*2.3. Procedure*

Before the data were collected, compliance with participant information standards, confidentiality, and ethics in data processing was guaranteed. Questionnaires were implemented on a Web platform, which enabled participants to fill them out online. A series of control questions were included to detect chance or incongruent answers, and any such cases were discarded from the study sample.

#### *2.4. Data Analysis*

The descriptive and confirmatory data analyses were done following the steps by Pérez-Fuentes et al. [51]; in addition, validation was performed in two stages following the steps by Álvarez-García et al. [52]. The first stage was intended to study the structure of the CBB. To approach this objective, the sample was first randomly divided into two independent homogeneous subsamples. The first (*n* = 605) was used as a calibration sample for confirmatory factor analysis (CFA) of the burnout model proposed (Figure 1). Then CFA was done for the proposed model, taking the following fit indices as measures: *χ2/gl*, comparative fit index (CFI), Tucker–Lewis index (TLI), root mean square error of approximation (RMSEA), with the confidence interval (CI) at 90%. The index *χ2/gl* was used considering values below 5 acceptable [53], Comparative Fit Index (CFI) and Incremental Fit Index (IFI) over or near 0.95, and RMSEA below or very near 0.06 [54]. As a general rule, good fit of the model would be found when ratio 2/GL (degrees of freedom)≤ 3; Goodness-of-fit index (GFI), adjusted goodness of fit index (AGFI), and TLI > 0.90; CFI > 0.95; and RMSEA ≤ 0.05. The advisable respecifications were made to the proposed model, which showed good fit indices, considering theoretical and statistical criteria (modification indices, estimation errors, standardized errors of measurement). The Akaike information criterion [55] was used for model selection based on the second subsample (*n* = 635), which was used as the validation sample to validate the respecified model. Cronbach's alpha [56] and split halves were used for the reliability analysis of the new scale.

**Figure 1.** Proposed Burnout model.

In the second stage, an analysis was done to support the proposed invariant factor structure across type of contract (permanent or temporary) and gender (male/female). First, both subsamples were checked to see the goodness of fit of these structures (model M0a, Permanent-Male, and model M0b, Temporary-Female). The four resulting nested models were evaluated: (a) Model 1. Both subsamples were considered simultaneously with free parameter estimation. (b) Model 2. Metric invariance was demonstrated. (c) Model 3. Scalar invariance was demonstrated. (d) Model 4. Strict invariance. There was no consensus criterion to determine the criteria to be used to evaluate the difference in fit of the nested models [57]. This study used ΔCFI to evaluate its fit. ΔCFI interprets the model as fully invariant if the value found is below 0.01 [58]. The analyses were performed using the SPSS version 23.0 Statistical Package for Windows (IBM, Armonk, NY, USA) and the AMOS 22 program (IBM, Armonk, NY, USA).
