Multidimensional Intergenerational Relationship and Social Participation among Late-Life Migrants in China: The Role of Depressive Symptom

Chinese adults will migrate to reunite with their children in later life, with social participation being an essential element of their successful adaptation. As children are older adults’ primary source of support, this study aimed to investigate both the direct and indirect efects of multiple dimensions of intergenerational relationship quality on social participation through the incidence of depressive symptom. A cross-sectional survey was conducted in Shenzhen, China, with late-life migrants aged 50years and above. Totally, 397 participants were included in this analysis. Te intergenerational relationship was measured using the In-tergenerational Relationship Quality Scale for Aging Chinese Parents, which covers four dimensions: structural-associational solidarity, afectual closeness, consensual-normative solidarity, and intergenerational conficts. However, as the structural-associational solidarity and afectual closeness dimensions did not have acceptable reliability, the four-dimension model was instead validated by using exploratory factor analysis and confrmatory factor analysis. Te same four-factor model was con-frmed, with the noted exception that item 10 (helping with housework) was deleted on the basis that it presented weak factor loading. Bootstrapping was used to examine the direct and indirect efects of the four dimensions on social participation through depressive symptom. Te results showed that only afective closeness and intergenerational confict were associated with social participation. Depressive symptom was a partial mediator of the correlation between afective closeness and social participation, whereas depressive symptom did not have any mediating efect on the other dimensions of intergenerational relationship quality and social participation. Tese fndings highlight the importance of afectual closeness and intergenerational confict in the social participation of late-life migrants in China.


Introduction
It is common for older adults to migrate internally within China to reunite with their children in other cities for taking care of their grandchildren and receiving care from their own children [1,2].In 2015, 17 million older adults moved across provinces in China to reunite with their children [3].Late-life migration not only disrupts older adults' social networks but also poses challenges with regard to establishing new social ties [4,5].Social participation is essential to facilitating the social integration of late-life migrants because it can broaden their social networks and promote their aging well in place [6][7][8][9][10].Social participation refers to engagement in outdoor social activities that lead to resource sharing and increased personal satisfaction [11].However, late-life migrants in China have limited opportunities for social participation [12].Over 80% of participants in two separate surveys reported that they seldom or never participated in any outdoor social activities [13,14].
Te results of some studies have indicated that family relationships can infuence older adults' level of social participation [11,15,16].Due to the inherent disruption that occurs with regard to their original social networks, children become the primary source of support for late-life migrants [17,18].In addition, given the importance of intergenerational relationships in Chinese culture [19,20], the quality of these relationships may infuence late-life migrants' level of social participation.However, previous studies have mostly focused on the infuence of intergenerational relationships on the physical and mental health of Chinese late-life migrants [21][22][23][24][25], without paying sufcient attention to their efects on these migrants' social adaptations.Intergenerational relationships are multidimensional and include solidarity and confict domains [26,27].Te existing literature on the intergenerational relationships of Chinese late-life migrants has mostly focused on the solidarity domain without considering the confict domain [21][22][23][24][25]. How diferent dimensions of intergenerational relationship quality were related to late-life migrants' social participation remains to be explored.Tus, this study aimed to investigate the relationships between diferent dimensions of intergenerational relationship quality and social participation among late-life migrants in China.

Intergenerational Relationship and Social Participation.
Intergenerational relationship quality comprises the solidarity and confict domains.Te solidarity domain includes fve dimensions: associational (frequency of interactions and types of activities in which the family engages), structural (size and geographic proximity of the family), functional (exchange of support and resources), afectual (degree of afection and positive sentiments), consensual (agreement on beliefs and values), and normative (commitment to familial roles and obligations) [26].Based on this model, Bai [28] developed and validated the Intergenerational Relationship Quality Scale for Aging Chinese Parents (IRQS-AP) for application to older Chinese adults and proposed a four-dimensional model of structural-associational solidarity, afectual closeness, consensual-normative solidarity, and intergenerational confict.Structural-associational solidarity refers to residential proximity to children, frequency of contact with children (face-to-face/through social media), and assistance with housework for children.Afectual closeness focuses on perceived closeness and intimacy with children and the amount of fnancial support received from them.Consensual-normative solidarity emphasizes the similarity of general opinions, opinions on caring for parents, and opinions on social issues.Finally, intergenerational confict stresses excessive demands from children, criticism from children, and strained feelings toward them.Tis study used Bai's model to examine the intergenerational relationship quality of late-life migrants in China.
Families play an important role in facilitating older adults' social participation [15].Having family support and family members involved in social groups can motivate older adults to participate more socially [11,16,29,30].Children are among the main sources of support for older adults [31].However, few studies to date have investigated the efects of intergenerational relationships on older adults' social participation.Only one study thus far has indicated that fnancial and emotional support from children helps to facilitate older Chinese adults' participation in social activities [32].However, focusing exclusively on social support does not help us understand the other dimensions of social relationships [33].Support from children constitutes only one of the dimensions of intergenerational support.Late-life migrants usually live with children and receive hands-on care from them; however, they will also inevitably come into confict with them [34,35].Terefore, intergenerational interactions have both positive and negative efects.Tis points to a clear need to comprehensively examine the infuence of the quality of intergenerational relationships on social participation.Some qualitative studies have indicated that the diverse dimensions of intergenerational relationship quality may have diferent relationships with social participation.Existing qualitative studies have shown that late-life migrants in China experience better social adaptation when they have intimate relationships with their own children.Tese studies have also reported a reduced willingness among late-life migrants to integrate into host cities when they have strained relationships with their children [2,14,[36][37][38][39].Tus, afectual closeness may be positively associated with social participation, whereas intergenerational confict may be negatively associated.Structural-associational solidarity focuses on living proximity and frequency of contact with children.It is common for Chinese late-life migrants to live with their children because they need to take care of their grandchildren or receive care from their own children [1,2].Living with children does not imply high-quality intergenerational relationships because of the possible conficts that arise from coresidence with children [40].Although coresidence increases the likelihood of older adults' having daily contact with children, these kinds of contacts usually consist of routine interactions, instead of the type of contacts that can actually promote intergenerational intimacy [1].In this case, although late-life migrants are likely to report high structural-associational solidarity, this does not guarantee more positive interactions with children.Terefore, structural-associational solidarity may not be related to social participation.For consensual-normative solidarity, late-life migrants report that they quarrel with their children because of disagreements about lifestyle and how grandchildren should be taken care of.Tis can cause them to feel distressed and reduce their interactions with others [13,39].Tus, it is possible that late-life migrants who have less consensual-normative solidarity with children may experience less social participation.

Te Mediating Efect of Depressive Symptom.
Te International Classifcation of Functioning, Disability, and Health (ICF) model was developed by the World Health Organization (WHO) and is applicable to all people, irrespective of their health conditions and cultural contexts [41].Te ICF model acknowledges that environmental factors (e.g., support and relationships) infuence individuals' functioning (e.g., psychological functioning) and 2 Health & Social Care in the Community participation (e.g., social participation).Psychological functioning and social participation are mutually correlated.Te ICF model also proposes that health conditions and sociodemographic factors are another factor infuencing people's functioning and participation.Tese items are regarded as control variables in this study (Figure 1).Te ICF model has been used to study the relationship between depressive symptom and social participation in older Chinese adults [42].Te results of a longitudinal study indicate that depressive symptom as a psychological function negatively infuences older adults' social participation [42].Terefore, based on the tenets of the ICF model, depressive symptom is a possible mediator between social relationships and social participation.Te quality of intergenerational relationships infuences older adults' psychological functioning, because of socioemotional selectivity theory stating that people emphasize emotional goals, such as deepening intimacy, when aging [43].Older adults tend to interact with intimate social partners (e.g., family) and attach importance to relationship quality when seeking to satisfy their emotional needs [43,44].Accordingly, afectual closeness and consensualnormative solidarity have been found to be negatively correlated with depressive symptom in older adults in China [17,[45][46][47].Conversely, intergenerational confict has been found to be positively associated with depressive symptom [17,45,47].Te efect of structural-associational solidarity on depressive symptom is complex.Te efect that living with children has on depression is dependent upon whether the older adults being assessed prefer to live with children [48].Although flial piety encourages coresidence with children, older Chinese adults are expressing a growing demand for independence and privacy.Tus, living with children has been found to reduce depressive symptom in some older Chinese adults [46], while others have reported an increase in depressive symptom when living with children [40].Late-life migrants in China usually live with their children regardless of whether this is their own personal preference or not [2,14,[37][38][39].It is likely that the efect that coresiding with children has on depressive symptom is cancelled out by the diferent degrees of willingness older adults express with regard to living with children.In addition, late-life migrants do not expect to have much frequent contact with their children, considering that they are busy at work; instead, they expect support when they are in need [37,39].A systematic review found that neither living with children nor the frequency of contact with children was associated with the incidence of depressive symptoms in Chinese late-life emigrants [49].Terefore, structuralassociational solidarity may not be related to depressive symptom.
Regarding the relationship between depressive symptom and social participation, the interactional theory of depression states that people with depressive symptom interact negatively with social partners [50].For example, individuals with depressive symptoms are less inclined to make contact and interact with others [51,52].A longitudinal study found that older adults with more depressive symptoms engaged in less social participation [53].Tus, depressive symptom is negatively associated with social participation among older Chinese adults [7,[54][55][56].Although depressive symptom has a reciprocal relationship with social participation, longitudinal studies have suggested that the efect of depressive symptom on social participation in older adults is stronger than the opposite efect [57,58].In summary, depressive symptom may mediate the relationship between the quality of intergenerational relationships and social participation among late-life migrants.

Te Present
Study.Tis study aimed to investigate the associations between diferent dimensions of intergenerational relationship quality and social participation and the possible mediating efect of depressive symptom.We hypothesized that the diferent dimensions of intergenerational relationship quality would have diferent relationships with social participation.
H1a: Structural-associational solidarity is not associated with social participation.H1b: Afectual closeness is positively associated with social participation.H1c: Consensual-normative solidarity is positively associated with social participation.H1d: Intergenerational confict is negatively associated with social participation.In addition, depressive symptom had diferent mediating efects on the association between intergenerational relationships and social participation.H2a: Depressive symptom is not a mediator between structural-associational solidarity and social participation.H2b: Depressive symptom mediates the association between afectual closeness and social participation (i.e., late-life migrants with higher afectual closeness will report fewer depressive symptom, which in turn causes them to have more social participation).H2c: Depressive symptom mediates the association between consensual-normative solidarity and social participation (i.e., late-life migrants with higher consensual-normative solidarity will report fewer depressive symptom, which in turn causes them to have more social participation).H2d: Depressive symptom mediates the association between intergenerational confict and social participation (i.e., late-life migrants with fewer intergenerational conficts will report fewer depressive symptoms, which will in turn lead to more social participation).

Participants and Sampling.
Te data for this study came from a cross-sectional study exploring the social capital of older adults migrating with their children in China.Tis original study was conducted by the second author.Te study participants were recruited between November 2019 and January 2020 in Shenzhen, a city in Southern China.After being listed as a special economic zone, Shenzhen was constructed as a leading Chinese commercial center, one that has subsequently experienced rapid economic development [59].Due to these circumstances, Shenzhen has attracted a large migrant population, which reached over 12 million in 2021 [60].It is also a city with an increasing number of older individuals.Te percentage of people in Shenzhen aged over 60 increased to 5.36% of the total population in 2021 [60].Although there are no updated statistics currently available on older migrants in Shenzhen, it is reported to be one of the main metropolitan areas where late-life Chinese migrants choose to settle [3].
Te participants for this study were selected using convenience sampling.Considering the varying levels of economic development throughout diferent districts in Shenzhen and the fact that residents may have diferent levels of socioeconomic status, participants were approached in two districts with more advanced development (i.e., Futian and Nanshan) and two districts with less urban development (i.e., Bao'an and Longhua).In each district, two social service agencies were contacted with 50 individual participants approached by each agency.Across the four districts, a total of 400 late-life migrants were willing to complete the survey.Te inclusion criteria were as follows: (1) people aged 50 years and above, which is the minimum ofcial retirement age in China; (2) people who had migrated to Shenzhen with their children; and (3) people who had resided in Shenzhen for over 6 months.After excluding three participants with missing data, 397 participants were included in the fnal analysis.

Data Collection.
Surveys were completed through faceto-face interviews conducted by the staf in the community service centers, which lasted for approximately one hour.Te staf received training from the corresponding authors on how the interviews should be conducted.Before the interviews, the participants were informed about the aims of the research, the background of the researchers, their right to withdraw, their guarantee of confdentiality, and the potential associated risks and benefts.Written or verbal consent was obtained as appropriate.Coupons as thanks gifts were provided to the participants after the interviews.Ethical approval was gained (HSEARS 20190912001).13-items, which evaluates the four dimensions of intergenerational relationships: structural-associational solidarity (4 items), afectual closeness (3 items), consensual-normative solidarity (3 items), and intergenerational confict (3 items) [28].Te participants were asked to rate the quality of their relationships with their children residing in Shenzhen.Participants with two or more children residing in Shenzhen were asked to think of a child whose birthday was approaching and to rate the quality of the relationship between that child and themselves [28].Responses were made on a fve-point Likert scale for each item with a response range of 1-5.Higher scores indicated better structural-associational solidarity, afectual closeness, consensual-normative solidarity, and more intergenerational confict.Cronbach's alpha for the entire scale was 0.764.Cronbach's alphas were 0.64 for structuralassociational solidarity, 0.698 for afectual closeness, 0.783 for intergenerational confict, and 0.772 for consensualnormative solidarity.4

IRQS-AP. IRQS-AP contains
Health & Social Care in the Community validated for use with older Chinese migrants [61,62].Te scale contains 15 items that investigate feelings and behaviors that have occurred over the past week (5 positive items and 10 negative items).Te participants chose yes (0) or no (1) for each item.Negative items were reverse coded.Te total score for GDS-15 ranges from 0 to 15, with a higher score representing a higher incidence of depressive symptoms (Cronbach's a � 0.71).

Social Participation.
Te frequency with which participants had engaged in each of the nine social activities cited in the questionnaire over the last 12 months was also assessed.Tese activities included sports and physical activities, meeting friends, educational and cultural activities, volunteer work, and so on.Responses were rated on a fvepoint Likert scale (0 � never to 4 � several times a week).Te highest frequency that occurred among the questions was adopted when participants who seldom participated in diferent activities scored higher than the participants who frequently engaged in certain activities [63,64] (Cronbach's a � 0.82).
Health-related factors were also considered, including selfrated health (0 � very poor to 4 � very good) and having any chronic disease (0 � no, 1 � yes).

Data Analysis.
Descriptive analyses were performed to describe the participants' characteristics using Statistical Package for the Social Sciences (SPSS) 23.Assumptions of conducting regressions were examined, and the data fulflled the requirements of homoscedasticity (the scatterplot showed that residuals were randomly scattered), independence (the Durbin-Watson value was 1.53, which was greater than 1.5), and no multicollinearity (variance infation factor (VIF) values ranged from 1.136 to 1.66, which were all less than 5).However, the data were slightly left-leaning since the skew values of some variables were smaller than −2, while the kurtosis met the requirement of the range between −10 and 10 (Table1) [65].
Because the reliability of the structural-associational solidarity and afectual closeness subscales was lower than 0.7, which would infuence the reliability of the results [66], the four-dimensional model needed to be validated.First, exploratory factor analysis (EFA) was performed in SPSS to examine whether the data ft the four-dimensional model using varimax rotation.Confrmatory factor analysis (CFA) was then conducted in Analysis of Moment Structures (AMOS) 23 to confrm the identifed model.Te following indices were used to assess the model ft [65]: (a) A chisquare index with smaller values suggested a better ft for the model.(b) A comparative ft index (CFI) and a goodnessof-ft index (GFI) with values greater than 0.9 indicated a good ft.(c) Te root mean square error of approximation (RMSEA) with values less than 0.08 showed an acceptable ft, while values below 0.05 indicated a good ft.Te probability of close ft (PCLOSE) should also be greater than 0.05 for this method.(d) Te standardized root mean square residual (SRMR) with 0.08 or less indicated a good ft.In addition, the reliability and validity of the model were evaluated using the methods recommended by Fornell and Larcker [67], which are commonly used to analyze the reliability and validity of measurement models [66].Te reliability of the model was examined using composite reliability [66,67].Tis method is more suitable for measuring the reliability of latent variables than Cronbach's alpha, which examines the extent to which observed variables constitute the latent variable and for which values greater than 0.7 show good reliability.Validity was assessed by examining convergent and discriminant validity.Good convergent validity was indicated by standardized factor loadings with values greater than 0.4 and an average variance extracted (AVE) value exceeding 0.5 [66,67].AVE examines the "average amount of variance that a construct explains in its indicators relative to the overall variance of its indicators" [66].Discriminant validity was assessed by comparing the square root of the AVE of each dimension of intergenerational relationship quality and the correlations among these constructs.When the square root of AVE was greater than the interconstruct correlations, this meant that "the latent variable explains more variance of the indicators than another latent variable" [66].Tus, discriminant validity was established.Te above analyses were conducted using StatWiki software [68].
Subsequently, the regression weights and indirect efects in the structural model were assessed using bootstrapping with AMOS.Tis is a resampling procedure that addresses the non-normality issue in the data and examines the mediating efects [65].Specifcally, 2000 bootstrap resamples and 95% bias-corrected confdence intervals were applied.An indirect efect was considered signifcant when the confdence interval did not exceed zero.

Demographic Backgrounds of Participants.
Te mean age of the 397 participants was 63.4 (SD � 6.7), and they had settled in Shenzhen for an average of seven years (SD � 5.3).Among them, 69% were female and 84.1% were married.Information regarding the other demographic factors is shown in Table 2. Participants averaged relatively low scores for depressive symptom (M � 3.39, SD � 2.58) and high scores for social participation (M � 3.48, SD � 1.50).

Validation of the Measurement Model.
First, EFA was performed.Te Kaiser-Meyer-Olkin value was 0.753, and Bartlett's test of sphericity reached statistical signifcance (p < 0.001), indicating that the data were suitable for factor analysis [69].A four-factor model with eigenvalues over 1 was obtained, which was the same as Bai's model (2017), except that Health & Social Care in the Community  1).Although the correlation of item 10 with the afectual closeness dimensions (r � 0.300) was close to that recorded for consensual-normative solidarity (r � 0.396), 0.32 is recommended as a threshold for minimum loading [70].Tus, item 10 was regarded as being loaded in the consensualnormative solidarity.CFA was then conducted to verify the structures.Te model showed an acceptable model ft (χ 2 �125.979,df � 59, p < 0.001; CFI � 0.960; GFI � 0.955; RMSEA � 0.054, PCLOSE � 0.311, 90% CI (0.041, 0.066); SRMR � 0.0513).Although the composite reliability of all dimensions was above 0.7, consensual-normality solidarity had issues with convergent validity.Te factor loading of item 10 was lower than 0.4.In addition, the AVE of this dimension was lower than 0.5 (Table 1), which indicated item 10 may not adequately explain the concept of consensual-normality solidarity.Tus, item 10 was deleted, and the validity of the adjusted model was checked using CFA.Te adapted model showed a better ft (χ 2 � 85.246, df � 48, p < 0.05; CFI � 0.977; GFI � 0.967; RMSEA � 0.044, PCLOSE � 0.717, 90% CI (0.028, 0.059); SRMR � 0.0430) and demonstrated good reliability because the composite reliability of each dimension was above 0.7 (Table 3).Regarding convergent validity, although the factor loading of item 9 was still lower than 0.4, the AVE values of all dimensions exceeded 0.5.Collier [65] suggested retaining the item if the AVE values reached the threshold of 0.5; however, models must also be validated in other samples.For discriminant validity, Table 3 shows that the square root of the AVE of each dimension (in bold) was greater than the interconstruct correlations, which demonstrates the good discriminant validity of the adapted model.

Structural Model.
Based on the adapted model, the composite score of each dimension was calculated for the correlation analysis before conducting the path analysis.Correlations among the diferent dimensions of intergenerational relationship quality, depressive symptom, and social participation were examined using Spearman's rho due to skewed data (Table 4).Structural association solidarity was not signifcantly correlated with social participation (r � 0.02, p � 0.77).However, afectual closeness (r � 0.27, p ≤ 0.001), consensual-normative solidarity (r � 0.11, p � 0.03), and intergenerational confict (r � −0.31, p ≤ 0.001) all had small but signifcant associations with social participation.Depressive symptom also had a weak but signifcant correlation with social participation (r � −0.30, p ≤ 0.001).Structural-associational solidarity did not have a signifcant correlation with depressive symptom (r � −0.01, p � 0.88), while the other three dimensions of intergenerational relationship quality were signifcantly related to depressive symptom, despite the associations being small (afectual closeness: r � −0.25, p ≤ 0.001, consensualnormative solidarity: r � −0.13, p ≤ 0.001, and intergenerational confict: r � 0.18, p ≤ 0.001).

Discussion
Tis study used Bai's IRQS-AP to investigate the direct and indirect efects of the multiple dimensions of intergenerational relationship quality on social participation by exploring the incidence of depressive symptom among late-life migrants in China.Tis study found that item 10 (helping with housework) did not ft well for this target group; therefore, this item was deleted, and an adapted model of intergenerational relationship quality was validated.Te adapted dimensions of intergenerational relationship quality were further applied in the subsequent path analysis.Te results showed that structural associational solidarity was not related to social participation, whereas afectual closeness and intergenerational confict were signifcantly associated with social participation, as   previously hypothesized (H1a, H1b, and H1d).However, consensual-normative solidarity was found not to be related to social participation in the multivariate analysis, which failed to support H1c. Regarding the mediating efect of depressive symptom, the results showed that depressive symptom did not mediate the relationship between structural-associational solidarity and social participation but had a signifcant partial mediating efect on the association between afective closeness and social participation, supporting H2a and H2b.Contrary to H2c and H2d, depressive symptom was not a mediator of consensualnormative solidarity, intergenerational confict, or social participation.Item 10 was originally included in the structuralassociational solidarity subscale because the scale was developed for community-dwelling older adults, and these older adults regard helping with housework as a chance to increase the frequency of intergenerational interactions [28].However, this study found that item 10 was loaded in terms of consensual-normative solidarity, which focused on agreement on beliefs and flial responsibility.Unlike community-dwelling older adults, late-life migrants have complex feelings about helping with housework.Instead of receiving care from their children, Chinese late-life migrants usually provide an extensive amount of assistance regarding household chores [1].In Ruan's study, late-life migrants even complained about being treated as a "babysitter for free" [2].Tus, this item relates to respondents' level of agreement with flial responsibility.Sharing housework is also an action through which parents provide care for their children, and which enables the children to more easily work full-time [36,71].Item 10 also correlated with the afectual closeness dimension to a certain extent.In this case, item 10 did not have a high factor loading in the consensualnormative solidarity dimension in the CFA process.Te problematic nature of item 10 and the fact that item 9 presented weak loadings in the afectual closeness dimension indicates the diferences that apply to the intergenerational interactions of late-life migrants and local older adults; therefore, Bai's model may need to be further validated among more Chinese late-life migrants.
Supplemented by the ICF model, this study showed that the structural dimension of intergenerational relationship quality was not related to social participation among late-life migrants in China.However, social participation was found to be correlated with positive and negative intergenerational interactions (i.e., afectual closeness and intergenerational confict).Consensual-normative solidarity signifcantly correlated with social participation in the bivariate analysis but not in the multivariate analysis when afectual closeness and intergenerational confict were included.Tese fndings indicate that afectual closeness and intergenerational confict may explain the correlation between consensualnormative solidarity and social participation.For example, late-life migrants may experience conficts with their children because of disagreements regarding lifestyle and grandparenting, causing them to report reduced social interactions [13,39].Future studies should examine whether afectual closeness and intergenerational confict mediate the relationship between consensual-normative solidarity and social participation.In addition, supplemented by Wangliu's [32] fnding that intergenerational support is benefcial to older adults' social participation, the fndings of this study further indicate that intergenerational confict is also related  Health & Social Care in the Community to social participation.It should be noted that solidarity and confict are not on opposite sides of this relationship and that improving solidarity does not necessarily mean reducing confict.Family relations can exhibit both high solidarity and confict or low solidarity and confict [27].Terefore, though a certain amount of correlation exists between both items, afectual closeness is a concept that is distinct from intergenerational confict, with both items being signifcantly related to social participation.Tese fndings suggest that improving afectual closeness with children may not be enough to facilitate greater social participation among latelife migrants and that reducing the intergenerational confict they experience is also necessary.
Moreover, this study found that afectual closeness was correlated with social participation by infuencing depressive symptom late-life migrants.Consistent with the results of previous studies [17,[45][46][47], afectual closeness was found to be related to the incidence of depressive symptom.Tis is because older Chinese adults attach great importance to their relationships with children.In particular, latelife migrants need support to ensure appropriate social adaptation.In this context, having close relationships with children enables them to feel supported and experience less emotional distress [2,14,[36][37][38][39].Te partial mediating efect of depressive symptom on afectual closeness and social participation indicates the presence of other mediators such as life satisfaction.Wangliu [32] found that intergenerational support was related to social participation through life satisfaction.Surprisingly, the signifcant association between intergenerational confict and depressive symptom disappeared in the multivariate analysis with the inclusion of afectual closeness, which contradicts the fndings of a number of previous studies [17,45,47].Zhou and Bai [47] found that among the dimensions of intergenerational relationship quality, afectual closeness had the strongest association with depressive symptom in older Chinese adults.Afectual closeness may have a greater infuence on late-life migrants' depressive symptom than intergenerational confict; therefore, the association between confict with children and depressive symptom was not found to be signifcant.It is also possible that intergenerational confict is related to social participation through other psychological functions, such as a sense of belonging to the host city.Te results of two qualitative studies have indicated that late-life migrants who experience confict with their children express their intention to return to their hometowns and report having fewer social interactions with others [37,39].However, further studies are needed to verify this relationship.Consistent with previous fndings [7,[54][55][56], depressive symptoms are negatively related to social participation.Tis indicates that the degree of social participation of late-life migrants is related to their psychological functioning.

5.1.
Limitations.Tis study had a number of limitations.First, this was a cross-sectional study, so the causal relationships among the variables were not tested, and a longitudinal design is needed to further verify the validity of the study's fndings.Second, this study used nonprobability sampling by conveniently sampling participants from social service agencies.It is likely that participants who seldom visited the agencies due to a decline in health or the presence of a disability were not approached.Tus, the fndings recorded are only applicable to late-life migrants with good physical capabilities.Finally, this study found that certain items in IRQS-AP were not applicable to late-life migrants; however, it is also worth noting that the participants for this study were recruited exclusively from Shenzhen.Tus, the validity of IRQS-AP for late-life migrants in terms of its potential application to other cities must be investigated further.

Conclusion and Implication
Tis study investigated the correlations between the subdimensions of intergenerational relationship quality and social participation among late-life migrants in China and the possible mediating efect of depressive symptom on these correlations.Current interventions designed to facilitate the social adaptation of late-life migrants in China mainly focus on providing peer-support groups [72].Tis study showed that afectual closeness and intergenerational confict signifcantly correlated with social participation.Tis calls for practitioners to pay greater attention to these two types of intergenerational interactions when attempting to facilitate greater social participation among late-life migrants.Practitioners should especially focus on dealing with depressive feelings when late-life migrants report feeling or experiencing less closeness with their children.Tis may help to positively infuence their level of social participation.

Table 4 :
Correlations for key variables included in the model: Spearman correlation (n � 397).

Table 5 :
Pathways to social participation.