Your heart is where your roots are? Place attachment and belonging among Polish and Lithuanian returnees

detailed Abstract The reintegration of return migrants has been an important issue in migration studies for several decades. While much research has been done to identify returnees’ strate‑ gies and their labour market situation in their country of origin, little has been devoted to their attachment, especially in quantitative studies. This paper seeks to address this gap, analysing predictors of place attachment and belonging among Polish and Lithuanian returnees from the United Kingdom. We consider autobiographi‑ cal factors connected to migration history and time spent in a place; relational factors linked to social networks, bonds and contacts; cultural factors connected to feeling “at home”; and economic and welfare factors primarily linked to economic stability and job opportunities. Our analysis is based on a web survey of Polish and Lithuanian returnees conducted in 2020 (n = 740). The results reveal that the influence of autobio‑ graphical, cultural and relational determinants on returnees’ attachment is greater than the influence of their labour market position.

examination of the factors that influence returnees' attachment to their places of residence in Poland and Lithuania, based on an online comparative survey of 740 Polish and Lithuanian returnees conducted in 2020. We assume that attachment to one's locality is an important aspect of reintegration, and seek to shed new light on this process.
While the notion of place attachment-as well as the similar concepts of embedding (Ryan, 2018;Ryan & Mulholland, 2015) and anchoring (b; Grzymala-Kazłowska & Phillimore, 2018; Grzymala-Kazłowska, 2018a)-has often been examined in the context of international mobility (Botterill & Hancock, 2019;Rishbeth & Powell, 2013;Trąbka, 2019), including forced migration (Boğaç, 2009;Fullilove, 2014) and job-related mobility of highly skilled professionals (Mulholland & Ryan, 2014;Piekut, 2013), it has not been thoroughly analysed with reference to return migration. This article aims to address this gap by examining which factors facilitate and constrain returnees' attachment to their place of re-settlement after return. In order to explore this issue, we take into account various economic, social and cultural variables. We argue that the process of establishing or re-establishing bonds with one's place of residence may proceed differently in different life domains and thus we suggest that these variables may be grouped thematically.
This article consists of five parts. Following this introduction, we discuss the theoretical framework to conceptualise our leading concepts of attachment and belonging and present the factors that we hypothesise would be connected with them. After that, we describe a brief outline of return mobility to our case countries, i.e. Poland and Lithuania. We then discuss our methodology, explaining the data collection process, the sample and our analytical strategy. The results of the linear regression model are then presented, followed by the discussion and general conclusions.

Theoretical framework: correlates and predictors of place attachment
Though they have received much attention in recent migration studies, the notion of place attachment and the similar concept of belonging (Antonsich, 2010;Yuval-Davis, 2006) remain somewhat fuzzy concepts that lack operational definition. The former is often broadly defined as "the bonding of people to all kinds of places at various scales" (Di Masso et al., 2019: 126), while the latter has been described as "[the] personal, intimate, feeling of being 'at home' in a place (place-belongingness)" (Antonsich, 2010: 645;Yuval-Davis, 2006: 197) or as " [an] individual's embeddedness in a collective" (Pfaff-Czarnecka and Toffin, 2011, 3). Given these definitions, attachment and belonging appear closely related and are sometimes used as synonyms. For instance, Arhin-Sam (2019: 37), who researched return migration to Ghana, claims that "belonging among returnees can…be perceived as the 'desire for attachments, be it to people, places, or modes of beings. '" In this paper, we speak about attachment to a place of residence, but we assume that this notion is not limited to bonds with a physical space. On the contrary, the ties we have with a particular place are connected to the social relations we have there, to our personal history in that place and memories collected there. In this sense, the notion of attachment is similar to placebelongingness as defined by Antonsich, and therefore we found his framework useful in the analysis of returnees' attachment. Moreover, we believe that the latter constitutes a dimension of the even broader notion of re(integration). While the majority of studies on the reintegration of returnees have so far been focused on the economic dimension and labour market performance, we argue that exploring its understudied emotional aspects, namely attachment, will complement the existing knowledge.
The different concepts describing the ties of migrants with their places of residence highlight their differentiated, segmented and multifaceted character (Antonsich, 2010;Lewicka, 2011;Ryan, 2018). With reference to place attachment, for example, Williams and Vaske (2003) differentiate between place dependence and place identity. While the former is based on more instrumental, functional bonds with a particular place, often resulting from the various opportunities and affordances that it offers, place identity represents deeper and more emotional ties with a place. With reference to belonging, Yuval-Davis (2006) described the politics of belonging and personal belonging, and the latter is conceptually closer to the notion of attachment. Based on her analytical scheme, Antonsich (2010) put forward a framework of different factors influencing place-belongingness and politics of belonging. Specifically, he suggested that analyses of belonging should take into account five groups of factors: autobiographical, relational, cultural, economic and legal.
Autobiographical factors relate to an individual's personal history in a given place, the memories and experiences that create their attachment to that place (ibid.). This is in line with numerous studies on place attachment which have demonstrated that time spent in a certain locality and having personal as well as family history in a locality are one of the strongest predictors of place-belonging (Hay, 1998;Lewicka, 2011). For returnees, their attachment can be enhanced or reduced depending on whether they were prepared for return (Cassarino, 2004) and the amount of time spent abroad (King, 2015). Therefore, we assume that time spent abroad (Hypothesis 1, H1), time that has passed since the return (H2) affect attachment upon return. Moreover, many CEE migrants led a transnational lifestyle, often visiting families and friends in their places of origin (Bell & Bivand-Erdal, 2015). Such occasional visits increase the likelihood that migrants will be well prepared for their return and smooth the way for the creation or recreation of a sense of attachment. For this reason we presuppose that the choice to return to one's previous place of residence (and not to some other locality in Poland or Lithuania) (H3) is connected with attachment.
The second group of factors that Antonsich (2010) lists, relational factors, concern people's social networks in a given place. Previous studies on place attachment prove that social relations are indeed among the strongest predictors of attachment. They include both strong ties (e.g. family and friends) as well as weak ties, such as neighbours or acquaintances from a local community (Granovetter, 1973). Where return migrants are concerned, as mentioned above, the emergence or re-emergence of attachment also depends on the ties they maintained with their country of origin during their time abroad, i.e. transnational practices (White, 2014). Furthermore, reintegration is not exclusively an individual process: for example, the educational performance of children (Vathi et al., 2016) and the sense of attachment of other family members can largely impact one's reintegration and belonging. Having in mind the role of relational factors, we assume that people who returned for family reasons will be more attached to their places of residence (H4), while those who lack social networks (H5) will be less attached to their localities.
In that context, it is worth noting that previous studies on return motivations unanimously suggest that cultural and emotional factors-e.g. missing one's country of origin or not feeling at home abroad-play a crucial role in returnees' decision to return (Czeranowska & Wermińska-Wiśnicka, 2021;Dzięglewski, 2020;Holmes & Burrows, 2012;Slany & Ślusarczyk, 2010). For instance, Holmes and Burrows (2012) apply the notion of emotional reflexivity to emphasise that in their decisions to stay or return migrants consider particularly familial and emotional factors. Apart from the willingness or obligation to be close to family members in the country of origin, they mention homesickness and a search for a feeling of belonging. This corresponds with the third dimension of Antonsich's framework: cultural factors. We therefore expect that people who moved back for such nostalgic reasons will be more attached to their place of residence upon return (H6). On the other hand, bearing in mind results indicating that after spending a significant period of time abroad many returnees adopt a more distanced and critical perspective on cultural practices in their country of origin (Dzięglewski, 2020), we expect that feeling different from people in the local community after return (H7) will hinder return migrants' attachment.
When considering the economic aspects of returnees' belonging-the fourth group of factors linked to belonging listed by Antonsich (2010), we should take into account not only income, but also a broader sense of stability and opportunities. For example, White (2014: 35) showed that returnees to Poland were "mostly integrated emotionally and culturally into Polish society", but "not well integrated into the Polish labour market". Failure to reintegrate into the country of origin-especially economically-can motivate "double return" to the migration country (White, 2014), i.e. re-settlement abroad and new transnational strategies, or result in a "trap of migration loop" for returnees that are more vulnerable to skill waste and other negative effects from migration experience and thus find it hard to get a satisfactory job both abroad and in their country of origin (Iglicka, 2010). Along the same lines, a survey of returnees in Lithuania revealed a number of difficulties that many returnees experience after coming back: finding a well-paid job, utilising skills acquired abroad, and readjusting to Lithuanian social norms (Žvalionytė, 2012). Thus, we assume that employment status will also affect attachment (H8). To date, no research has been done to determine how economic factors affect whether Lithuanian returnees feel attached to their place of residence after return.
Economic stability is also connected with housing conditions. It has been demonstrated that migrants, especially those who maintain transnational ties with their countries of origin, often invest in property there (Boccagni, 2020), and that owning a house or apartment is an important pull factor for returnees. Correspondingly, living in an owned house (as opposed to a rental home) is significantly correlated with place attachment (Lewicka, 2011). We therefore suppose that owning a property in one's place of residence is a predictor of attachment (H9). Lastly, in qualitative studies (Dzięglewski, 2020) returnees often mention the low level of social services and benefits compared to their countries of emigration, which may contribute to their sense of insecurity. Therefore, we hypothesise that the evaluation of social services and welfare state institutions will also affect attachment (H10).
Antonsich (2010) refers to citizenship and residence permits as legal factors that influence migrants' belonging. These statuses afford a sense of security by ensuring access to certain rights such as social benefits. We argue that these factors are less relevant in the context of return migration to Lithuania and Poland because the vast majority of returnees have maintained the citizenship of their country of origin while living abroad and have access to welfare services and benefits. Therefore, our analysis excludes such legal aspects.
Likewise, this article focuses on returnees' sense of attachment to their places of residence, and not on the politics of belonging described by Yuval-Davis (2006) and Antonsich (2010). While we acknowledge the critical role of the discursive practices of inclusion/exclusion and the relations of power in maintaining boundaries between those who "deserve" to belong and those who do not, we assume that in the case of returnees it is crucial to gain an understanding of how their intimate attachment emerges or, as the case may be, re-emerges.
Based on the framework we have described and the results of previous studies on place-belonging, attachment and return migration, we assume a number of hypotheses, grouped thematically below. This approach may shed new light on return migration, as the majority of literature on the subject has focused on motivations for return and types of return (Cerase, 1974;Ghosh, 2000), often providing only holistic descriptions of the re-entry into the society of origin, without taking into account the various processes taking place in different spheres of life. Below, we summarise our theorydriven hypotheses, clustered into four groups of factors: (1) autobiographical factors; (2) relational factors; (3) cultural factors and (4) economic and welfare factors.
Firstly, in the group of autobiographical factors we assume that:

H1.
Length of stay abroad is negatively correlated with attachment to one's place of residence after return. H2. The more years have passed since their return, the more returnees feel attached to their place of residence. H3. Returning to one's previous place of residence (not another locality in Poland or Lithuania) is positively correlated with one's attachment after return.
Secondly, in the group of relational factors we presuppose that:

H4.
A lack of social networks in one's place of residence upon return is negatively correlated with attachment to the place of residence. H5. Migrants who returned because of family reasons are more likely to score higher on attachment to their place of residence.
Thirdly, in the group of cultural factors, we presume that: H6. Nostalgia for one's country of origin is positively correlated with attachment after return. H7. Feeling different from others after return is correlated with lower attachment.
Finally, in the domain of economic and welfare factors we hypothesise that: H8. Failure to integrate economically (unemployment) is negatively correlated with level of attachment. H9. Owning property in one's place of residence in Lithuania/Poland is a predictor of attachment. H10. Preferring health and education services/benefits in one's home country (in comparison with the UK) is positively correlated with the attachment.
In the subsequent sections of this article, we operationalise and test these hypotheses, assessing which of the factors outlined in this conceptual framework have the strongest influence on Polish and Lithuanian returnees' attachment to their place of residence following return from the UK. We believe the novelty of our approach lies in the fact that this conceptual framework has only been used in qualitative studies, determining only which factors are important for re-integration, but not which are most important and what is the strength of their impact.

Research context: return mobility to Poland and Lithuania
The massive post-accession outflow of migrants from CEE countries and the economic crisis several years later sparked the interest of researchers and policy makers in migrants' intentions to settle abroad or return to their home countries. The number of returnees to Poland rose in the years following Poland's accession to the EU, peaking after the economic crisis of 2008 that hit the UK and Ireland particularly hard. Meanwhile, the unemployment rate in Poland decreased from 18% in 2005 to 7% in 2008 and has remained relatively low since then, which may have encouraged migrants to return (Fries-Tersch et al., 2020). Most returnees to Poland have spent a relatively short time (one to a few years) abroad. The number of returnees to Poland between 2004 and 2014 is estimated at a minimum of 587,000 (Karolak, 2016). For Polish migrants in the UK (which has been a major destination country for Poles, along with Germany), another wave of returns took place after the Brexit referendum; between 2017 and 2018, the number of Poles living in the UK decreased from 922,000 to 832,000 (ONS 2019).
Looking at the age structure of returnees to Poland, we see that young adults dominate. Up to 2009, returnees predominantly fell into the age group of 20-29; in the subsequent decade, returnees tended to be older, and the dominant age category was 30-39 (Fries-Tersch et al., 2020).
Outward mobility from Lithuania is among the highest in all EU Member States, and has been for several decades. It is estimated that almost 700,000 people have left Lithuania since 2004 (Žvalionytė, 2012). This has resulted in a significant decline of the Lithuanian population and a relatively large diaspora around the world. However, according to Statistics Lithuania, approximately 188,000 emigrants have declared their return to Lithuania since 2001. As not all migrants declare changes in their country of residence (Žvalionytė, 2012), the numbers of emigrants from and returnees to Lithuania are likely to be even higher in reality.
For many years, returning Lithuanians accounted for the largest share (on average about 82%) of immigration flows to Lithuania (European Migration Network, 2021). Although the percentage of returnees in the inflow has been decreasing since 2016 as more people from other countries are moving to Lithuania, migration statistics show an increase in the absolute numbers of returnees in recent years. In 2019, return migration reached 20,000, the highest number in the past three decades. Almost half of these were returnees from the United Kingdom (European Migration Network, 2021), which is also the most common country of emigration for Lithuanians according to Eurostat (2021).
Much as in Poland, return mobility to Lithuania primarily consists of young working age migrants (Fries-Tersch et al., 2020). In 2017, 60% of all returnees were between 20 and 39 years old. Of those, around 40% were of very young working age (20-29). The proportion of returnees aged 65 years and older was less than 1%. Taking into account that outward migration from Lithuania is also dominated by young movers, these trends lead to the conclusion that many Lithuanians return in the first few years after leaving the country.

Methods and data
We used data from an CAWI survey to test which factors best explain and predict the extent to which returnees from the UK feel attached to their place of residence following return. We measured the attachment with the question: "How attached do you feel to your current city/town of residence in Lithuania/Poland?". Respondents were asked to use a 1-10 scale, from not attached at all to very attached. This is in line with how place attachment, as a subjective feeling, is usually measured. Both in quantitative and qualitative studies, self-reported measures prevail (see: Lewicka, 2011). The survey was conducted between May and August 2020, generating a total of 740 responses from Polish (525) and Lithuanian (215) returnees who had spent at least one year in the UK.
Respondents were recruited via social media sites such as Facebook, including online groups for Polish and Lithuanian migrants and returnees and Facebook ads targeting people meeting the study criteria (Pötzschke & Braun, 2017), as well as through personal networks. The survey included biographical questions about returnees, their migratory experience, relationships, experiences upon return to Lithuania and Poland, employment history, the evaluation of social and welfare services in both the UK and country of origin, place of residence, and intentions to re-migrate. No incentives were used to recruit participants.
Given that the survey was conducted using a convenience sample, we cannot claim that the results presented in this paper are representative of all returnees from the UK to Lithuania and Poland. Nevertheless, to ensure that our findings are not biased due to the composition of our sample (see Table 1 below), we employed a regression analysis, which controls for different respondent characteristics. Specifically, we employed the following ordinary least-squares regression model: where the dependent variable Y is the self-reported level of attachment to the place of residence by returnee i. Xi is a set of biographic variables, including the number of years spent in the UK, the number of years since return to Lithuania or Poland, and whether the person returned to the same city or town they had lived in before migrating to the UK. Wi represents relational factors, e.g. whether the respondent indicated that they lacked a social network/friends (yes/no) and whether they had returned for family-related reasons (yes/no). Zi concerns cultural factors, namely "nostalgic return" Yi = β0 + β1Xi + β2Wi + β3Zi + β4Vi + εi (missing the country of origin as the reason for return-yes/no) and whether respondents indicated that they experienced difficulties stemming from feelings of being different from people in the community after return (yes/no). Economic and welfare factors (Vi) included labour market status (employed/inactive/unemployed) after return (at the moment of participation in the survey), owning property in the country of origin (yes/no) and an index of preferences regarding welfare and social services. The index was constructed from twelve variables comparing various services and social benefits in the respondents' home country and the UK (variables included in the comparison: healthcare services, dentistry services, mental health services, education, childcare services, counselling and skill training support for job seekers, long-term care for the elderly, long-term care for people with disabilities, childcare support, support for maternity/paternity leave, unemployment benefits, and public support to people with low incomes). The model included socio-demographic controls (age, gender, education level and legal status in the UK).
As Table 1 shows, our sample consists of a heterogeneous population of returning migrants from the UK to both Poland and Lithuania, though certain groups may be underrepresented due to the sample size. In our survey sample, the average returnee . Polish respondents tended to be slightly more educated than Lithuanian ones. About 70% of both Poles and Lithuanians in our sample were employed after returning to their country of origin. The high percentages of unemployed people in both subsamples can be explained by the fact that many people only start looking for work after their return rather than prior to it. Furthermore, returnees often have savings, so they can afford to take more time to find suitable employment (Karolak, 2020).
In both sub-samples, length of stay in the UK is nearly equally distributed across the various categories and the majority of migrants returned to the same city/town they lived in before migration. About half of the sample listed family-related reasons (more frequently in the Polish sub-sample) and missing their home country (more frequent in Lithuanian sub-sample) as reasons for their return. After returning, approximately onequarter of each sample experienced difficulties relating to a lack of friends and feeling different from people around them. Also, respondents quite strongly preferred various services and benefits in the UK, with a mean index value of approximately 2 in both subsamples (minimum value 0 for a person who does not prefer any services or benefits in the country of origin; maximum value 12 for a person who prefers all the services and benefits listed in the country of origin). Table 1 reveals that the samples of Polish and Lithuanians return migrants are very similar in terms of the descriptive statistics (which may be partly due to the fact that both sub-samples were recruited via the internet and social media). While the country of origin variable was initially included in the model, it was not significant, meaning that no differences in terms of factors that influence attachment were found between Lithuanians and Poles. For this reason, the analysis below does not make comparisons between the subsamples, focusing instead on the exploration of factors that influence attachment based on the theoretical framework in the whole sample.
The article is focused on returns from the UK only. Since the enlargement of the EU in 2004, it has been the main destination country for both Poles and Lithuanians, and since the Brexit referendum a significant decrease of these migrants has been noted (see Sect. 3); thus we considered it to be a timely phenomenon worth exploring. Nevertheless, this focus on the UK may limit generalisation of the data regarding returnees to Lithuania and Poland from other countries. As far as Antonsich's (2010) theoretical framework makes it possible to demonstrate the differentiated process of building and rebuilding a sense of belonging in various life spheres, the variables in different thematic groups are intertwined, and sometimes it is hard to assign them to a single group. For instance, feeling different from others has both a cultural and a social dimension; owning a property is important from an economic point of view, but may also be connected with having a personal or family history in a place, i.e. with autobiographical factors. In order to confirm Antonsich's model and the relations between the predictors of attachment, more sophisticated analyses need to be performed, which was not possible in this paper due to the nature of our data.
Furthermore, like the majority of surveys, ours measures variables at one particular moment, which does not reflect their dynamics in time. In order to explore how attachment and other variables change, a longitudinal study would have to be conducted, which is beyond the scope of the project.
Another limitation lies in the fact that simply knowing the strength of returnees' attachment and its main predictors does not tell us anything about their future plans and decisions. It is certainly possible for a person to be extremely attached to their place of residence for the reasons described in this article, but nevertheless be motivated to leave for other reasons-pragmatic reasons, for example. It is also likely that there are returnees who do not feel attached to Poland or Lithuania, but remain for the sake of family or because they perceive few opportunities abroad. White's research on double returns (2014) suggests that challenges experienced on the labour market may be a substantial push factor for returnees. Indeed, analysis of the future plans of returnees demonstrates that although attachment is a factor in these plans, the factors predicting a sense of attachment are not the same as the predictors of re-emigration (c.f. other papers in this paper cluster).

Findings: the preponderance of cultural and autobiographical factors
In this section, we present the results of the regression analysis, which indicate that the majority of the factors considered are indeed significant in predicting returnees' attachment to their place of residence. Feeling different from others after return presented the greatest obstacle to the building or rebuilding of a sense of attachment. It was followed by missing one's country as one of the reasons for return, which facilitates attachment. Two other factors of particular importance were returning to the city or town of origin (instead of to another locality in Poland or Lithuania) and lack of social networks in the place of residence, which impedes attachment. In the following section, we verify our hypotheses and interpret the results, discussing their theoretical implications and formulating suggestions with reference to policy making (Table 2).

Autobiographical factors
The regression analysis confirmed that the majority of the autobiographical factors that we assumed would be significant predictors of attachment to the place of residence were indeed revealed as such. With the exception of the first hypothesis (H1), in which we assumed that the length of stay abroad would be negatively correlated with attachment to the place of residence after return, all our hypotheses were confirmed. This means that, as hypothesised, the more years have passed since their return, the more returnees feel attached to their place of residence (H2). The analysis also corroborated our assumption that returning to the place of previous residence (and not to a different locality in Poland or Lithuania) positively affects the attachment after return (H3); this proved to be the strongest predictor in this group of variables and one of the strongest in the model as a whole.
The result confirming the significance of the role of time is consistent with other studies on place attachment, in which length of stay is one of the crucial predictors of bonds with a locality (Hay, 1998;Lewicka, 2011). Time allows people to get to know a place, to find favourite spots and to create rituals and routines that make them feel at home. The fact that the amount of time spent abroad was not correlated with attachment can be explained by the transnational ties that many Polish and Lithuanian migrants maintain (Bell & Bivand-Erdal, 2015;Ryan et al., 2009). Regular visits, holidays spent in the country of origin and the maintenance of contact with friends and family may prevent or slow the erosion of affinity with a place. In fact, some studies suggest that, as they say, absence makes the heart grow fonder: being away can make people appreciate what they have left behind and hence may even increase affection for one's home (Case, 1996), as nostalgic feelings can lead migrants to idealise their home country. Unsurprisingly, returning to the city/town of a previous residence (be it a place of origin or not) was a very strong predictor of attachment, which suggests that while time spent in a location does indeed significantly facilitate attachment, having a personal history in a given place is an even greater factor. It comes as no surprise, as childhood or youth memories are known to be positively connected with place attachment (Hay, 1998;Morgan, 2010). Additionally, we may argue that respondents who did return to another locality in Poland or Lithuania, not to their places of origin (approximately 30% of our sample, most of whom were living in major cities, as these offer opportunities for career-building and self-development), need more time to develop a sense of familiarity and belonging. Table 2 Regression model: attachment to the place of residence *p < 0.05; **p < 0.01; ***p < 0.001

Variable Coefficient (standard errors in parentheses)
Spent more than five years, but less than ten years in the UK (ref: spent over a year, but less than five years) − 0.095 (0.232) Spent more than ten years in the UK (ref: spent over a year, but less than five years) 0.362 ( Adjusted R-squared 0.2923

Relational factors
Relational factors and social networks are also significant for the formation of attachment bonds with the places of residence of returnees. Lacking social networks after return was correlated negatively with attachment (H4). However, contrary to what we assumed, family-related reasons to return were not a significant predictor of attachment (H5). With reference to social networks, many studies have demonstrated that social capital in a given place is closely connected with attachment to it, and some authors claim that the social aspects of place attachment are a crucial and inherent part of it (cf. Lewicka, 2011). At the same time, establishing or recreating social networks upon return is not always easy, as shown by the fact that 23% of our sample named it as a difficulty. The role of family ties makes this picture more complex and nuanced, as it appears that family-related reasons for return (which were reported by more than 50% of the sample) were not a significant predictor of attachment. The family-related reasons covered in the survey among others included family reunification, providing care for elderly family members and the preference to have one's children receive Polish education. The fact that family-related reasons are key motivational factors for many migrants to return (Dzięglewski, 2020;Slany & Ślusarczyk, 2010), but do not play a vital role in the building of attachment upon return, sheds new light on the divergence between motivations for return and the actual feelings or behaviours of returnees following return. We may assume that some respondents felt obliged or pressured to return for these reasons, as a result of which they do not feel attached to their current place of residence. Were it not for their families, they might well have stayed in the UK or chosen a different locality in Poland or Lithuania.

Cultural factors
Overall, cultural factors were by far the most important in terms of effect strength, providing evidence for their crucial role as predictors of attachment. Migrants who returned because they missed Poland/Lithuania (approximately 40% in our sample) felt a greater sense of attachment to the place they lived in, which supports our sixth hypothesis (H6). This finding proves not only that homesickness is a powerful motivation for return, but also that it makes people prone to developing or restoring attachment to their place of origin. Moreover, we can assume that the homesickness experienced by respondents abroad was also connected with their sense of alienation or uprootedness in the UK due to cultural differences, lack of familiar landscapes, unaccustomed food or different leisure habits. Indeed, many qualitative studies identified the motive of not feeling at home in one's destination country, together with the more general feeling of homesickness for one's homeland, as pivotal in the process of deciding to return (Dzięglewski, 2020;Filimonau & Mika, 2019;Slany & Ślusarczyk, 2010). The above analysed role of family ties, together with nostalgia and not feeling the sense of belonging in the UK confirm that also in the context of CEE migrants, emotional reflexivity (Holmes & Burrows, 2012) is crucial migrants' decision-making process.
In the hypothesis 7 we assumed that individuals who reported feeling different from the people around them as a difficulty they encountered after return would have lower attachment scores (H7). Our analysis confirmed this hypothesis as well; in fact, this factor proved to be the strongest of all the predictors in this study. Thus, not feeling at ease and accepted upon return-which 25-30% of respondents reported-constitute a major hurdle in the formation of bonds with one's place of residence. Interestingly, the percentage of respondents who experienced such problems is similar in the subsamples of returnees who returned to their previous place of residence and returnees who returned to a different locality in Poland and Lithuania, although it is slightly higher among returnees living in small towns than in bigger cities. This finding brings to mind the role of reverse culture shock (Gullahorn & Gullahorn, 1963) and similar concepts describing the cultural changes that take place during the process of migration. Migrants may have adopted certain aspects of the British lifestyle, changed their value systems and aspirations, and these changes-while imperceptible during short visits-become a source of alienation or uprootedness upon return.

Economic factors
The final group of variables, which we labelled "economic and welfare, " were of a more practical nature. Surprisingly, our initial assumption that labour market status would correlate with returnees' attachment to their place of residence was not confirmed (H8). It would appear that being inactive or unemployed has no significant effect on one's bonds with one's locality. The fact that a given place offers professional opportunities and economic stability is not crucial in people's development of attachment to it, in contrast to the cultural and social factors discussed above. This may have to do with a finding in several other studies among returnees, where returnees reported that they needed a period of rest after return and therefore refrained from entering the labour market, living on economic remittances or savings that they brought with them to their country of origin (cf. Fihel & Grabowska-Lusinska, 2014;Fihel & Grabowska-Lusińska, 2010). Another explanation may lie in the fact that the majority of the respondents in both subsamples were employed.
However, owning a property in the place of residence in Poland or Lithuania-also a factor concerning economic stability-proved to be a very strong predictor of attachment (H9). This finding is not surprising, considering that owning a house or apartment has not only a financial dimension, but is often a very personal matter, particularly if the residence was inherited, or built with the intention of returning. As Lewicka (2011) points out, two phenomena overlap here: on the one hand, people tend to invest in a property in the place they are attached to, and on the other, the sense of stability resulting from ownership of a property fosters belonging. In either way, the confirmation of H9 indicates the importance of having a home-base in the country of origin for returnees, corroborating the findings of several other studies (Trąbka & Wermińska-Wiśnicka, 2020).
Finally, consistent with expectations, preferring health and education services/benefits in the home country (over those in the UK) positively affects attachment (H10). While this is in line with numerous studies about migrants preferring and using certain services (e.g. healthcare, see : Osipovič, 2013) in their countries of origin, we must note that the overall evaluation of services and benefits by both groups of migrants strongly favoured the UK. We may draw the conclusion that although state welfare institutions facilitate attachment, as they contribute to a sense of safety, stability and better quality of life, their quality in Poland and Lithuania is not satisfactory for returnees.

Conclusions
The aim of this paper was to investigate autobiographical, relational, cultural and economic (cf. Antonsich, 2010) predictors of attachment to the place of residence in Poland and Lithuania among returnees from the UK. Our goal was to present a different perspective: instead of identifying returnees' motivations for return and the challenges they encountered, as other researchers have already done, we wanted to determine which factors facilitate their sense of attachment, and which factors hinder it. Our analysis allowed us to draw several conclusions that complement and develop the existing knowledge on the process of migrants' establishment and re-establishment of bonds with their countries of origin following their return. Thus, this article also contributes to the knowledge about return migration and uncovers a deeper layer of reintegration: factors affecting and predictors of attachment to a place of residence after return.
The results of our regression analysis demonstrated that cultural factors play a crucial role in shaping return migrants' attachment, followed by autobiographical and relational factors. Surprisingly, labour market status, which was portrayed in the previous studies as a considerable challenge for returnees from Western Europe to CEE countries, proved to be not significant for their sense of attachment to their place of residence. However, other economic factors, such as owning property or appreciating social and welfare services, did turn out to be significant. Referring to the two types of attachment put forward by Williams and Vaske (2003), we see that among returnees, "place identity" (i.e. deeper and more intimate bonds) prevails over "place dependence" (i.e. instrumental, functional ties resulting from the various opportunities available in a particular place).
We argue that the role of attachment to a place of residence should not be underestimated, as it goes hand in hand with the quality of everyday life and wellbeing. It has also been proven that feeling attached to one's locality makes people more likely to engage in activities benefiting the place and the community (Lewicka, 2005).
Finally, the knowledge obtained regarding the predictors of attachment enables us to formulate several recommendations, for policy makers as well as for individuals considering return, that could make the transition smoother. Firstly, the importance of social and welfare services paired with their critical evaluation in comparison with their counterparts in the UK suggest that in order to attract returnees and facilitate their attachment to their homeland, the quality of these services should be improved. Secondly, migrants considering return should take into account not only the pragmatic, economic factors, but also the density and quality of social networks in their future place of residence and re-settlement, as well as the likelihood that they will fit in culturally. Taken together, these are the factors that conspire to make one feel at home.