Introduction

Traumatic events are often experienced during development, and can include natural disasters, traffic accidents, the death of relatives, and physical abuse (Adams et al. 2018; Harville et al. 2015; He et al. 2013; Wang et al. 2015). Individuals can exhibit a variety of negative psychological responses to traumatic events (Hovens et al. 2012; Kukihara et al. 2014; Lee et al. 2012; McLean et al. 2015). Post-traumatic stress symptoms (PTSSs) are the symptoms of trauma and stressor-related disorders that are delayed and persistent, and are a result of threats or catastrophic events (American Psychiatric Association [APA] 2013). One recent study reported that the prevalence of PTSD in adolescents during the two years following the Wenchuan earthquake ranged from 13.5% to 23.3% (Fan et al. 2015). A recent study also reported that the prevalence of PTSD among adolescents a decade after the Wenchuan earthquake ranged from 1.3% to 78.3% (Wang et al. 2018).

In addition to PTSSs, trauma can result in behavioral problems such as alcoholism, aggressive behaviors, sleep problems, and internet addiction (IA) (Dalbudak et al. 2014; Dutra and Sadeh 2018; Hsieh et al. 2016; Strom et al. 2012; Tang et al. 2018; Watt et al. 2012; Yates et al. 2012; Zhou and Wu 2020). Although trauma-induced stress can be alleviated by resorting to network communication and games (Yu 2017; Tang et al. 2014), this has the potential to lead to IA. Here, IA refers to a behavior addiction in which individuals use the internet uncontrollably in a way that affects their life, study, or work, and may damage their physical and mental health (Gao and Chen 2006). Recently, post-traumatic IA has attracted widespread attention (Contractor et al. 2017a, b; Lee et al. 2017; Yang et al. 2017), and it has been reported that 23.8% of adolescents had a tendency toward IA after a traumatic event (Duan et al. 2013).

While both PTSSs and IA are common negative reactions to trauma, the relationship between the two is not clear. According to the theory of basic psychological needs (Deci and Ryan 2000), there are three basic psychological needs: those related to autonomy, capacity and relationships. When these needs are satisfied, individuals are more likely to exhibit positive behaviors (e.g., engaging in study). Conversely, this can easily lead to negative behaviors (e.g., excessive time spent using the internet). The loss-of-compensation hypothesis proposes that adolescent development might be challenged by external factors that limit the degree to which their basic psychological needs are met. In such cases, the internet may provide a safe place where they can meet these needs undisturbed. In other words, when teenagers fail to meet their basic psychological needs in the real world, they may turn to the internet to compensate for unsatisfied needs (Gao and Chen 2006; Zhang et al. 2017), which could lead to IA (Kardefelt-Winther 2014; Young 1998). However, the emergence of PTSSs may signal that the basic psychological needs of adolescents have not met in situations in which they experience severe adversity (Corrales et al. 2016). In such cases, adolescents may turn to the internet to achieve psychological satisfaction, and when conscious self-control is diminished (Bandura 1991), IA may develop. Indeed, an empirical study found that IA was significantly associated with higher PTSSs severity (Lee et al. 2017), and indicated that PTSSs can increase IA.

However, which dimensions of PTSSs play a key role in IA is still unclear. To advance our understanding of this issue, some studies have investigated the effects of specific symptom clusters of PTSSs on IA. For example, Contractor et al. (2017a) found that the NAMC symptom cluster had a significant predictive effect on IA, but that intrusive and avoidance symptoms were not related to IA. Thus, distinct PTSSs symptom clusters appear to have different effects on IA. The authors proposed that these different effects could be attributed to factors that moderate the relationship between PTSS symptom clusters and IA. According to previous studies, beliefs may influence an individual’s definitions and concepts of adversity, and their corresponding behaviors (Shek 2005; Shek et al. 2003). Adversity beliefs may be one important moderator of this relationship. Adversity beliefs are beliefs about the cause of, consequences of, and appropriate coping behavior regarding adversity (Shek et al. 2003). According to the model of resilience by Garmezy and Masten et al., some individuals are better able to cope with adversity than others (Garmezy et al. 1984; Masten et al. 1988). This could be due to adversity beliefs, which are considered to be an important psychological resource (Zhang et al. 2014), an important internal factor in individual psychological resilience (Olson et al. 1979), and a protective factor with respect to psychological adaptation and problem behaviors in adolescents (Shek 2004, 2005; Shek et al. 2003). Individuals with positive adversity beliefs are more optimistic, have higher self-efficacy, and regard misfortune as an opportunity for individual development (Shek 2005; Zhao et al. 2013). For individuals with severe PTSSs, positive adversity beliefs might help them to effectively regulate their emotions and behaviors to avoid the development of IA. In contrast, individuals with negative adversity beliefs are more pessimistic and have a relatively low sense of self-efficacy and control (Shek 2005; Zhao et al. 2013). Considering this evidence, we proposed that adversity beliefs moderate the relationship between PTSSs and IA.

While more studies have indicated that PTSSs are related to greater levels of IA and that positive adversity beliefs are related to lower IA, fewer studies have examined the interactive relationship between PTSS symptom clusters and positive adversity beliefs. To address this, we examined and compared the role of four PTSS symptom clusters in IA, and explored the moderating role of adversity beliefs in the relationship between PTSS symptom clusters and IA in adolescents following the Wenchuan earthquake. At 8.0 magnitude, this was the most destructive earthquake in China after the Tangshan earthquake in 1976. Up to 100,000 km2 were damaged by the earthquake, and this included 10 worst-hit cities, 41 heavy-hit cities, and 186 small cities that were classified as general disaster areas. More than 60,000 people died and more than 370,000 people were injured.

Methods

Procedures and Participants

This large-scale study investigated adolescents’ post-traumatic reaction 9.5 years after the Wenchuan earthquake. We recruited participants from Wenchuan county and Dujiangyan city, which were worst-hit by this earthquake. Ethical approval was granted by the Research Ethics Committee of the Faculty of Psychology, Beijing Normal University, prior to the study. We then informed the local bureau of education about the aim of the study and asked them to select and invite several schools to take part. After obtaining approval from the school principals, we selected classes that had around 40 students. Informed consent was obtained from all participants. Students in the selected classes were informed regarding the aim of investigation. After all students had confirmed that they understood the aim of the study, they provided written informed consent. We then sent the questionnaire to them to complete. Trained postgraduate students supervised the conduction of assessments, and provided counseling services to those who needed it.

A total of 776 students participated in this study (mean age = 14.55 years, standard deviation [SD] = 1.47 years; 10 students did not report their age). Of these participants, 430 participants (55.4%) were female, 337 participants (43.4%) were male, and 9 students did not report their sex.

Measures

Negative Life Events

Zhou et al. (2018) revised the Youth Life Event Scale (Liu et al. 1997) to assess negative life events in adolescents. The original scale has 27 items covering six dimensions, such as interpersonal relationships, learning pressure, punishment, loss, health adaptation, and ‘others’. Considering the frequent aftershocks, debris flows, and landslides in Wenchuan earthquake-stricken areas, Zhou et al. (2018) added the dimension of disaster events (four items, including strong earthquakes and debris flows) to the original scale. A 6-point Likert scale was used to assess each item, whereby 1 represented “never” and 6 represented “seriously”. The Cronbach’s α was 0.96 for this scale.

PTSSs

Zhou et al. (2017) revised the DSM-5 PTSD checklist (Weathers et al. 2013) to measure PTSSs. The scale has 20 items, including intrusive symptoms (5 items), avoidance symptoms (2 items;), hyperarousal symptoms (6 items), and negative alterations in mood and cognition symptoms (NAMC; 7 items). A 5-point Likert scale was used to assess the checklist, whereby 0 represented “not at all” and 4 represented “extremely”. The cutoff score for identifying probable PTSD cases was 34 in our study. The revised checklist has good reliability and validity (Cronbach’s α = 0.90; Zhou et al. 2017) for use with young people after the earthquake.

Positive Adversity Beliefs

We used the Chinese adversity beliefs scale (Shek 2005) to assess adversity beliefs in adolescents after the Wenchuan earthquake. The scale includes 9 items that are scored from 1 “strongly disagree” to 6 “strongly agree” and was found to have good applicability in Chinese participants (Shek 2005; Zhao et al. 2013, 2017). The scale has two subscales that assess positive and negative adversity beliefs. In this study, Cronbach’s α for the total scale was 0.74, Cronbach’s α for positive adversity beliefs was 0.84 (7 items), and Cronbach’s α for negative adversity beliefs was 0.59 (2 items). Considering the aim of this study and the reliability of the subscales, we only used data from the positive adversity beliefs items.

IA

We adopted the Chinese internet addiction scale, developed by Chen et al. (2003), to assess IA in adolescents. The scale includes 26 items that are divided into the five following dimensions: tolerance (4 items), withdrawal symptoms (5 items), forced internet access (5 items), time management (5 items), and interpersonal health (7 items). For each item, the scale ranges from 1 (“strongly disagree”) to 4 (“strongly agree”). A higher total score indicates a more severe IA. The Cronbach’s α of this scale was 0.97.

Data Analysis

We used SPSS 17.0 software to conduct data analysis. To examine the moderating role of positive adversity beliefs in the relationship between PTSS symptom clusters and IA, we conducted the following regression procedures according to Wen et al. (2002) suggestions. First, the PTSS symptom clusters and positive adversity beliefs were centered to avoid multicollinearity between the main and interaction effects. We considered sex, age, left-behind experiences, and negative life events as covariates and controlled these for each analysis. Next, we included age, sex, left-behind experiences, and negative life events in the first hierarchical regression to test their effect on IA. The second hierarchical regression only included the effect of PTSS symptom clusters and positive adversity beliefs on IA. The third hierarchical regression model primarily examined the effect of interaction items between PTSSs symptom clusters and positive adversity beliefs on IA. After determining the significance of interactions between PTSS-specific symptom clusters and positive adversity beliefs, simple slope analysis was adopted to test the specific moderating effect of positive adversity beliefs.

Results

Correlations between the Variables

Table 1 shows the correlation results. We found a significant positive correlation between sex and intrusive symptoms, and between sex and hyperarousal symptoms. However, there were no significant relationships between sex and any other variables. Left-behind experiences were significantly correlated with intrusive, NAMC, and hyperarousal symptoms, and with IA. Age was correlated with intrusive symptoms. Negative life events were significantly correlated with positive adversity beliefs, all PTSSs symptoms, and IA. Positive adversity beliefs were negatively correlated with IA, and all four PTSS symptom clusters were positively correlated with IA.

Table 1 Descriptive statistics and correlations between the main variables
Full size table

Testing the Moderating Role of Positive Adversity Beliefs

As shown in Table 2, sex, age, and left-behind experiences had no significant predictive effect on IA, but negative life events significantly and positively predicted IA. While the avoidance symptom cluster had no significant role in IA, the other three symptom clusters had significant positive effects on IA, and adversity beliefs negatively predicted IA. The interaction effect of positive adversity beliefs and PTSS-related NAMC symptoms on IA was marginally significant. The interaction effects between positive adversity beliefs and the other three symptom clusters on IA were not significant. These results suggest that except for the avoidance symptoms cluster, all PTSS symptom clusters increased IA, and that positive adversity beliefs could reduce IA. Positive adversity beliefs moderated the relationship between PTSS-related NAMC symptom clusters and IA, but did not moderate the relationships between other PTSS symptom clusters and IA.

Table 2 Regression analysis results: the role of PTSD and adversity in IA
Full size table

To examine the significance of the moderating effect of positive adversity beliefs, we divided the adolescents into two groups: those with high positive adversity beliefs (M + 1SD) and those with low positive adversity beliefs (M-1SD). Then, we used a simple slope test (Preacher et al. 2006) to examine our aim. Figure 1 shows the results of the simple slope test. The NAMC PTSS symptoms had a weaker predictive effect on IA when there were higher/more positive adversity beliefs (β = 0.24, p = 0.009) than when there were lower/less positive adversity beliefs (β = 0.47, p < 0.001). This suggests that positive adversity beliefs buffer the positive relationship between PTSS NAMC symptoms and IA.

Fig. 1
figure 1

The moderating role of adversity beliefs in the relationship between negative alterations in mood, cognitive symptoms, and internet addiction

Full size image

Discussion

In this study, after controlling for sex, age, life experience, and negative life events, we found that intrusive, NAMC, and hyperarousal PTSS symptoms were significantly and positively related to IA. This suggests that PTSSs is a risk factor for IA (e.g., Hsieh et al. 2016). Additionally, positive adversity beliefs were negatively associated with IA, which indicates that positive adversity beliefs are protective factors for IA. Positive adversity beliefs also buffered the relationship between NAMC symptoms and IA.

We found a positive correlation between intrusive, NAMC, and hyperarousal PTSS symptom clusters and IA. This finding may be attributable to the specific characteristics of each symptom cluster. Individuals with intrusive symptoms often have experiences that are similar to the initial traumatic event, such as flashbacks and nightmares, which can be accompanied by painful emotional reactions (APA 2013). To avoid these intrusive memories and thoughts, individuals often resort to the virtual world to divert their attention and alleviate psychological distress (Schimmenti and Caretti 2010).

Adolescents with hyperarousal symptoms generally exhibit hypervigilance and somatic symptoms of difficulty sleeping and concentrating (APA 2013). They may use the internet to relieve these symptoms, which can lead to a higher risk of IA (Moreno et al. 2015). Individuals may relieve negative emotions by browsing the internet (Elhai et al. 2016; Yang et al. 2017), and individuals with NAMC symptoms may therefore resort to internet use. In addition, for those with psychiatric symptoms, negative mood may impair an individuals’ ability to regulate internet use, which may lead to IA (Larose et al. 2003).

One interesting finding was that the PTSS avoidance symptom cluster was not significantly associated with IA. Individuals with more severe avoidance symptoms tend to avoid contact with people, events, and sites related to trauma and are reluctant to live in their reality (APA 2013); these avoidance behaviors may mean that individuals avoid trauma cues that could be found on the internet.

We also found that positive adversity beliefs were negatively associated with IA. Adolescents with positive adversity beliefs are able to explain the meaning of adversity (such as traumatic events) positively and form positive beliefs that help them to reframe their understanding of the post-traumatic self, other people, and the world, thus helping them face the post-traumatic world in a more positive way. All of these factors could help reduce the use of virtual networks to escape from reality, thus decreasing the chances of IA.

We also found that positive adversity beliefs played a marginally significant moderating role in the relationship between PTSS’ NAMC symptoms and IA. Specifically, compared with fewer positive adversity beliefs, NAMC symptoms had a weaker positive predictive effect on internet addiction when individuals held more positive adversity beliefs. Individuals with more positive adversity beliefs are more optimistic and possess self-efficacy, and consider misfortune to be an opportunity for development (Shek 2005; Zhao et al. 2013). Thus, these individuals may be better at effectively managing their own emotions and perceptions, which may reduce their dependence on the internet to escape reality (Lin et al. 2011; Mohsen et al. 2013; Yang et al. 2017). Therefore, in this study, even if participants with strong positive adversity beliefs had NAMC symptoms they appeared to have relatively weak effects on IA. In contrast, individuals with fewer positive adversity beliefs are more pessimistic, have relatively low self-efficacy, and believe that their own fate is out of their hands (Shek 2005; Zhao et al. 2013). Negative mood and cognition may increase negative coping behavior (Amir et al. 1997), and these individuals might use more external forces (e.g., the internet) to escape real life after traumatic events.

Positive adversity beliefs were found to have no significant moderating role in the relationship between PTSS’ avoidance symptoms, intrusive symptoms, and hyperarousal symptoms and IA. One potential explanation is that, regardless of how positive one’s adversity beliefs are, once an individual has these symptoms they become more irritable, re-experience more emotion and scenes triggered by reminders of the traumatic event, or are less likely to face reality (APA 2013). As a consequence, they may resort to the virtual world by browsing the internet, which can ultimately lead to IA (Tzavela et al. 2015).

This study has several limitations that should be noted. First, although there are many potential moderators between PTSSs and IA we only examined the moderating effects of adversity belief. Future studies should select and test the moderating effect of other variables. Second, this study only focused on adolescents following the Wenchuan earthquake, and so caution should be exercised when generalizing to other samples. Moreover, while there are adversity beliefs that are specific to particular individuals in various cultures, we did not examine this. Future cross-cultural studies could examine the role of culture. Nevertheless, this study found that PTSSs are a risk factor for IA, positive adversity beliefs were a protective factor against IA, and that positive adversity beliefs buffered the relationship between NAMC symptoms and IA. These findings extend those of previous studies on the relationship between PTSSs and IA by examining the role of specific PTSS symptom clusters in IA, and our findings also contribute to the theory of IA. From a clinical perspective, school psychologists should aim to enhance positive adversity beliefs in adolescents. This may help them to reframe the post-traumatic world positively and decrease PTSSs, thus reducing the incidence of IA.