China’s rapid urbanization has attracted millions of rural migrant workers to urban centers in pursuit of economic advancement. However, most of these workers have to leave their children in their hometowns because they cannot afford to raise them in urban settings (Wang et al., 2022). This has resulted in 41.77 million children known as “left-behind children” (LBC) (United Nations Children’s Fund, 2023). LBC are rural children left at home for extended periods, typically over 6 months, when one or both of their parents migrate to urban areas for work (Zhao et al., 2020). The psychological well-being of LBC has become an increasing concern and attracted considerable public attention in recent years. Research has demonstrated that parental migration is associated with increased risk of poorer mental health among LBC (Fellmeth et al., 2018; Hu et al., 2022; Lin et al., 2025). A well-documented manifestation of this effect is their markedly lower subjective well-being compared to other children (Li and Liu, 2013; Ye et al., 2020). This pattern has been consistently identified in studies conducted across Europe, Africa, America, and Asia (Graham and Jordan, 2011; Jordan and Graham, 2012; Mazzucato et al., 2015; Bălţătescu et al., 2023). As a multidimensional construct encompassing both individuals’ cognitive and affective assessments of their lives (Diener, 2000), subjective well-being serves as a critical measure of mental health (Diener et al., 1999). Therefore, identifying the factors and mechanisms that shape LBC’s subjective well-being is essential for developing effective interventions.

Perceived discrimination, defined as the perception of being treated differently or unfairly owing to membership or classification within a certain group (Pascoe and Richman, 2009), functions as a chronic stressor (Thoits, 2010) and has been identified as a significant threat to mental health (Pascoe and Richman, 2009; Schmitt et al., 2014). Within the context of LBC in China, children are labeled as “left-behind,” leading them to perceive both individual-level and group-level discrimination. The absence of parental care exposes these children to stigmatization, such as being ridiculed as “unwanted children,” which fosters individual-level discrimination. Additionally, media coverage highlighting behavioral issues among some LBC has reinforced negative public stereotypes, further branding them as “problem children,” and contributing to group-level discrimination (Zhao et al., 2016, 2020; Hu et al., 2022). Consistent with previous studies (Pascoe and Richman, 2009; Schmitt et al., 2014; Adedeji et al., 2025), higher levels of perceived discrimination were associated with lower levels of subjective well-being among LBC (Shen et al., 2009; Zhang et al., 2020). In addition, this association was not moderated by sex (Pascoe and Richman, 2009; Zhao et al., 2016; Emmer et al., 2024). The stress process model offers a theoretical framework for understanding the association between perceived discrimination and subjective well-being (Pearlin et al., 1981; Pearlin, 1989; Wheaton, 2010; Turner, 2013). Specifically, this model suggests that exposure to stress may lead to increased risk of poorer mental health either directly through the activation of the stress process or indirectly through the depletion of psychosocial resources (Pearlin, 1989; Wheaton, 1994; Turner, 2013). Importantly, empirical evidence indicates that although there are sex differences in stress exposures, these pathways tend to operate similarly for both sexes, with limited support for sex as a moderating factor (Pascoe and Richman, 2009; Turner, 2013; Anderson et al., 2022). Within this model, stress generally encompasses life events and chronic strains, such as chronic illness and perceived discrimination (Pearlin, 2010). Psychosocial resources generally include social resources (e.g., social support) and personal resources (e.g., mattering and self-esteem). Although the stress process model has received substantial empirical support (Cairney et al., 2013; Aneshensel and Mitchell, 2014), its application among vulnerable populations, such as LBC, remains limited. Given the central role of psychosocial resources in this framework, it is essential to examine specific mechanisms–particularly social support–through which discrimination may affect the subjective well-being of LBC.

Social support is defined as spiritual and material support (e.g., love, care, respect) obtained from social relationships like family, colleagues, groups, organizations, and communities (Cobb, 1976). Previous research has consistently found that greater social support is associated with higher levels of subjective well-being (Yalcin, 2015; Siedlecki et al., 2014) and is especially vital for LBC lacking parental care. According to the stress process model, perceived discrimination can undermine social support, reducing the protective resources available to cope with stress and thereby increasing the risk of lower levels of subjective well-being. In addition, the social support deterioration deterrence model also provides insight into the potential mediating pathway between perceived discrimination and mental health outcomes via social support. This model posits that a stressor such as perceived discrimination (Vines et al., 2017) may inhibit the protective effects of social support on an individual’s mental health (Kaniasty and Norris, 1993; Norris and Kaniasty, 1996). Indeed, empirical studies across socio-cultural groups (Prelow et al., 2006; Jia and Liu, 2017), including left-behind children (Fu et al., 2024), demonstrate that perceived discrimination significantly diminishes social support, which in turn leads to poorer mental health. Therefore, social support may partially explain the association between perceived discrimination and subjective well-being among LBC.

In addition to eroding social resources, exposure to stress may be also associated with lower levels of subjective well-being through eroding personal resources (Pearlin, 1989; Wheaton, 1994). Psychological resilience is defined as the adaptability and ability to remain healthy despite experiencing stress and adversity (Ungar, 2012). While resilience is a protective factor again poorer mental health (Fritz et al., 2018; Ungar and Theron, 2020), including among LBC (Han et al., 2021), it can be eroded by chronic stressors, as suggested by the stress process model. Symbolic interactionist theory (Brownfield and Thompson, 2005) provides a specific mechanism for this effect: perceived discrimination (a form of negative social feedback) promotes the internalization of stigma (Fernández et al., 2015), which in turn lowers self-evaluation (Szalacha et al., 2003), thereby resulting in reduced psychological resilience (Olsson et al., 2003). Indeed, studies on LBC support this pathway, showing that higher levels of perceived discrimination were associated with reduced resilience, which is in turn correlated with poorer school adjustment and more severe emotional and behavioral problems among LBC (Han and Long, 2020; Lv et al., 2021). Therefore, psychological resilience may partially explain the relationship between perceived discrimination and subjective well-being among LBC.

Social support and psychological resilience can each independently partially explain the association between perceived discrimination and subjective well-being, and social support may also influence psychological resilience, resulting in a potential chain mediation effect. Specifically, the association between perceived discrimination and subjective well-being may be partially explained by a path involving the depletion of social resources (social support) leading to the erosion of personal resources (psychological resilience). The strong link between social support and psychological resilience is well-established in the literature. According to the resilience framework (Kumpfer, 2002), stable and diverse social ties can provide external support resources for individuals to adapt to adversity, and subsequently produce positive outcomes (Noh and Park, 2022). Extensive evidence suggests that social support is closely related to psychological resilience and, particularly, resilience development in children (Olsson et al., 2003; Zimmerman, 2013; Masten, 2018). Research has also shown that higher levels of social support are strongly correlated with higher levels of resilience among LBC (Huang et al., 2022; Nguyen et al., 2023). Recent studies indicate that resilience may partially explain the relationship between social support and mental health (i.e., subjective well-being and satisfaction with life) (Fan and Lu, 2020; Yıldırım and Tanrıverdi, 2021). Furthermore, existing research proposes a serial mediation model in which the effects of stigma on mental health outcomes among people living with HIV are sequentially explained by a path through lower levels of social support leading to reduced psychological resilience (Yan et al., 2019; Yuan et al., 2024), highlighting their potential role in explaining the relationship between discrimination and subjective well-being. For LBC, we hypothesize a similar serial pathway. Based on the above-mentioned information, we hypothesized that the association between perceived discrimination and subjective well-being would be partially explained by a path through lower levels of social support leading to reduced psychological resilience.

In summary, guided by the stress process model, the current study aims to further our understanding of the mechanisms underlying the relationship between perceived discrimination and subjective well-being among LBC by exploring the role of social support and psychological resilience using a serial mediation model. Based on previous studies and theories, we hypothesized that: (H1) perceived discrimination is negatively correlated with subjective well-being among LBC; (H2) social support would partially explain the association between perceived discrimination and subjective well-being; (H3) psychological resilience would partially explain the relationship between perceived discrimination and subjective well-being; and (H4) the association between perceived discrimination and subjective well-being would also be partially explained by the path through lower levels of social support leading to reduced psychological resilience. Finally, consistent with evidence that the pathway from exposure to stress to subjective well-being via the depletion of psychosocial resources operates similarly across sex (Pascoe and Richman, 2009; Turner, 2013; Anderson et al., 2022), we also hypothesized that sex would not moderate any of the indirect effects of social support or psychological resilience in the association between perceived discrimination and subjective well-being (H5).

This study, applying the stress process theory to the context of LBC and discrimination, utilized structural equation modeling to examine the mechanisms linking perceived discrimination to subjective well-being among Chinese LBC. Findings showed that higher perceived discrimination was significantly associated with lower subjective well-being. This association was partially explained by social support and psychological resilience directly, as well as a path involving lower levels of social support leading to reduced psychological resilience. These findings have important implications for mental health education, discussed near the end of this concluding section.

Consistent with Hypothesis 1 and previous studies (Pascoe and Richman, 2009; Schmitt et al., 2014; Zhang et al., 2020; Hu et al., 2022), we found that higher perceived discrimination was significantly associated with lower levels of subjective well-being among LBC. According to the stress process theory, LBC are often exposed to both blatant and subtle forms of discrimination, such as verbal or physical harassment, being ignored, or excluded (Zhao et al., 2016, 2020), which functions as a chronic stressor associated with adverse mental health outcomes. The negative link between perceived discrimination and subjective well-being among LBC can be understood through mechanisms such as feelings of rejection (Leary et al., 1995) or negative internalization (Jost and Banaji, 1994). These discrimination experiences negatively impact individuals’ mental health by threatening their need for acceptance and inclusion (Wirth and Williams, 2009; Ma et al., 2022). Our results further support the notion that perceived discrimination poses a significant risk to the psychological well-being of LBC, highlighting the importance of addressing systemic stigma for this vulnerable population.

Consistent with the stress process model, which posits that exposure to stress can increase the risk of poorer mental health through the depletion of social resources, the current study found that the association between perceived discrimination and subjective well-being was partially explained by social support. This supports Hypothesis 2 and is consistent with previous studies (Prelow et al., 2006; Jia and Liu, 2017; Fu et al., 2024). This result also verified the social support deterioration deterrence model (Kaniasty and Norris, 1993; Norris and Kaniasty, 1996). One possible explanation is that perceived discrimination implies a rejection or exclusion of the targeted group (Schmitt et al., 2014). LBC, who perceive themselves as being discriminated against, may be inclined to avoid interpersonal interactions to protect themselves from social rejection (Zhang et al., 2020), which limits their access to potential sources of social support. In addition, perceptions of everyday injustice or negative feelings of rejection can reduce individuals’ help-seeking, preventing the establishment of social networks between youth and society (Cristini et al., 2011; Mao et al., 2022). When individuals perceive inadequate social support (perceived unwillingness to help, listen, etc.), various mental health problems (anxiety, depression, burnout, etc.) may arise (Guilaran et al., 2018; Xiang et al., 2020; Chang et al., 2022).

Psychological resilience also partially explained the association between perceived discrimination and subjective well-being in LBC, providing support for Hypothesis 3 and aligning with the stress process model’s view of personal resources as mediators. This finding also aligns with previous research (Han and Long, 2020; Lv et al., 2021). According to symbolic interaction theory, LBC who are exposed to negative social stereotypes and prejudices–such as being perceived as unloved, unwanted, or problematic–may internalize these external evaluations. Discrimination is not just an external event but can be internalized (Fernández et al., 2015), damaging LBC’s self-concept and eroding the internal coping capacities (e.g., resilience) needed to manage adversity (Lv et al., 2021). Although psychological resilience is a protective factor against poorer mental health (Fritz et al., 2018; Ungar and Theron, 2020), perceived discrimination, as a chronic stressor, is linked to sustained stress, which may in turn be associated with reduced psychological resilience (Vargas et al., 2021), thereby resulting in lower subjective well-being (Zhou and Chen, 2024).

The results of the current study show the serial mediation effect of social support and psychological resilience between perceived discrimination and subjective well-being among LBC, confirming Hypothesis 4. These findings are consistent with recent studies supporting a serial mediation model in which social support and psychological resilience jointly influence mental health outcomes (Yan et al., 2019; Yuan et al., 2024). These findings also align with the resilience framework (Kumpfer, 2002), illustrating how perceived discrimination–acting as a chronic stressor–undermines both external support systems and internal psychological resilience among LBC, ultimately impairing their capacity to maintain well-being. Critically, these findings provide empirical support for the stress erosion hypothesis (Pearlin, 1989; Wheaton, 1994). The relationship between stress and psychosocial resources is complex. While the well-known stress buffering hypothesis posits that resources can mitigate the negative influences of stress, the stress erosion hypothesis argues that persistent and chronic stressors can overwhelm these resources. Our model demonstrates this erosion process in action: perceived discrimination as a chronic stressor erodes the LBC’s external support system (social support), which in turn leaves them without the necessary relational resources to build or maintain their internal coping capacities (psychological resilience), ultimately leading to the lower levels of subjective well-being.

Furthermore, the multi-group structural equation model revealed that sex did not moderate the indirect effects of social support and psychological resilience on the association between perceived discrimination and subjective well-being, nor the direct and total effects of perceived discrimination on subjective well-being, supporting Hypothesis 5. This finding suggests that the mechanism of stress erosion–whereby perceived discrimination depletes social support, which in turn erodes resilience–functions similarly for both left-behind boys and girls in our sample. This result suggests that the constant and widespread discrimination experienced by LBC may universally deplete their resources of social support and resilience, affecting their psychosocial well-being in a comparable manner regardless of sex.

While the context of Chinese LBC is unique–rooted in internal rural-to-urban migration and the hukou system–the mechanisms identified in this study are consistent with findings from other vulnerable populations. Research among immigrant children and ethnic minority youth has documented that both social support (Oppedal, 2011; Katsiaficas et al., 2013; Fernández et al., 2015), and psychological resilience (Vargas et al., 2021; Cénat et al., 2022) partially explain the relationship between discrimination and health outcomes. Furthermore, a serial mediation pathway involving social support and resilience has also been identified in people living with HIV (Yuan et al., 2024). This suggests that the stress erosion process, wherein chronic discrimination depletes external and internal resources, may represent a pathway impacting the mental health of stigmatized populations across different cultures.

Our findings have several practical implications. Perceived discrimination was negatively associated with subjective well-being among LBC. Therefore, to improve their well-being, eliminating discrimination and prejudice against LBC is the first step. In cases where discrimination cannot be completely eliminated, teachers and caregivers should pay more attention to LBC’s experiences of discrimination and guide them to formulate correct attributions regarding perceptions of discrimination. This would help eliminate the negative emotional experiences caused by perceived discrimination (Ma et al., 2022). Meanwhile, interventions aimed at improving self-acceptance (Fernández et al., 2015) and enhancing a sense of control (Szalacha et al., 2003) can help prevent the internalization of prejudice and discrimination. Our findings highlight the important roles of social support and psychological resilience in the relationship between perceived discrimination and subjective well-being. Thus, tailored psychosocial interventions that aim to enhance LBC’s social network and improve their resilience could facilitate the mitigation of the detrimental effects of discrimination on mental health. To build psychological resilience, group psychological counseling focusing on emotional regulation and coping skills can be utilized to enhance the psychological resilience of LBC in combating internalized discrimination (Ma, 2021). Resilience-based intervention programs (e.g., setting positive expectations, improving self-efficacy), which are incorporated into regular school hours, have also shown positive outcomes (Qu et al., 2024). Meanwhile, effective programs enhancing social support are warranted. Educating LBC about the importance of social support and the availability of associations that can provide it is a key element in an intervention program (Fernández et al., 2015). Promoting regular communication between LBC and their parents through digital platforms (Xie et al., 2024), as well as organizing activities to enhance social bonding at the level of family caregivers, school teachers, peers, and the community (Su et al., 2017), can help LBC build a strong social support network. Crucially, our serial mediation finding implies that these interventions must be integrated. As our stress erosion model shows, strengthening social support may be a prerequisite for building resilience. Interventions that only teach coping skills (internal) to a child who remains socially isolated (external) will likely fail. We advocate for integrated programs that, for example, use group therapy to simultaneously build coping skills and foster a new peer-support network, addressing both mediators in our model.

This study also has some potential limitations. First, although the instruments showed good reliability and validity, the current study used a single evaluation method, and all the data were based on self-reports of LBC. The use of paper-and-pencil surveys with verbal explanations for younger pupils, while intended to improve comprehension, may have introduced interviewer effects or social desirability bias. Future studies should use multi-source data (e.g., peer or teacher reports). Second, our study design was cross-sectional, which cannot establish a causal relationship; longitudinal or experimental studies should be performed to verify our results. Future longitudinal research with at least three time points is crucial to pinpoint when the erosion of social support begins and subsequently triggers a decline in psychological resilience within this population. Third, apart from demographic characteristics, we were unable to account for other unmeasured confounders that may have influenced the associations among perceived discrimination, social support, psychological resilience, and subjective well-being, such as duration of parental absence and peer attachment. Finally, the sample size and limited geographical scope (i.e., only a single county in southwest China) limit the extent to which the findings of this study can be generalized. Thus, there is still some uncertainty about whether our findings can be generalized to all LBC in Sichuan Province or elsewhere in the country.

The current study adds further support to previous findings regarding the negative impact of perceived discrimination on individuals and delineates the path of this impact through a combination of social and individual factors. The findings highlight the importance of considering social support and resilience in the effect of perceived discrimination on subjective well-being among LBC. Reducing discrimination, enhancing psychological resilience, and providing increased social support are effective measures for intervening in the mental health of LBC.