1. Introduction
In Any sphere, whether individual, community or organizational, health has been considered an essential factor in the social dimension of sustainability [
1,
2,
3] and especially in the search for a safe return to the new normality [
4] since it is not possible to aspire to long-term development without conditions that guarantee the health of the community. For decades, Amartya Sen [
5], philosopher and economist, has affirmed that the only path to sustained development in a nation, region, or community is the cultivation of crucial capabilities that empower citizens: education, decent work, fair wages, health, and gender equity. Likewise, and recently, in the report on the progress of the Sustainable Development Goals, the warning continues not to lower our guard in the social aspects of sustainability. The Inter-American Development Bank [
6], in its report “How to Rebuild Education after the Pandemic,” [
6] the Inter-American Development Bank found that by 2021, 165 million students in Latin America and the Caribbean were disconnected from their school life, more than anywhere else in the world. The evidence collected in the report shows that the effects of school disengagement mean that students have not accumulated significant human capital skills, which will have immediate and long-term consequences. Returning to the classroom requires institutions willing to address academic challenges exacerbated in confinement [
7,
8] and provide a nurturing, safe, and stimulating space for students to grow personally and socioemotionally. It insists on the call for social sustainability, paying particular attention to issues such as those mentioned in Goal 2 on health and well-being, Goal 4 on quality education, and Goal 8 on decent work and economic growth [
1]. Forging a healthy occupational context is essential when talking about socially sustainable well-being in young people, both in their academic work as well as in their first job opportunities [
9]. Today there is greater awareness of the importance of social aspects and their influence on occupational well-being because interaction networks build a social capital that is fundamental for decent and socially sustainable workplaces [
10].
Aspects such as teleworking, technostress [
11,
12,
13], academic activity at a distance, and conditions of loneliness were a stressor [
14] for the educational community during confinement, bringing with it effects such as anxiety, [
15,
16], neurosis [
17], and occupational burnout [
4,
18,
19]. As described by Uribe Prado [
20] and other research [
21,
22], taking up Maslach [
23], burnout syndrome consists of the permanence and development of psychosomatic symptoms and illnesses after a stressor event, manifested in emotional fatigue, cynicism, and occupational efficiency. The Mexican version of Uribe Prado’s MBI scale [
20] pays special attention to the somatization factor. Anecdotal and case evidence expressed by university psychological counselors has shown that it is imperative to pay emotional attention to students returning to the classroom post-COVID-19 since signs of emotional exhaustion are beginning to arise in students [
4]. In the current context, occupational burnout went from being a predominant issue in health care workers (doctors and nurses) to a growing issue in the educational field and, increasingly, in students [
24].
The post-COVID-19 face-to-face return to the classroom is a challenge for universities as they must ensure socially sustainable health conditions regarding emotional health. In this sense, it is not convenient to take intervention initiatives without first having a reliable and valid diagnosis that gives us an accurate x-ray of the current socioemotional situation of university students who return to the classroom [
25]. Therefore, instrumental resources that allow us to carry out such diagnoses become essential. In the peak period of the SARS-CoV2 pandemic (2020–2021), multiple investigations focused on the measurement of occupational burnout in the school context, and there has been a growing focus and readjustment of traditional instruments with a focus on students. The Maslach Burnout Inventory (MBI) [
26] has proven to be a viable and reliable instrument in different contexts around the world, and it has its student version, the Maslach Burnout Inventory-Student Survey (MBI-SS) [
27,
28,
29].
Although the validation of PLS-SEM instruments on burnout has grown, it has been less in students and more in terms of exploration in Latin American students. Therefore, this study aims to validate, from a predictive approach, measures of school burnout in Latino university students from Mexico and Colombia. Consequently, the distinctive contribution of the research is the contribution to diminishing two gaps regarding burnout scales in the school context: (1) exploring the validation of the MBI-SS measurement model in the Latin context, (2) further exploring EMEDO measures in the university setting, and (3) the use of PLS-SEM as a non-parametric statistical technique to validate and distinguish the preponderant dimension with the highest predictive strength in both measures in the Mexican-Colombian context. This research seeks to answer the implicit question: Is there validity and reliability in adapting the burnout scale for Latino college students? In order to answer this question, the following sections are developed: (a) a review of the literature on validation of the burnout scale for students; (b) Materials and Methods; (c) Results; (d) Discussion; (e) Conclusion, implications, and limits of the research.
2. Review of the Literature: Validation of MBI-SS Dimensions
We performed a systematic search in the Web of Science (WOS) for the last 14 years where research on occupational burnout that has used the Maslach Burnout Inventory (MBI) has had a growing trend, and this has been exponential, especially in the context of confinement by COVID-19. The MBI-SS consists of 15 items in three dimensions: burnout, cynicism, and professional efficacy [
27,
28,
29]. Although studies have confirmed its validity for university student populations, different versions have emerged, such as Chinese, Japanese, Portuguese-Brazilian, Italian, German, and French [
25,
27,
30,
31]; studies on its use with Latin American students are still proportionally limited. In the years 2020, 2021, and 2022, 35% of the global production on MBI is concentrated, and, according to the search criteria used, research specifically related to teachers and students reaches almost 30% of the global total. Concerning research on MBI in students in Latin America, there are only 60 research studies (1.6%), with Brazil having 34 products, Mexico with 9, Chile with 7, and Colombia with 5 (
Figure 1).
According to the systematic review carried out, it is noteworthy that more than a decade ago, emphasis was placed on valuing the contribution of second-generation statistics for the validation of various instruments that measure occupational burnout, which allows overcoming the weaknesses of previous techniques where the use of Cronbach’s alpha reliability criterion and various multivariate techniques such as factor analysis and regression was common [
32,
33]. Today, it is possible to measure unobservable variables using indirect observation measures or indicators using the Structural Equation Modeling (SEM) technique [
34]. A widely used SEM technique in the validation of MBI-SS is Covariance Based (CB-SEM), in students, for example, in Korea [
35,
36], Japan [
31], Iran [
37,
38], USA [
39]. This second-generation statistical technique has been useful to confirm or reject the theory, prioritizing the use of regressions based on the sum of scores and presupposing strict conditions on the sample that, in many cases, become unrealistic when measuring social behavior [
34].
Within the SEM techniques, the one based on Partial Least Squares (PLS) allows the complexity of the use of multi-constructs and direct and indirect multi-relationships. Its measurement philosophy distinguishes PLS-SEM. It focuses on the exploration and detection of latent variables with a preponderant variance within the dimensions of the models, focusing, therefore, on the variance of a final construct and its predictive power concerning the constructs that influence it [
40]. An example of this research is that carried out by Manzano García and Ayala Calvo [
41] who validated a burnout construct in nursing professionals in Spain, or the work of Lin and colleagues [
42] who validated the dimensions of the MIB as a construct in the Taiwanese tourism sector personnel. Furthermore, recently, Permarupan and colleagues [
43] validated with PLS-SEM the depersonalization dimension of the MIB in Malaysian nursing staff.
Specifically in the educational sector, representative studies include Acaray et al. [
44], who validated organizational cynicism as a construct in the educational sector in Istanbul, Turkey, and Morales-Sánchez and colleagues [
45], who validated and correlated the Spanish version of the Athlete Burnout Questionnaire as a construct in adolescent athletes in Malaga, Spain. Other research was focused on students, such as the work of Wang et al. [
12], where they used and validated the Copenhagen Burnout Inventory in the university context in China. In the same educational sector, Parmar et al. [
43] validated the dimensions of the MBI in a sample of private university teachers in Pakistan. Regarding current Burnout Research based explicitly on the Maslach Burnout Inventory-Student Survey (MBI-SS), several recent works have also demonstrated the usefulness of PLS-SEM to validate its dimensions. For example, in 2017, Acaray et al. [
44] with Turkey students, and in 2021 the research of Smith and Emerson [
22] where the three dimensions of the MBI-GS(S) were validated for US undergraduate accounting students. Later, the same group of researchers in 2022, Emerson, Hair, and Smith [
21] correlated and confirmed the means of the MBI-GS(S) in a convenience sample of 1119 students pursuing various majors in business at four US universities using SmartPLS 3.3.7 program [
46].
Table 1 specifies the tendency of research on college attrition to confirmatory statistics, and the few studies focused on prediction using second-generation techniques such as PLS-SEM, especially the lack of such studies in the Latino context.
The PLS-SEM technique, by its nature, has shown to contribute to reflecting and explaining the psychosocial phenomena of the population in the educational sector realistically. PLS-SEM, unlike other resources, allows assessing the partial weights of validity and reliability of each dimension as a construct, with robust criteria beyond Cronbach’s alpha. Moreover, its prediction-centered approach has the robustness to predict intra- and extra-sample between subdimensions of the same model and, likewise, of one model with others. According to Hair and colleagues [
34,
64], in PLS-SEM, seven statistical conditions that test the reliability of a measurement model can be considered. Based on these criteria, the statistical hypotheses of this research are derived:
Hi. 1. There is individual reliability in the MBI-SS and EMEDO indicators in the Latin context.
Hi. 2. There is internal consistency in the MBI-SS and EMEDO constructs in the Latin context.
Hi. 3. There is reliability in the constructs of the MBI-SS and EMEDO in the Latin context.
Hi. 4. There is convergent validity in the constructs of the MBI-SS and EMEDO in the Latin context.
Hi. 5. There is discriminant validity between the MBI-SS and EMEDO constructs in the Latin context.
Hi. 6. There is relevant explained variance in the dependent constructs of the MBI-SS and EMEDO in the Latin context.
Hi. 7. There are relevant effects in the relationships of the MBI-SS and EMEDO con-structs in the Latin context.
Although this work’s objective is not to assess the structural model or the contrast of relational hypotheses but to assess the measurement model using PLS-SEM, it is necessary to establish a model of relationships. Therefore, this model does not represent the hypothesis system of the work but justifies the logical connection between the variables used. These relationships are initially established based on the logic of the original instrument by Schaufeli and colleagues [
27,
28,
29] and the Mexican version of the MBI by Uribe Prado [
20], where the ultimate variable is inefficacy. As well as considering the correlation values between the dimensions as constructs of the MBI-SS obtained in recent works [
40,
41,
43], resulting in the relational model shown in
Figure 2. This connection is necessary for validating the measures’ internal consistency in PLS-SEM.
6. Conclusions, Implications, and Limits
The question implicit in the objective of this research was: Is there validity and reliability in adapting the burnout scale for Latino college students? In the first instance, in the validated instrument, there is individual reliability in the indicators, internal consistency, reliability, convergent and discriminant validity among the constructs. Likewise, an explained variance and relevant effects in the relationships of the constructs. Therefore, it can be concluded that within the limits of the sample analyzed, the proposed instrument is valid, reliable, and useful to monitor the state of emotional health and thus procure a socially sustainable return, in terms of mental health, for Mexican-Colombian university students.
The evidence showed that in many Latino university students who return to face-to-face activity, school discouragement prevails, which is strongly influenced by emotional exhaustion that generates psychosomatic diseases in the student population. Emotional exhaustion showed a close and essential relationship with the generation of diseases experienced in the analyzed sample. In emotional exhaustion, aspects such as pressure and lack of energy due to burnout were relevant, generating somatizations among which depression, anxiety, and panic stand out. Therefore, it is recommended to manage emotional containment strategies in this return and to diagnose and channel cases where there is any disorder.
Another element that prevails within this burnout syndrome in university students is family cynicism, which is multifactorial driven at home by stress in the face of the COVID-19 crisis. The item “It is a challenge for my family to understand my school and personal problems” stood out regarding family cynicism. The above reflects that university students experience disunity with their families. Strategies that seek and encourage communication and bonding between parents and children are recommended.
At the end of the questionnaire, a qualitative question was left where students who wanted help were invited to leave their contact data. Thirty percent of the students left their contact information. This reflects an opportunity for intervention, but, on the other hand, it also means an urgency for students who wish to receive support. It is recommended to open spaces for students to ask for help and to follow up as soon as possible since not attending to them at a moment of vulnerability could result in a double impact on their openness to their current situation.
The present study has limits which, in turn, become an opportunity for future lines of research. On the one hand, the sample is not representative, so it is not possible to generalize the results. On the other hand, the predictive scope allowed by the PLS-SEM technique was only tested within the sample analyzed in both countries. As future lines of research, we invite future researchers to continue with the validation of school attrition factors in other Latin contexts and to use advanced PLS techniques to test hypotheses such as multigroup invariance, and to explore the predictive power of the model outside the sample, among others. It is also important to note as a limit that this analysis is limited to exploring relationships within the subdimensions of the model. It would be important to continue exploring relationships with other variables to approach the issue of school attrition in a multicausal and multifactorial manner. Finally, as a limit, it is recognized that the validations were in the context of the pandemic, it would be important to analyze the validity of the data outside the crisis by COVID-19.
In summary, mental health is a factor of social sustainability, and only in a way where students are guaranteed the healthy containment of their emotions can we move towards a sustainable education in return to our universities.