Examining longitudinal relationship among effort reward imbalance,coping strategies and academic burnout in Korean middle school students

Abstract

The present study examined the longitudinal relationship between effort-reward imbalance as a stressor and academic burnout as a strain. The study also examined the moderation effect of coping strategies, a problem-focused coping and an emotion-focused coping, in the relationship between effort-reward imbalance as a stressor and middle school students’ academic burnout as a strain using multi-group latent growth modeling (LGM) analysis. The results indicated a significant relationship between the initial status of effort-reward imbalance and the initial status of academic burnout. The results also indicated a significant relationship between the change rate of effort-reward imbalance and that of academic burnout. In addition, problem-focused coping strategies had a moderation effect on the relationship between effort-reward imbalance and academic burnout longitudinally. However, the emotion-focused coping strategy did not have a moderation effect. The authors discuss specific findings and practical implications for teachers and school psychologists.

Keywords

academic burnout coping effort-reward imbalance multi-group latent growth modeling stressor

Stress is a significant risk factor that is associated with maladaptive behaviors, various health problems, and psychological maladjustment in adolescence. Generally, the source of a stress is called a stressor. Academic stress is one of the strongest stressors and includes stress related to testing, grades, future performance expectations, parent expectations, and future goals (Omizo, Omizo, & Suzuki, 1988). When stress arising from stressors is not solved, burnout occurs (Maslach & Schaufeli, 1993). To be more precise, burnout is a psychological syndrome associated with chronic stress (Maslach, Schaufeli, & Leiter, 2001). While research on burnout has been conducted widely in the employment context, burnout has also received attention for a long period of time in the fields of education and psychology (Lee, Puig, Kim, Shin, & Lee, 2010; Schaufeli, Martinez, Marques-Pinto, Salanova, & Bakker, 2002).

Academic burnout is the set of psychological syndromes that occurs due to chronic academic stress and academic pressures. Academic burnout consists of three dimensions: Emotional exhaustion (being exhausted due to study demands), cynicism (having a cynical and negative attitude towards one’s schoolwork), and academic inefficacy (feeling incompetent as a student). Academic burnout can be associated with mental health problems (Zhang, Gan, & Cham, 2007), suicidal behaviors (Rubenstein, Halton, Kasten, Rubin, & Stechler, 1998), and depression (Lee & Larson, 2000) among adolescents. School is a place where students gain knowledge, but it is also a place where they can grow and develop as an individual. The latter has been lacking in South Korean schools because students spend most of their time studying and they experience great pressure due to the competitive college entrance examination (Lee, Puig, Lea, & Lee, 2013; Park, Lee, Choi, Jin, & Lee, 2010). In 2009, almost all middle school graduates in South Korea went on to high school (Organisation for Economic Co-operation and Development, 2009). High academic pressure can be seen in South Korean adolescents as a previous study indicates that South Korean students spend about twice as much time doing schoolwork than American students (Lee & Larson, 2000). South Korean students rated academic stress to be the most stressful factor in their lives (Hwang, 2006; Kim, Kim, Kim, et al., 2000). Before students move into high school where greater levels of academic pressure may occur, it is necessary to study middle school students so that practical interventions can be made to help them prevent severe academic burnout. Also, these practical interventions will allow them to practice using the coping strategies they learn. Therefore, it is necessary to understand the theory and explore useful interventions for students in the academic domains.

Among various theoretical models that explain academic burnout, the Effort Reward Imbalance (ERI; Siegrist, 1998) model is most predominant (Lee, Puig, & Lee, 2012). According to Lee, Puig, and Lee (2012), effort-reward imbalance was applied to school settings to study students’ burnout. In this study, effort has been defined as time or energy students use for academic performance, and reward refers to good grades, esteem for academic activities, and support from significant others related to their school work (e.g., teachers and friends).

Students pursue the balance between their inputs (e.g., time and effort) and their outputs (e.g., status, stability of academic grade, and appreciation). Even if their inputs and outputs are not completely equalized, students will expect to gain rewards as they invest efforts in studying. According to the ERI model, a reciprocal relationship between efforts and rewards is important for psychological well-being. Students who have high efforts and low rewards feel the lack of reciprocity between ‘costs’ and ‘gains’. Based on the theory of reciprocity, individuals expect to gain rewards for their efforts. If there is an imbalance between effort and reward, psychological well-being among students is threatened (Fukuda, Yamano, Joudoi, et al., 2010). Therefore, factors of the ERI (i.e., effort, reward) could be considered as a significant stressor for the academic burnout of students. In this study, based on the study of Lee, Puig, and Lee (2012), effort is defined as time or energy the students use for academic performance, and reward refers to good grades, esteem for academic activities, and support from persons related to school works (e.g., teachers and friends).

In order to explore the intervention for people who experience burnout, several burnout researchers (e.g., Wallace, Lee, & Lee, 2010) reported that coping strategies could moderate the relationship between academic stressor (ERI in the present study) and strains (burnout symptoms). That is, effective coping strategies (i.e., problem-focused coping) may play an important role in buffering the effects of stressors on the burnout symptoms. On the other hand, the relationship between ERI (stressor) and burnout (strain) is partly accounted for by the function of how negatively students cope with stressful situations (i.e., avoidance coping strategies). The ways children and adolescents cope with stressors have potentially functioned as important moderators which influence on strains for adjustment and psychopathology (Compas, Connor-Smith, Saltzman, Thomsen, & Wadsworth, 2001). While problem-focused coping strategies buffered the effects of academic stressor on burnout symptom (Brenner, Sorbom, & Wallius, 1985), emotion-focused coping, including avoidance, may worsen stress (Chan, 1994). Recently, several researchers (e.g., Lowe & Bennett, 2003) suggested utilizing coping strategies as a psychological intervention between the stressor–strain relationships.

Traditionally, coping strategies were categorized into two dimensions: Problem-focused coping and emotion-focused coping (Folkman & Lazarus, 1984). Problem-focused coping refers to a strategy that solves and manages the problem actively. Emotion-focused coping includes effort or strategies for decreasing stressful and unpleasant feelings by using relaxation, substances, social activities or defense mechanisms such as avoidance (Rothmann & van Rensburg, 2002). In most cases when dealing with burnout, the problem-focused coping strategy is negatively related to burnout, while the emotion-focused coping is positively associated to burnout (Griva & Joekes, 2003). According to Leiter (1991), workers from a mental hospital employing the problem-focused coping experienced less burnout symptoms than workers who employed the emotion-focused coping. In general, it is well known that problem-focused coping strategies buffer the effects of job stress (Brenner et al., 1985). Thus, this study focused on the moderation effect of problem-focused coping in the relationships between ERI and academic burnout. In order to examine this hypothesis, the present study conducted two separate multi-group analyses in the problem-focused coping group and the emotion-focused coping group.

In short, the purpose of this study was to examine the longitudinal relationship between ERI as a stressor and academic burnout as a strain among boys and girls from the ages 12- to 13-years-old. Using multi-group latent growth modeling (LGM), we determined the developmental trajectory of academic burnout as well as the determinants of initial academic burnout and the growth of academic burnout. Moreover, the present study explored the moderation effects of two coping strategies, problem-focused coping and emotion-focused coping, to see the longitudinal relationship between stressor based on effort-reward model and middle school students’ academic burnout. The hypotheses were as follows: a) Academic burnout would be longitudinally related to stressor(s) (effort reward imbalance; ERI); b) Academic burnout would be differently associated with stressor(s) (effort reward imbalance; ERI) by coping strategies (e.g., problem-focused and emotion-focused coping strategies) over time.

Method

Participants

The data were collected from a major city (e.g., Seoul) in South Korea by a self-reported paper survey. The sample consisted of 405 middle school students who attended a school that had roughly representative characteristics in metropolitan South Korea in terms of parents’ education level and students’ GPA. The father’s education level were as follows: Elementary school (N = 7, 1.5%), middle school (N = 15, 3.5%), high school (N = 124, 29.4%), university (N = 179, 42.8%), and graduate school (N = 42, 10.1%). The mother’s education level was also collected: Elementary school (N = 5, 1.2%), middle school (N = 28, 6.5%), high school (N = 147, 35.2%), university (N = 156, 37.3%), and graduate school (N = 31, 7.4%). The participants were asked three times and the surveys were conducted at the end of a semester; this comprised six-month interval from the previous surveys being conducted. The first survey was conducted in December (end of fall semester) right before the final exam in seventh grade (N = 405, 100%), the second was carried out in June (end of spring semester) right before the final exam in eighth grade (N = 375, 92.59%), and the third was collected in December (end of fall semester) right before the final exam in eighth grade (N = 388, 95.80%). A total of 412 questionnaires were distributed at first. Finally, a total of 405 survey sets were used for the present study due to the exclusion of seven samples resulting from the absence of identifiers in the second and third grade. All participants were in the seventh grade (i.e., 12-years-old) at the first, and in the eighth grade at the second and third survey. The sample consisted of 176 (43.5%) male and 229 (56.5%) female participants.

Measures

Maslach burnout inventory-student survey (MBI-SS)

This scale, designed to measure the burnout level of students, contains 15 items which assess the dimensions of Emotional Exhaustion (five items), Cynicism (four items) and Academic Efficacy (six items). Cronbach’s alphas of Emotional Exhaustion (e.g., I feel emotionally drained by my studies.) were $0.87 (α_{T 1}), 0.87 (α_{T 2}), and 0.88 (α_{T 3}$ ), Cynicism (e.g., I have become less enthusiastic about my studies.) were $0.83 (α_{T 1}), 0.83 (α_{T 2}), and 0.85 (α_{T 3}$ ), and Academic Efficacy (e.g., I believe that I make an effective contribution to the classes that I attend.) were $0.84 (α_{T 1}), 0.83 (α_{T 2}), and 0.85 (α_{T 3}$ ). Students responded depending on their level of agreement with each item. MBI-SS is a five-point Likert measure ranging from 1 (totally disagree) to 5 (totally agree). High scores on Exhaustion and Cynicism dimensions and low perception of Academic efficacy means the state of burnout. Academic efficacy was reversed as academic inefficacy for measurement of academic burnout. Shin and colleague (2011) examined the factorial validity of the Korean version of the MBI-SS. They confirmed that the three-factor model of burnout is more appropriate compared to the one factor model. Additionally, they found that factor loadings were equivalent regardless of academic setting (i.e., middle and high school) and gender (i.e., male and female) using multi group analyses.

Effort-reward inventory (ERI)

The Effort-Reward Inventory includes two subscales: Effort (e.g., I often feel pressure to spend many hours on studying) and reward (e.g., Teachers like me because I do well in school). Two subscales originally developed for the adult population consist of 17 items in the effort-reward inventory, six items for measuring effort and 11 items for measuring reward (Siegrist, Starke, Chandola, et al., 2004). Using factor analysis, however, Lee and her colleagues (2009) cross-validated and modified the effort-reward inventory for middle school students. As a result, two items from the reward subscale were removed.

Based on the ratio of summed items measuring ‘Effort’ to summed items measuring ‘Reward’ (ERR; Effort-Reward Ratio), the expected students to experience a critical effort-reward imbalance (a ratio >1.0) were easily recognized. A ratio is calculated according to the formula e/(rxc) where ‘e’ is the sum score of the effort scale, ‘r’ is the sum score of the reward scale, and ‘c’ defines a correction factor for different numbers of items in the nominator and denominator. Therefore, a value close of 0 will indicate a condition where there is low effort and relatively high reward, and a value beyond 1 will indicate high level of effort that is not met by the rewards. The effort- reward inventory is a five-point Likert scale from ‘Strongly disagree’ to ‘strongly agree’. Cronbach’s alphas of Effort were $0.77 (α_{T 1}), 0.80 (α_{T 2}), and 0.81 (α_{T 3}$ ). Cronbach’s alphas of Reward were $0.81 (α_{T 1}), 0.80 (α_{T 2}), and 0.75 (α_{T 3}$ ).

The ways of coping checklist (WCC)

The Ways of Coping Checklist (WCC; Vitaliano, Russo, Carr, Maiuro, & Becker, 1985) was cross-validated and revised by Kim (2006) to examine how coping styles operate as a measure of coping methods. It includes problem-focused coping and emotional-focused coping. The problem-focused coping refers to a coping style dealing with the situation directly and seeking to modify or eliminate the source of problems (Vitaliano, Russo, Car, Maiuro, & Becker, 1985). The emotion-focused coping can be defined as handling emotions aroused by a situation and responses whose primary function is to manage the emotions aroused by stressors (Moos & Billings, 1982). In WCC, students were asked to give ratings on a five-point Likert format (1 = never, 5 = always). Reliabilities for the subscale were α = 0.85 (problem-focused coping), α = 0.80 (emotion-focused coping) in the present study.

Data analysis

PASW18.0 was used to estimate the descriptive statistics and inter-correlations of three dimensions for basic understanding of academic burnout symptoms across three time waves. Using latent growth modeling (LGM; Muthen & Muthen, 1998–2006) with AMOS 18.0, we simultaneously analyzed the data both at the level of means and at the level of individual variation across means (variance). LGM estimates different growth components, such as intercept and slope (change), from data consisting of at least three measurements. LGM proposes more flexible and various possibilities to analyse predictors and outcomes of diverse growth components. LGM has numerous advantages over traditional methods for constructing complex and dynamic models that assess change. LGM is able to test both linear and nonlinear growth functions, and to determine which predictor variables affect the rate of development. Full information maximum likelihood (FIML) was used for statistical analysis with missing data. When MCAR (Missing Completely At Random) or MAR (Missing At Random) occurs, FIML is more efficient for estimating than either listwise or pairwise deletion method (Arbuckle, 1996). In this study, the model included two latent constructs (i.e., effort-reward imbalance and academic burnout). In order to measure effort-reward imbalance, three subscales (i.e., effort-reward imbalance for Self-Esteem, effort-reward imbalance for Academic Promotion, and effort-reward imbalance for Academic Security) were utilized as observed variables (see Figure 1). We also measured academic burnout indirectly by three observed variables (i.e., Emotional exhaustion, Cynicism, and Academic Inefficacy).

Figure 1.

The research model.

Results

Descriptive analysis

Descriptive statistics for the measures of academic burnout and effort- reward imbalance slightly increased over time throughout the study period. The means (M) of Emotional Exhaustion were $M_{T 1} =$ 2.60 ( $SD$ = 0.87), $M_{T 2} = 2.82 (SD = 0.88)$ , $M_{T 3} = 2.72 (SD = 0.90) .$ The means (M) of Cynicism were $M_{T 1} =$ 2.45 ( $SD$ = 0.90), $M_{T 2} = 2.56 (SD = 0.91)$ , $M_{T 3} = 2.69 (SD = 0.94) .$ The means (M) of Academic Inefficacy were $M_{T 1} =$ 2.85 ( $SD$ = 0.72), $M_{T 2} = 2.91 (SD = 0.69)$ , $M_{T 3} = 2.94 (SD = 0.72) .$ The means (M) of ratio of effort-reward imbalance for Self-Esteem were $M_{T 1} = 1.09$ (SD = 0.52), $M_{T 2} =$ 1.10 ( $SD$ = 0.53), $M_{T 3} = 1.13$ ( $SD$ = 0.50). The means (M) of effort-reward imbalance for Academic Promotion were $M_{T 1} =$ 0.90 ( $SD$ = 0.41), $M_{T 2} = 0.89 (SD = 0.38)$ , $M_{T 3} = 0.94 (SD = 0.39)$ The means (M) of effort-reward imbalance for Academic Security were $M_{T 1} = 1.17$ ( $SD$ = 0.54), $M_{T 2} = 1.16 (SD = 0.52)$ , $M_{T 3} = 1.20 (SD = 0.56) .$ The range of inter-correlations among dimensions of academic burnout and effort-reward imbalance were from −0.15 (p < 0.001) to 0.68 (p < .001). Table 1 shows inter-correlations among dimensions of academic burnout and effort-reward imbalance. As shown in Table 1, kurtosis and skewness were calculated for normality and the normality was confirmed according to suggestions of Kline (2015).

Longitudinal association between stressor and academic burnout

In latent growth model analysis (LGM), the basic measurement model consists of two latent constructs. One is an intercept which represents the initial state of variables; the other is a slope which means the change rate of the constructs. In this study, the model included two latent constructs (i.e., academic burnout and effort-reward imbalance). After reviewing the change of mean in each time point, this study attempted to determine whether a change model or a non-change model was employed (Kim, Kim, & Jin, 2008). In other words, the current study tried to confirm whether there were longitudinal changes in the stressor and academic burnout or not. If not, latent growth model analysis is not necessary because there is no longitudinal effect. In general, in order to check it, most of the studies used suitable fit indexes for continuous data by Bentler (1990) to test model fit; Non-Normed Fit Index (NNFI, also known as TLI) > 0.90, Comparative Fit Index (CFI) > 0.90, and standard root mean square residual (RMSEA) < 0.08 (McQuitty, 2004). Each model fit index was suggested in Table 2. As shown in Table 2, model fit indexes of the linear change model was more reasonable than non-change models at the three time points. That is, academic burnout and effort-reward imbalance significantly increased over time. In addition, as shown in Table 2, there were intra-changes of individuals in effort-reward imbalance because variance of the initial status and change rates were significant. It could be interpreted that the longitudinal change occurred within a student. However, in case of academic burnout, variance of the initial status was only significant; it indicated that intra-differences existed in the initial status only. This means that there are no intra-differences (within differences) in change of academic burnout for the three time points despite the inter-differences (between differences) of academic burnout in the initial time points.

Table 1.

Cross-sectional and longitudinal correlations among variables.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21
1. EE T₁	1
2. CY T₁	0.61***	1
3. AI T₁	0.46***	0.57***	1
4. EE T₂	0.47***	0.35***	0.28***	1
5. CY T₂	0.48***	0.62***	0.48***	0.55**⁸	1
6. AI T₂	0.39***	0.46***	0.64***	0.43***	0.57***	1
7. EE T₃	0.46***	0.26***	0.29***	0.62***	0.41***	0.34***	1
8. CY T₃	0.49***	0.54***	0.47***	0.48***	0.66***	0.49***	0.55***	1
9. AI T₃	0.42***	0.50***	0.62***	0.40***	0.54***	0.68***	0.38***	0.58***	1
10. EF T₁	0.47***	0.22***	0.11*	0.38***	0.23***	0.20***	0.34***	0.24***	0.14*	1
11. ES T₁	−0.43***	−0.50***	−0.60***	−0.26***	−0.41***	−0.48***	−0.23***	−0.36***	−0.49***	−0.06	1
12. AP T₁	−0.38***	−0.44***	−0.46***	−0.26***	−0.34***	−0.34***	−0.24***	−0.29***	−0.31***	−0.17***	0.53***	1
13. AS T₁	−0.51***	−0.43***	−0.36***	−0.32***	−0.40***	−0.34***	−0.30***	−0.38***	−0.30***	−0.44***	0.39***	0.38***	1
14. EF T₂	0.29***	0.11*	0.02	0.61***	0.23***	0.14**	0.52***	0.27***	0.18***	0.56***	−0.01	−0.05	−0.24***	1
15. ES T₂	−0.41***	−0.49***	−0.57***	−0.34***	−0.53***	−0.56***	−0.33***	−0.43***	−0.54***	−0.09	0.64***	0.43***	0.32***	−0.02	1
16. AP T₂	−0.35***	−0.29***	−0.37***	−0.38***	−0.42***	−0.39***	−0.31***	−0.35***	−0.36***	−0.25***	0.34***	0.49***	0.27***	−0.22***	0.51***	1
17. AS T₂	−0.39***	−0.34***	−0.35***	−0.44***	−0.41***	−0.39***	−0.40***	−0.36***	−0.36***	−0.34***	0.31***	0.29***	0.51***	−0.40***	0.33***	0.34***	1
18. EF T₃	0.28***	0.04	0.02	0.36***	0.11*	0.04	0.52***	0.23***	0.08	0.56***	0.04	−0.01	−0.19***	0.62***	−0.02	−0.15***	−0.25***	1
19. ES T₃	−0.33***	−0.42***	−0.51***	−0.33***	−0.45***	−0.49***	−0.32***	−0.46***	−0.61***	−0.08	0.65***	0.41***	0.35***	−0.08	0.70***	0.40***	0.32***	0.03	1
20. AP T₃	−0.27***	−0.31***	−0.36***	−0.27***	−0.33***	−0.29***	−0.31***	−0.31***	−0.39***	−0.07	0.37***	0.40***	0.25***	−0.17***	0.41***	0.46***	0.32***	−0.04	0.54***	1
21. AS T₃	−0.25***	−0.22***	−0.20***	−0.31***	−0.30***	−0.24***	−0.30***	−0.37***	−0.28***	−0.24***	0.17***	0.20***	0.36***	−0.28***	0.17***	0.19***	0.36***	−0.37***	0.17***	0.19***	1
Skewness	0.08	0.22	0.00	0.02	0.33	0.32	0.05	0.23	0.06	0.16	0.36	0.04	−0.39	0.31	0.20	−0.17	−0.28	0.19	0.37	−0.03	−0.36
Kurtosis	−0.31	−0.49	0.13	−0.42	−0.20	0.46	−0.38	−0.26	0.44	−0.30	−0.09	−0.35	−0.22	−0.04	−0.13	0.00	−0.36	−0.36	−0.14	−0.37	−0.01

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

Note: AI = Academic Inefficacy; AP = Academic Promotion; AS = Academic security; CY = Cynicism; EE = Emotional exhaustion; EF = Effort; ES = Self-esteem.

Multi-group analysis for moderation effect of coping strategies

In order to conduct multi-group analyses, we confirmed invariance tests by comparing model 1, 2, and 3 for moderation effects. For better model fit indices, the error terms were correlated, such as the correlation among the same measures on different times. The baseline model analysis represented that the data fitted the necessary assumptions in this study. The baseline model in this study was relatively a reasonable fit for problem-focused coping groups and emotion-focused coping groups except for TLI: TLI = 0.88, CFI = 0.91, RMSEA = 0.07. After identifying the best fitting baseline model, the sample was divided into a high problem-focused coping group (N = 196) and a low problem-focused coping group (N = 195) by using the median value. The median-split was conducted in the current study based on previous research for examining moderation effect (Kim, Kim, Yang, et al., 2016; Slaney & Brown, 1983). The sample was also divided into a high emotion-focused coping group (N = 208) and a low emotion-focused coping group (N = 183) using the same method. In order to test the differences between the high and low group in each coping strategy, this study conducted two independent analyses. To test whether one common model can be applied to the data of the two groups (i.e., a high problem-focused coping group and a low problem-focused coping group; a high emotion-focused coping group and a low emotion-focused coping group), it is required to have invariance tests such as configural invariance, mean invariance of intercept, mean invariance of slope, and cross-group equality constraint.

As shown in Table 3, two groups of problem-focused coping were compared to test configural invariance. The configural invariance indicated good fit statistics (CFI = 0.90, RMSEA = 0.07). The next step was to test the mean invariance of intercepts. The model (Model 2) was constrained on intercept of effort-reward imbalance and intercept of academic burnout. The difference in χ² values between Model 1 and Model 2 was statistically significant. Although model fit indexes of Model 1 (baseline model) were better than Model 2, the differences of fit indexes were very small: ΔCFI ≤ −0.01 (Hu & Bentler, 1999). This implied that the means of intercepts were equal across groups. The third step was to test the difference in χ² values between Model 2 and Model 3 (i.e., Model whose mean of each slope was constrained) and to compare model fit indexes in order to verify mean invariance of slopes. The difference in χ² values between Model 2 and Model 3 is statistically significant. The difference in model fit indexes of Model 2 and Model 3 did not exist. The results of the analysis of Model 3 also showed an acceptable fit (CFI = 0.89, RMSEA = 0.07). Therefore, it could be concluded that the research model on this study fitted the data of two groups equally.

Table 2.

Model fit indexes, mean, and variance of the non-change model and the change model.

	χ²	p	TLI	CFI		Initial status		Change rate
	χ²	p	TLI	CFI	RMSEA	Mean	Variance	Mean	Variance
Academic Burnout
Non-change	88.28 (df = 22)	0.000	0.92	0.96	0.09	17.35***	11.14***
Change	79.29 (df = 20)	0.000	0.93	0.97	0.09	17.08***	9.48***	0.24*	0.09
ERI
Non-change	132.38 (df = 25)	0.000	0.92	0.95	0.10	12.93***	18.91***
Change	114.20 (df = 23)	0.000	0.92	0.96	0.10	12.99***	30.07***	−0.08*	0.18***

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

Table 3.

Model fit indexes for testing invariance of two problem-focused coping groups.

	Problem-focused coping groups			Emotion-focused coping groups
	χ²	CFI	RMSEA	χ²	CFI	RMSEA
Model 1	705.94*** (df = 244)	0.90	0.07	689.28*** (df = 244)	0.90	0.07
Model 2	746.47*** (df = 246)	0.89	0.07	696.45*** (df = 246)	0.90	0.07
Model 3	749.99*** (df = 248)	0.89	0.07	704.51*** (df = 248)	0.90	0.07

Note: Model 1: Configural invariance; Model 2: Mean invariance of intercept; Model 3: Mean invariance of slope.

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

To examine significant differences of path coefficients between two problem-focused coping groups, three paths within the model were constrained for invariance and it was compared with Model 3. Table 4 presents the fit indexes and standardized path coefficients for the constrained model. Model fit indexes of all constrained models were almost unchanged. However, the path from the intercept of effort-reward imbalance to the slope of academic burnout was statistically different. In short, the path was positively significant in the low problem-focused coping group whereas the path was not significant in the high problem-focused coping group. Except for the path, the differences of other paths were not significant statistically.

Table 4.

Standardized path coefficients and model fit indexes of model constrained on path for invariance in the problem-focused coping groups.

Parameter	High PFC	Low PFC	Δ χ²	High EFC	Low EFC	Δ χ²
E i → B i	0.71***	0.88***	2.21	0.77***	0.85***	0.42
E i → B s	0.23	0.55***	8.21**	0.44**	0.59***	3.03
E s → B s	1.11***	1.25***	0.68	1.16***	1.28***	4.29
All constrained			20.68***			8.28**

Note: B i = intercept of academic burnout; B s = slope of academic burnout; E i = intercept of effort-reward imbalance; E s = slope of effort-reward imbalance; EFC = Emotion-focused coping; PFC = Problem-focused coping.

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

In case of the high problem-focused coping group, the path from the intercept of effort-reward imbalance to the intercept of academic burnout was significant (β = 0.71, p < 0.001). The slope of effort-reward imbalance had an effect on the slope of academic burnout positively (β = 1.11, p < 0.001). However, the path from the intercept of effort-reward imbalance to the slope of academic burnout was not significant. On the other hand, in the low problem-focused coping group, the path from the intercept of effort-reward imbalance to the slope of academic burnout was positively significant (β = 0.55, p < 0.001). This result shows that the problem-focused coping strategy functions as a moderator on the relationship between effort-reward imbalance and academic burnout.

As shown in Table 3, two groups of emotion-focused coping were also compared to test configural invariance. The model fit indexes of both two groups were reasonable. Fit indexes of the model testing configural invariance was satisfied (CFI = 0.90, RMSEA = 0.07). The next step was to test the mean invariance of intercepts like testing the problem-focused coping group. The model (Model 2) was constrained on mean of each intercept to compare with the model of configural invariance (Model 1). The difference in χ² values between Model 1 and Model 2 is statistically significant. The difference in model fit indexes of Model 1 and Model 2 did not exist. Therefore, Model 2 fitted the data better than Model 1. The third step was to test difference of in fit indexes between Model 2 and Model 3 (i.e., Model whose mean of each slope was constrained) and to verify mean invariance of slope. As shown in Table 3, the difference in χ² values between Model 2 and Model 3 is statistically significant. The results showed that ΔRMSEA (0.00) and ΔCFI (0.00) in the mean invariance of slope model revealed appropriate fit statistics as well. This implies that the mean invariance of slope is equal across groups. Model fit indexes of all constrained models were almost unchanged and there were no differences of paths of the model between the low and high emotion-focused coping groups.

To examine significant differences of path coefficients between two emotion-focused coping groups, two paths within the model were also constrained for invariance like problem-focused coping groups and it was compared with Model 3. Table 4 presents the fit indexes and standardized path coefficients for the constrained model. Model fit indexes of all constrained model were almost unchanged. The intercept of effort-reward imbalance had an effect on the intercept of academic burnout positively in both groups ( $high group : β$ = 0.77, p < 0.001, low group: β = 0.85, p < 0.001). The path from intercept of effort-reward imbalance to the slope of academic burnout was also positively significant ( $high group : β$ = 0.44, p < 0.01, low group: β = 0.59, p < 0.001). The path between the slope of effort-reward imbalance and the slope of academic burnout was also significant ( $high group : β$ = 1.16, p < 0.001, low group: β = 1.28, p < 0.001). In this result, there were no moderating effects in emotion-focused coping groups.

It is important to note that some of the standardized coefficients exceeded 1.0. Jöreskog (1999) suggested that regression coefficients could arrive at levels higher than 1. Several researchers also provided the explanation for higher coefficients due to multicollinearity (Grewal, Cote, & Baumgartner, 2004) and the Heywood case (Dillon, Kumar, & Mulani, 1987). However, these do not apply in the current study. Roben, Cole, and Armstrong (2013) similarly reported regression coefficients greater than 1 in their article using latent growth modeling. In this study, the change scores of effort-reward imbalance and change scores of academic burnout are highly correlated to each other (r = above 0.75). This may explain why the path coefficients between the slope of effort-reward imbalance and the slope of academic burnout in LGM reached levels higher than one (1.11).

Discussion

The purpose of this study was to examine the longitudinal relationships between effort-reward imbalance as a stressor and academic burnout as a strain among middle school students. Moreover, this study investigated whether coping strategies would moderate the longitudinal relationship between effort-reward imbalance and academic burnout.

First, the results indicated that effort-reward imbalance was significantly associated with academic burnout over time. These results support previous research which cross-sectionally examined the positive relationship between ERI and burnout (Schulz, Damkröger, Heins, et al., 2009). Although the previous research (e.g., Bakker, Killmer, Siegrist, & Schaufeli, 2000) statistically demonstrated the positive relationship between the variables, the results were limited in the area of job burnout and cross-sectional data. Through the results of the present study, the relationship between two variables was widened into academic burnout research.

Second, coping strategies interacted with effort-reward imbalance and academic burnout. These results are in line with Rijk, Blanc, Schaufeli, and Jonge (1998), Siu, Spector, and Cooper (2006), and Yip, Rowlinson, and Siu (2008). Problem-focused coping plays an important role in psychological well-being and enhances the capacities of individuals to deal with stressful events or situations. As a result, it contributes to decrease strain such as burnout. It can be explained by the theory of Folkman and Lazarus (1980). According to Folkman and Lazarus (1980), coping strategies which individuals use differ depending on each situation. In cases related with one’s work, individuals tend to use problem-focused coping strategies to solve problems. In other hands, when the situation is considered as a changeable event, individuals are likely to employ problem-focused coping strategies. Emotion-focused coping strategies are often used by a person who experiences a stressful event which is unchangeable. Thus, two theories imply that employing more problem-focused coping is beneficial to reduce development of psychological strain because academic activities can be changeable.

The present study has the following theoretical and practical implications for educators including teachers and school psychologists. First, the results of this longitudinal study can be used to promote understanding of how much students suffer from the imbalance between their efforts and rewards. Based on the results of this study, problem-focused coping is more helpful when practitioners are intervening to reduce academic burnout. As students move into higher grades at school, academic related workloads increase. In order to cope with academic stressors, removing the stressors is not very realistic, therefore, utilizing suitable coping strategies for decreasing and minimizing levels of academic burnout is needed. Teachers and school psychologists could develop various programs using problem-focused strategies by helping students learn how to plan and organize work-times and break-times. Chapman and Mullis (1999) suggested that learning coping strategies is effective for students from programs based on individual differences. The programs should consider taking individual characteristics into an account when teaching coping strategies to handle demands and resources that are most appropriate for the students. As students plan to balance out the amount of time to do school work, and time for resting or non-related schoolwork, the students will be able to manage excessive amounts of schoolwork more efficiently. Increased efficiency may allow them to prevent or reduce the level of academic burnout.

Next, this result is consistent with early theoretical studies in that problem-focused coping is related with the alleviated levels of stressor and academic burnout. In addition to these previous studies, the present study examined the development of relationships between effort-reward imbalance and academic burnout as well as the development of each variable of this study. Despite the recent surge in interest on academic burnout, longitudinal studies on academic burnout are still limited. Using longitudinal data, this study focused not only on the interpersonal differences but also on the intra-differences (within-differences) of effort-reward imbalance and academic burnout. Given that effort-reward imbalance and academic burnout are phenomenon that students suffer from their psychological domain, it is important to understand both the intrapersonal and interpersonal differences to prevent the development of effort-reward imbalance and academic burnout. Therefore, the present study is meaningful in improving school settings and educational environment by providing longitudinal evidences.

Lastly, these findings would contribute to examining an individual’s both intra-changes (within oneself) and inter-changes (between or among groups) in the relationship between ERI and academic burnout. By using LGM, this study found that there were significant changes within the students, which had an effect on severity of academic burnout. Based on this study, practitioners would differently approach and manage students’ psychological problems depending on each student’s levels of effort-reward imbalance and academic burnout.

Limitations and suggestions for future studies

The current study also has several limitations. First, this study used self-report questionnaires to measure psychological constructs. Although these measurements are frequently used and common for the variables in this study, the study nevertheless has an interpretative problem. Self-report questionnaires may overestimate or underestimate effort-reward imbalance and academic burnout phenomena. Future studies should apply alternative measurements (e.g., behavioral observation) to estimate these variables of psychological domains. Second, participants in this study consist of only students from the metropolitan (e.g., Seoul) regions in South Korea. Education levels and academic ability levels may be different in each region in South Korea because educational resources are concentrated in a few areas of the country. Due to the relatively heavy polarization in access to academic resources among different regions in Korea, education levels and academic performances of students may differ accordingly. For example, the urban areas are in a compact mass with private educational institutes. The rural students’ academic stress and burnout levels may be different from those of students in Seoul area from which the participants of the current study were selected. Therefore, generalization of the findings/results of this study across all regions in Korea should be approached with caution. To generalize these longitudinal results, replication studies should be conducted with more representative samples of students. Future research needs to determine if significant similarities or differences exist in effort-reward imbalance and academic burnout among students in other regions. Third, the model fit indices for the moderation analyses were not improved to the satisfactory level. Fourth, the current study was based on the longitudinal data that did not consider ordering effect. Future studies should collect data counterbalancing the scales in order to control ordering effect. Fifth, this study used the median split method for the multi-group analysis. This approach has a limitation because the inventory measuring coping strategies does not have cut-off scores. Further studies should be approached differently in order to confirm the moderation effect of coping on the relationship between academic burnout and effort-reward imbalance. Finally, in this study, there was a lack of previous studies for generalization of the longitudinal causal relationship between effort-reward imbalance and academic burnout. Therefore, future research needs to examine the longitudinal causal relationship between effort-reward imbalance and academic burnout using a different longitudinal analysis method, such as autoregressive cross-lagged modeling.

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 work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2017S1A5B5A02026765).

Author's biographies

Boyoung Kim is a Research Professor in the department of Education at Korea University. She is interested in human development, career counseling, and the promotion of Psychological well-being. She enjoys taking care of her two dogs, Tofu and Blacky.

Eunjoo Kim is a M.A Candidate in the department of Education majoring in counseling at Korea University Graduate School. Her school counseling interests include career development and career barriers, foreign international students and interpersonal relationships.

Sang Min Lee is a Professor in the department of Education at the Korea University. His research interests include student's academic burnout syndrome, counselor's burnout syndrome, and measurement issues in counseling. He enjoys playing golf and spending time with his family.

References

Arbuckle

J. L.

(1996) Full information estimation in the presence of incomplete data. In: Marcoulides

G. A.

Schumacker

R. E.

(eds) Advanced structural equation modeling, Mahwah, NJ: Lawrence Erlbaum Associates, Inc, pp. 243–277.

Bakker

A. B.

Killmer

C. H.

Siegrist

Schaufeli

W. B.

(2000) Effort–reward imbalance and burnout among nurses. Journal of advanced nursing 31(4): 884–891. doi:10.1046/j.1365-2648.2000.01361.

Bentler

P. M.

(1990) Comparative fit indices in structural models. Psychological Bulletin 107(2): 238–246. doi: 10.1037/033-2909.107.2.238.

Brenner

S. O.

Sorbom

Wallius

(1985) The stress chain: A longitudinal confirmatory study of teacher stress, coping and social support. Journal of Occupational Psychology 58: 1–13. doi:10.1111/j.2044-8325.1985.tb00175.

Chan

D. W.

(1994) The Chinese Ways of Coping Questionnaire: Assessing Coping in Secondary Teachers and Students in Hong Kong. Psychological Assessment 6(2): 108–116. doi: 10.1037/1040-3590.6.2.108.

Chapman

P. L.

Mullis

R. L.

(1999) Adolescent coping strategies and self-esteem. Child Study Journal 29(1): 69–69.

Compas

B. E.

Connor-Smith

J. K.

Saltzman

Thomsen

A. H.

Wadsworth

M. E.

(2001) Coping with stress during childhood and adolescence: problems, progress, and potential in theory and research. Psychological Bulletin 127: 87–127. doi: 10.1037/0033-2909.127.1.87.

Dillon

W. R.

Kumar

Mulani

(1987) Offending estimates in covariance structure analysis: Comments on the causes of and solutions to Heywood cases. Psychological Bulletin 101(1): 126–135. doi: 10.1037/0033-2909.101.1.126.

Folkman

Lazarus

R. S.

(1980) An analysis of coping in a middle-aged community sample. Journal of health and social behavior 21(3): 219–239. doi: 10.2307/2136617.

10.

Folkman

Lazarus

R. S.

(1984) If it changes it must be a process: Study of emotion and coping during three stages of a college examination. Journal of personality and Social Psychology 48: 150–170. doi: 10.1037/0022-3514.48.1.150.

11.

Fukuda

Yamano

Joudoi

Mizuno

Tanaka

Kawatani

Watanabe

(2010) Effort-reward imbalance for learning is associated with fatigue in school children. Behavioral Medicine 36(2): 53–62. doi: 10.1080/08964281003774919.

12.

Grewal

Cote

J. A.

Baumgartner

(2004) Multicollinearity and measurement error in structural equation models: Implications for theory testing. Marketing Science 23(4): 519–529. doi: 10.1287/mksc.1040.0070.

13.

Griva

Joekes

(2003) UK teachers under stress: Can we predict wellness on the basis of characteristics of the teaching job? Psychology-and-Health 18: 457–471. doi: 10.1080/08870440310001471.

14.

Bentler

P. M.

(1999) Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling 6: 1–55. doi: 10.1080/10705519909540118.

15.

Hwang

H. J.

(2006) A study on the stress and its relativeeffects of variables among children and adolescents. Journal of Elementary Education 19(1): 193–215.

16.

Jöreskog, K. G. (1999). How large can a standardized coefﬁcient be? Retrieved February 2, 2005, from http://www.ss icentral.com/lisrel/column2.htm.

17.

Kim, B., Kim, S. R., Yang, N. Y., Yaung, H., Ha, G. Y., Yang, J. Y., Lee, B., & Lee, S. M. (2016). Longitudinal relationships between planned happenstance skills and life adjustment and the moderating role of career barriers. Journal of Career Development, 0894845316681078.

18.

Kim, Y. J. (2006). The relationships among communication in parent-child, stress coping and adolescents’ school adjustment. Unpublished master’s thesis, Seoul Woman’s University, Seoul.

19.

Kim

Jin

(2008) A longitudinal study of the effects of media use on the evaluation of the leading candidate in Korean 2007 Presidential election. Korean Journal of Journalism & Communication Studies 44: 76–107.

20.

Kim

K. H.

Kim

D. I.

Kim

B. H.

Kim

C. D.

Kim

H. S.

Nam

S. I.

Cho

H. I.

(2000) School counseling andguidance, Seoul, South Korea: Hakjisa Press.

21.

Kline

R. B.

(2015) Principles and practice of structural equation modeling 4th ed. New York, NY: Guilford Press.

22.

Lazarus

R. S.

Folkman

(1984) Stress, appraisal, and coping, New York: Springer.

23.

Lee

Larson

(2000) The Korean “examination hell”: Long hours of studying, distress, and depression. Journal of Youth and Adolescence 29: 249–272. doi: 10.1023/A:1005160717081.

24.

Lee

Puig

Kim

Shin

Lee

S. M.

(2010) Academic burnout profiles in Korean adolescents. Stress and Health 26(5): 404–416. doi: 10.1002/smi.1312.

25.

Lee

Puig

Lea

Lee

S. M.

(2013) Age related differences in academic burnout of Korean adolescents. Psychology in the Schools 50(10): 1015–1031. doi: 10.1002/pits.21723.

26.

Lee

Puig

Lee

S. M.

(2012) The effect of the demand control and effort reward imbalance models on the academic burnout of Korean adolescents. Asia Pacific Journal of Education 32(1): 113–123. doi: 10.1080/02188791.2012.655373.

27.

Lee

Y. B.

Lee

S. M.

Lee

(2009) Development of Korean academic burnout scale. Korea Educational Review 15(3): 59–78.

28.

Leiter

M. P.

(1991) Coping patterns as predictors of burnout: The function of control and escapist coping patterns. Journal of Organizational Behaviour 12: 123–144.

29.

Lowe

Bennett

(2003) Exploring coping reactions to work-stress: Application of an appraisal theory. Journal of Occupational and Organizational Psychology 76: 393–400. doi: 10.1348/096317903769647247.

30.

Maslach, C., & Schaufeli, W. B. (1993). Historical and conceptual development of burnout. Professional Burnout: Recent Developments in Theory and Research, edited by Schaufelial., Washington, D. C.: Taylor and Francis. 1-18.

31.

Maslach

Schaufeli

W. B.

Leiter

M. P.

(2001) Job burnout. Annual Review of Psychology 52: 397–422. doi: 10.1146/annurev.psych.52.1.397.

32.

McQuitty

(2004) Statistical power and structural equation models in business research. Journal of Business Research 57(2): 175–83. doi: 10.1016/S0148-2963(01)00301-0.

33.

Moos

R. H.

Billings

A. G.

(1982) Conceptualizing and measuring coping resources and processes . In: Goldberger

Breznitz

(eds) Handbook of stress: Theoretical and clinical aspects, New York: Free Press, pp. 212–230.

34.

Muthen, L., & Muthen, B. O. (1998–2006). Mplus Version 4.2. http://www.statmodel.com/index2.Html.

35.

Organisation for Economic Co-operation and Development (2009) Education at a glance 2009: OECD indicators, Paris, France: Author.

36.

Omizo

M. M.

Omizo

S. A.

Suzuki

L. A.

(1988) Children and stress: An exploratory study of stressors and symptoms. The School Counselor 3: 267–275.

37.

Park

I. K.

Lee

S. M.

Choi

B. Y.

Jin

Lee

(2010) Relationship between the big five personality factors and academic burnout. Korean Journal of Social and Personality Psychology 24(1): 81–93.

38.

Rijk

A. E.

Blanc

P. M.

Schaufeli

W. B.

Jonge

J. J.

(1998) Active coping and need for control as moderators ofthe job demand-control model: effects on burnout. Journalof Occupational and Organisational Psychology 71(1): 1–18.

39.

Roben

C. K.

Cole

P. M.

Armstrong

L. M.

(2013) Longitudinal relations among language skills, anger expression, and regulatory strategies in early childhood. Child development 84(3): 891–905. doi: 10.1111/cdev.12027.

40.

Rothmann

Van Rensburg

(2002) Psychological strengths, coping and suicide ideation in the South African Police Services in the North West Province. South African Journal of Psychology 28(3): 39 49. doi: 10.4102/sajip.v28i3.64.

41.

Rubenstein

J. L.

Halton

Kasten

Rubin

Stechler

(1998) Suicidal behavior in adolescents: Stress and protection in different family contexts. American Journal of Orthopsychiatry 68(2): 274–284.

42.

Schaufeli

W. B.

Martinez

I. M.

Marques-Pinto

Salanova

Bakker

(2002) Burnout and engagement in university students: A cross-national study. Journal of Cross-Cultural Psychology 33(5): 464–481. doi: 10.1.1.459.8952.

43.

Schulz

Damkröger

Heins

Wehlitz

Löhr

Driessen

Behrens

Wingenfeld

(2009) Effort–reward imbalance and burnout among German nurses in medical compared with psychiatric hospital settings. Journal of psychiatric and mental health nursing 16(3): 225–233. doi: 10.1111/j.1365-2850.2008.01355.

44.

Shin

Puig

Lee

J. H.

Lee

S. M.

(2011) Cultural validation of the Maslach Burnout Inventory for Korean students. Asia Pacific Education Review 12(4): 633–639. doi: 10.1007/s12564-011-9164-y.

45.

Siegrist

(1998) Adverse health effects of effort-reward imbalance at work: theory, empirical support and implications for prevention . In: Cooper

C. L.

(ed.) Theories of Organizational Stress, Oxford: Oxford University Press).

46.

Siegrist

Starke

Chandola

Godin

Marmot

Niedhammer

Peter

(2004) The measurement of effort–reward imbalance at work: European comparisons. Social science & medicine 58(8): 1483–1499. doi: 10.1016/S0277-9536(03)00351-4.

47.

Siu

O. L.

Spector

P. E.

Cooper

C. L.

(2006) A three phase study to develop and validate a Chinese coping strategies scales in Greater China. Personality and Individual Differences 41: 537–48. doi: 10.1016/j.paid.2006.02.012.

48.

Slaney

R. B.

Brown

M. T.

(1983) Effects of race and socioeconomic status on career choice variables among college men. Journal of Vocational Behavior 23: 257–269. doi: 10.1016/0001-8791(83)90040-4.

49.

Vitaliano

P. P.

Russo

Carr

J. E.

Maiuro

R. D.

Becker

(1985) The ways of coping checklist: Revision and psychometric properties. Multivariate behavioral research 20(1): 3–26. doi: 10.1207/s15327906mbr20011.

50.

Wallace

Lee

(2010) Job stress, coping strategies, and burnout among abuse-specific counselors. Journal of Employment Counseling 47: 111–122. doi: 10.1002/j.2161-1920.2010.tb00096.x.

51.

Yip

Rowlinson

Siu

O. L.

(2008) Coping strategies as moderators in the relationship between role overload and burnout. Construction Management and Economics 26: 869–880. doi: 10.1080/01446190802213529.

52.

Zhang

Gan

Cham

(2007) Perfectionism, academic burnout and engagement among Chinese college students: A structural equation modeling analysis. Personality and Individual Differences 43(6): 1529–1540. doi: 10.1016/j.paid.2007.04.01.