The study, published in the journal Discover Education, was conducted by Zafer Algül and Osman Nejat Akfırat of Kocaeli University. Rather than treating burnout as a single global score, the team modeled it as a multidimensional phenomenon, distinguishing job-related burnout, physical and emotional burnout, alienation or depersonalization directed at students, and alienation directed at colleagues and administrators. Self-efficacy, too, was broken into four facets—academic, professional, social, and intellectual—while teachers’ appraisals of their work environment were captured through four dimensions of perceived work-related motivational conditions: physical facilities, in-school factors, out-of-school factors, and professional development and prestige.

The theoretical backbone of the study is a resource-informed occupational stress perspective, drawing on the job demands–resources (JD-R) tradition. In this framing, personal resources such as self-efficacy beliefs and contextual resources such as supportive working conditions can buffer the strain that leads to burnout. What distinguishes the new work is its insistence on domain specificity: the researchers hypothesized that a teacher’s confidence in managing instructional demands would not necessarily predict the same outcomes as confidence in navigating social relationships within the school, and that different environmental appraisals would likewise attach to different burnout facets.

Data were collected between May and June 2022 through both online surveys and paper-and-pencil questionnaires administered across 57 schools in a large metropolitan area in western Türkiye. After screening for multivariate outliers using Mahalanobis distance, the analytic sample for structural equation modeling was 740 teachers. Missing data, which were minimal at less than 0.13 percent per item, were handled with full-information maximum likelihood estimation, and the models were fitted in Python using the semopy package with robust sandwich-type standard errors to accommodate mild non-normality in the Likert-type responses.

The measurement phase of the study was rigorous. Confirmatory factor analyses supported a 12-factor structure covering the four self-efficacy facets, four motivational condition dimensions, and four burnout facets. Convergent validity was supported by average variance extracted values exceeding 0.50, composite reliability exceeded 0.70 across all latent dimensions, and the heterotrait–monotrait ratio peaked at 0.74, comfortably below the conventional 0.85 threshold for discriminant validity problems. A single-factor model—a common diagnostic for shared method variance—fit the data poorly by comparison, with a substantial difference in CFI of 0.506, suggesting that the multidimensional structure was not an artifact of self-report methodology alone.

The structural results were striking for their selectivity. Of all the paths connecting self-efficacy facets to appraisals of work conditions, only two emerged as statistically significant: social self-efficacy was positively associated with in-school factors such as collegial collaboration and administrative support, while professional self-efficacy was positively associated with out-of-school factors such as community expectations and broader perceptions of the teaching profession. The models explained only 5 to 12 percent of the variance in the motivational condition dimensions, a finding the authors interpret as evidence that factors outside the model—leadership practices, school climate, workload, and organizational support—likely play a larger role in shaping how teachers appraise their circumstances.

Direct links between self-efficacy and burnout were similarly concentrated. Professional self-efficacy, which reflects confidence in managing core instructional and professional role demands, showed small but reliable negative associations with job-related burnout and with student alienation and depersonalization. Social, academic, and intellectual self-efficacy, by contrast, showed no significant direct associations with any burnout facet once covariates were accounted for. This pattern suggests that professional efficacy may operate as a targeted protective resource for the specific strains—exhaustion with the job itself and emotional distancing from students—that most closely mirror the demands of the instructional role.

The clearest and strongest findings concerned the environment. Perceived professional development and prestige opportunities were negatively associated with job-related burnout, while in-school factors showed negative associations with both physical and emotional burnout and, most strongly, with colleague and administrator alienation, where the standardized coefficient reached −0.507. Teachers who saw their school’s internal conditions—collaboration, communication, administrative support, school climate—more favorably reported substantially less relational distancing from the people around them. Out-of-school factors, meanwhile, were largely unrelated to burnout once the other dimensions were in the model. The burnout dimensions were explained to a moderate degree, with the model accounting for 17 to 28 percent of their variance.

Notably, bootstrapped tests of indirect associations, based on 5,000 resamples, did not reach statistical significance. In other words, the data did not support the idea that appraisals of working conditions act as a pathway transmitting the effect of self-efficacy beliefs onto burnout. The authors are careful to emphasize that the findings represent direct cross-sectional associations rather than causal mediation, and that the cross-sectional design cannot establish temporal ordering—burnout, after all, could just as plausibly erode self-efficacy and darken appraisals of working conditions over time.

A distinctive contribution of the study is its comparison of public and private school teachers. Before any group comparisons could be made, the researchers established measurement invariance hierarchically, testing configural, metric, and scalar models. The changes in approximate fit indices were minimal—change in CFI no greater than 0.006 and change in RMSEA no greater than 0.002—supporting scalar invariance and licensing meaningful latent mean comparisons. Those comparisons revealed that private school teachers scored higher on social and intellectual self-efficacy, with standardized mean differences of 0.31 and 0.28 respectively. Crucially, however, a multi-group SEM comparing unconstrained and path-constrained models showed almost no fit decrement (change in CFI of 0.003), indicating that the pattern of structural associations was statistically indistinguishable across the two sectors. In plain terms, public and private school teachers differed in baseline levels of some resources, but the architecture linking beliefs, conditions, and burnout was the same for both groups.

The authors acknowledge a series of limitations that temper interpretation. Because unique school identifiers were not retained in the de-identified dataset, the analysis could not correct for the likely clustering of teachers within schools, meaning some standard errors may be underestimated and small effects treated with caution. The data were collected in a post-disruption period following earlier phases of the COVID-19 pandemic, the sample came from a single region rather than a nationally representative frame, gender was assessed in binary form, and no formal multiple-testing correction was applied across the many structural paths. Sensitivity analyses—re-running the models with and without outlier-flagged cases, comparing maximum likelihood with robust estimation, and toggling covariate adjustment—confirmed that the focal associations were stable.

Even with these caveats, the practical implications are pointed. The findings suggest that burnout-prevention programs built exclusively on individual resilience training or generic self-efficacy boosting may miss the mark. Because the most consistent burnout associations involved teachers’ appraisals of school-embedded conditions, interventions such as protected time for collaborative planning, structured peer support teams, transparent workload allocation, mentoring, and credible, competence-linked professional recognition may be more promising targets. The invariance of the structural pattern across public and private schools suggests these strategies could apply across sectors, with delivery adapted to local constraints rather than redesigned from scratch.

The study also carries a methodological message for the research field: averaging burnout into a single score can obscure precisely the domain-specific relationships that matter for theory and intervention. A teacher’s professional confidence may guard against exhaustion and withdrawal from students while doing little for physical and emotional fatigue, and no personal belief may substitute for a collegial, well-supported school environment when it comes to maintaining relationships with colleagues and administrators. As the authors conclude, understanding and preventing teacher burnout likely requires attending simultaneously to what teachers believe about their capabilities and to what they experience, day to day, in the schools where those capabilities are exercised.