How Resilience and Social Support Promote PostTraumatic Growth after Wenchuan Earthquake: An Integrated Perspective

People could recover and obtain growth after traumatic experience. Two vital protective factors for posttraumatic growth (PTG) are resilience and social support. However, it remains unclear how the two factors affect PTG. Some research demonstrated that resilience could indirectly promote PTG though social support, while other researchers found that social support could indirectly promote PTG through resilience. Integratively considering, the present study tested whether these two perspectives simultaneously tenable. We recruited 2336 high school students from the Wenchuan earthquake-stricken area to participate in our survey. Scales of Connor-Davidson Resilience, Perceived Social Support and Post-traumatic Growth were joined in one questionnaire. After ltering the invalid questionnaires, 2150 samples (aged 15–18 years old, M = 16.21, SD = .79; 844 females) remained for further analyses. The results demonstrate that resilience, social support and PTG are signicantly correlated with each other. Resilience partly mediates the relationship between social support and PTG, meanwhile social support partly mediates the relationship between resilience and PTG. It supports the integrated perspective that resilience and social support can directly promote PTG and indirectly promote PTG through each other. Therefore, in the psychological assistance of people endured adversity, we can cultivate their internal resilience and help them reconstruct their external social support resources, so that they would integrate internal and external resources to achieve more post-traumatic growth.


Introduction
About thirteen years ago, a big earthquake hit Wenchuan, resulting in over eighty-seven thousand people's deaths or missing. Years later, with the experience of this adversity, the induced post-traumatic stress disorder symptoms could be transformed into post-traumatic growth (PTG) based on the time effect (Goenjian et al., 2000). PTG refers to the positive psychological changes of individuals after experiencing traumatic events (Jayawickreme & Blackie, 2014). The positive psychological changes mainly include improved self-awareness, life values, interpersonal experiences in addition to other positive outcomes (Tedeschi & Calhoun, 1996). Many protective factors have been found in promoting PTG, such as personality traits (e.g., optimism, Bozo Plenty of evidence revealed that resilience is positively associated with PTG (Coifman, Bonanno, & Rafaeli, 2007;Kilmer, 2006;Nishi, Matsuoka, & Kim, 2010). It is commonly acknowledged that resilience is an individual's capability to cope with and adapt to di cult situations. When an individual is dealing with pressure or depression, the recovery capability will be promoted (Pidgeon, Ford, & Klaassen, 2014). How does resilience promote PTG? Firstly, when the level of individual resilience is high, it is usually accompanied by optimism and positive countermeasures. These characteristics can help individuals to deal with negative life events in a successful way so as to maintain their good mental and physical health (Connor & Davidson, 2003). Secondly, resilience can help individuals identify the signi cance of the predicament with positive experience of the past to alleviate psychological pressure, encouraging individuals to positively construct their lives and ultimately achieve PTG (Prati & Pietrantoni, 2009). Finally, they are more optimistic in their expectations for the future, with employing a positive method to solve problems (Burton, Pakenham, & Brown, 2010). In conclusion, individuals with high resilience show more positive and optimistic attitudes towards past negative events and future affairs, thus contributing to PTG. Based on the evidence provided, resilience is hypothesized as a protective factor in promoting PTG.
In addition to resilience, a large number of studies have found that social support as an interpersonal Perceived social support refers to the emotional experience and satisfaction of individuals in social relations, including being respected, supported and understood (Sarason, et al., 1991). Social support can improve interpersonal relationships and enhance intimacy with others, which also helps to increase the possibility of emotional sharing and cognitive reassessment and plays a key role in in uencing the Social support, therefore, is widely known as an interpersonal promoting factor for PTG.
It is easy to understand that resilience and social support could promote PTG. However, the real question is what the relationship between resilience and social support is. On the one hand, previous studies found that the higher the level of individual resilience is, the higher the degree of perceived social support will be  Fig. 1A. In this effect pattern, resilience is an independent variable and social support is a mediating variable in the relationship between resilience and PTG.
On the other hand, social support can also better promote resilience. When encountering adversity, individuals could get assistance from their families, schoolmates, and friends. These social supports serve as social resources for developing resilience (Gilligan, 2000). Previous research proposed a dynamic perspective (Freitas & Downey, 1998) and developmental perspective (Aldwin & Sutton, 1998) for resilience trajectories, both of which emphasized social support's function on resilience. Our previous study also veri ed the mediating role of resilience in the relationship between social support and PTG (Xin, Bai, Chen, Zhu, & Liu, 2019), as shown in Fig. 1B. In this effect pattern, social support is an independent variable and resilience is a mediating variable in the relationship between social support and PTG.
Given that both mediating models ( Fig. 1A and Fig. 1B) have considerable evidence, the attention is drawn to whether the two mediating models are simultaneously tenable in the sampling population in Wenchuan earthquake-stricken areas, especially the adolescents. For adolescents who are in the critical period of physical and mental development, the plasticity of cognitive development is high, which provides a great opportunity to achieve PTG. When the earthquake happened, they were children about six or seven years old. After twelve years, they might have formed better resilience, got more social support, and achieved greater PTG. Thus, we attempt to test whether these two mediating models could simultaneously tenable in this population and this may provide a reference to psychological assistance intervention for people hit by adversity.

Participants
With a cluster sampling strategy, 2336 students from two high schools in Wenchuan earthquake-stricken areas were selected for this survey. The research project was approved by the Institutional Review Board of Psychology Department of Southwest University of Science and Technology. The study conforms to the provisions of the Declaration of Helsinki. All the participants gave informed consent to participate in the study. The participants would be excluded for a further analysis if they satisfy one of the three criteria: over a half of the questionnaire is un nished; one or more of the scales are lled with the same answer; the hometown of the participant is not located in Wenchuan earthquake-stricken areas. Having gone through a screening process, 2150 participants (aged 15-18 years old, M = 16.21, SD = .79, 844 females and 1306 males, 1108 senior rst grade and 1042 senior second grade) were included for further analyses.

Materials and Measurements
Resilience Scale.
Using the three-factor scoring method of Yu and Zhang (2007), the scores are divided into three factors: tenacity, strength, and optimism. There are 25 question items (e.g., "When things are changed, I can adapt myself") with a ve-point scale (0 = not at all, 4 = almost always). The higher the total score is, the higher the level of resilience will be. The original internal consistency coe cient of the English version is .89 and the retest reliability coe cient is .87, while the internal consistency coe cient of the Chinese version is .91. The Cronbach's α in present study is .91.
Perceived Social Support Scale. Social support was measured by the Perceived Social Support Scale (Blumenthal et al., 1987). Considering that the participants of the present study were high school students, "leaders and colleagues" in the sub-scale of the other support were changed into "teachers and classmates". Thus, the scale consists of three sub-scales: family support, friend support and other support (teachers, classmates, relatives). Each sub-scale has four questions and a total of 12 items, e.g., "when I encounter problems, some people (teachers, classmates, relatives) will accompany me". Participants rated on a 7-point scale (1 = disagree, 7 = agree). The total score for perceived social support is aggregated based on the data collected. The higher the score is, the more social support s/he perceives. The original Cronbach's α for the total scale is .88 and it is .92 in the present study.
Post-traumatic Growth Scale.
The Post-traumatic Growth Inventory (Tedeschi & Calhoun, 1996)  . In this study, gender and age were signi cantly correlated with PTG (Table 1). Therefore, when examining the mediation model, gender and age were controlled. Moreover, we also retested the mediation models with Mplus 7.4 so that we can compare the model t indices between mediation models.

Common Method Bias Test
As the main method of data acquisition relies on self-reporting scales, it may lead to common method bias. In the process of data collection, the corresponding control was carried out, such as an anonymous test, reversed narrative questions and different response sentences. To examine common method bias, Harman's single factor test was used (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). All items relevant to the study were subjected to an exploratory factor analysis. The results show that 8 factors can be obtained, while no single factor accounts for the majority of the covariance among the variables (the biggest factor loads value is 22.35%, less than the critical value of 40%). Therefore, no signi cant common method bias exists in the current study.

Test of the Mediating Models
Regression analysis showed that resilience had a signi cant effect on PTG (Eq. 1, Table 2) and on social support (Eq. 3, Table 2). Social support also had a signi cant effect on PTG (Eq. 2, Table 2). When the effects of both resilience and social support on PTG were simultaneously considered, the coe cients of resilience and social support were also signi cant (Eq. 5, were signi cant (Table 3), con rming that social support could partly mediate the relationship between resilience and PTG. Furthermore, we used the mediation command of Mplus with 5000 bootstraps and con rmed the mediation model again, and the Bayesian Information Criteria for the model t is 11352.015. Note: N = 2150, * P < .05, ** P < .01, *** P < .001.  (Table 3), con rming that resilience could partly mediate the relationship between social support and PTG. Thus, the hypothesized mediating models were simultaneously demonstrated (Fig. 2). Furthermore, we used the mediation command of Mplus with 5000 bootstraps and con rmed the mediation model again, and the Bayesian Information Criteria for the model t is 8754.942.

Discussion
The results of this study show that both resilience and social support can positively predict PTG. Meanwhile, social support had a signi cant mediating effect on the relationship between resilience and PTG, and resilience also had a signi cant mediating effect on the relationship between social support and PTG. The results demonstrated that resilience and social support could promote PTG directly and indirectly through each other. If we take the model t indices into account, the model of resilience as mediator is better than the model of social support as mediator.

Integrative Perspectives of PTG
The present study has identi ed an integrated way of both intrapersonal and interpersonal perspectives of PTG. Intrapersonally, individuals actively adjust themselves with their own resilience when facing adversity, and then directly achieve PTG. Interpersonally, individuals receive social support from the external environment, such as their families, friends and schoolmates, and passively achieve PTG with social support. Circulatively, the intrapersonal impetus of resilience could indirectly promote PTG through the interpersonal motive of social support; the interpersonal motive of social support also could indirectly promote PTG through intrapersonal impetus of resilience. . However, there was no previous research that examined the two mediating models simultaneously in the same sampling population as shown in Fig. 2. It suggested that we should arrange resilience training and construct effective social support together, rather than focusing on only one of them. In this way, the intrapersonal and interpersonal perspectives could coordinate with each other and circulatively promote PTG.
Although the results support the integrative perspective of resilience and social support in promoting PTG, we still want to underline that resilience as a mediator is better than independent variable according to the model t indices. It is more likely that the participants developed PTG through social support and partly mediated by resilience. That is to say, intrapersonal resilience might be activated by social support and further promoting PTG. This realizing path of PTG might be the mainstream while resilience as an independent variable on PTG is the tributary. Therefore, reconstructing the social support system is more important and effective than directly cultivating an individual's resilience in the psychological assistance of adversity victims.

Contributions and Limitations
The main contribution of the present study is that it enables a simultaneously test on the two mediating models that resilience and social support could mediate their relationships with PTG. Speci cally, resilience as an intrapersonal impetus and social support as an interpersonal resource will co-function on PTG, and the function may be indirectly realized through reciprocal action. It provides an integrated way of both intrapersonal and interpersonal perspectives to further improve the mental health of adolescents in disaster areas. That is, it is necessary to pay more attention to the cultivation of positive psychological quality (e.g., resilience) and also reconstructing social support (such as parental support, teacher assistance, students' mutual assistance). In this way, when encountering stress, adolescents can not only make it through with the help of external support, but also make use of the positive psychological quality to help themselves.
However, future studies still can overcome some limitations in this study. Firstly, future studies can be expanded to a larger population to reexamine the two mediating models simultaneously. Although the models with a large sample of over two thousand were tested, the participants were mainly Grade One or Grade Two high school students. The sampling objects are relatively centralized. Therefore, the future study could reexamine the model in a larger population of a wider age range. Secondly, longitudinal studies would be helpful to check the causality of the model. The present study validated the model in cross-sectional data, and the causality might not be so stable and effective and could be further tested in a longitudinal study.

Conclusion
With a large sampling pool of adolescents in the Wenchuan earthquake-stricken area, two mediating models were tested, in which resilience and social support partly mediated their relationships with PTG. The results demonstrated an integrated way of both intrapersonal and interpersonal perspectives of PTG. It encourages efforts in cultivating resilience and reconstructing social support simultaneously so as to promote PTG for the victims of adversity. Figure 1 The hypothesized mediating models, PTG represents post-traumatic growth.