Introduction

In China, with the accelerating pace of urbanization, many people come to cities from rural areas to obtain more job opportunities and earn more money. Accompanying the large-scale transfer of rural labor to cities, an increasing number of underage children followed their parents into cities and became “migrant children (MC)”. In August 2021, the Ministry of Education released the Statistical Bulletin on the Development of Education in China in 2020, which showed that there were 14.30 million children of migrant workers in compulsory education (Ministry of Education of the People’s Republic of China, 2021). The process of migration and living in cities is potentially a severely stressful occurrence for MC and their parents (McCarthy, 1998). Recent studies have shown that MC are often perceived as outsiders and more likely to be bullied than their non-migrant peers, and their prevalence of depression was significantly higher than that of general children (Cui & ** behavior will be initiated, how much effort will be expended, and how long it will be sustained in the face of obstacles or adversity experiences (Bandura, 1977). High self-efficacy beliefs have a positive impact on motivational processes even in extreme adversity. By activating affective, motivational, and behavioral mechanisms in the face of adversity, self-efficacy beliefs can promote one’s resilience. Tsibidaki’s (2021) study showed that students with higher levels of self-efficacy had lower levels of anxiety, maintained their personal and family balance, and showed greater resilience when experiencing an uncommon negative situation such as the COVID-19 pandemic. A study of adolescents with depressed parents found that self-efficacy was not only a strong predictor of sustained mental health but also resilience to mood and behavior (Collishaw et al., 2016).

Currently, few studies have directly examined the effect of PER on children’s self-efficacy. The model of parents’ emotional socialization holds that parents’ emotional socialization behavior, especially their emotional expression and their response to children’s emotions, have a direct impact on children’s emotional awakening and emotion regulation learning (Hajal & Paley, 2020). Since the ability to regulate emotions is the foundation of children’s development and play functions in many fields (Paley & Hajal, 2022), PER may influence children’s self-efficacy. Oliver and Paull (1995) found that parental emotional warmth helps children promote self-efficacy. Franco and Rodrigues (2018) also found that academic self-efficacy and social self-efficacy were positively correlated with parental involvement and positive affect. Parents who are good at regulating their emotions are more capable of using supportive socialized strategies when confronting challenging parenting situations (Ravindran et al., 2022), which may help to establish children’s secure attachment and help them perceive more warmth and support from their parents’ positive responses, thus improving their self-efficacy. As previously mentioned, individuals with higher (vs. lower) self-efficacy will generate more resilience. Therefore, we infer that self-efficacy may mediate the relationship between PER and MC’s resilience.

The relations between PER, self-efficacy, hope, and MC’s resilience

Hope is defined as “the perceived capability to acquire pathways to desired goals, and motivate oneself through agency thinking to use those pathways” (Snyder, 2002). According to Snyder’s hope theory, hope represents the universal tendency to make conscious efforts to achieve a purpose or goal. Children’s hopes are influenced by the environment around them (McDermott & Hastings, 2000). A hopeful family and hopeful parents may help children learn to think hopefully. In the process of their development, children may make mistakes on the path to achieving their goals. When mistakes are made, children may experience negative emotions such as anxiety, frustration, distress, etc. At this moment, parents who can effectively regulate their emotions are more likely to respond in a supportive way to their children’s negative emotions. They encourage their children to express emotions appropriately and discuss problem-solving strategies with them (Fabes et al., 2001). They also praise their children for their efforts to pursue goals, not just for their achievements, which is the most crucial thing parents can do to foster a child with high hope (McDermott & Hastings, 2000). Additionally, parental supportive responses can help the child correct mistakes and continue his or her efforts. Therefore, PER may have a positive effect on children’s hope.

As a factor for an individual’s growth, hope is closely related to resilience (Shorey et al., 2007). Existing studies have shown that adolescents with high hope are more likely to invest effort in their goals (Sharabi et al., 2012). They strive to find pathways to reach their goals and maintain mental energy for them, especially in situations of obstacles or impediments. Compared to low-hope individuals, high-hope individuals have greater psychological well-being, less psychological distress (Snyder et al., 1991), diminished stress responses, more effective emotional recovery (Ong et al., 2006), and fewer externalizing and internalizing problems (Hagen et al., 2005). Thus, hope may mediate the relationship between PER and MC’s resilience.

Studies have shown that there is also a positive relationship between self-efficacy and hope. For example, Duggleby et al. (2009) found that improving an individual’s self-efficacy may foster hope. In light of self-efficacy theory, an individual’s self-efficacy has a positive effect on various motivational variables (Bandura, 1977). Individuals with high self-efficacy have clear goals and can continue working hard in the face of failure or frustration; they are more likely to achieve success, and thus enhance their level of hope. In contrast, people with low self-efficacy have unclear goals and lack motivation in the process of pursuing their goals, which leads to a lower level of hope (Snyder, 2002). He (2015) found that self-efficacy had a positive relationship with hope in Chinese MC. A meta-analysis has found that positive expectations (including self-efficacy, hope, and optimism) predicted resilience to stress (Gallagher et al., 2020). In MC’s family, effective parental emotion regulation can create a harmonious family atmosphere, enhance the self-efficacy of MC in adapting to the city, make them full of hope for themselves and their families, and improve their resilience against negative events.

Current Study

Although previous studies have examined the relationship between PER with children’s emotional regulation, mental health, and social competence (Holzman et al., 2022; Meng et al., 2020), few of them have explored how PER influences MC’s resilience. Based on a previous analysis, the following four research hypotheses are presented: H1. Parental emotion regulation positively predicts MC’s resilience; H2a. Self-efficacy significantly mediates the relationship between PER and MC’s resilience; H2b. Hope significantly mediates the relationship between PER and MC’s resilience; H3. Self-efficacy and hope have a chain mediating effect in the relationship between PER and MC’s resilience.

The relationship between PER and MC’s resilience and the chain mediating effect of self-efficacy and hope will be verified using the structural equation model. The hypothesized model is shown in Fig. 1.

Fig. 1
figure 1

The hypothesized mediating model

Full size image

Methods

Participants and procedure

This study was a cross-sectional survey of a convenience sample. First, we chose Guiyang City, Guizhou Province as our survey destination, because Guizhou Province is one of the provinces with the largest number of rural migrant workers in China, and Guiyang is the capital of this province, which gathers rural migrant workers from all over the province; second, we selected two public schools and two private schools where there were MC enrolled, and these schools were located in different districts of the city; third, the investigated schools have MC and non-MC, and they were in the same classes, so we conducted the survey by testing the entire classes. According to the definition of MC: “children under 18 years of age who have left their original residence and migrated to big cities for at least 6 months”(Shuang et al., 2022), no MC participant was excluded from the sample.

After matching the questionnaires of the students and their parents, the identity of the MC was determined based on the household registration information filled in by their parents. The MC’s participants were 745 parent–child dyads. The children were in the fifth to ninth grade with a mean age of 12.9 years (SD = 1.5, age range = 9 to 17 years) and were evenly distributed between the sexes (49.8% boys and 50.1% girls, one of whom did not report gender information). In China, the general age of students in the 5th to 9th grade is 10 to 15 years old, but there is also a small number of students under 10 or over 15 years of age, especially among migrant children. In our survey data, the number of 9-year-old students is 1, 17-year-old students are 3, and 16-year-old students are 16, representing 2.5% of the total number. Although this is not typical in China, it is normal.

Under the supervision of their head teacher (after training), the children were asked to complete the Adolescent Resilience Scale, the General Self-Efficacy Scale, and the Children’s Hope Scale in their classroom independently using half an hour. The Parental Emotion Regulation Scale was taken home by students after school and completed by their mother or father, then returned to school the next day and submitted to their head teacher. Within the parents’ sample, 50.7% were fathers and 48.9% were mothers (3 parents did not report gender). 6.1% of fathers and 2.5% of mothers had a college degree or higher, and 78% of fathers and 85.4% of mothers had completed junior high or under education. According to local minimum wage standards and different levels of personal income tax thresholds, family monthly income was assessed with five categorical responses, less than 1600 RMB accounts for 16.3%, 1600–3500 RMB accounts for 37%, 3500–5000 RMB accounts for 27.7%, 5000–8000 RMB accounts for 13.2% and more than 8000 RMB accounts for 5.8%, 1.2% parents (n = 9) did not report their family monthly income. Table 1 shows the demographic information of the participants.

Table 1 Demographic characteristics of participants (n = 745)
Full size table

Data was collected in December 2018. This study was approved by school principals and the Ethics Committee of the School of Psychology of Guizhou Normal University, China. Written informed consent was obtained from participating students and their parents. Before completing the survey, all participants were informed about the purpose of the survey and their right to decline. Due to limited research funding, except to pay a labor fee to the head teacher of each class for their organizational work in the investigation, the participants did not receive an incentive for participation.

Measures

Parental emotion regulation

The Parental Emotion Regulation Scale (PERS) was used to assess parents’ emotion regulation ability. The scale was developed by Pereira et al. (2017) and the Chinese version was revised by Lu et al. (2022). It aims to assess the regulation of negative emotions in the context of parenting (Pereira et al., 2017). The revised version of the PERS consisted of 13 items (for example, I am attentive to my child’s emotions and try to understand them) and was divided into four dimensions: orientation to child’s emotions, avoidance of child’s emotions, emotional control, and acceptance of emotions. The items were rated on a 5-point Likert-type scale (0 = never, 1 = rarely, 2 = sometimes, 3 = often, 4 = always). The items were averaged to create a PERS score, with a higher score indicating better regulation of emotional ability. Confirmatory factor analysis showed that the Chinese version of PERS in this study had good construct validity (𝜒2 = 248.97, df = 59, CFI = 0.929, TLI = 0.906, RMSEA = 0.069), and Cronbach’s α coefficient of the PER was 0.86.

Self-efficacy

Self-efficacy was measured using the Chinese version of the General Self-efficacy Scale developed by Schwarzer and Jerusalem (1995), which was translated and revised by Wang et al. (2001). It relates to the self-belief of a person when he or she encounters setbacks or difficulties. There were 10 items on the scale (eg, I can handle whatever comes my way), and participants had to evaluate themselves using a 4 Likert scale (1– not at all true; 4 – exactly true). A higher total score indicates greater self-efficacy. Confirmatory factor analysis showed that the Chinese version of the General Self-Efficacy Scale in this study had an acceptable construct validity (𝜒2 = 247.58, df = 35, CFI = 0.905, TLI = 0.878, RMSEA = 0.095), and the internal consistency coefficient was 0.86.

Hope

Hope was measured by adopting the Chinese version of the Children’s Hope Scale, which was revised by Zhao and Sun (2011). The self-report scale consisted of six items (e.g., I think the things I have done in the past will help me in the future): three items to measure agency thinking, and the other remaining three to measure pathway thinking. The children’s hope score was created by averaging the six items. Items were rated on a 6-point Likert-type scale ranging from 1 (none of the time) to 6 (all the time), with higher scores indicating a greater sense of hope and hopefulness. In this study, the confirmatory factor analysis showed that the Chinese version of the Children’s Hope Scale had an acceptable construct validity (𝜒2 = 50.723, df = 8, CFI = 0.959, TLI = 0.923, RMSEA = 0.089), and had a Cronbach’s α of 0.78, reached an acceptable level of psychological measures.

Resilience

The Resilience Scale for Chinese Adolescents was used to assess MC’s resilience. The scale was compiled by Hu and Gan (2008). It consisted of 27 questions and was classified into five factors: goal planning, help-seeking, family support, affect control, and positive thinking. It includes questions such as “Compared to the result, the process is more beneficial to growth”. Of the 27 items, 15 were rated on a 5-point Likert-type scale ranging from 1 (completely out of line) to 5 (completely in line), and the remaining 12 were reverse scored. The mean was taken for all items, with a higher score indicating better resilience. The confirmatory factor analysis showed that the Resilience Scale for Chinese adolescents in this study had an acceptable construct validity (𝜒2 = 1265.469, df = 314, CFI = 0.824, TLI = 0.804, RMSEA = 0.067), and the Cronbach α coefficient was 0.88.

Data analysis and common methods biases test

SPSS 21.0 (IBM-SPSS Inc., Chicago, IL) was used for missing data, descriptive statistics, and correlation analysis, and Mplus 8.0 (Muthén & Muthén, 1998–2004) was adopted to test the mediating model. The total number of missing for the four scales, which are used to measure parents and MC, respectively, is 283, which represents 0.6% of all measurement data, Little’s MCAR test: 𝜒2 = 5563.42, df = 4924, p < 0.001; indicates that the missing data are not random. Among missing data, the total number of missing data in MC is 199 (including 3 scales to measure MC), Little’s MCAR test: 𝜒2 = 3148.89, df = 2677, p < 0.001; the total number of missing data in parents is 84 (including 1 scale to measure parents), Little’s MCAR test: 𝜒2 = 283.32, df = 229, p = 0.008 < 0.01. The specific missing results of MC and their parents are shown in Table 1. To avoid the effect of common method biases interfering with our results, we took steps, such as some items reverse scored, and children and their parents completed the questionnaire anonymously. In addition, we used the Harman single factor test to measure the common method’s degree of variation (Podsakoff et al., 2003). The results showed that the variance explanation rate of the first extracted factor was 18.49%, much below the critical value of 40%. Therefore, it could be considered that there was no obvious common method deviation problem in this study.

Results

Preliminary analyses

After matching the data of the children and their parents, we performed a Pearson’s correlation analysis on PER, self-efficacy, hope, and resilience. The results are presented in Table 2. Specifically, PER has a significant positive correlation with self-efficacy, hope, and resilience. We also found a significant positive correlation between self-efficacy, hope, and resilience.

Table 2 Descriptive statistics and interrelations among all variables
Full size table

Testing for the hypothesized model

The above analysis showed that PER, self-efficacy, hope, and resilience have a significant pairwise correlation between them. Under control of grade, gender, only child, family income, parental gender, and education level as covariables, our next step was to further examine the other variables’ mutual influence with a mediating effect test.

A structural equation model was built to test our hypotheses (see Fig. 2). The fit of the model using the original items was poor (χ2/df (1760) =2.85, RMSEA = 0.050, CFI = 0. 762, TLI = 0. 753, SRMR = 0.056). To improve the quality of the model fit and reduce random errors, the parcel strategy was applied. For parental emotion regulation, hope, and resilience, the items included in the same dimension were packaged into a parcel. For self-efficacy, three parcels were created using an item-to-construct balance approach (Wu & Wen, 2011). First, we tested the direct effect of parental emotion regulation on resilience. The results indicated a perfect model fit (χ2/df (49) = 2.84, RMSEA = 0.050, CFI = 0. 952, TLI = 0. 939, SRMR = 0.041). We further test the mediating effect. Based on the original model, we added self-efficacy and hope as mediating variables and found that the hypothesized model showed a good fit (χ2/df (139) = 3.17, RMSEA = 0.054, CFI = 0.927, TLI = 0.916, SRMR = 0.048). The results showed that (1) parental emotion regulation positively predicted children’s self-efficacy (β = 0.172, p < 0.01) and resilience (β = 0.112, p < 0.01) significantly, but did not significantly predict children’s hope (β = 0.046, p > 0.05). (2) self-efficacy significantly positively predicted hope (β = 0.646, p < 0.001) and resilience (β = 0.289, p < 0.001). (3) Hope significantly positively predicted resilience (β = 0.288, p < 0.001). For details, see Fig. 2.

Fig. 2
figure 2

The mediation model results. Note: Factor loadings and path coefficients were standardized. PER1-PER4 are four parcels of parental emotion regulation; SE1-SE3 are three parcels of self-efficacy; HO1-HO2 are two parcels of hope; RE1-RE5 are five parcels of resilience. Almost all path coefficients and factor loadings (except for the dotted arrow) are significant at the 0.01 or 0.001 level. *p < .05, **p < .01, ***p < .001

Full size image

The bias-corrected bootstrap** procedure (5000 repeat sampling times) was used to test the different paths’ significance of the mediation effect, and the confidence interval was set at 95%. If the confidence interval does not include 0, it means that the mediating effect is significant. As shown in Table 3, the direct effect of parental emotion regulation on MC’s resilience was statistically significant (BC95% CI = 0.112 [0.033, 0.195]), representing 57.44% of the total effects. The total indirect effect of parental emotion regulation on children’s resilience was significant (BC95% CI = 0.083 [0.035, 0.136]) and represented 42.56% of the total effects. In the serial mediation model, the mediating effect of PER on resilience through children’s self-efficacy was statistically significant (β =0.044, BC95% CI [0.016, 0.083]), accounting for 22.56% of total effects. Furthermore, the chain mediating effect of PER on children’s resilience through self-efficacy and hope was also significant (β = 0.028, BC95% CI [0.011, 0.054]), accounting for 14.35% of the total effects. However, the mediating effect of parental emotion regulation on children’s resilience through hope was not significant (β = 0.012, BC95% CI [−0.007, 0.037]).

Table 3 Standardized effects and 95% confidence intervals
Full size table

We then compared the indirect effect values of the two paths. The 95% bootstrap confidence interval of the difference between indirect effects included zero (95% CI = 0.016 [−0.010, 0.054]). This result indicates that there is no significant difference between the two indirect paths.

Discussion

Parental emotion regulation and MC’s resilience

The results showed that PER could positively predict MC’s resilience, supporting Hypothesis H1. Studies have confirmed that emotion regulation is clearly connected with resilience, and resilience often appears through a person’s emotion regulation (Kay, 2016). Given that the parents of MC also need to adapt to life and work in the city, their resilience is also very important, and this requires them to regulate their emotions in the face of adversity. Morris et al. (2007) proposed an observational learning paradigm, arguing that parents can provide models of emotional display that are imitated by children. Thompson (1994) argued that constant exposure to parents’ emotion regulation may lead the child to imitate their parents’ way of managing emotions, ultimately making them take a similar emotion regulation style when confronted with their own emotionally eliciting situations. In MC’s family, with the help of observation, imitation, and internalization, MC may foster positive emotions similar to those of their parents, increasing their resilience. In contrast, when parents exhibit difficulties in emotion regulation, they tend to attribute malicious intentions to their child’s mental states(e.g., think their child was deliberately disobeying them), which in turn compromises their child’s social functioning (Wang, 2022) and induces children’s internalization and externalization problems (Crespo et al., 2017).

The mediating role of MC’s self-efficacy and hope

Consistent with hypothesis H2a, the results showed that MC’s self-efficacy significantly played an independent mediating role in the relationship between parental emotion regulation and children’s resilience. According to emotion socialization theory (Eisenberg et al., 1998; Morris et al., 2007), PER is related to the development of children’s competence through various mechanisms. Parents who are good at regulating emotions are more likely to use supportive social strategies when their children are emotional (Blandon, 2015). They validate their children’s emotional experiences, talk about emotions with children, and assist their children in clarifying and regulating emotions (Havighurst & Kehoe, 2017). These positive responses set healthy behavioral examples for their children. From their parents’ responses, children may learn how to effectively deal with negative emotions and develop their constructive co** abilities (Eisenberg et al., 1998). Meanwhile, children may gain confidence in parents’ encouragement to strengthen their self-efficacy and believe that they can cope with difficulties in life, thus enhancing their resilience in adversity.

The result showed that the independent mediating role of hope was not significant and hypothesis H2b is not supported. However, the path coefficient of PER to children’s self-efficacy and the path coefficients of self-efficacy to hope and hope to resilience are significant. The result means PER can explain children’s resilience through the chain mediating effect of “self-efficacy to hope”, Hypothesis H3 is verified. For hypothesis H2b, Yadav’s research (2010) had similar results to ours. He found that parents with high emotion regulation ability were more likely to provide emotional support to children, while parental emotional support was a relatively low predictor of children’s hope when compared to informational and tangible social support. This may indicate that PER does not directly have a significant impact on MC’s hope but has a significant indirect impact on it with the help of other factors (e.g. self-efficacy). The reason may be that hope is more inclined to the future, while PER is more focused on the present, so the direct influence of PER on hope is not significant, and it needs to be indirectly influenced by self-efficacy. For hypothesis H3, the result is consistent with the study by Li and Yin (2015), which found that children’s self-efficacy predicted their hope significantly. Individuals with high self-efficacy can develop reasonable methods to achieve their goals and have a higher level of motivation to achieve their goals. When faced with obstacles, they would sustain their efforts, find alternative ways, and show a higher level of hope. The results of our research suggested that PER may improve the level of MC’s hope and ultimately promote the development of their resilience mainly through enhancing MC’s self-efficacy.

Implications

This study has several important implications. From a theoretical point of view, first, although many studies have examined the effects of PER on children’s mental health and social development, few studies have specifically explored the effects of PER on children’s resilience from the perspective of resilience theory. Our results showed that PER positively predicts MC’s resilience, providing a new perspective to understand the association between PER and children’s resilience and enriching the research content of resilience theory. Second, we explored the underlying mechanisms of the relationship between PER and MC’s resilience and found that the independent mediating role of self-efficacy and the chain mediating role of self-efficacy and hope between PER and MC’s resilience, contribute to a deeper understanding of how PER influences children’s resilience. The findings suggest that PER not only directly influences children’s resilience through parent-child interaction but also indirectly influences children’s resilience through the mediating role of children’s self-efficacy and hope. This indicates that PER is an important support resource for children’s resilience. From a practical point of view, our study shows that improving the urban adaptation of MC by enhancing their resilience, PER and children’s self-efficacy may be key influencing variables. This finding will be helpful to remind people to take effective intervention measures to make MC’s parents realize the importance of their emotional regulation, and help them optimize their emotional regulation, thereby providing strong support resources for improving MC’s resilience, and ultimately facilitating MC’s urban adaptation.

Limitations and future directions

Although this study has theoretical and practical implications, it also has certain limitations. First, this study is cross-sectional survey research. However, the regulation of parental emotions that influence children’s resilience may be a dynamic developmental process. The results cannot reflect the developmental characteristics of the effect of PER on MC’s resilience. Furthermore, the cross-sectional design prevents any conclusions about the direction of the effects. Although PER may predict child resilience, it is equally plausible that child resilience contributes to PER. Therefore, the follow-up research design may be the better way to explore the relationship between PER and MC’s resilience. Second, the study only considered self-efficacy and hope as mediating variables between PER and children’s resilience and did not consider other variables and other effects. In fact, besides self-efficacy and hope, the effect of PER on MC’s resilience may be influenced by other variables, such as children’s emotion regulation, parent-child attachment, etc. In addition, indirect influence may also include a moderating effect. With these limitations in mind, more variables and indirect effects should be introduced in future research to further reveal the impact mechanism of PER on the resilience of MC. Third, the participants in this study came from the same region in China, so the sample’s representativeness was limited for Chinese MC. Thus, future studies should consider the diversity of samples, such as choosing students and parents from different provinces or regions of China. Fourth, this study did not collect the migration history of MC, but only defined that MC should live in the city for at least 6 months. In future research, we should consider the migration history of MC to explore their resilience more accurately.

Conclusions

In general, this study expands previous research on the impact of PER on children’s psychological and social development. Parental emotion regulation ability positively influences the resilience of MC, and children whose parents can effectively regulate their own emotions have a relatively higher level of resilience. In addition, children whose parents had higher emotion regulation competence showed higher self-efficacy, which in turn promoted their resilience. PER can also influence MC’s resilience through the chain mediating effect of “self-efficacy to hope”. In general, these findings indicate that parental emotion regulation can influence MC’s resilience through psychological factors (such as self-efficacy and hope).