Exploring the dynamics of cross-boundary interactions in Qinglinkou, China: the perspective of networks of second-home owners

ABSTRACT Cross-boundary interactions between second-home owners and local are complex over time – networks form and evolve within second-home owners and between owners and locals, each with its deliberately selective inclusion and exclusion. However, little attention has been paid to this phenomenon in the literature. This study, based on social network analysis alongside qualitative interviews, explores the dynamics of interactions between second-home owners and locals by analysing the networks formed by second-home owners in Qinglinkou, China. The ways in which second-home owners maintain and strengthen pre-existing networks with other owners and forge new links with locals, shape the cross-boundary interactions between the two groups. This study contributes to understanding the dynamics of cross-boundary interactions that are interwoven into the networks that second-home owners establish and maintain over time, and offers additional insights into the fragility of integration and high risk of segregation between the two groups in second-home destinations.


Introduction
A major challenge in second-home destination management, is to facilitate cross-boundary interactions between second-home owners and local communities (Barnett, 2014;Volo, 2011).The two groups can become highly intertwined through many socio-economic activities (Chen, 2020;Hjalager et al., 2011), which results in the exchanges, interrelationships, integration and/or segregation between them (Kondo et al., 2012;Liu-Lastres & Cahyanto, 2021).But conflicts between secondhome owners and locals and social movements led by locals against owners are commonplace in many destinations (Gallent et al., 2023;Qeidari et al., 2023).Cross-boundary interactions are key to second-home destination management.
Cross-boundary interactions constitute a longstanding issue in second-home studies (Larsson & Müller, 2019;Raun et al., 2023;Stapa et al., 2013).Past research has focused on the outcome of cross-boundary interactions and how second-home owners and local communities perceive each other (Solana-Solana, 2010;Waller & Sharpley, 2018).Such work indicates, either implicitly or explicitly, that cross-border interactions are underpinned by the networks of owners in second-home destinations.Selective inclusion and exclusion of network actors by second-home owners can reflect how they manage relationships with other owners and locals in second-home destinations (Honkanen et al., 2016;Huijbens, 2012).Networks that second-home owners maintain and foster in destinations may facilitate cross-boundary interactions with local communities in a positive manner, or the opposite (Halfacree, 2012;Kondo et al., 2012;Wu & Gallent, 2021).However, past studies on cross-boundary interactions between second-home owners and local communities have paid scant attention to owners' networks in second-home destinations.
This study aims to explore the dynamics of interactions between second-home owners and locals by analyzing the networks of owners in second-home destinations.To the best of the authors' knowledge, no other study has investigated cross-boundary interactions from the perspectives of owners' networks.Networks can provide a microscopic perspective from which to examine 'what is going on' within the dynamics of cross-boundary interactions by exploring how owners manage their networks with other owners and locals.Such knowledge can provide a glimpse into how to better align the interests of all actors involved in second-home destination and improve destination management.

Networks of second-home owners and cross-boundary interactions
The networks of second-home owners in second-home destinations can basically be divided into two types: (1) those between second-home owners and locals; and (2) those among secondhome owners.Firstly, second-home owners can maintain family links with locals, which may be prevalent, especially in places with a long history of second-home tourism (Brida et al., 2009;Müller, 2007).These networks tend to be stable and long-lasting.Even if second-home owners do not have family links with local communities, they may travel to second-home destinations on a regular basis.This mobility provides owners with the opportunity to connect with local communities (Nordin & Marjavaara, 2012).Interactions mainly occur when second-home owners purchase local properties or consume hospitality services/products, which are always provided by locals (Rye, 2011).Second-home owners and locals may also form bonds through non-economic activities, such as engaging in voluntary or hobby-based activities (Chen & Chen, 2018).Secondly, secondhome owners are likely to have strong networks with other owners in second-home destinations.They may already know other owners before purchasing second homes.In some instances, they may be persuaded to rent or buy second homes by people who have already bought second homes there (Wu et al., 2018).These 'familiar' second-home owners often provide advice to potential buyers on purchasing and relocation and assist them in adapting to the new environment.In addition, second-home owners tend to build new relationships with other owners with the same socio-economic characteristics (Nouza et al., 2018).
Networks of second-home owners in second-home destinations can shed light on cross-boundary interactions.Strong networks between second-home owners and locals are highly likely to help introduce owners into wider local communities (Kietäväinen et al., 2016).Through these, secondhome owners may gain a sense of belonging to the place (Stedman, 2006), which may act as a catalyst for them to integrate into local communities.While existing studies have paid very little attention to networks among second-home owners (Niu & Gibson, 2020), they have offered two main arguments regarding cross-boundary interactions.Firstly, the existence of a growing secondhome owner group may result in fewer ties between owners and locals.The most extreme example of this is second-home enclaves where second-home owners only bond with other owners and engage in few interactions with locals other than those of an economic nature (Monterrubio et al., 2018).Secondly, locals often have negative perceptions of the growing number of second-home owners, because they may fine the 'invasion' of their communities objectionable (Hall & Müller, 2004;Lorenzen, 2021), which may affect their willingness to interact with secondhome owners (Qeidari et al., 2023).Both arguments emphasise that networks of second-home owners in second-home destinations can significantly affect cross-boundary interactions.However, the underlying logic of how second-home owners manage their networks in second-home destinations and how their networks mutually affect their behaviours in interacting with locals remains unclear.Key aspects of these networks can help to decipher cross-border interactions from the perspective of second-home owners.

Studies of networks of second-home owners in second-home destinations
The networks of second-home owners in second-home destinations consist of two key facets: (1) network actors (i.e.immediate or extended family, friends and neighbours) (Müller, 2007); and (2) relationships between network actors (i.e.second-home owners and local villagers are often likely to forge strong economic ties, but as second-home owner spend more time in the area, they may build deeper connections with locals that go beyond the economic realm by participating in community events or initiatives, as well as cultural exchanges; relationships between owners are often based on mutual support and information exchange) (Keogh et al., 2022;Miletić et al., 2018).Existing research on networks of second-home owners have given some discussion on network actors and relationships, but it remains lacking in two regards.
Firstly, past studies have primarily underscored the static nature of the networks that secondhome owners have in second-home destinations.These networks can be dynamic: different actors (second-home owners or locals) join in, their relationships then evolve, and new networks are formed (Dovidio et al., 2011;Jack et al., 2010).It is not uncommon for second-home owners to maintain pre-existing networks with locals or other owners before relocating to second-home destinations (Hui & Yu, 2009).These networks can be extended through nurturing weak relationships formed through intermediaries and through daily greetings and gatherings (Wu & Gallent, 2021).Pre-existing and extended networks of second-home owners in second-home destinations can shed light on the dynamics of these networks and thus of cross-boundary interactions.This requires researchers to explore more deeply than the static characteristics of networks of second-home owners to unpack the dynamics of cross-boundary interactions in second-home destinations.
Secondly, most second-home studies have adopted a qualitative approach to describe the networks of second-home owners (Gallent, 2015;Rinne et al., 2015;Rye, 2011).Such an approach can offer a detailed picture of the internal working structure of networks (Wu et al., 2021), but key elements of networks, including identifying network actors, differences between actors in terms of advantages or constraints arising from their position within networks and the relationships generated, remain quantitatively under-researched.Exploring networks in a quantitative manner can provide greater objectivity and accuracy (Rydin, 2013).Social Networks Analysis (SNA), one of the major approaches used to systematically analyse networks which is extensively used in the social sciences (Czernek-Marszałek, 2018;Romeiro & Costa, 2010), can provide more quantitative insights into which network actors are interconnected and how, and what kind of relationships actors maintain in terms of their positions.A mixed-methods approach, which integrates SNA with qualitative data, can bridge the systematic characteristic of owners' networks with qualitative interpretations that second-home owners and locals attach to owners' networks.
Next, the empirical part of this paper, using mixed-methods (SNA and qualitative methods), unpack the dynamics of cross-boundary interactions by analysing the networks of owners in second-home destinations.

Study context
Qinglinkou village is situated approximately 200 kilometres from Chengdu, the capital city of Sichuan province (see Figure 1).It emerges as a rural tourist destination that is popular among tourists from nearby cities, but since 2011, Qinglinkou has gradually become a favoured destination for second-home owners.Due to its historic buildings, convenient transport links and stunning scenery, many tourists make day trips to the village, some of whom have purchased houses in Qinglinkou as second homes.To date, 11 have been purchased as second homes out of a total of 68 houses in Qinglingkou's two main streets.The transition from tourists to local housing purchasers is a common driver that turns tourism destinations into second-home destinations (Tuulentie, 2007).
The evolving networks that second-home owners have in Qinglinkou are ideally suited to the purpose of this study.Firstly, all second-home owners have pre-existing networks with other owners or locals before relocating to Qinglinkou, which is suitable for understanding how second-home owners manage the pre-existing links and how the networks affect the behaviours of both groups during the stages of relocation and subsequent settlement.This is a key facet to understanding the dynamics of cross-boundary interactions between owners and local communities.Besides that, as observed in many second-home destinations, second-home owners often have some kind of pre-existing links to the destination, such as family roots (e.g.inheriting a family member's house and converting it into a second home, or purchasing a second home close to family members) (Brida et al., 2009;Flemsaeter & Setten, 2009;Hui & Yu, 2009;Müller, 2007;Qviström et al., 2016).It is also not uncommon for individuals to acquire a second home because friends or acquaintances already have a second home in the vicinity, such as in the case of Wexford, in Ireland (Quinn, 2004) and Sanya, in China (Wu et al., 2015).In other words, new second-home owners follow the herdone buys and others follow.The existence of strong or weak pre-existing ties between second-home owners and their destinations is prevalent in numerous second-home destinations across various settings and contexts.In this instance, the findings derived from Qinglinkou can be widely generalised.
Table 1 shows the pre-existing networks of second-home owners.Pre-existing networks largely underpin the successful transformation of these 11 s-home owners from tourists to housing purchasers.In 2012, S1 and his family went on a day trip to Qinglinkou and were impressed by the scenic landscapes, so decided to buy a house there as a second home.S2 also started out as a visitor to Qinglinkou but became a second home buyer in 2013.Since S1 and S2 have settled in Qinglinkou, they often invite their friends or relatives to visit their second homes, some of whom then became purchasers.Thus, these second home buyers are the friends or relatives of S1 and S2 and purchased local properties with their assistance, except for S6 who acquired property sales information from S11.In this way, all the second-home owners are part of this pre-existing network, and its key actors are S1 and S2.At this stage, apart from S1 and S2, interactions between the remaining owners are minimal or non-existent, but they can nonetheless engage with one another via S1 or S2.At the early stage of relocation, pre-existing networks dominate the interactions of second-home owners in Qinglinkou.
Secondly, second-home owners have strengthened their connections with other owners and established more links with locals.Second-home owners often stay in Qinglinkou for weekends or summer holidays.Their daily activities in Qinglinkou are recreationalthey often gather to catch up, dine out, or travel in groups to nearby villages and towns.Through engaging in these activities, second-home owners have become progressively more familiar with one another.Compared with the initial phase of relocation when they relied on S1 and S2 to make contact with other owners, they now forge strong and direct relationships with them.In addition, they have developed harmonious relationships with locals through recurrent visits.Second-home owners also actively participate in local affairs.Second-home owners have a WeChat 1 group through which they exchange their thoughts on local affairs.Second-home owners regularly share their opinions with the head of Qinglinkou village committee and of Qinglinkou village heritage bureau.
The generalisation of Qinglinkou in exploring the dynamics of cross-boundary interactions at second-home destinations can be twofold.The pre-existing connections among second-home owners and between owners and locals exemplify a prevailing trend across second-home destinations, where familial and social ties, to a certain extent, inform second-home acquisitiona trend that positions Qinglinkou within a wider and observable pattern.The social dynamics of the village, characterised by the reinforcement of pre-existing networks and the establishment of new local ties, reflects a fundamental aspect inherent in many second-home destinationsthe growing social cohesion between owners and the potential for them to integrate with local communities.The pre-existing and extended networks that second-home owners have in Qinglinkou render it an appropriate and generalisable choice for gaining insight into the social intricacies of crossboundary interactions in second-home destinations.

Methodology
To examine the dynamics of interactions between second-home owners and the local community, this study deployed a mixed-method approach to analyse the networks of second-home owners in Qinglinkou.It involved a sequential exploratory design comprising two phases of data collection: a quantitative tool (questionnaire surveys) followed by a qualitative method (semi-structured interviews).Questionnaire surveys were used to collect the data for the SNA to assess the networks of second-home owners in Qinglinkou.Semi-structured interviews were conducted to gain more insights into the internal working structures of networks identified through SNA, which mainly concern the logic behind the inclusion and exclusion of network actors and relationships between actors.A mixed-methods approach allowed the researchers to make full use of the advantages of quantitative and qualitative methods to acquire a more comprehensive and nuanced understanding of the dynamics of cross-boundary interactions in Qinglinkou.

Data collection
Fieldwork was undertaken in August 2019 and July 2021.Data were collected in two stages: the first stage involved a questionnaire survey, while the second stage comprised the semi-structured interviews.As the number of second-home owners (11 in total) in Qinglinkou is relatively small, we asked all of them to complete the questionnaire surveys.We were first introduced by the head of the Qinglinkou village committee to S1 and S2.The snowballing technique was successfully used to reach the remaining 9 s-home owners via S1 and S2.
The data for network analysis was collected via two different questionnaires.Questionnaire One was designed for second-home owners, while Questionnaire Two was for local residents.We first sent Questionnaire One to all the second-home owners in Qinglinkou and Questionnaire Two to all the locals mentioned by the second-home owners in Questionnaire One.Questionnaire One consisted of two parts.Respondents were asked to: (1) list second-home owners/locals in Qinglinkou whom they previously knew; and (2) list second-home owners/locals whom they got to know after relocating to Qinglinkou.Following the collection of Questionnaire One from the secondhome owners, Questionnaire Two was then distributed to locals mentioned in Questionnaire One.Questionnaire Two consisted of three parts.Respondents were asked to: (1) confirm their pre-existing links with the second-home owner; (2) confirm their extended links with the second-home owner; and (3) confirm whether there was a link with other locals mentioned in Questionnaire One.A total of 35 questionnaires were collected: 11 s-home owners filled in Questionnaire One, while 24 locals completed Questionnaire Two.
In-depth interviews were conducted with 11 s-home owners and 24 locals who completed the questionnaires mentioned above.Interviews were used to explore the internal working structure of pre-existing and extended networks.The interview scripts covered 5 aspects: (1) the history of Qinglinkou; (2) second-home development in Qinglinkou; (3) the daily activities engaged in by secondhome owners or locals in Qinglinkou; (4) interactions between these two groups; and (5) their perceptions of each other.Each interview lasted between 20 and 60 min.Confidentiality was maintained throughout, with respondents identified as either 'second-home owners' (labelled S1, S2, S3 …) or 'locals' (labelled L1, L2, L3 …).
This study draws on a sample of 11 s-home owners and 24 local villagers.The sample can be generalised in two ways.On the one hand, the characteristics of these 11 s-home owners are consistent with traditional perceptions of second-home owners that have been studied in many contexts, such as destinations in Europe and North America, with a long history of second-home development (Gallent et al., 2005).More specifically, the first is the occupations of the 11 s-home owners, which include teachers, doctors and business professionals.This suggests a strong middle-class profile, which is widely recognised as a feature of second-home owners.Secondly, second-home ownership is often linked to retirement or future retirement plans (Norris & Winston, 2010), and these 11 s-home owners are no exception.They are either nearing retirement or have already retired.They use rural housing in Qinglinkou as a second home primarily for the purpose of retirement.Furthermore, the pursuit of an idyllic lifestyle serves as another significant motivation behind the demand for second homes (Adamiak, 2016).The 11 s-home owners, all of whom live in cities, are dissatisfied with the fast pace of life, poor living conditions and environmental pollution in the metropolis.Thus, Qinglinkou is regarded a desirable place that allows these second-home owners to escape the rapid-pace of urban life and seek solace in a more tranquil setting.Thirdly, the 11 s-home owners usually spend time in their second homes to relax and enjoy their weekends or long vacations.Hence, the sample shows a high level of alignment with their fellow owners in many other contexts in terms of demographics and occupation, motivations for second-home ownership, and length of stay, and is therefore broadly generalisable.
On the other hand, the sample of 24 local villagers can be seen as generalisable because it offers a comprehensive overview of the specific types of locals who tend to establish connections with second-home owners and the channels through which these connections are formed.L1 and L2 are the locals, with whom second-home owners were acquainted before relocating to Qinglinkou.Second-home owners often maintain connections with these local acquaintances in second-home destinations (Wagner, 2014).L5, a member of the Qinglinkou village committee, is responsible for providing public services to second-home owners in Qinglinkou.Second-home owners usually prefer to maintain harmonious relationships with local authorities (Tuulentie & Kietäväinen, 2020).L6 is the owner of a local grocery shop, which illustrates how second-home owners typically form connections with local communities through the channel of local retailers (Wu & Gallent, 2021).The remaining villagers reside in close proximity to the second-home owners, which offers an opportunity to enhance cross-boundary connections through daily greetings and casual conversation over time (Volo, 2011).The foregoing discussion gives an indication of the historical, political, economic and geographical contextsthe most prevalent circumstances identified in previous studieswithin which local villagers can interact with second-home owners.On this basis, the survey sample of 24 local villagers is sufficient to provide a relatively comprehensive understanding of the main types of locals who are likely to interact with second-home owners, and the general nature of cross-border relationships.Thus, findings from the sample can be applied to a wide range of settings.

Data analysis
Two types of dataquestionnaire data and interview datawere collected for this study.The data analysis commenced by collating the database for the SNA.The questionnaire data were used as inputs for the SNA to identify pre-existing and extended networks of second-home owners in Qinglinkou.Each link between second-home owners/locals was assigned a value of either zero or one: zero indicates that there was no link between the two actors; one indicates that there was a link between them.
According to the questionnaires collected, 11 s-home owners and 2 locals were included in the preexisting networks that the second-home owners have at the early stage of relocation, while 11 s-home owners and 24 locals were involved in extended networks.Centrality refers to the control that one actor in the network may have over the flow of information, depending on their position in the network (Landherr et al., 2010).Centrality contributes to understanding who key network actors are and what kind of relationships they maintain.Three primary measures of centrality measuresdegree, closeness and betweennessare used to indicate the extent to which a network actor is directly connected to many others, is indirectly close to many others, and serves as a key broker between many other actors (Haythornthwaite, 1996).The measures are to unpack the systematic characteristic (network actors and relationships between actors) of owners' networks.More detailed explanations of each measure, together with the relevant equations, are shown in Table 2.
The second step involved analysing the interview data.All the interviews were transcribed and coded using Nvivo12.Thematic analysis helped identify key themes concerning how secondhome owners and locals interpret cross-boundary interactions using their own frames of reference.Thematic analysis revolved around the influx of second-home owners into Qinglinkou, the way in which second-home owners adapted to Qinglinkou, interactions among second-home owners, cross-boundary interactions between second-home owners and locals, and their perceptions of each other.Quotations were collected that illustrated both the positive and negative aspects relating to each broad theme.For this paper, we focused principally on these themes to draw out references to 'pre-existing networks, the influx of second-home owners and cross-boundary interactions' (concerning how pre-existing networks were utilised to facilitate this influx of second-home owners and the interactions generated) and 'extended networks and interaction dynamics' (concerning the relationships that second-home owners built with other locals, the extension of pre-existing networks between owners and the internal working structures of such relationships).
5. Pre-existing networks, the influx of second-home owners and cross-boundary interactions 5.1.The form and structure of pre-existing networks: social network analysis Network visualisation can convey meaningful information about the form and structure of networks.Figure 2 visualises the pre-existing networks of second-home owners in Qinglinkou.11 s-home owners and 2 locals are involved in these pre-existing networks.The most important network actors are S1 and S2, who were the first to purchase second homes in Qinglinkou.S1 purchased his second home in 2012, while S2 bought his home in 2013.Both play a crucial role in helping potential buyers acquire second homes, by providing information on local housing and coordinating house purchases.As depicted in Figure 2, only S1 and S2 had direct contact with locals at an early stage, whereas all the other owners, except for S6, only had indirect ties with locals via S1 and S2.
Table 2.The detailed explanation of each measure and relevant equations.

Measure
Concept Equation

Degree centrality
The degree centrality of a node is the number of edges connected to that node (Freeman, 1978).
Where I ij is the value of the tie between node i and node j (the value is either 0 or 1: I ij = 1 means there is a tie between node i and node j, I ij = 0 means there is no tie between them); n is the number of nodes (Freeman, 1978).

Closeness centrality
The closeness centrality of a node is the sum of the shortest distances between that node and other nodes in the network (Freeman, 1978).
Where d ij denotes the length of the shortest path between node i and node j, and n is the number of nodes (Freeman, 1978).

Betweenness centrality
The betweenness centrality for each node is the number of these shortest paths that pass through that node (Freeman, 1978).
Where g jk is the number of shortest paths necessary to reach node k from node j; g jk (i) is the number of shortcuts from node j and node k, which also cross point i; n is the number of nodes (Freeman, 1978).According to the interviews with the second-home owners, S3, S4, S8, S9 and S10, aided by S1, obtained information about house sales from S1 and concluded the purchase of a home, while S6, S7 and S11 were primarily assisted by S2.Without making direct contact with S1 or S2, S6 was already acquainted with S11.This suggests that pre-existing networks play a key role in secondhome relocation, which is consistent with other forms of migration mobility (Wu et al., 2018).
Although second-home owners have pre-existing networks with locals, these links are somewhat fragile as the communities of second-home owners and locals are introverted, as supported by Granovetter's (1973) claim that interconnectors often take the form of 'weak ties'.Table 3 demonstrates the centrality rankings of the pre-existing networks.S1 and S2 have the highest degree and betweenness centrality, demonstrating that S1 and S2 are the most connected within the pre-existing networks and have the most frequent contact with others.S1 also has the smallest degree of closeness, implying that S1 has the shortest distance from all other actors and thus maintains the closest relationships with other actors in this network.The aforementioned centrality measures suggest that S1 and S2 are situated in the most central position and play a crucial role in maintaining the pre-existing network.Access to information about second-home destinations through acquaintances is crucial in achieving second-home relocation (Abbasian & Müller, 2019).Both S1 and S2 play an important part in establishing connections between other owners and locals and helping them with home acquisitions and renovations, subsequent relocations and adapting to the new environment.Second-home relocation can be made easier by preexisting networks, as key network actors make the move feasible.Although S1 and S2 bonded weakly with locals, they occupy powerful positions within the pre-existing networks, being located between groups (Gallent, 2014).However, the network analysis does not provide much insight into how pre-existing networks function to facilitate the relocation of second-home owners, nor does it provide a more in-depth look at the cross-boundary interactions between network actors.Therefore, it is necessary to examine the interview data in order to obtain these kinds of additional insights.

Pre-existing networks and cross-boundary interactions: second-home relocation
S1 and S2 were the earliest second-home owners in Qinglinkou.Their move into Qinglinkou resulted from their existing connections with the place and the local community.About 40 years previously, S1's father worked in a nearby village, while S2's mother taught at the local elementary school.Thus, both S1 and S2's parents already had ties to Qinglinkou.When S1 and his family visited Qinglinkou in 2012, he recalled visiting the Qinglinkou fair as a youngster and purchasing tofu from L2's shop.This childhood experience prompted S1 to purchase a local residence for his retirement (Wagner, 2014).He asked locals if they were interested in selling their homes, and he eventually found one to purchase.In 2013, S2 and his family also spent a weekend in Qinglinkou.The landscape impressed S2 and his family so much that they decided to purchase a home there.They unexpectedly encountered L1, a student of S2's mother.With the assistance of L1, S2 bought a local house as his second home in Qinglinkou.
We spend a weekend in Qinglinkou, and now we have finally become residents.My mum had wonderful memories of this village, and we loved the local environment.We have a magical bond with Qinglinkou.(S2, secondhome owner, August 2019) S1 and S2 and their networks drove the growth in second homes in Qinglinkou.Both frequently invited friends and family to spend weekends at their homes.These informal gatherings allowed these visitors to gain a greater understanding of Qinglinkou.Seduced by the idyllic lifestyle, some of them attempted to purchase local houses as second homes, but without many local links, potential purchasers primarily obtained housing information via S1 and S2, including information about which houses were available for purchase, the pros and cons of those houses, and the price they could expect to pay.In some instances, they participated in housing transactions, such as negotiations with local property owners.Thus, these potential buyers became the owners of second homes in Qinglinkou, as S4 explained: S1 kept an eye on the housing information in Qinglinkou for us … After purchasing the house, we always asked S1 for advice about how to renovate our house.He introduced us to a great renovation studio and told us where we could buy renovation materials.I may not have bought a house in Qinglinkou without S1's help.(S4, secondhome owner, July 2021) Following the acquisition of a second home, these pre-existing networks remained functional.New owners often need to deal with housing issues, including recruiting a refurbishment team and purchasing interior decoration materials.Due to their lack of local ties, they relied heavily on the advice given by S1 and S2 to address these issues.Many of the second-home owner interviewees emphasised the significance of maintaining contact with S1 or S2 to gain a greater sense of security and belonging in the early stages of relocation (Stedman, 2006).Despite the increased affinity between second-home owners, they were comparatively isolated from the local community.Other locals, excluding L1 and L2 who knew S1 and S2, acted rather indifferently towards these newcomers.Locals described their initial interactions with them as being oriented by economic interests, revolving around housing transactions, home renovations and the purchase of groceries.Their interactions were business-oriented and rarely led to deeper connections.L4, a construction worker, described his initial encounter with a second-home owner: I was hired by one of the second-home owners for housing renovations.They were very friendly and generous.They were strict about the renovation quality.We knew each other, but I would describe them as acquaintances.Basically, these second-home owners and I are just employers and employees.(L4, local, August 2019) The roles played by locals as service providers and second-home owners as recipients may make it difficult for them to establish closer bonding (Hannonen et al., 2015).In addition to these intangible differences, tangible differences in all aspects of life acted as a barrier to cross-boundary interactions between second-home owners and locals.Second-home owners lived in quite luxurious dwellings with ostentatious interior and exterior decoration, in stark contrast to the traditional local dwellings.This may affect the extent to which locals were willing to bond with second-home owners at the early stage of second-home relocation.Second-home owners also noticed unexpected indifference from locals at this stage and felt excluded from the local community and its practices.
6. Extended networks and cross-boundary interactions: the coexistence of community integration and segregation 6.1.The form and structure of extended networks: social network analysis L5 is one of the Qinglinkou village committee members and is acquainted with every second-home owner, while L6 is the owner of a local grocery shop and has some business contact with the second-home owners and other locals.S1 and S2 continue to play a crucial role in connecting locals and other owners to the extended networks.Figure 3 indicates that locals are now significantly more connected to second-home owners than they were previously, suggesting that there is potential for the enhancement of relationships between second-home owners and locals (Tuulentie & Kietäväinen, 2020).
Table 4 shows the results of the centrality measures of the extended networks.L5 has the highest degree and betweenness centrality, followed by L6.This indicates that both these locals have extensive contact with other actors and perform a connecting role between them.Their professions, which afford them more opportunities to interact with other actors, could be one possible explanation for this.Both L5 and L6 have the lowest closeness centrality, indicating that they are centrally located in the extended networks.The degree and closeness centrality of second-home owners in the extended networks are much lower than those of the locals.As second-home owners have predominantly strong relationships with other owners and a limited number of specific locals, it is not surprising that they are at a disadvantage when it comes to forming relationships with other actors, particularly locals.This helps to explain the isolated networks that second-home owners have in many areas (Wu et al., 2015).S1 and S2 rank fourth and fifth in terms of betweenness centrality, largely resulting from their crucial positions in establishing and maintaining relationships with other owners and locals.In contrast to pre-existing networks, a greater number of locals play a significant role in the extended networks and have increased contact with second-home owners.
Overall, the above results suggest that compared with the pre-existing networks, the extended networks have undergone substantial growth over time.Second-home owners were initially less connected with locals, but cross-boundary interactions are enhanced over time, which is not always adequately documented in existing studies.S1 and S2 continue to facilitate connections with other second-home owners and locals, but their significance has been challenged at this stage.L5 and L6 are interconnectors between the communities of second-home owners and locals, albeit that they have weak bonds with second-home owners.However, it remains unclear how extended networks are developed and how locals become increasingly incorporated into them, a subject which is discussed in the next section.

Establishing relationships between second-home owners and locals
Qinglinkou is a small village where second-home owners and locals frequently greet each other on a daily basis, although they did not previously know one another.As a result of recurrent visits and increased encounters, cross-boundary interactions become extended from the economic to the social spheres.Some second-home owners sometimes buy gifts for locals with whom they maintain good relationships.In exchange, locals frequently gift them locally grown agricultural produce.Although these items are inexpensive, these acts of kindness contributes to strengthening their relationships.Several second-home owners employ L3, a local villager, as their housekeeper.He looks after their front door keys and contacts them if any housing issues need to be resolved.He emphasised that he does this voluntarily: It is not a big problem for me.The only thing I do is to check on their house occasionally and call them if needed.Second-home owners are always nice.Every time they come back to Qinglinkou, they bring some precious gifts when they visit me.(L3, local, August 2019) Over time, local villagers have gradually regarded second-home owners as an integral part of the rural community and become more willing to integrate the owners into some aspects of community life.Locals occasionally seek advice from second-home owners, especially in matters about which they think that these owners may be more knowledgeable.For example, local villager L5 sought advice from S2 on purchasing property in Chengdu.In addition, it is not uncommon for locals to invite second-home owners to participate in community events and activities in Qinglinkou, such as weddings and funerals.Second-home owners, who place considerable value on nurturing a harmonious relationship with local villagers, also show a keen interest and willingness to engage in these events, as S1 described in the following excerpt: If they invite us to their events, we are happy to attend.We also need to send some gifts if we are not in Qinglinkou as an apology for not being able to attend the event.(S1, second-home owner, August 2019) Activities such as joining in with local events help to facilitate the integration of second-home owners into the local community.This phenomenon has also been observed in numerous second-home destinations, as has the common thread of the outcomes of such cross-boundary interactionsfostering deeper connections between second-home owners and the local community (Robertsson & Marjavaara, 2015).

Enhancing the pre-existing networks among second-home owners
Qinglinkou plays host to second-home owners, whose presence is often short-term, seasonal or periodic.These forms of 'periodic absence', to some extent, enable social ties between owners and locals to remain relatively weak, although they attempt to establish and sustain harmonious relationships.The main interactions that second-home owners have in Qinglinkou are with other owners who they share many similar characteristics with.At the early stage of relocation, most secondhome owners only knew S1 or S2, but they quickly became acquainted with one another.This resulted primarily from the frequent gatherings organised by S1 and S2.The houses belonging to S1 and S2 serve as 'social' spaces for other owners.Second-home owners often gather to play mahjong.In addition, some traditional leisure activities popular among second-home owners, such as going fishing or taking day trips to nearby destinations (Naess et al., 2019), are frequently organised in Qinglinkou.Such social and recreational activities significantly contribute to binding second-home owners together, as evidenced by S2: In Qinglinkou, we often hang out together.We may have known each other before, but not that well.We are always busy with many recreational activities in Qinglinkou, which keeps us closely bonded.(S2, secondhome owner, August 2019) Extending pre-existing networks among second-home owners helps to enhance their sense of belonging and attachment to Qinglinkou.Yet, multiple and extensive interactions among owners may exacerbate their detachment from the local community.Numerous locals identified class and lifestyle disparities with affluent second-home owners who appear to only engage in leisure activities and do not need to work.In contrast, the daily routines of locals include earning a livelihood, with few scheduled recreational activities: Their (second-home owners') parties usually last until midnight, but we often go to bed quite early.The noise they make is very annoying.Many residents have reported it to the village committee several times.(L7, local, July 2021) The increasing number of second-home owners pose a threat to the fabric of the local community and amplifies its heterogeneity.For instance, locals who maintain close relationships with secondhome owners may be treated with contempt by other locals who refer to them as 'servants'.It is not uncommon for locals to engage with second-home owners selectively, focusing on activities from which they may directly or indirectly benefit.Otherwise, it is rare to see them interact with each other.L6 described his contact with second-home owners as follows: Many second-home owners often buy groceries from my shop.We know each other and have casual chat, but this is mainly because they are my customers, and we have little personal bonding.(L6, local, July 2021) Unsurprisingly, as many second-home owners described, implicit boundaries have been increasingly reinforced in Qinglinkou, which can create a degree of segregation between them and the local community (Wu & Gallent, 2021).Second-home owners are also in the equivocal position of being both insiders and outsiders: they are physically and economically connected to Qinglinkou, but spiritually separated from the local community.

Discussion and conclusions
This study unpacked the dynamics of cross-boundary interactions by analysing the networks of second-home owners in Qinglinkou.The dynamics of cross-boundary interactions between second-home owners and locals become complex over time, with networks forming and evolving among second-home owners and between owners and locals, each involving deliberately selective inclusion and exclusion.Pre-existing and extended networks of second-home owners in Qinglinkou work together to shape cross-boundary interactions.Pre-existing networks facilitate the relocation and arrival of second-home owners and their adaptation to Qinglinkou (Nefedova et al., 2016).S1 and S2 assist other owners with their relocation.The chain of movement to Qinglinkou, with new owners following the herd, accords with the traditional narrative of migration patterns characterised as: 'one moves, others follow' (Choldin, 1973, p. 166).However, pre-existing links within second-home owners are relatively introvertedtwo owners act as interconnectors between second-home owners and locals.In this instance, such weak links challenge the significance of pre-existing links in deepening these cross-boundary interactions (Brida et al., 2009;Larsson & Müller, 2019).
The extended networks that second-home owners have in Qinglinkou include the relationships they have developed with locals and the strengthening of those with other owners.Although second-home owners are less connected to locals, their relationship has improved over timemore connections have been formed between the two groups.S1 and S2 remain the key actors in the extended networks as in the pre-existing networks, but L5 and L6 occupy a central position in the extended networks.Through boundary-crossing and inclusive practices, cross-boundary interactions are facilitated in Qinglinkou through 'past dense person-to-person interactions overlaid into more complex, extended and open networks' (Gallent, 2014, p. 179).Beyond this, the study also suggests that the relationship between the two groups is based on loose acquaintances rather than close community bonds.Thus, it is questionable whether the network practices described above are sufficient to sustain an enduring cross-boundary relationship.
Although a harmonious relationship between second-home owners and locals in Qinglinkou has developed, interactions among second-home owners remain the primary constituent of the social life of the latter.Due to the recurrent visits of the second-home owners, more new contacts are established between them.Closer, more functional and advantageous friendships are established through gatherings involving second-home owners.S1 and S2 remain the most important actors in bonding with other owners and enhancing the pre-existing networks, as the second-home owner community progressively coalesces.This phenomenon, characterised by a rapidly expanding and increasingly interconnected group of second-home owners, is not uncommon in many secondhome destinations (Kim, 2018;Müller, 2007).The increase in the number of second-home owners makes locals feel threatened and slightly alters local attitudes toward second-home owners.The outcome is a reluctance to further bond with second-home owners and an enhancement of the implicit boundaries that are identified by both groups.This finding illustrates the fragility of the cross-boundary relationship between second-home owners and locals (Armstrong & Stedman, 2013).
Although this study draws on Qinglinkou to depict the dynamics of cross-boundary interactions by analysing the networks of second-home owners, the findings derived can be generalised in two main ways to enhance the existing literature on second homes.Firstly, this study examines the inherent logic of how second-home owners manage their networks in second-home destinations over time, which addresses the first research gap.While second-home owners in Qinglinkou possess pre-existing links with either locals or other owners before relocation, they gradually develop new links with the local community through daily greetings and participating in local events, as well as creating a more connected community of second-home owners via recurring visits and gatherings.These growing networks can also be observed in numerous second-home destinations (Nordin & Marjavaara, 2012;Wu & Gallent, 2021).The networks that second-home owners form and develop with other owners and locals offer a micro perspective that can be used to gain further insight into the dynamics of cross-border interactions in second-home destinations.Secondly, this study suggests that second-home owners tend to form close and strong links with other owners, and maintain loose and weak ties with a very small number of locals; thus, this results in increasing social segregation and weakening social integration between the two groups.The growing community of second-home owners, as well as the corresponding changes in local perceptions of an increase in second homes, are accompanied by deliberately selective inclusion and exclusion by both second-home owners and locals.This corresponds to the second research gap.This finding can be generalised to illustrate and understand a challenge that is prevalent in many second-home destinationsthe frequently observed social segregation between second-home owners and local communities (Mamonova & Sutherland, 2015;Van Noorloos & Steel, 2016).Thus, the research findings can be widely applied.
This study makes three main contributions.Firstly, it applies a mixed-methods approach to explore the networks of second-home owners in Qinglinkou, thereby offering comprehensive insights into not only the systematic characteristics of second-home owners' networks, but also the internal working structure of these networks (Wu et al., 2021).The second contribution involves revealing the underlying dynamics of cross-boundary interactions.The increasing numbers of encounters with locals and the acceptance of local customs contribute to crossboundary bonding, but when pre-existing networks between second-home owners grow stronger, their ties with locals weaken to some extent.As second-home owners continue to periodically reside in Qinglinkou, their interactions with locals may become dynamic over time.This broadens the previous understanding of cross-boundary interactions in second-home destinations as static (Overvåg & Berg, 2011).Thirdly, this study argues that the complexity of crossborder interactions is manifested in the coexistence of segregation and integration between second-home owners and locals.Although both groups become more connected over time, socio-economic differences between them still permeate all aspects of life: class, identity and socio-economic status result in integration being fragile and a high risk of segregation (Barnett, 2014;Rye, 2011).
The characteristics of cross-boundary interactions between the two groups, how second-home owners manage their networks in importing regions, and the segregation and integration generated can provide insights into the management of second-home destinations and guidance for policymakers.The study has some limitations and two avenues are suggested for future research.Firstly, as many second-home studies focus on the interaction outcomes, it could be worth investigating the motivations behind the interactions between the two groups.Secondly, this study focuses explicitly on the interaction dynamics at second-home destinations through a network lens.Scant attention has been paid to the impacts on the local community caused by those interactions.Further research could thus be undertaken into the nexus between such interactions and local impacts.

Figure 2 .
Figure 2. Visual representation of pre-existing networks.

Figure 3
Figure 3 depicts a visualisation of the extended networks that second-home owners have developed in Qinglinkou.The most connected actors in the extended networks are L5, L6, S1 and S2.

Figure 3 .
Figure 3. Visual representation of extended networks.

Table 1 .
Pre-existing links between second-home owners in Qinglinkou.

Table 3 .
Centrality measures of pre-existing networks.

Table 4 .
Centrality measures of extended networks.