Trajectories of academic burnout in adolescents after the Wenchuan earthquake: A latent growth mixture model analysis

Abstract

Changes in academic burnout in adolescents have attracted much research attention; however, most studies assume that adolescent academic burnout is characterized by overall homogenous change and overlook the heterogeneity of burnout change. To address this issue, this study examined distinct latent trajectories of academic burnout in adolescents following the Wenchuan earthquake, China. Adolescents were surveyed at 1 (T1), 1.5 (T2), 2 (T3), and 2.5 years (T4) after the earthquake. Self-reported questionnaires were administered to 391 participants aged 12- to 19-years-old. The results identified three academic burnout trajectories: Increasing (3.9%), low (85.4%), and decreasing (10.7%). Additionally, intrusive post-traumatic stress disorder (PTSD) symptoms were more likely in the increasing group, avoidance PTSD symptoms were more likely in the decreasing group, and PTSD hyperarousal symptoms were more likely in the decreasing group, but less likely in the increasing group. These findings indicate that adolescents experienced heterogeneous academic burnout changes following the earthquake.

Keywords

academic burnout adolescents post-traumatic stress disorder trajectories

Earthquakes are mass traumatic events that not only damage civil infrastructures, but also present victims with physical, psychological, and social challenges. Individuals who have experienced an earthquake may report physical injury and/or psychological disorders (e.g., Bianchini et al., 2015; Chui et al., 2017; Jin, Sun, Wang, An, & Xu, 2018). In addition to post-earthquake physical and mental problems, adolescents may also experience academic-related problems such as academic burnout (e.g., Ying, Wang, Lin, & Chen, 2016; Zhou, Zhen, & Wu, 2017). Academic/school burnout is a syndrome characterized by emotional exhaustion, cynicism, and academic inefficacy owing to the continued failure to successfully manage academic stress (e.g., Schaufeli, Martinez, Pinto, Salanova, & Bakker, 2002). In academic burnout, emotional exhaustion reflects fatigue rather than stemming from interactions with others; cynicism reflects indifference or a detached attitude toward study in general, not necessarily in relation to others (e.g., Schaufeli et al., 2002); and academic inefficacy reflects feelings of incompetency owing to reduced academic accomplishments.

Academic burnout in adolescents following an earthquake can be explained by the demand-resources model (e.g., Bakker & Demerouti, 2007; Demerouti, Bakker, Nachreiner, & Schaufeli, 2001). According to this model, academic activities and other occupations can be categorized in terms of two factors: demands and resources. If an individual's resources are insufficient to meet the demands of engaging in study or work, then burnout may occur (e.g., Cadime, Pinto, Lima, Rego, Pereira, & Ribeiro, 2016; Salmela-Aro & Upadyaya, 2014a; Salmela-Aro & Vuori, 2015; Zhou et al., 2017). An earthquake not only disrupts school infrastructure (e.g., Zhao, Taucer, & Rossetto, 2009), but also creates an uncertain and insecure school setting (e.g., Ying et al., 2016). This may reduce adolescents' academic resources to study and prevent them from meeting their academic demands, which may result in academic burnout (e.g., Ying et al., 2016; Zhou et al., 2017).

However, following an earthquake, victims may receive substantial social support from governments, social groups, and other sources (e.g., Guo, He, Qu, Wang, & Liu, 2017; Ke, Liu, & Li, 2010), which can help them to rebuild their homes. Post-disaster reconstruction of public service facilities and infrastructures in affected areas can substantially improve the mental health of victims (People's Network, 2011). This can help adolescents to gain access to academic resources and thus meet academic demands. In accordance with the demand-resources model (Bakker & Demerouti, 2007; Demerouti, Bakker, Nachreiner, & Schaufeli, 2001), it is likely that academic burnout in adolescents will show temporal changes.

Unfortunately, few studies have examined academic burnout changes in adolescents following trauma experiences (e.g., Narimani, Mohammadi, Rad, & Bytamar, 2017). Many studies on academic burnout have focused on adolescents who have not experienced trauma, and show inconsistent results. For example, some studies have suggested that academic demands increase with age (e.g., Wigfield & Wagner, 2005), and therefore adolescent academic motivation tends to decline over time (e.g., Lepper, Corpus, & Iyengar, 2005; Otis, Grouzet, & Pelletier, 2005). This leads to decreased academic engagement (e.g., Wigfield, Eccles, Schiefele, Roeser, & Davis-Kean, 2006) and a greater likelihood of academic burnout. In line with this proposition, some empirical studies show that adolescent academic burnout increases over time (e.g., Bask & Salmela-Aro, 2013; Kim, Lee, Kim, Choi, & Lee, 2015; Salmela-Aro & Tynkkynen, 2012; Wang, Chow, Hofkens, & Salmela-Aro, 2015). Conversely, other studies on adolescents indicate that academic burnout shows moderate stability over time (e.g., Parker & Salmela-Aro, 2011; Salmela-Aro, Savolainen, & Holopainen, 2009), as it may be related to personality characteristics (e.g., Walburg, 2014).

Such inconsistent findings can be attributed to the limitation of the variable-centered methods used. These methods “assume that individuals come from a single population and that a single growth trajectory can adequately approximate an entire population” (e.g., Jung & Wickrama, 2008, p. 302), and overlook individual differences in academic burnout (e.g., Lee et al., 2010). Indeed, most young people adequately manage adolescence; some experience difficulties in adapting to the transitions and changes of this life stage (e.g., Salmela-Aro et al., 2009), which leads to individual differences in changes in adolescent academic burnout (e.g., Rosellini, Coffey, Tracy, & Galea, 2014). For adolescents in particular, individual differences in obtaining resources and coping strategies may be apparent following a natural disaster (e.g., Fan, Long, Zhou, Zheng, & Liu, 2015); thus, there may be individual differences in adolescent academic burnout. To examine individual differences in academic burnout changes, a person-centered approach should be used (e.g., Mäkikangas & Kinnunen, 2016); this type of approach assumes that the population is heterogeneous in terms of the mean level or changes in the studied phenomenon (e.g., Laursen & Hoff, 2006).

Latent growth mixture modeling is a useful approach that can be used to assess the distinct latent trajectories of adolescent academic burnout. Although some studies have assessed the trajectories of academic performance or activities in children and adolescents (e.g., Bonneville-Roussy, Bouffard, & Vezeau, 2017; Fu, Chen, Wang, & Yang, 2016; Garon-Carrier et al., 2018; Ladd, Ettekal, & Kochenderfer-Ladd, 2017), only one study, to our knowledge, has used a person-centered approach to examine trajectories of academic burnout in adolescents without a history of trauma experiences (e.g., Salmela-Aro & Upadyaya, 2014b). This study identified four latent trajectories: Low-stable, high-decreasing, low-increasing, and low-strongly increasing. Although this study demonstrated the heterogeneity of academic burnout in adolescents, it focused on adolescents without a history of trauma experiences and thus it is unclear whether the findings can be generalized to adolescents who have experienced natural disasters. To build on these previous findings, the first aim of the present study was to examine trajectories of academic burnout in adolescents following an earthquake. Such an assessment enables an examination of the applicability of the demand-resources model to dynamic processes within a post-trauma context. It may also provide schools and teachers with educational proposals for long-term academic intervention following natural disasters.

An additional salient question concerns which predictors differentiate the trajectories of academic burnout in a heterogeneous group of adolescents following an earthquake. Recent studies on trauma have shown that after a traumatic event, post-traumatic stress disorder (PTSD) is common and may increase the likelihood of burnout (e.g., An, Yuan, Liu, Zhou, & Xu, 2018; Zhou et al., 2017). Zhou et al. (2017) examined the relation of PTSD to academic burnout among adolescents following an earthquake, and found that PTSD exacerbated the severity of academic burnout. To explain this, they suggested that individuals with PTSD may use substantial individual resources to relieve the distress; this leads to resource depletion as the individual seeks to cope with academic challenges, which finally results in a mismatch between individual resources and demands and generates academic burnout (e.g., Zhou et al., 2017).

However, the role of PTSD in trajectories of academic burnout in adolescents remains unclear. In addition, whether distinct PTSD symptom clusters are associated differently with academic burnout trajectories has not yet been assessed. To address these questions, the second aim of this study was to examine the role of distinct PTSD symptom clusters in differentiating latent trajectories of academic burnout in adolescents following an earthquake.

Method

Procedures and participants

This study began 1 year after the Wenchuan earthquake. This devastating earthquake, which occurred on 12 May 2008, had a magnitude of 8.0 on the Richter scale and resulted in more than 69,000 deaths and 370,000 injuries. This study focused on the Wenchuan and Maoxian counties in Sichuan province, as these were the areas most severely affected by the earthquake. The local education authorities were first contacted and informed of the aims and methods of the investigation; the authorities indicated that back-up psychological services could be provided if required. On gaining their approval, four middle schools in the two counties were chosen. With the help of the principals and school psychologists, several classes in each school were selected; each class included approximately 40 students. All students in the selected classes were attending school on the assessment date, and all agreed to participate in the study and to complete self-report questionnaires.

This project was approved by the Research Ethics Committee of Beijing Normal University, the local education authorities (in this case, county departments of education), and the principals of the participating schools. Written informed consent forms were obtained from school principals and class teachers. In China, research projects that are approved by local education authorities and school administrators and are deemed to provide services for students do not require parental consent. The study purpose and the autonomy of students were emphasized to participants before the survey. Written informed consent was obtained from each participant, and participants were free to withdraw from the survey at any time. The assessments were conducted at different time points under the supervision of trained psychology postgraduate students.

Assessment of the adolescents took place at four time points: May 2009 (1.0 year post-earthquake; Time 1, T1); November 2009 (1.5 years post-earthquake, T2); May 2010 (2.0 years post-earthquake, T3); and November 2010 (2.5 years post-earthquake, T4). There were 638 adolescents in the first wave measurement (T1). However, as some students dropped out or graduated from school, there was some dropout in each follow-up survey. A total of 611 (95.8%) adolescents from the original sample of 638 completed the second wave survey (T2), 320 (50.2%) adolescents completed the third wave measurement (T3), and 230 (36.1%) adolescents participated in the fourth wave measurement (T4). Barzi and Woodward (2004) have suggested that if the dropout rate in a longitudinal study is over 60%, longitudinal data are not useful and can lead to invalid results. Less than 40% (230/638 = 36.1%) of the original sample participated in the fourth wave measurement, which indicated that the data were not useful. However, as there is a lack of longitudinal data on adolescents following the Wenchuan earthquake, we wished to use as much of the longitudinal data as we could. Therefore, in accordance with analytic methods used in previous longitudinal studies (Bachem, Levin, & Solomon, 2018), we selected those participants who had participated in at least three waves of measurement; this produced final participant totals of 391, 376, 307, and 229 at T1, T2, T3, and T4, respectively. A distribution of attrition test for the final sample indicated that the attrition rates by gender [χ2(1) = 0.64, p = 0.726] and age [t(389) = −1.42, p = 0.157] were randomly distributed, suggesting that attrition rate had a non-significant effect on the results. The mean age of the adolescents was 15.28 (SD = 1.81) years at the first measurement wave, and age ranged from 12.0 to 19.0 years. Of the 391 participants, 226 (57.8%) were female.

Measures

Academic Burnout

Academic burnout was measured using the Academic Burnout Inventory (ABI) developed by Hu and Dai (2007). The ABI consists of 21 items that measures a four-dimensional construct (emotional exhaustion, physical exhaustion, alienation between teachers and students, and academic inefficiency). Each item is rated on a five-point Likert scale ranging from 0 (never) to 4 (always). Total scores for the overall inventory range from 0–84. Scores are calculated by summing the total for all items. High scores indicate high academic burnout levels. In this study, the mean ABI scores were 31.26 (SD = 12.80), 31.06 (SD = 11.99), 34.22 (SD = 12.49), and 33.08 (SD = 12.63) for T1, T2, T3, and T4, respectively. The ABI showed good internal consistency in a previous study of adolescents following an earthquake (Zhou et al., 2017). In the current study, the ABI exhibited good validity [χ²(182) = 564.765, CFI = 0.92, TLI = 0.91, RMSEA (90% CI) = 0.057 (0.052–0.063), SRMR = 0.048] and good internal consistency for the total scale (Cronbach's alphas were 0.87, 0.85, 0.89, and 0.88 for T1, T2, T3, and T4, respectively).

PTSD

The 17-item Child PTSD Symptom Scale (Foa, Johnson, Feeny, & Treadwell, 2001) was used to assess adolescent PTSD. Adolescents rated the frequency of symptoms from 0 (not at all) to 3 (almost always). Subscale scores ranged from 0–15 for intrusion symptoms [mean value: 4.30 (SD = 3.08)], 0–21 for avoidance symptoms [mean value: 5.58 (SD = 3.64)], and 0–15 for hyperarousal symptoms [mean value: 5.06 (SD = 3.18)]. In this study, the scale exhibited good validity [χ²(114) = 370.649, CFI = 0.93, TLI = 0.91, RMSEA (90% CI) = 0.059 (0.053–0.066), SRMR = 0.048] and good internal consistency for the total scale (Cronbach's alpha = 0.90), intrusion symptoms subscale (Cronbach's alpha = 0.80), avoidance symptoms subscale (Cronbach's alpha = 0.73), and hyperarousal symptoms subscale (Cronbach's alpha = 0.75).

Data analysis

Little's missing completely at random (MCAR) test was used first to analyse missing response values. The results revealed that the data were not missing completely at random, χ²(87) = 126.255, p = 0.004. We also tested the differences in academic burnout from the first assessment between subjects in the longitudinal sample and subjects who did not follow up. Attrition analysis results showed a significant difference [t(389) = −5.07, p < 0.001], which suggested that the academic burnout attrition data were non-randomly distributed. Therefore, robust maximum likelihood estimation (MLR) was used to handle missing data.

The Mplus 7.0 software (Muthén & Muthén, 2012) was used to analyse the data. To examine the potential role of age and gender in academic burnout, linear conditional latent growth mixture modeling (LGMM) with age and gender as covariates was used first to identify trajectories of academic burnout over time. Then, corresponding time scores in the linear conditional model were used to determine the slope of academic burnout (e.g., 0, 1, 2, 3), as the four measurement waves were conducted at 1, 1.5, 2, and 2.5 years, respectively. As there were four time points, and potential curvilinear patterns of academic burnout change were also considered, the LGMM with a quadratic factor was chosen to allow detection of curvilinear trajectories in addition to linear patterns. To avoid multicollinearity between the linear and quadratic slopes, centered time scores were used to remove convergence problems.

Initially, we examined a series of linear conditional LGMMs (i.e., age and gender as covariates) and linear + quadratic conditional LGMMs with age and gender as covariates from 1- to 6-class solutions. To determine the optimal number of latent classes, the solutions were evaluated and compared according to fit statistics, interpretability, and theoretical considerations. A good model fit is indicated by a low Bayesian information criterion (BIC), low Akaike information criterion (AIC), low sample size-adjusted Bayesian information criterion (adjBIC), higher entropy, a significant Lo–Mendell–Rubin adjusted likelihood ratio test (ALMR-LRT) result, and a significant bootstrap likelihood ratio test (BLRT) result. The values of the model fit indices indicate whether the class solution has a good fit and is a credible classification (e.g., Akaike, 1987; Lo, Mendell, & Rubin, 2001; Nylund, Asparouhov, & Muthén, 2007).

Subsequently, the relation between PTSD symptom clusters and academic burnout trajectories was tested. To predict class membership, PTSD symptom cluster covariates were not included in the latent growth mixture model, as they would have interfered with the mixture solution. Then, the most likely class membership variable for academic burnout trajectories was exported to SPSS; we ensured that the LGMM results matched the participants one by one in SPSS. Next, the most likely class membership variable in the multinomial logistic regression analyses was used to assess the role of PTSD in differentiating academic burnout trajectories.

Results

Trajectories of academic burnout

The fit indices for 1 to 6 class solutions for academic burnout are shown in Table 1. The linear model showed that the BLRT was significant for the 2- to 4-class solutions, but the ALMR-LRT and BLRT were not significant for the 5- to 6-class solutions. As both the ALMR-LRT and BLRT should be significant for an optimal solution, a solution with one fewer classes than the 5- to 6-class solutions seemed acceptable. However, one of the 4-class groups in the linear model had 0.8% (N = 3) participants; therefore, a 3-class solution was considered instead. Similarly, in the linear + quadratic model, the BLRT was significant for the 2- and 4- class solutions, but the ALMR-LRT and BLRT were not significant for the 5- to 6-class solutions. Additionally, one of the 4-class groups in the linear + quadratic model included only 6 (1.6%) participants, indicating that this class should be excluded and a 3-class solution considered for the linear + quadratic model. Furthermore, compared with the 3-class solution in the linear model, the 3-class solution in the linear + quadratic model showed lower AIC and adjBIC; thus, the 3-class solution in the linear + quadratic model was selected as the optimal solution.

Table 1.

Quadratic and linear changes in academic burnout over time

	Linear Weights only						Linear + Quadratic Weights
Model	AIC	BIC	adj BIC	Entropy	ALMR-LRT	BLRT	AIC	BIC	adj BIC	Entropy	ALMR-LRT	BLRT
1 Class	9659.62	9710.98	9669.73	–	–	–	9662.02	9737.09	9676.80	–	–	–
2nd Class	9638.61	9701.82	9651.06	.79	25.57**	27.01***	9638.97	9729.84	9656.87	.79	29.80**	31.04***
3rd Class	9633.03	9708.10	9647.81	.76	10.96	11.57***	9618.53	9725.19	9639.53	.75	27.30	28.45***
4th Class	9631.44	9718.35	9648.55	.79	11.75	12.40*	9611.48	9733.96	9635.60	.81	14.43	15.04*
5th Class	9630.46	9729.23	9649.91	.79	6.61	6.98	9613.70	9751.98	9640.93	.79	5.55	5.78
6th Class	9632.75	9743.36	9654.52	.82	8.17	8.62	9605.09	9759.17	9635.43	.72	9.75	10.16

Note: ***p < 0.001; **p < 0.01; *p < 0.05.

Table 2.

The relation of PTSD symptom clusters to trajectories of academic burnout

	Increasing group vs. Low group		Decreasing group vs. Low group		Increasing group vs. Decreasing group
	OR	95% CI	OR	95% CI	OR	95% CI
T1 Intrusive symptom	1.47**	1.14∼1.89	0.95	.82∼1.09	1.56***	1.17∼2.07
T1 Avoidance symptom	0.90	0.70∼1.15	1.21**	1.06∼1.37	0.74*	0.58∼0.98
T1 Hyperarousal symptom	0.65**	0.46∼0.92	1.10	0.95∼1.28	0.59***	0.41∼0.85

Note: *p < 0.05; **p < 0.01; ***p < 0.001.

For three latent trajectories of academic burnout in the linear + quadratic model (with age and gender as covariates), the intercept was 9.83, 22.06, and 12.39 for group 1, 2, and 3, respectively. The linear slope was 22.96 (SE = 4.80, p < 0.001) and the quadratic slope was 5.79 (SE = 9.30, p = 0.533) for group 1; the linear slope was −6.72 (SE = 5.02, p = 0.002) and the quadratic slope was −8.13 (SE = 10.04, p = 0.418) for group 2; and the linear slope was 5.49 (SE = 4.57, p = 0.230) and the quadratic slope was −16.00 (SE = 8.63, p = 0.063) for group 3. Figure 1 shows the patterns of academic burnout trajectories in adolescents. Group 1 was labeled the ‘Increasing group’ (N = 15, 3.9%) and was characterized by an initial low level and an increasing tendency for burnout, followed by an increasing tendency for burnout from T2 to T4. Group 2 was labeled the ‘Decreasing group’ (n = 41, 10.7%) and was characterized by an initial high level and decreasing tendency for burnout from T2 to T4; the level of academic burnout at T4 was lower. Group 3 was labeled the ‘Low group’ (N = 328, 85.4%) and was characterized by a constantly low tendency for burnout across T2 to T4.

Figure 1.

Three latent trajectories of academic burnout.

The role of PTSD symptom clusters in differentiating trajectories of academic burnout

In the table 2, we found that intrusive, avoidance, and hyperarousal PTSD symptom clusters played a significant role in distinguishing different trajectories of academic burnout. Compared with the low group, intrusive PTSD symptoms (OR = 1.47, 95% CI = [1.14–1.89]) were more likely, and hyperarousal PTSD symptoms (OR = 0.65, 95% CI = [0.46–0.92]) less likely, in the increasing group; avoidance PTSD symptoms were more likely in the decreasing group (OR = 1.21, 95% CI = [1.06–1.37]). Compared with the decreasing group, intrusive PTSD symptoms were more likely (OR = 1.56, 95% CI = [1.17–2.07]), and avoidance and hyperarousal PTSD symptoms less likely (OR = 0.74, 95% CI = [0.58–0.98]; OR = 0.59, 95% CI = [0.41–0.85], respectively), in the increasing group. These results indicated that intrusive PTSD symptoms were more likely in the increasing group, PTSD hyperarousal symptoms were more likely in the decreasing and low groups, and avoidance PTSD symptoms were more likely in the decreasing group.

Discussion

To our knowledge, this is the first study to examine the trajectories of adolescent academic burnout after natural disasters. After controlling for age and gender, we found three distinct latent trajectories characterized by increasing (3.9%), low (85.4%), and decreasing tendency (10.7%) for burnout. These findings extend Salmela-Aro and Upadyaya's (2014b) results for adolescents with no experience of trauma, and suggest that academic burnout shows a heterogeneous distribution in adolescents following an earthquake. In addition, we found that distinct PTSD symptom clusters had different roles in distinguishing the latent trajectories of adolescent academic burnout.

Specifically, we found that most adolescent academic burnout was low over time, which is consistent with Salmela-Aro and Upadyaya's (2014b) findings and further supports Salmela-Aro and colleagues' (2009) conclusion that most adolescents have no problems in managing their adolescence. Help and support from governments, social groups, and other sources were provided to victims over a long period following the Wenchuan earthquake (e.g., Ma et al., 2011; Ni, Chow, Jiang, Li, & Pang, 2015). Therefore, most adolescents received sufficient resources to meet the demands of academic activities, and thus may show lower academic burnout over time. Nevertheless, we found two additional smaller trajectories, increasing and decreasing academic burnout, indicating that the tendency toward change in academic burnout was not homogenous following the natural disaster.

We found that 3.9% of adolescents showed an increasing tendency for academic burnout over time. This may be attributed to an increase in academic demands over time. Older children face more challenging academic workloads and thus greater academic task demands (Finn, 1989); adolescents experience more academic stress, which may lower engagement and increase burnout (e.g., Durán, Extremera, Rey, Fernández-Berrocal, & Montalbán, 2006; Maslachi, Jackson, & Leiter, 1996). In addition, parents and teachers of older children place more importance on academic study, leading them to exercise more control over adolescents and limit their freedom (e.g., Chai & Gong, 2015). This may lead to a mismatch between autonomy needs and the environment, which eventually results in increased likelihood of burnout.

Although some adolescents in this study showed an increase in academic burnout, others (10.7%) showed decreasing academic burnout over time. Some adolescents found it difficult to cope with the disrupted school infrastructure and associated uncertainty in the short time since the earthquake, and the insufficient academic resources to meet their academic demands may have led to academic burnout (e.g., Ying et al., 2016; Zhou et al., 2017). However, improvements in post-disaster work, public service facilities, and infrastructures in the area affected by the earthquake had positive effects on victims' mental health (People's Network, 2011). Some adolescents were able to obtain the academic resources needed to meet academic demands, resulting in a gradual decrease in post-disaster burnout.

We also examined the role of PTSD symptom clusters in differentiating distinct trajectories of academic burnout and found that the increasing academic burnout group showed more intrusive PTSD symptoms. The presence of intrusive symptoms indicates that traumatic events can be re-experienced in various ways (American Psychiatric Association, 2013), as individuals ruminate about the trauma-related event (e.g., Bauwens & Tosone, 2014; Morris, Shakespeare-Finch, Rieck, & Newbery, 2005). This process may deplete adolescents' cognitive resources for academic activities, which may lead to a mismatch between academic resources and demands and a greater likelihood of burnout over time.

In addition, we found that the decreasing or low academic burnout groups showed more PTSD hyperarousal symptoms. Dekel, Ein-Dor, and Solomon (2012) have suggested that individuals in a hyperaroused state show greater participation in activities, which increases their appreciation of life, their perception of new possibilities, and their personal strengths (e.g., Tiamiyu et al., 2016). Such adolescents may prefer to engage in academic activities and experience positive emotions in relation to these activities; thus, their academic burnout may decrease or remain low over time rather than increasing.

An interesting finding is that the decreasing academic burnout group was more likely to show the PTSD avoidance symptom cluster. This may be because individuals engaging in avoidance often make deliberate efforts to avoid thoughts, memories, feelings, activities, objects, situations, or people who arouse recollections of the traumatic event, and may avoid talking about it (American Psychiatric Association, 2013). This may distract their attention from the trauma to some extent; therefore, adolescents may find it helpful to focus on their studies. This tendency might increase adolescent engagement in academic activities and decrease academic burnout over time.

Several study limitations should be noted. First, academic burnout trajectories were assessed during the 2.5 years since the Wenchuan earthquake. However, the first wave measurement was 1-year post-earthquake, which means that burnout during the year after the earthquake was not assessed. Thus, the academic burnout trajectories assessed were discontinuous owing to a lack of data in the 12 months following the earthquake. Second, we focused on PTSD symptom clusters as predictors of academic burnout and did not assess other psychological factors. Third, as this longitudinal study spanned four time points over 2.5 years after the Wenchuan earthquake, participants might have experienced other traumatic events during these 2.5 years that could have affected the change in academic burnout. However, we did not examine experiences of other traumatic events. Fourth, because of the study design limitation, we did not assess the demographics of target regions when we carried out the investigation. Additionally, although we used MLR to handle missing data in the LGMM analysis, some missing data remained. Moreover, PTSD was assessed using only the Child PTSD Symptom Scale. Future studies should consider the use of other measures of PTSD symptoms (e.g., DSM-5-based PTSD scales).

Despite these limitations, this is the first study to examine trajectories of academic burnout in adolescents following trauma events. The findings indicate that adolescent academic burnout does not remain stable over time; rather, traumatic events and their academic burnout trajectories are heterogeneous. This study has demonstrated the applicability of the resource-demand model for explaining the dynamics of long-term academic burnout following natural disasters. More importantly, it also extends previous work on academic burnout trajectories in adolescents with no experience of trauma experiences compared to adolescents who have experienced trauma, and provides a new perspective for longitudinal studies of academic burnout in adolescents. The clinical implications of the findings are as follows. Although most adolescents were in the low academic burnout group, some experienced delayed-onset academic burnout. Therefore, educational interventions for academic burnout are needed, so that teachers can help students to focus on their beliefs and their understanding of themselves and their environment (Wilson & Buttrick, 2016), which will increase their engagement in academic activities. The results showed that specific PTSD symptom clusters may affect burnout trajectories. Therefore, school psychologists should focus on relieving PTSD symptoms when treating adolescent academic burnout. The implementation of a teacher-delivered universal school-based program to relieve PTSD symptoms (Berger, Gelkopf, Heineberg, & Zimbardo, 2016) should be considered. Additionally, because academic burnout trajectories are heterogeneous, there is a need for specific psychological or educational interventions to reduce academic burnout in different adolescent groups.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Major Project of Beijing Social Science Fund, China (Grant Number: 15ZDA11); the General Research Program of Education Department of Zhejiang, China (Grant Number: Y201840094).

Author biographies

Xiao Zhou is a Researcher in the Department of Psychological and Behavioral Sciences at Zhejiang University. His research focuses on posttraumatic reactions among adolescents following earthquake and cancer patients. His PhD in Developmental and Educational Psychology is from School of Psychology at Beijing Normal University.

Rui Zhen is an Associate Researcher at the Institute of Psychological Science at Hangzhou Normal University. Her research focuses on academic engagement and burnout in adolescents. She achieved her PhD in Educational Psychology from the School of Psychology at Beijing Normal University.

Xinchun Wu is a Professor in the Faculty of Psychology at Beijing Normal University. His interest focuses on trauma psychology, educational psychology, and clinical psychology. He achieved his PhD in Educational Psychology from the Department of Psychology at Beijing Normal University.

References

Akaike

(1987) Factor analysis and AIC. Psychometrika 52: 317–332. doi:10.1007/bf02294359.

American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders, 5th ed. Washington, DC: American Psychiatric Association.

Yuan

Liu

Zhou

(2018) Dispositional mindfulness mediates the relationships of parental attachment to posttraumatic stress disorder and academic burnout in adolescents following the Yancheng tornado. European Journal of Psychotraumatology 9: 1472989 doi:10.1080/20008198.2018.1472989 .

Bachem

Levin

Solomon

(2018) Trajectories of attachment in older age: interpersonal trauma and its consequences. Attachment & Human Development 18: 1–9. doi:10.1080/14616734.2018.1479871.

Bakker

A. B.

Demerouti

(2007) The job demands-resources model: State of the art. Journal of Managerial Psychology 22: 309–328. doi:10.1108/02683940710733115.

Barzi

Woodward

(2004) Imputations of missing values in practice: results from imputations of serum cholesterol in 28 cohort studies. American Journal of Epidemiology 160: 34–45. doi: 10.1093/aje/kwh175.

Bauwens

Tosone

(2014) Posttraumatic growth following Hurricane Katrina: The influence of clinicians' trauma histories and primary and secondary traumatic stress. Traumatology 20: 209–218. doi:10.1037/h0099851 .

Bask

Salmela-Aro

(2013) Burned out to drop out: Exploring the relationship between school burnout and school dropout. European Journal of Psychology of Education 28: 511–528. doi:10.1007/s10212-012-0126-5.

Berger

Gelkopf

Heineberg

Zimbardo

(2016) A school-based intervention for reducing posttraumatic symptomatology and intolerance during political violence. Journal of Educational Psychology 108: 761–771. doi:10.1037/edu0000066

10.

Bianchini

Roncone

Giusti

Casacchia

Cifone

Pollice

(2015) PTSD growth and substance abuse among a college student community: Coping strategies after 2009 L'aquila earthquake. Clinical Practice and Epidemiology in Mental Health 11: 140–143. doi: 10.2174/1745017901511010140.

11.

Bonneville-Roussy

Bouffard

Vezeau

(2017) Trajectories of self-evaluation bias in primary and secondary school: Parental antecedents and academic consequences. Journal of School Psychology 63: 1–12. doi:/10.1016/j.jsp.2017.02.002.

12.

Cadime

Pinto

A. M.

Lima

Rego

Pereira

Ribeiro

(2016) Well-being and academic achievement in secondary school pupils: The unique effects of burnout and engagement. Journal of Adolescence 53: 169–179. doi: 10.1016/j.adolescence.2016.10.003.

13.

Chai

Gong

(2015) Secondary school students'engagement in learning mathematics: The effects of perceived support from mathematics teachers and mathematics self-concept. Chinese Journal of Special Education, 6: 79–85.

14.

Chui

C. H.

Ran

M. S.

R. H.

Fan

Zhang

Y. H.

Fang

D. Z.

(2017) Predictive factors of depression symptoms among adolescents in the 18-month follow-up after Wenchuan earthquake in China. Journal of Mental Health 26: 36–42. doi:10.1080/09638237.2016.1276542

15.

Dekel

Ein-Dor

Solomon

(2012) Posttraumatic growth and posttraumatic distress: A longitudinal study. Psychological Trauma: Theory, Research, Practice, and Policy 4: 94–101. doi:10.1037/a0021865.

16.

Demerouti

Bakker

A. B.

Nachreiner

Schaufeli

W. B.

(2001) The job demands-resources model of burnout. Journal of Applied Psychology 86: 499–512. doi:10.1037/0021-9010.86.3.499 .

17.

Durán

Extremera

Rey

Fernández-Berrocal

Montalbán

F. M.

(2006) Predicting academic burnout and engagement in educational settings: Assessing the incremental validity of perceived emotional intelligence beyond perceived stress and general self-efficacy. Psicothema 18: 158–164.

18.

Fan

Long

Zhou

Zheng

Liu

(2015) Longitudinal trajectories of post-traumatic stress disorder symptoms among adolescents after the Wenchuan earthquake in China. Psychological Medicine 45: 2885–2896. doi:10.1017/S0033291715000884.

19.

Finn

J. D.

(1989) Withdrawing from school. Review of Educational Research 59: 117–142. doi:10.3102/00346543059002117.

20.

Foa

E. B.

Johnson

K. M.

Feeny

N. C.

Treadwell

K. R.

(2001) The Child PTSD Symptom Scale: A preliminary examination of its psychometric properties. Journal of Clinical Child Psychology 30: 376–384. doi: 10.1207/S15374424JCCP3003_9

21.

Chen

Wang

Yang

(2016) Developmental trajectories of academic achievement in Chinese children: Contributions of early social-behavioral functioning. Journal of Educational Psychology 108: 1001–1012. doi:10.1037/edu0000100.

22.

Garon-Carrier

Boivin

Lemelin

J. P.

Kovas

Parent

Sãguin

J. R.

Dionne

(2018) Early developmental trajectories of number knowledge and math achievement from 4 to 10 years: Low-persistent profile and early-life predictors. Journal of School Psychology 68: 84–98. doi:10.1016/j.jsp.2018.02.004.

23.

Guo

Wang

Liu

(2017) Post-traumatic stress disorder and depression among adult survivors 8 years after the 2008 Wenchuan earthquake in China. Journal of Affective Disorders 210: 27–34. doi:10.1016/j.jad.2016.12.009.

24.

Dai

(2007) A research on middle school students' learning burnout structure [in Chinese]. Psychological Science (Chinese) 30: 162–164.

25.

Jin

Sun

Wang

(2018) The relationship between PTSD, depression and negative life events: Ya'an earthquake three years later. Psychiatry Research 259: 358–363. doi:10.1016/j.psychres.2017.09.017

26.

Jung

Wickrama

(2008) An introduction to latent class growth analysis and growth mixture modeling. Social and Personality Psychology Compass 2: 302–317. doi:10.1111/j.1751-9004.2007.00054.x.

27.

Liu

(2010) Social support and Quality of Life: a cross-sectional study on survivors eight months after the 2008 Wenchuan earthquake. BMC Public Health 10: 573. doi:10.1186/1471-2458-10-573 .

28.

Kim

Lee

Kim

Choi

Lee

S. M.

(2015) Longitudinal analysis of academic burnout in Korean middle school students. Stress and Health 31: 281–289. doi:10.1002/smi.2553.

29.

Ladd

G. W.

Ettekal

Kochenderfer-Ladd

(2017) Peer victimization trajectories from kindergarten through high school: Differential pathways for children's school engagement and achievement?. Journal of Educational Psychology 109: 826–841. doi:10.1037/edu0000177.

30.

Laursen

B. P.

Hoff

(2006) Person-centered and variable-centered approaches to longitudinal data. Merrill-Palmer Quarterly 52: 377–389. doi:10.1353/mpq.2006.0029.

31.

Lee

Puig

Kim

Y. B.

Shin

Lee

J. H.

Lee

S. M.

(2010) Academic burnout profiles in Korean adolescents. Stress and Health 26: 404–416. doi:10.1002/smi.1312.

32.

Lepper

M. R.

Corpus

J. H.

Iyengar

S. S.

(2005) Intrinsic and extrinsic motivational orientations in the classroom: Age differences and academic correlates. Journal of Educational Psychology 97: 184–196. doi:10.1037/0022-0663.97.2.184 .

33.

Mendell

N. R.

Rubin

D. B.

(2001) Testing the number of components in a normal mixture. Biometrika 88: 767–778. doi:10.1093/biomet/88.3.767.

34.

Mäkikangas

Kinnunen

(2016) The person-oriented approach to burnout: A systematic review. Burnout Research 3: 11–23. doi:10.1016/j.burn.2015.12.00.

35.

Liu

Qiu

Wang

Huang

(2011) Risk indicators for post-traumatic stress disorder in adolescents exposed to the 5.12 Wenchuan earthquake in China. Psychiatry Research 189: 385–391. doi:10.1016/j.psychres.2010.12.016.

36.

Maslachi

Jackson

S. E.

Leiter

M. P.

(1996) Maslach burnout inventory manual, 3rd ed. Palo Alto, CA: Consulting Psychologists.

37.

Morris

B. A.

Shakespeare-Finch

Rieck

Newbery

(2005) Multidimensional nature of posttraumatic growth in an Australian population. Journal of Traumatic Stress 18: 575–585. doi:10.1002/jts.20067.

38.

Muthén

L. K.

Muthén

B. O.

(2012) Mplus statistical modeling software: Release 7.0, Los Angeles, CA: Muthén & Muthén.

39.

Narimani

Mohammadi

Rad

N. A.

Bytamar

J. M.

(2017) The role of moral intelligence and risk perception in predicting students' academic burnout and procrastination. Journal of School Psychology 6: 124–145. doi:10.22098/jsp.2017.571.

40.

Chow

M. C. M.

Jiang

Pang

S. M. C.

(2015) Factors associated with resilience of adult survivors five years after the 2008 Sichuan earthquake in China. PloS one 10: e0121033. doi:10.1371/journal.pone.0121033.

41.

Nylund

K. L.

Asparouhov

Muthén

B. O.

(2007) Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling 14: 535–569.. doi:10.1080/10705510701575396.

42.

Otis

Grouzet

F. M.

Pelletier

L. G.

(2005) Latent motivational change in an academic setting: A 3-year longitudinal study. Journal of Educational Psychology 97: 170–183. doi:10.1037/0022-0663.97.2.170.

43.

Parker

P. D.

Salmela-Aro

(2011) Developmental processes in school burnout: A comparison of major developmental models. Learning and Individual Differences 21: 244–248. doi:10.1016/j.lindif.2011.01.005.

44.

People's Network. (2011). Jiabao Wen's talk in a symposium about reconstruction and recovery of post-Wenchuan earthquake. Retrieved from http://politics.people.com.cn/GB/1024/14590742.html.

45.

Rosellini

A. J.

Coffey

S. F.

Tracy

Galea

(2014) A person-centered analysis of posttraumatic stress disorder symptoms following a natural disaster: Predictors of latent class membership. Journal of Anxiety Disorders 28: 16–24. doi: 10.1016/j.janxdis.2013.11.002

46.

Salmela-Aro

Savolainen

Holopainen

(2009) Depressive symptoms and school burnout during adolescence: Evidence from two cross-lagged longitudinal studies. Journal of Youth and Adolescence 38: 1316–1327. doi:10.1007/s10964-008-9334-3.

47.

Salmela-Aro

Tynkkynen

(2012) Gendered pathways in school burnout among adolescents. Journal of Adolescence 35: 929–939. doi: 10.1016/j.adolescence.2012.01.001

48.

Salmela-Aro

Upadyaya

(2014a) School burnout and engagement in the context of demands–resources model. British Journal of Educational Psychology 84: 137–151. doi: 10.1111/bjep.12018

49.

Salmela-Aro

Upadyaya

(2014b) Developmental trajectories of school burnout: Evidence from two longitudinal studies. Learning and Individual Differences 36: 60–68. doi:10.1016/j.lindif.2014.10.016.

50.

Salmela-Aro

Vuori

(2015) Student well-being and dropout at school – Preparing for work life. In: Vuori

Blonk

Price

(eds) Sustainable working lives: Managing work transitions and health throughout the life course, New York, NY: Springer.

51.

Schaufeli

W. B.

Martinez

I. M.

Pinto

A. M.

Salanova

Bakker

A. B.

(2002) Burnout and engagement in university students: A cross-national study. Journal of Cross-Cultural Psychology 33: 464–481.

52.

Tiamiyu

M. F.

Gan

Kwiatkowski

Foreman

K. C.

Dietrich

Elliott

Elhai

J. D.

(2016) Relationships between latent factors of posttraumatic stress disorder and posttraumatic growth. The Journal of Nervous and Mental Disease 204: 344–348. doi:10.1097/NMD.0000000000000490 .

53.

Walburg

(2014) Burnout among high school students: A literature review. Children and Youth Services Review 42: 28–33. doi:10.1016/j.childyouth.2014.03.020.

54.

Wang

M.-T.

Chow

Hofkens

Salmela-Aro

(2015) The trajectories of student emotional engagement and school burnout with academic and psychological development: Findings from Finnish adolescents. Learning and Instruction 36: 57–65. doi:10.1016/j.learninstruc.2014.11.004.

55.

Wigfield

Eccles

J. S.

Schiefele

Roeser

Davis-Kean

(2006) Development of Achievement Motivation. In: Damon

Eisenberg

(eds) Handbook of child psychology, (6th Ed. Vol.3.) Social, emotional, and personality development, New York, NY: John Wiley, pp. 933–1002.

56.

Wigfield

Wagner

A. L.

(2005) Competence, motivation, and identity development during adolescence. In: Elliot

A. J.

Dweck

C. S.

(eds) Handbook of competence and motivation, New York, NY: Guilford Press, pp. 222–239.

57.

Wilson

T. D.

Buttrick

N. R.

(2016) New directions in social psychological interventions to improve academic achievement. Journal of Educational Psychology 108: 392–396. doi:10.1037/edu0000111.

58.

Ying

Wang

Lin

Chen

(2016) Trait resilience moderated the relationships between PTG and adolescent academic burnout in a post-disaster context. Personality and Individual Differences 90: 108–112. doi:10.1016/j.paid.2015.10.048.

59.

Zhao

Taucer

Rossetto

(2009) Field investigation on the performance of building structures during the 12 May 2008 Wenchuan earthquake in China. Engineering Structures 31: 1707–1723. doi:10.1016/j.engstruct.2009.02.039.

60.

Zhou

Zhen

(2017) Posttraumatic stress disorder symptom severity and control beliefs as the predictors of academic burnout amongst adolescents following the Wenchuan Earthquake. European Journal of Psychotraumatology 8: 1412227. doi:10.1080/20008198.2017.1412227 .