (0.330), thus highlighting how next-generation outcomes guide migrants’ sense of identity.

4.2.3. Paths Affecting Place Identity

Based on the standardized path coefficients in Figure 3, four of the five direct paths to place identity, which are personal characteristic (0.352), socioeconomic characteristics (0.608), land and housing factors (0.490), and social relations (0.768), were all significant, while the direct path from family features was not. Among these four, social relations showed the largest positive effect (0.768), corroborating the hypothesis that restructuring social relations forms the principal catalyst for place identity among development-induced migrants. Socioeconomic characteristics followed at 0.608, then land and housing factors at 0.490, while personal characteristic lagged at 0.352.

Although family features exhibited no significant direct effect, they still manifested indirectly (Section 4.2.4). Indeed, the ultimate key might lie in family co-residence enhancing migrants’ social networks. Consequently, policies simply promoting family togetherness alone may not directly raise place identity unless they also reinforce social relations or align with supportive socioeconomic conditions.

In addition, the distinct categories of social relations matter. Enhancing geographical ties to local residents and institutions yields a substantial boost to place identity by bridging cultural gaps and expanding resource access. Conversely, high-intensity home ties may slow the development of local community membership if migrants rely too heavily on support systems from their place of origin instead of establishing new local networks.

Land and housing factors positively affect place identity not only directly but also through partial mediation by social relations. Satisfactory housing conditions increase perceived comfort and “sunk cost”, further encouraging local involvement. Lastly, the effect of personal characteristics, though significant, is mainly driven by “migrant generation”, indicating that generational differences in adaptation outweigh typical demographic attributes such as age or education level.

4.2.4. The Mediating Role of Social Relations

To verify social relations’ mediating role in shaping place identity, we applied the bootstrap approach (2000 resamples, 95% confidence intervals) to multiple pathways (

Table 7. Mediation effect test.

Mediating Path	Effect Value	95% CI	p
Personal characteristic → social relations → place identity	−0.384	(−0.019, 0.031)	0.037
family features → social relations → place identity	0.024	(0.015, 0.073)	0.004 ***
socioeconomic characteristics → social relations → place identity	0.019	(0.027, 0.043)	0.017 **
land and housing factors → social relations → place identity	0.032	(0.003, 0.065)	0.006 ***

Note: ** and *** indicate significance at the 0.05 and 0.01 levels, respectively.

). Except for the link from personal characteristic to place identity, where no mediation occurred, social relations showed significant mediation in the other three paths, with the strongest indirect effect found in the link from land and housing factors to place identity. This implies that, for rural migrants, improved land or housing conditions often channel their positive impact on place identity primarily through social relations. Meanwhile, socioeconomic characteristics also showed partial mediation, whereas family features had a full mediating effect, implying that family-based drivers of place identity largely depend on social relations as a bridging mechanism.

Notably, the direct impact of personal characteristic on place identity remained significant, but no mediating role was found for social relations, indicating that the migrant generation in particular directly shaped migrants’ perceptions of identity without working through social relations. In contrast, family features did not affect place identity outright but functioned strongly through social relations, emphasizing that household composition and co-residence must reinforce social bonding networks to cultivate place identity. Meanwhile, socioeconomic characteristics and land/housing elements also partially mediated place identity, influencing it both directly and by stimulating broader social ties. Among these, land and housing factors displayed the largest indirect effect size, highlighting that stable housing, farmland, and other tangible resources not only elevate subjective well-being but also serve as catalysts for expanding local social networks, further strengthening place identity.

5. Conclusions and Discussion

Despite the growing concern about social reconstruction in development-induced migration projects, research on how migrants’ social relations shape place identity during forced relocation remains inadequate. This study categorized social relations into geographical, home-tied, and carried-over dimensions. By integrating these with personal traits, family features, socioeconomic features, and land–housing factors, it systematically explored the mechanisms through which they impacted place identity.

Forced displacement severs long-standing ties in rural “acquaintance societies”, creating a profound identity rupture. In this context, geographical social relations emerge as the catalyst for rebuilding belonging. Unlike voluntary movers, who often pre-plan networks at destination [37], development-induced migrants rely almost entirely on post-relocation, host-community engagement to reestablish emotional and practical support. Our mediation analysis confirms that without these new local ties, even favorable personal, family, or material conditions fail to translate into strong place identity—a dynamic unique to the forced context.

Comparative evidence underscores the distinctiveness of China’s centralized, state-led resettlement model. In the Lesotho Highlands and Manwan Dam projects, SIAs highlighted grassroots, community-driven initiatives as critical to social recovery, yet often lacking in top-down coordination [83]. Ghana’s Akosombo case similarly suffered from protracted isolation when planners prioritized technical compensation over social-network reconstruction [84]. By contrast, China’s model relocates entire villages as a whole, which ensures infrastructural parity but has historically underemphasized the importance of embedded social relations. Our findings reveal that only when state provision is paired with deliberate social-capital investment (e.g., facilitated introductions, community events) do geographical ties truly flourish, shortening the integration timeline [85].

Urban shantytown transformation projects share key social integration challenges with dam resettlement. In Shandong, China, courtyard demolition induced significant declines in residents’ neighborhood trust and routine exchanges, echoing the “acquaintance-society” disruptions observed among rural dam migrants [86]. Both contexts demonstrate that quantity of contacts is insufficient: quality and embeddedness of ties must be cultivated, lest displaced populations experience prolonged place-attachment gaps.

These findings suggest that successful resettlement depends on weaving social capital considerations into every step of the relocation process. First, formal housing and land allocations should be accompanied by structured social networks in which village committees pair newly arrived households with long-term residents, fostering the embedded ties that are most conducive to place identity. Secondly, because entire families integrate more smoothly than fragmented ones, relocation policies should prioritize whole-household moves. Thirdly, to sustain these new ties, local governments need to diversify interaction, such as regular cultural festivals and vocational-training workshops that bring migrants and host residents into regular contact. Finally, integration policies must be attuned to generational differences: first-generation migrants benefit from orientation courses that ease cultural unfamiliarity, whereas younger members require extracurricular opportunities that embed them in local schools and peer networks.

Although this research offers theoretical and methodological innovations, it has limitations. First, its sample concentrated on long-distance out-migrants from the Three Gorges, and findings may not generalize to other contexts of forced migration or distinct cultural settings. Second, the study used cross-sectional data, covering conditions 20 years post-relocation but lacking longitudinal insights into evolving psychosocial states. Future work might employ multiwave data or combined qualitative–quantitative approaches. Third, deeper psychological attributes, such as subjective well-being, trauma, and cultural conflict, were not fully explored here. Investigating these factors could refine our knowledge of forced migrants’ settlement experiences and strengthen the policy relevance of place identity research.