Validation and Reliability of the German Version of the School Burnout Inventory
Abstract
Abstract. This study investigates the validity and reliability of the German version of the School Burnout Inventory (SBI-G) in 1,570 secondary-school students (Mage = 14.11, SD = 0.78; 51.7 % girls). Results indicate that school burnout consists of two correlated but separate dimensions including (1) exhaustion at school, (2) cynicism toward the meaning of school and sense of inadequacy. The study revealed that school burnout can be measured as a two-factor model, which provided good reliability and validity indices. Further, we verified concurrent validity, finding that students suffering from general stress also reported overall school burnout as well as exhaustion, cynicism, and inadequacy. Students who exhibited cynicism and inadequacy also reported lower levels of behavioral, emotional, and cognitive school engagement, while exhausted students reported lower emotional school engagement but higher cognitive school engagement.
Zusammenfassung. Diese Studie untersucht die Validität und Reliabilität des School Burnout Inventory in deutscher Sprache (SBI-G) mit 1.570 Sekundarschülern (MAlter = 14,11, SD = 0,78; 51,7 % Mädchen). Die Ergebnisse zeigen, dass sich School Burnout aus zwei korrelierten, aber separaten Dimensionen zusammensetzt, darunter (1) Erschöpfung in der Schule und (2) Gefühl der Unzulänglichkeit und Zynismus gegenüber der Schule. Demnach kann Schul-Burnout als Zwei-Faktoren-Modell mit einem Gesamtscore für Schul-Burnout gemessen werden. Das Modelle zeigt gute Reliabilitäts- und Validitätsindizes. Des Weiteren wurde die konkurrente Validität verifiziert, wobei sich herausstellte, dass Schülerinnen und Schüler, die unter allgemeinem Stress litten, auch über Schul-Burnout sowie Erschöpfung, Zynismus und Unzulänglichkeit berichteten. Schülerinnen und Schüler, die Zynismus und Unzulänglichkeit aufwiesen, berichteten auch über ein niedrigeres Niveau an verhaltensbezogenem, kognitivem und emotionalem schulischen Engagement, während erschöpfte Schülerinnen und Schüler ein niedrigeres emotionales schulisches Engagement, aber ein höheres kognitives schulisches Engagement angaben.
School-related stress and school burnout are prevalent among students in secondary educational settings and increase over the school years (Bask & Salmela-Aro, 2013; Inchley et al., 2016). Both concepts are interconnected and overlap considerably (Salmela-Aro et al., 2008). While school-related stress occurs when students feel overwhelmed by schoolwork, worry about their educational performance, and feel pressured to achieve good grades (Pascoe et al., 2019), school burnout is engendered by chronic stress and is measured along three dimensions, which include exhaustion from schoolwork, feeling inadequate as a student, and portraying a cynical and detached attitude toward school (Salmela-Aro et al., 2009; Schaufeli et al., 2002).
Various international studies have identified school-related stress and school burnout as risk factors that jeopardize students’ well-being and academic development (Beisenkamp et al., 2009; Hoferichter et al., 2021; OECD, 2017; Pascoe et al., 2019). In detail, school burnout is related to depressive symptoms (Gerber et al., 2015; Salmela-Aro et al., 2009), low educational aspirations and achievement (Kaplan et al., 2005; Liu & Lu, 2010; Salmela-Aro et al., 2008; Vasalampi et al., 2009), low school motivation and high levels of absenteeism (Fiorilli et al., 2017), early school drop-out (Korhonen et al., 2014), and scholastic alienation among secondary-school students (Natvig et al., 1999).
To capture school burnout, Salmela-Aro and colleagues (2009) introduced a valid and reliable measure, the School-Burnout-Inventory (SBI), first used in the Finnish and Swedish context. According to their study, burnout can either be measured applying a three-factor model, which includes the components exhaustion at school, cynicism toward the meaning of school, and sense of inadequacy at school; or as an overall factor by measuring a second-order model. Since then, this instrument has been investigated in other languages, e. g., French (Meylan et al., 2015), English (May et al., 2020), Italian (Fiorilli et al., 2017), and German (Herrmann et al., 2019). However, some of these studies failed to replicate the three-component structure (De Stasio et al., 2014) and rather proposed a two-factor model by eliminating the factor inadequacy (Herrmann et al., 2019).
Concerning the dimensionality of burnout, researchers put forward various suggestions, including conceptualizing burnout by two-factors exhaustion and depersonalization attitutes/cynicism (Herrmann et al. 2019; Kalliath et al., 2000). Others suggested burnout as a unidimensional phenomenon comprised only of exhaustion (Aiken et al., 2002; Halbesleben & Bowler, 2007; Halbesleben & Buckley, 2004), which, however, was discouraged by a meta-analysis (Purvanova & Muros, 2010). Yet others have suggested applying the “exhaustion+1” rule, which includes exhaustion in combination with either depersonalization or low personal accomplishment (Brenninkmeijer & VanYperen, 2003). Most of this research, except that by Hermann et al. (2019), is based on the Maslach Burnout Inventory (MBI, Maslach & Jackson, 1981), which was originally applied to the work context and only recently was adapted to the school context (MBI-SS, Schaufeli et al., 2002). By introducing a new concept and instrument of school burnout, the SBI promises to move research on the dimensionality of burnout forward and presents a short scale that can be easily handled by school students.
Previous studies on school-related stress and burnout indicated differences related to gender and school type. In detail, girls and students from higher-track schools tend to report higher levels of stress and exhaustion, while students from lower-track schools exhibit higher levels of cynicism (Ge et al., 1994; Herrmann et al., 2019; Kiuru et al., 2008; Salmela-Aro et al., 2009; Salmela-Aro et al., 2008). Yet, the tracking system of a country may be unique, so that results on relating school types and burnout may differ across countries. A recent study in the German context found that the general stress level of students from lower-track schools increases from early to middle adolescence but not for students from higher-track schools (Kulakow et al., 2021).
To validate the SBI-G in the German context, we investigated students’ perceived general stress (Cohen et al., 1983; Klein et al., 2016) and their emotional, behavioral, and cognitive school engagement (Fredericks et al., 2005) on a sample of 1,570 secondary-school students. We thereby took gender and school type into account when analyzing the validity and reliability of the SBI-G.
Perceived Stress and School Engagement Related to School Burnout
According to the transactional stress model, students feel stressed if they feel threatened or challenged by a stimulus that demands their internal and external resources (Lazarus & Folkman, 1984). General stress has proved to be a strong predictor of school burnout, if students feel that demands exceed their resources and prevail over a longer period (Lin & Huang, 2013; Morrison & O’Conner, 2005).
Furthermore, students who feel burned out, and as such overwhelmed by schoolwork, also exhibit lower school engagement (May et al., 2020; Salmela-Aro et al., 2009; Teuber et al., 2020; Vasalampi et al., 2009). School engagement is a multifaceted construct comprised of behavioral, emotional, and cognitive components (Fredricks et al., 2004). More specifically, behavioral engagement reflects students’ involvement in school-related activities, whereas emotional engagement represents positive and negative reactions toward school and social relationships at school. Cognitive school engagement encompasses willingness and effort to understand complex ideas and handle challenging tasks (Fredrickson et al., 2004).
Most research on school burnout, stress, and engagement applied either global composite scores of different engagement measures or investigated only two components, i. e., behavioral and emotional school engagement (May et al., 2020; Raufelder et al., 2014; Tuominen-Soini & Salmela-Aro, 2014; Vasalampi et al., 2009). In general, school engagement predicts positive academic outcomes (Lee, 2014; Li & Lerner, 2011) and is inversely related to test anxiety (Raufelder et al., 2015), general stress (Grützmacher & Raufelder, 2019; Raufelder et al., 2014), and burnout in educational settings (Schaufeli et al., 2002; Tuominen-Soini & Salmela-Aro, 2014).
The original validation study of the SBI (Salmela-Aro et al., 2009) indicated that school engagement was negatively related to overall school burnout as well as its dimensions cynicism and inadequacy, whereas exhaustion was not significantly related to school engagement, exhaustion being the first stage of burnout (Gustavsson et al., 2010; Maslach & Leiter, 2016; Maslach et al., 2001). Similarly, other studies also assumed a negative relationship between engagement and burnout (Kiuru et al., 2009; Schaufeli & Bakker, 2004). There have been proposals to measure engagement as the opposite of burnout in a single instrument (Maslach & Leiter, 1997). Although, conceptually speaking, engagement presents the positive antithesis of burnout, both concepts have to be operationalized in their own right as their structure differs (Schaufeli et al., 2002). Conceptually, engagement presents an overall framework comprised of a positive approach and enthusiasm toward studying (mapping onto emotional engagement), dedication and absorption in schoolwork (mapping onto behavioral engagement), and challenge and inspiration (mapping onto cognitive engagement) (Salmela-Aro et al., 2009; Salmela-Aro & Upadyaya, 2012; 2013; Schaufeli et al., 2002). Although dimensions of burnout and engagement are intertwined, if one conceptually links the dimensions of engagement with the dimensions of burnout, feeling exhausted may present the opposite of enthusiasm, while cynicism may present the opposite of dedication, and inadequacy may present the opposite of feeling inspired or challenged by school work (cf. Schaufeli et al., 2002). However, this conceptual approach has been neither fully verified nor falsified empirically. Whether it is actually an opposing relationship or possibly a causal one demands longitudinal investigations, for example, in random intercept cross-lagged panel designs. Moreover, previous studies often used different instruments to assess school engagement (2 vs. 3 subscales) and school burnout (MBI vs. SBI). Because the original study examined school engagement as a predictor of school burnout, we adopted this design for the present study. These conceptual antitheses may explain why students who engage in schoolwork are less likely to feel burned out. However, overly engaged individuals may be at risk of experiencing burnout, because their engagement requires too much energy, presenting a strain rather than a gain to them.
In fact, the mechanisms between school burnout and engagement are rather complex, the profile analysis by Tuominen-Soini and Salmela-Aro (2014) discovered. The authors investigated high-school students over 6 years and found four relatively stable profiles of students: engaged, engaged-exhausted, cynical, and burned-out. Similarly, Salmela-Aro et al. (2016) presented profiles of Finnish and US high-school students who were highly engaged and exhausted.
In sum, the research indicates that students who exhibit high levels of general stress are more likely to experience school burnout and show less school engagement. Additionally, students who feel burned out by school also tend to engage less in school activities as they feel detached from school and portray a cynical attitude toward school.
Hence, the aims of the current study are twofold and propose to (1) validate the SBI in German (SBI-G) and (2) identify how students’ general stress level, behavioral, emotional, and cognitive school engagement is related to the three dimensions of burnout, controlling for the effects of gender and school track. In accordance with the literature, we formulated four hypotheses. In arranging the variables in the model, we draw on the original validation study by Salmela-Aro and colleagues (2009) .
Hypotheses
Hypothesis I: Based on the original validation study (Salmela-Aro et al., 2009), a three-factor model with moderate intercorrelation patterns among the components (exhaustion at school, cynicism toward the meaning of school, sense of inadequacy at school) as well as a second-order model to assess overall school burnout are expected to fit the data well.
Hypothesis II: General stress is strongly related to overall school burnout and the three dimensions of the SBI-G.
Hypothesis III: In line with the original validation study (Salmela-Aro et al., 2009), we expected that the overall school burnout is negatively related to behavioral, emotional, and cognitive school engagement. Specifically, we assumed that cynicism and inadequacy would be negatively related to all three components of school engagement, whereas exhaustion would not be significantly related.
Method
Participants
The present study is based on questionnaire data from 1,570 students attending grades 7 and 8 (Mage = 14.11, SD = 0.78; 51.7 % girls; nclasses = 103) from six lower-track schools (“Regionale Schule”) (n = 553) and 13 higher-track schools (“Gymnasium”) (n = 1,031) in the federal state of Mecklenburg-Western Pomerania in northern Germany. In this federal state, there are two ability‐grouped secondary school types: academic or higher-track schools (“Gymnasium”) and nonacademic or lower-track schools (“Regionale Schule”). These types differ in their didactic concepts, achievement orientation, and school culture (Maaz et al., 2008). This study was conducted in autumn 2018 in the classroom via a paper-pencil questionnaire. Written consent was obtained beforehand from students and parents who were informed about the anonymity and voluntariness of the study. Permission was also given by the county Department of Education. Because only 4.7 % of the state’s population has a migration background, we did not consider students’ migration background in this study (Statista, 2020).
Procedure
Constructing the German Version of the School Burnout Inventory (SBI)
The original 9 items of the SBI (Salmela-Aro et al., 2009) were translated into German and refined in a multiple-step procedure (Appendix A1). Since the original items in the validation study are presented in English, these were used in the present study as a basis. In a first step, two professional translators adapted all items using backtranslation (Brislin, 1970, 1980). In detail, one translator translated the measure from English into German, whereas the second translator translated the measure back into English, without knowing the original items. Subsequently, both English versions were compared to evaluate the quality of the translation. During this process, a bilingual Finnish/German scientist was present to support the translation and give feedback on the translated items. In a second step, a pilot study with 56 7th- and 8th-grade students from secondary schools in Greifswald was conducted to check the wording and syntax of the items as well as the range of answers (Tashakkori & Teddlie, 1998). Consequently, the final SBI-G was prepared for the validation and reliability study. According to the original SBI, the SBI-G comprises 9 items (α = .82) with three subscales: (a) exhaustion at school (four items; e. g., “I feel overwhelmed by my schoolwork”; α = .72), (b) cynicism toward the meaning of school (three items; e. g., “I feel that I am losing interest in my schoolwork”; α = .74), and (c) sense of inadequacy at school (two items; e. g., “I often have feelings of inadequacy in my schoolwork”; α = .54). All SBI-G items were introduced to the participants in the following way: “Please choose the answer that best describes your situation (estimation from the previous month)”. Students could rate their response on a 6-point Likert scale ranging from 1 (completely disagree) to 6 (strongly agree).
Measures
To validate the study, we applied the following well-established and validated German instruments for the age cohort of secondary-school students.
School Engagement
School engagement measures were developed by Fredericks and colleagues (2005) and distinguish three subscales: The behavioral school engagement (BSE) subscale (α = 0.67) includes four questions such as “I pay attention in class”; the emotional school engagement (ESE) subscale (α = 0.72) is comprised of four questions, e. g., “I feel happy in school”; the cognitive school engagement (CSE) subscale (α = 0.76) with questions such as “I check my schoolwork for mistakes” consists of five items. Answers were rated on a five-point Likert scale from (1) strongly disagree to (4) strongly agree.
Perceived Stress Scale
This scale captures students’ self-reported feelings of stress during the last month (Cohen et al., 1983; Klein et al., 2016), including ten items, e. g., “In the past month, how often have you been angered by things that were outside of your control?” and a good reliability of α = 0.79. Answers were rated on a five-point Likert scale from (1) strongly disagree to (4) strongly agree.
Statistical Analyses
All statistical analyses were computed with the Mplus statistical package (Version 8.5; Muthe´n & Muthe´n, 1998 – 2017). We used the maximum likelihood robust (MLR) estimator, which is implemented in Mplus and which is robust against violation of normality assumptions (Yuan & Bentler, 2000). Missing data were completely at random, as Little’s MCAR test confirmed [χ2 (99) = 121.83; p > .05], and therefore handled – by default – with full information maximum likelihood (FIML) in Mplus (Schafer & Graham, 2002). To account for the multilevel structure of the data (i. e., students nested in classes), the type is a complex feature of Mplus was used for all analytical steps. This feature computes standard errors of parameter estimates that take into account stratification and nonindependence of observations because of cluster sampling (Asparouhov, 2005;Muthén & Muthén, 1998 – 2017).
The statistical analyses in the present study are based on the analytical procedure in the original validation study by Salmela-Aro and colleagues (2009) . We initially estimated the structure of the SBI-G by comparing three different theoretical models: (1) a one-factor model (M1), following the assumption that all the SBI-G items can be described by one latent factor; (2) a three-factor model (M2), which supposes that the SBI-G items underlie three correlated latent factors (exhaustion, cynicism, and inadequacy); and (3) a second-order-factor model (M3), which assumes that the three first-order factors could be explained by a second-order factor assessing overall school burnout; in addition (4) a bi-factor model (M4), which assumes a general factor onto which all items load and three orthogonal (uncorrelated) subdimensions (grouping factors) (see Dunn & McCray, 2020; Reise et al. 2007, 2010; Yung et al. 1999). The latter model is particularly valuable for evaluating the empirical plausibility of subscales, because “a bifactor model is capable of retaining a unidimensional conceptualization while also acknowledging the unintended and meaningless covariance that can occur between particular items in a scale because of wording effects and can thus present spurious evidence of multidimensionality” (Hyland et al., 2013, p. 3). M1 and M2 were compared by the χ2-difference test (Satorra & Bentler, 2001). Because M2 and M3 include the same number of estimated parameters, no χ2-difference test could have been computed. Figure 1 depicts all four theoretical models.
Further, we employed confirmatory factor analysis to determine the reliability and validity of the nine SBI-G items. To estimate the reliability of the coefficients, we measured the squared correlation between items and factors (see Bollen, 1989; Liukkonen & Leskinen, 1999). In detail, standardized validity coefficients (i. e., standardized factor loadings) give insight into the structural association between the factor and the items, also described as the validity of the items (Bollen, 1989). Furthermore, we estimated the factor-score scale reliabilities, indicated by the squared correlations between the factor-score scale and the latent factor as well as Cronbach’s alphas, to evaluate the internal consistency of the inventory.
Finally, we included meaningfully related variables of school burnout, namely, school engagement and perceived stress, to the best fitting model to provide evidence for the concurrent validity of the SBI-G. Gender and school track were considered as control variables and added to the model.
Model fit was estimated with the following primary-fit indices as recommended by Hu and Bentler (1999): the χ2 test of model fit, root mean square error of approximation (RMSEA), comparative fit index (CFI), and standardized root mean square residual (SRMR). The CFI and the RMSEA range from 0 to 1, while for the CFI a value closer to 1 represents a good model fit. In contrast, an RMSEA value closer to 0 exemplifies that the data fit the model well. An acceptable fit to the data is usually indicated with CFI and TLI values greater than .90 and RMSEA values lower than .08. A good fit to the data is specified with CFI and TLI values greater than .95 and RMSEA values lower than .05 or .06 (Hu & Bentler, 1999).
Results
Factorial Validity: Structure of the SBI-G
Table 1 presents the bivariate correlations, means, and variances for the school-burnout variables.
Confirmatory Factor Analyses
Following the hypothesis-testing approach of the four above-mentioned theoretical models (M1, M2, M3, M4), we ran a confirmatory factor analysis (CFA) to test M1. The analysis showed that the one-factor model (M1), which assumes that there is one latent factor underlying all the SBI-G items, exhibited the following fit: (χ2(26) = 286.63, p < .001, CFI = .91, RMSEA = .08 (.07-.09), SRMR = .05). Subsequently, a second CFA (M2), assuming that all items underlie three correlated latent factors, i. e., exhaustion, cynicism, and inadequacy were computed. The fit indices revealed a good model fit (χ2 (23) = 160.40, p < .001, CFI = .95, RMSEA = .06 (.05-.07), SRMR = .04). Both models were compared by a χ2-difference test suggesting that the fit indices of M2 were superior to that of M1: χ2 (3) = 102.17, p < .001. In M2, each factor functions as a covariate of the other factors (see Figure 1). The range of the factor loadings was from .41 to .69 for exhaustion, from .57 to .78 for cynicism, and from .58 to .64 for inadequacy. Because of the standardized factor loading, the explained variance of each item is presented by the squared factor loading which ranges from .17 to .62. Positive correlations were found for all three factors: cynicism and exhaustion (r = .73, p < .001), inadequacy and exhaustion (r = .82, p < .001), inadequacy and cynicism (r = .92, p < .001). The internal consistency reliability, calculated as Cronbach’s alpha, indicated good reliability scores for exhaustion (α = .72) and cynicism (α = .74) but only acceptable reliability scores for a sense of inadequacy at school (α = .54). However, following Kopp and Lois (2012), the critical value of Cronbach’s alpha is α > 0.50, indicating that the subscale of inadequacy can still be considered as acceptable.
Subsequently, we ran the third theoretical model M3 (see Figure 1), assuming that the three first-order factors could be explained by a second-order factor assessing overall school burnout. As mentioned above, M2 and M3 are data-equivalent models, which means that no χ2-difference test can be applied. The residual variance of the factor “inadequacy at school” was fixed to zero, because – like the original study (Salmela-Aro et al., 2009) – its estimates were negative. M3 showed good fit indices: χ2 (23) = 156.85, p < .001, CFI = .95, RMSEA = .06 (.05-.07), SRMR = .04.
We then ran the fourth theoretical model M4 (see Figure 1) to test whether a unidimensional scoring scheme is preferred and whether the creation of subscales is appropriate (Reise et al., 2010). The fit indices revealed a good fit (χ2 (15) = 65.22, p < .001, CFI = .98, RMSEA = .05 (.04-.06), SRMR = .02). However, the items of inadequacy and cynicism loaded onto a general factor, which supports a unidimensional scoring scheme, whereas two of the items of exhaustion loaded more strongly onto the respective grouping factor (see Table 2), which indicates that the creation of a subscale is appropriate.
Because with the unidimensional scoring scheme M1 showed a) the lowest fit indices compared to all other models, b) two item loadings of exhaustion, which implied creating subcales, and because of the high correlations between the factors of inadequacy and cynicism in M2, and c) the low reliability score of inadequacy, an additional two-factor model (M5) was conceptualized. This model consisted of one latent factor presenting exhaustion and one latent factor presenting both cynicism and inadequacy items. The fit indices of M5 were good (χ2 (25) = 175.50, p < .001, CFI = .95, RMSEA = .06 (.05-.07), SRMR = .04). Consequently, we compared M5 with M2 by a χ2-difference test, demonstrating that model M5 was superior to that of M2 (χ2 (2) =14.84, p > .05). Because a second-order factor model needs a minimum of three first-order factors, another second-order factor model with only two subscales (exhaustion, inadequacy and cynicism as common factor) could not be computed.
In sum, evaluating the fit indices and conducting χ2-difference tests to compare the models revealed that M4 represents the best data fit. However, the bifactor model suggests that the items of inadequacy and cynicism load more strongly onto a general factor, whereas two of the items of exhaustion load more strongly onto the respective grouping factor. While M1 with a general factor was found to not represent the data well and M4 suggested that the creation of the subscale exhaustion might be appropriate, M5 was constructed with two subscales (exhaustion, cynicism and inadequacy). M5 was found to be superior to M2. In other words, the model with two closely associated factors best depicts students’ school burnout. In the following, we proceeded to test the reliability and validity of the models M1, M2, M3, M4, and M5.
Reliabilty and Validitiy
In the next step, we examined the item reliabilities and validities for M1, M2, M3, M4, and M5. Further, we separately explored three one-factor models for exhaustion at school, cynicism toward the meaning of school, and sense of inadequacy at school. Table 3 provides the item reliability and factor loadings (standardized validity coefficients). Compared to M2, M3, M4, and M5, M1 demonstrated the lowest reliability and validity coefficients. The results also show that all items were revealed to be good indicators of one latent SBI-G factor (M4). In line with the results of the above-described factorial validity, M2, M3, M4, and M5 were superior to M1.
To evaluate the internal consistency of the three SBI-G scales, we calculated the factor-score scales reliabilities as well as Cronbach’s alphas for direct sums of items. Table 4 demonstrates the coefficients for factor-score scales, reliabilities for factor-score scales, and Cronbach’s alphas. By applying a regression model, Mplus estimates the coefficients for factor-score scales. The findings indicate good internal consistency for all factor-score scales, which tend to be higher than Cronbach’s alpha reliabilities. Particularly, the factor-score scales of inadequacy in M1, M2, M3, and M4 were good compared to the relatively low Cronbach’s alpha. M2 and M3 demonstrate almost similar coefficients. In sum, M5 showed the best factor-score scale reliability. And Cronbach’s alpha for the common factor of cynicism and inadequacy in M5 is good.
Concurrent Validity
To test concurrent validity, we regressed variables of similar (or dissimilar) and/or related constructs (Campbell et al., 1996), namely, emotional, behavioral, and cognitive school engagement and perceived stress on the school burnout scales of M5. Furthermore, because gender and school track proved to relate differently to school-related stress (Grützmacher & Raufelder, 2019; Hoferichter & Raufelder, 2021; Mezulis et al., 2010), we entered these variables as control variables. The fit indices for M5 with predictor variables were likewise acceptable (χ2 (67) = 421.64, p < .001, CFI = .91, RMSEA = .06 (.05-.06), SRMR = .05); regression coefficients are presented in Figure 2. The results demonstrate that the more students perceive stress, the higher their school burnout. In turn, the more engaged students are at school, the lower their scores of cynicism and inadequacy. Furthermore, emotional engagement is negatively and cognitive engagement positively associated with exhaustion. Gender and school track were not significantly related to students’ school burnout. Overall, these findings indicate concurrent validity for the SBI-G.
Discussion
This study developed and validated a German version of the SBI, a self-report measure that can be employed to assess students’ feelings of burnout along three dimensions: (a) exhaustion at school, (b) cynicism toward the meaning of school, and (c) sense of inadequacy at school. Analyses indicate that school burnout can be measured either as a summary score measuring school burnout or as two factors comprised of exhaustion and cynicism/inadequacy.
In accordance with the original validation study by Salmela-Aro et al., 2009, we tested three separate models: (1) A one-factor model in which all items of the SBI-G underly one latent factor; (2) a three-factor model that includes three correlated latent factors including exhaustion, cynicism, and inadequacy; and (3) a second-order factor model indicating that the three first-order factors explain a second-order factor that assesses overall school burnout. In addition, we computed (4) a bifactor model for evaluating the empirical plausibility of subscales. In contrast with Hypothesis I and the original SBI validation study (Salmela-Aro et al., 2009), we found that model 2 (three-factor model) showed a critical correlation >.90 between the subscales cynicism and inadequacy. Furthermore, model 4 indicated that all items (except for 2 items of exhaustion) were loading on a unidimensional factor, which supports model 1. However, model 1 with a unidimensional one-factor solution showed only an acceptable model fit. Consequently, we computed (5) a model with only two subscales (exhaustion and a common factor of cynicism and inadequacy), because two items of exhaustion loaded higher on the grouping factor of model 4, which supports the use of a subscale. M5 fitted the data well and indicated the best reliability and validity indices, demonstrating that school burnout can be measured as a two-factor model with two correlated factors, or as an overall score of school burnout. The SBI-G also revealed a good concurrent validity, which was tested by relating school burnout to general stress, emotional, behavioral, and emotional school engagement. In detail, students who self-reported high levels of general stress also exhibited higher levels of exhaustion, cynicism, and inadequacy at school (see two-factor model) as well as higher overall school burnout (see second-order model), confirming Hypotheses 2 and previous studies (Lin & Huang, 2013; Morrison & O’Conner, 2005).
As expected, and formulated in Hypothesis III, students who generally felt burned out by school also indicated less engagement in school on a behavioral, emotional, and cognitive level, which confirms previous research (May et al., 2020; Salmela-Aro et al., 2009; Vasalampi et al., 2009). In line with the findings by Salmela-Aro et al., 2009, we found that all three components of school engagement were negatively related to cynicism and inadequacy at school. However, in the study by Salmela-Aro et al., 2009, cynicism and inadequacy were operationalized as two distinct factors, whereas in the present study we included both components in one factor because of the results of the CFAs. Interestingly, exhausted students also reported low emotional school engagement but high cognitive engagement, whereas there was no significant relationship between exhaustion at school and behavioral school engagement. Because we applied a multifaceted measure of school engagement, we were able to derive detailed information on how emotional, cognitive, and behavioral school engagement relates to the three respectively two dimensions of school burnout, while previous research primarily created a global composite score of school engagement including components of vigor, dedication, and absorption (May et al., 2020; Salmela-Aro et al., 2016; Teuber et al., 2020; Vasalampi et al., 2009). While Salmela-Aro et al., 2009found no significant relationship between exhaustion and school engagement, the current findings also suggest that the exhaustion students experience in relation to school work does not map on the behavioral level of their school engagement, which includes, for example, how students behave during class. However, when it comes to emotional involvement in class and school activities, exhausted students are less likely to feel happy about school and may even feel distant and detached from school emotionally. In fact, students who feel stressed or burned out by school also exhibit a negative attitude toward school and are less satisfied with school (Aypay, 2017; Hoferichter et al., 2021). In an examination of students’ cognitive involvement in school, the current study indicates that exhausted students also tend to cognitively engage more in school activities, i. e., exhausted students exhibit more effort and willingness to understand and handle complex and challenging tasks. This finding relates to person-oriented research that has, among others, detected a profile of highly exhausted and engaged students (Salmela-Aro et al., 2016; Tuominen-Soini & Salmela-Aro, 2014). This research also shows that engaged-exhausted students tend to worry about their school performance, which in turn might encourage them to engage in their school work on a cognitive level and consequently feel more exhausted (Tuominen-Soini & Salmela-Aro, 2014).
This study also revealed no gender differences related to the level of school burnout, which is surprising as previous studies found that girls report higher levels of stress and burnout (Ge et al., 1994; Kiuru et al., 2008; Salmela-Aro et al., 2008, 2009). Considering different school types, the current study did not reveal differences in students’ burnout across lower- and higher-track schools. This is surprising, since a previous German study (Kulakow et al., 2021) had found that lower-track school students exhibit higher levels of stress in the course of their school career than higher-track school students. However, this study did not focus on school burnout but on general stress, which may explain why we could not replicate the findings. In the Finnish context, Salmela-Aro et al., 2009 found that, compared to students from vocational schools, upper secondary-school students report a higher level of exhaustion, cynicism, and inadequacy; the German tracking system varies, so that results may be different.
Overall, the SBI-G is a reliable and valid instrument to assess secondary students’ school burnout in a German-speaking context. The SBI-G also indicated good concurrent validity in the context of general stress and emotional, behavioral, and cognitive school engagement. This instrument can therefore support research on school burnout, and its use may serve to identify factors that mitigate and prevent school burnout among secondary-school students.
Strength, Limitations, and Future Directions
The current study uses a large sample of students who report on their school burnout, general stress, and school engagement. By following the original validation study of the SBI by Salmela-Aro et al., 2009, we were able to verify the SBI-G as a valid and reliable measure. When validating the instrument, we followed the approach by Salmela-Aro et al., 2009, and in addition, we conducted a bifactorial model that provides information about the necessity of the subscales. However, this study is limited by the use of cross-sectional data and only a sample of secondary students from northern Germany. Therefore, future studies are warranted to investigate various age groups of students from both rural and urban areas to replicate the current findings as well as to combine self-report data with biological stress markers to investigate and portray the complex mechanisms of stress and burnout.
We would like to thank all the students, parents, and teachers who have supported this research.
Literatur
(2002). Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction. JAMA, 288 (16), 1987 – 1993. https://doi.org/10.1001/jama.288.16.1987
(2005). Sampling weights in latent variable modeling. Structural Equation Modeling, 12 (3), 411 – 434. https://doi.org/10.1207/s15328007se m1203_4
(2017). A positive model for reducing and preventing school burnout in high school students. Educational Sciences: Theory and Practice, 17, 1345 – 1359. https://doi.org/10.12738/estp.2017.4.0173
(2013). Burned out to drop out: Exploring the relationship between school burnout and school dropout. European Journal of Psychology of Education, 28, 511 – 528.
(2009). LBS-Kinderbarometer Deutschland 2009. Wir sagen euch mal was [
LBS Children’s Barometer Germany 2009. We’ll tell you something ]. RDN Verlags GmbH & Co KG.(1989). Structural equations with latent variables. Wiley.
(2003). How to conduct research on burnout: Advantages and disadvantages of a unidimensional approach in burnout research. Occupational and Environmental Medicine, 60, 16 – 20.
(1970). Back-translation for cross-cultural research. Journal of Cross-Cultural Psychology, 1 (3), 185 – 216. https://doi.org/10.1177/135910457000100301
(1980).
Translation and content analysis of oral and written material . In H. C. TriandisJ. W. Berry (Eds.), Handbook of cross-cultural psychology: Methodology (Vol. 2, pp. 389 – 444). Allyn and Bacon.(1996). Self-concept clarity: Measurement, personality correlates, and cultural boundaries. Journal of Personality and Social Psychology, 70, 141 – 156. https://doi.org/10.1037/0022-3514.70.1.141
(1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24, 385 – 396. https://doi.org/10.2307/2136404
(2014, September). School burnout in a group of Italian high school students [Poster presentation]. ECER, Porto, Portugal. https://doi.org/10.1449/78365
(2020). The place of the bifactor model in confirmatory factor analysis investigations into construct dimensionality in language testing. Frontiers in Psychology, 11, 1357. https://doi.org/10.3389/fpsyg.2020.01357
(2017). School burnout, depressive symptoms and engagement: Their combined effect on student achievement. International Journal of Educational Research, 84, 1 – 12. https://doi.org/10.1016/j.ijer.2017.04.001
(2005).
School engagement . In K. A. MooreL. Lippman (Eds.), What do children need to flourish? Conceptualizing and measuring indicators of positive development (pp. 305 – 321). Springer.(2004). School engagement: Potential of the concept, state of the evidence. Review of Educational Research, 74, 59 – 109. https://doi.org/10.3102/00346543074001059
(1994). Trajectories of stressful life events and depressive symptoms during adolescence. Developmental Psychology, 30, 467 – 483.
(2015). The relationship between mental toughness, stress, and burnout among adolescents: A longitudinal study with Swiss vocational students. Psychological Reports, 117, 703 – 723.
(2019). Das reziproke Zusammenspiel von Stresserleben, sozialen Beziehungen mit Peers und Lehrkräften und schulischer Leistung im Verlauf von früher zu mittlerer Adoleszenz [The reciprocal interaction of stress, social relationships with peers and teachers and academic achievement over the course of early to middle adolescence]. Psychologie in Erziehung und Unterricht, 66, 118 – 130. https://doi.org/10.2378/peu2019.art11d
(2010). Early career burnout among nurses: Modelling a hypothesized process using an item response approach. International Journal of Nursing Studies, 47 (7), 864 – 875. https://doi.org/10.1016/j.ijnurstu.2009.12.007
(2007). Emotional exhaustion and job performance: The mediating role of motivation. Journal of Applied Psychology, 92, 93 – 106. https://doi.org/10.1037/0021-9010.92.1.93
(2004). Burnout in organizational life. Journal of Management, 30 (6), 859 – 879. https://doi.org/10.1016/j.jm.2004.06.004
(2019). Do girls take school too seriously? Investigating gender differences in school burnout from a self-worth perspective. Learning and Individual Differences, 69, 150 – 161. https://doi.org/10.1016/j.lindif.2018.11.011
(2021). The development of school well-being in secondary school: High academic buoyancy and supportive class- and school climate as buffers. Learning and Instruction, 71, 101377. https://doi.org/10.1016/j.learninstruc.2020.101377
(2021). Kann erlebte Unterstützung durch Lehrkräfte schulische Erschöpfung und Stress bei Schülerinnen und Schülern abfedern? [Can experienced teacher support buffer school exhaustion and stress in students? ]. Zeitschrift für Pädagogische Psychologie, 36, 101 – 114. https://doi.org/10.1024/1010-0652/a000322
(1999). Cutoff criteria for fit indexes in covariance structural analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1 – 55. https://doi.org/10.1080/10705519909540118
(2013). The factor structure and composite reliability of the profile of emotional distress. The Cognitive Behaviour Therapist, 6. https://doi.org/1010.1017/S1754470X13000214,
E15 .(Eds.). (2016). Growing up unequal: Gender and socioeconomic differences in young people’s health and well-being. Health Behaviour in School-Aged Children (HBSC) Study: International report from the 2013/2014 survey. World Health Organization.
(2000). A test of the Maslach Burnout Inventory in three samples of healthcare professionals. Work & Stress, 14, 35 – 50.
(2005). School related stress in early adolescence and academic performance three years later: The conditional influence of self-expectations. Social Psychology of Education: An International Journal, 8, 3 – 17.
(2008). Peer group influence and selection in adolescents’ school burnout: A longitudinal study. Merrill-Palmer Quarterly, 54, 23 – 55.
(2009). Peer group homogeneity in adolescents’ school adjustment varies according to peer group type and gender. International Journal of Behavioral Development, 33 (1), 65 – 76. https://doi.org/10.1177/0165025408098014
(2016). The German version of the perceived stress scale: Psychometric characteristics in a representative German community sample. BMC Psychiatry, 16, 1 – 10. https://doi.org/10.1186/s12888-016-0875-9
(2012). Sozialwissenschaftliche Datenanalyse. Eine Einführung [
Data analysis in social sience. An introduction ]. VS.(2014). Learning difficulties, academic well-being and educational drop out: A person-centred approach. Learning and Individual Differences. 31, 1 – 10. https://doi.org/10.1016/j.lindif.2013.12.011
(2021). School-related pressure and parental support as predictors of change in student stress levels from early to middle adolescence. Journal of Adolescence, 87, 38 – 51. https://doi.org/10.1016/j.adolescence.2020.12.008
(1984). Stress, appraisal, and coping. New York: Springer.
(2014). The relationship between student engagement and academic performance: Is it a myth or reality? The Journal of Educational Research, 107 (3), 177 – 185.
(2011). Trajectories of school engagement during adolescence: Implications for grades, depression, delinquency, and substance use. Developmental Psychology, 47, 233 – 247. https://doi.org/10.1037/a0021307
(2013). Life stress and academic burnout. Active Learning in Higher Education, 15, 77 – 90. https://doi.org/10.1177/1469787413514651
(1999). The reliability and validity of scores from the children’s version of the Perception of Success Questionnaire. Educational and Psychological Measurement, 59, 651 – 664.
(1981). The measurement of experienced burnout. Journal of Occupational Behavior, 2, 99 – 113. https://doi.org/10.1002/job.4030020205
(2016). Understanding the burnout experience: Recent research and its implications for psychiatry. World Psychiatry, 15 (2), 103 – 111. https://doi.org/10.1002/wps.20311
(1997). The truth about burnout. Jossey Bass.
(2001). Job burnout. Annual Review of Psychology, 52, 397 – 422. https://doi.org/10.1146/annurev.psych.52.1.397
(2008). Educational transitions and differential learning environments: How explicit between‐school tracking contributes to social inequality in educational outcomes. Child Development Perspectives, 2 (2), 99 – 106. https://doi.org/10.1111/j.1750‐8606.2008.00048.x
(2020). School Burnout Inventory: Latent profile and item response theory analyses in undergraduate samples. Frontiers in Psychology, 11, 188. https://doi.org/10.3389/fpsyg.2020.00188
(2015). Validation française du School Burnout Inventory. European Review of Applied Psychology, 65, 285 – 293. https://doi.org/10.1016/j.erap.2015.10.002
(2010). Gender differences in the cognitive vulnerability-stress model of depression in the transition to adolescence. Cognitive Therapy and Research, 34, 501 – 513. https://doi.org/10.1007/s10608-009-9281-7
(2005). Predicting psychological distress in college students: The role of rumination and stress. Journal of Clinical Psychology 61 (4), 447 – 460.
(1998 – 2017). Mplus user’s guide (7th ed.). Author.
(1999). School-related stress and psychosomatic symptoms among school adolescents. Journal of School Health, 69, 362 – 368. https://doi.org/10.1111/j.1746-1561.1999.tb06430.x
(2017). PISA 2015 results (Volume III). Author.
(2019). The impact of stress on students in secondary school and higher education. International Journal of Adolescence and Youth, 25 (1), 104 – 111. https://doi.org/10.1080/02673843.2019.1596823
(2010). Gender differences in burnout. Journal of Vocational Behavior, 77, 168 – 185. https://doi.org/10.1016/j.jvb.2010.04.006
(2015). The perceived role of parental support and pressure in the interplay of test anxiety and school engagement among adolescents: Evidence for gender-specific relations. Journal of Child and Family Studies, 24, 3742 – 3756. https://doi.org/10.1007/s10826-015-0182-y
(2014). The interplay of perceived stress, self-determination and school engagement in adolescence. School Psychology International, 35, 405 – 420. https://doi.org/10.1177/0143034313498953
(2010). Bifactor models and rotations: Exploring the extent to which multidimensional data yield univocal scale scores. Journal of Personality Assessment, 92, 544 – 559. https://doi.org/10.1080/00223891.2010.496477
(2007). The role of the bifactor model in resolving dimensionality issues in health outcomes measures. Quality of Life Research, 16, 19 – 31. https://doi.org/10.1007/s11136-007-9183-7
(2009). School Burnout Inventory (SBI) reliability and validity. European Journal of Psychological Assessment, 25, 48 – 57. https://doi.org/10.1027/1015-5759.25.1.48
(2008). The role of educational track in adolescents’ school burnout. British Journal of Educational Psychology, 78, 663 – 689. https://doi.org/10.1348/000709908X281628
(2016). Integrating the light and dark sides of student engagement using person-oriented and situation-specific approaches. Learning and Instruction, 43, 61 – 70. https://doi.org/10.1016/j.learninstruc.2016.01.001
(2009). Depressive symptoms and school burnout during adolescence: Evidence from two cross-lagged longitudinal studies. Journal of Youth and Adolescence, 38, 1316 – 1327. https://doi.org/10.1007/s10964-008-9334-3
(2013). School burnout and engagement in the context of demands–resources model. British Journal of Educational Psychology, 84 (1), 137 – 151. https://doi.org/10.1111/bjep.12018
(2012). The schoolwork engagement inventory: Energy, dedication and absorption (EDA ). European Journal of Psychological Assessment, 28, 60 – 67. https://doi.org/10.1027/1015-5759/a000091
(2001). A scaled difference chi-square test statistic for moment structure analysis. Psychometrika, 66, 507 – 514. https://doi.org/10.1007/BF0229
(2002). Missing data: Our view of the state of the art. Psychological Methods, 7 (2), 147 – 177. https://doi.org/10.1037/1082-989x.7.2.147
(2004). Job demands, job resources, and their relationship with burnout and engagement: A multi-sample study. Journal of Organizational Behavior, 25(3), 293 – 437. https://doi.org/10.1002/job.248
(2002). The measurement of engagement and burnout: A two sample confirmatory factor analytic approach. Journal of Happiness Studies, 3 (1),, 71 – 92. https://doi.org/10.1023/A:1015630930326
(2002). Burnout and engagement in university students: A cross-national study. Journal of Cross-Cultural Psychology, 33 (5), 464 – 481.
(2020). Ausländeranteil in Mecklenburg-Vorpommern bis 2019 [Share of foreigners in Mecklenburg-Western Pomerania 2019]. Retrieved from https://de.statista.com/statistik/daten/studie/258844/umfrage/auslaenderanteil-in-mecklenburg-vorpommern/
(1998). Mixed methodology: Combining qualitative & quantitative approaches. Sage.
(2020). Burnout und Engagement: Eine Pilotstudie zum Transfer des Job-Demands-Resources-Modells in den deutschen Schulkontext [Burnout and engagement: A pilot study on the transfer of the job-demands-resources model to the German school context]. Prävention Gesundheitsforschung, 15, 269 – 274. https://doi.org/10.1007/s11553-019-00752-1
(2014). Schoolwork engagement and burnout among Finnish high school students and young adults: Profiles, progressions, and educational outcomes. Developmental Psychology, 50, 649 – 662. https://doi.org/10.1037/a0033898
(2009). Adolescents’ self-concordance school engagement, and burnout predict their educational trajectories. European Psychologist, 14, 332 – 341. https://doi.org/10.1027/1016-9040.14.4.332
(2000). Three likelihood-based methods for mean and covariance structure analysis with nonnormal missing data. Sociological Methodology, 30, 167 – 202. https://doi.org/10.1111/0081-1750
(1999). On the relationship between the higher order factor model and the hierarchical factor model. Psychometrika, 64, 113 – 128. https://doi.org/10.1007/BF02294531
Appendix
