Teacher Confidence in Professional Training: The Predictive Roles of Engagement and Burnout

Teachers’ work engagement positively impacts teachers’ attitudes towards their job. Nevertheless, teachers may experience burnout during their career, which negatively impacts their professional learning opportunities. In this study we investigated the relationship between teachers’ levels of burnout, work engagement, and their confidence in in-service training in a sample of Italian teachers. We expected that burnout mediated the relationship between work engagement and teachers’ confidence in training. A total of 481 teachers completed self-report questionnaires about engagement and burnout, with an ad hoc Confidence in Training Index developed to assess their attitudes towards professional development courses. The mediation analysis confirmed that the teachers’ levels of burnout mediated the relationship between their work engagement and their confidence in in-service training. Findings suggest that teacher confidence in policies about professional training should be evaluated by taking into account their level of engagement and burnout.


Introduction
Despite most research on human capital focusing on profit and nonprofit organizations, governmental institutions are required to address human capital in order to effectively deliver their services. Indeed, human capital approaches encourage investing in people as capital, thus influencing the organizational practices aimed at developing human resources and the organization in general [1,2]. In the context of educational institutions, teachers represent the most relevant human capital, as they impact the quality of the educational services delivered to students, families, and the community [3,4]. In this regard, for this reason, policies aimed at recognizing and addressing teacher development are crucial for the organizational performance of schools, as well as for teacher well-being. Thus, the attitudes towards in-training experiences play a major role in the effectiveness of human capital development initiatives. This link is particularly relevant when considering the peculiar and specific professional experience of teachers.
Indeed, recent findings suggest that teachers begin their professional careers with high levels of work engagement and commitment, which give them the energy to cope with stressful events. Nonetheless, teaching is known to be a stressful occupation and teachers' exposure to multiple sources teachers necessarily perceive it as effectively helping them to do their job. Numerous authors maintain that effective professional development interventions can lead to a change in teachers' educational practices, provided the teachers have a positive attitude towards the training activities (e.g., [30,31]).
Despite the close relationship between effective training and teachers' professional development, no previous studies have focused on teachers' trust in formally organized training or the extent to which they view it as a help to their everyday educational practice. First, given the current lack of instruments for assessing teachers' attitudes towards their in-service training, we set out in the current study to develop an ad hoc measure and to evaluate its psychometric characteristics. Secondly, we investigated the relationships between teachers' levels of work engagement, burnout, and confidence in the in-service training they had received. More specifically, we expected that teachers' burnout scores (in the personal, work-related, and student-related domains) would be negatively associated with their work engagement (in terms of vigor, dedication, and absorption). Furthermore, we expected that poor confidence in their in-service training would be positively associated with burnout and negatively associated with work engagement. Crucially to our research design in the current study, we expected that burnout would mediate the relationship between work engagement and teachers' confidence in training. Specifically, we expected that teachers with low work engagement would display high levels of burnout that, in turn, would negatively influence their confidence in training.

Procedure
We collected the data in ten schools in Northern, Central and Southern Italy. We contacted the teachers after initial approval by their school principals. We administrated all the questionnaires individually to the teachers who completed the paper-pencil questionnaires. This study was conducted in accordance with the requirements of privacy and informed consent laid down by Italian law (Law Decree DL-196/2003). Teachers were assured that participation was optional and that their responses would remain anonymous. The current study is part of wider ranging research that include further variables and instruments.

Measures
Three self-report questionnaires were administered: the Copenhagen Burnout Inventory [25], Utrecht Work Engagement [32], and the Index of Confidence in Training.
Copenhagen Burnout Inventory (CBI): The CBI ( [25]; Italian version by [21]) is a 19-item inventory measuring three sub-dimensions of burnout (personal burnout, work-related burnout, and student-related burnout). It is composed of three subscales with a total of 19 items, each rated on a 5-point Likert scale. The first dimension is composed by six items concerning the physical and psychological fatigue and overall exhaustion experienced by an individual (e.g., How often do you feel tired?; How often do you feel worn out?). The second dimension is composed of seven items concerning the physical and psychological fatigue experienced by respondents due to their teaching work (e.g., Do you feel burnt out because of your work?; Do you feel worn out at the end of the working day?). Finally, the third dimension include six items evaluating the physical and psychological fatigue experienced by people who work with students (e.g., Do you find it hard to work with students?; Are you tired of working with students?). Participants respond using a 5-point scale, ranging from 1 (never) to 5 (always). Alpha coefficients were 0.92 for the full-scale CBI, 0.90 for personal burnout, 0.72 for work-related burnout, and 0.82 for student-related burnout.
Index of Confidence in Training (ICT): The Index of Confidence in Training (ICT) was designed through a three-step process: (1) analysis of the literature and identification of the theoretical framework; (2) definition of the items and development of the index; and (3) testing the model of measurement according to confirmatory factor analysis procedures. First, an analysis of both the theoretical and empirical literature addressing teachers' professional development was undertaken. After an extensive review (e.g., [14,15,17,18,[29][30][31]), we selected six topics as the start-point for measuring teachers' confidence in attempted training (i.e., how training has impacted on teaching style and practices, learning motivation, engagement at work, discovering new knowledge, changing mind, and self-confidence). In the second step, a focus group composed of experienced teachers and trainers generated statements focusing on the teachers' confidence in the attempted training. The authors examined and discussed the proposed items in accordance with the main core of the under-construction index. The item generation process resulted in an 11-item scale (see Appendix A). The items were preceded by the following introductory statement: "Making a balance of the different training moments in the course of your experience, can you indicate to what extent you agree with the following statements?". Participants were asked to express their level of agreement on a 4-point Likert scale, where 1 = disagree, 2 = little agree, 3 = enough agree, and 4 = agree. The sum of the agreement levels indicated the perceptions of training and it was labeled as the Index of Confidence in Training. Finally, (step three) the model of measurement composed of two latent factors (personal growth in training and professional growth in training) was estimated by using confirmatory factor analysis ( [34]; see for details [35,36]). The results of the analysis (χ2 (35) = 113.6, p < 0.05, NFI = 0.936, NNFI = 0.954, CFI = 0.955, RMSEA = 0.070, and 90% CI = 0.055-0.084) supported the adoption of the model as a mean to assess teachers' confidence in training. Factor loadings ranged from 0.60 ("they made me feel less motivated about teaching"-ITEM 2 reverse, i.e., the numerical scoring scale runs in the opposite direction) to 0.75 ("they increased my confidence in my own competencies"). In the present study, the reliability alpha coefficients were 0.815 for personal growth in training and 0.812 for professional growth in training. Reliability for the unidimensional cumulate score was 0.832.

Strategy of Data Analyses
In keeping with current literature, the conceptual structure to be tested was modelled with three latent endogenous variables and eight empirical indicators. Starting from the left to the right (see Figure 1), work engagement was measured by vigor, dedication, and absorption, and it was considered as the explanatory variable. Then, teachers' burnout levels (as the mediator) was conceptualized as estimated by personal burnout, work-related burnout, and student-related burnout. Finally, the confidence in training was used as the target latent variable and it was measured by means of personal and professional growth in training. To estimate the hypothesized model, we used Amos 23.0 [37]. According to a multifaceted approach [38], the following criteria were employed to evaluate the goodness of the tested models: χ2 likelihood, Tucker and Lewis Index (TLI), Comparative Fit Index (CFI), Normed Fit Index (NFI), and the Root Mean Square Error of Approximation (RMSEA), with the corresponding confidence intervals. We considered the model as acceptable for TLI, NFI, and CFI values greater than 0.95 [39] and for RMSEA values lower than 0.08 [40]. In keeping with current practices in SEM (e.g., [41]), we estimated the confidence limits using both Monte Carlo simulation and bootstrapping methods with a set of random samples (k = 500). Mediation analysis was conducted by decomposing the total effects among the variables in direct and indirect effects. We also computed Mahalanobis' distance (p < 0.001) to identify the multivariate outliers, finding that no cases needed to be removed from the dataset. Finally, we assessed the data for normality of distribution. None of the score sets displayed kurtosis or skewness values exceeding the recommended limits (−1, +1), and consequently the Maximum Likelihood method [42] was adopted to estimate the parameters for the SEM analysis.

Descriptive Statistics and Zero-Order Correlations among Teachers' Engagement, Burnout, and Confidence in Training
According to the explanatory aims of the current research, the scores for work engagement and professional burnout are firstly presented, in a disaggregated fashion, according to the standard cut-off points in order to offer a portrait of the overall psychological condition of the participants. Then, the aggregate cumulate scores were used in the multivariate statistics and we attempted to model the data. The results for work engagement are summarized in Table 1, whereas the levels of burnout are reported in Figure 2.  In general, the participants can be described as having a quite good level of engagement and moderate levels of professional burnout. The analysis of the descriptive data suggested that approximately half of the sample reported "high" or "very high" levels of engagement on all three empirical dimensions of engagement. Table 2 summarizes the results of the correlational exploratory analysis. Exploration of zero-order correlations revealed that the three subscales of work engagement (vigor, dedication, and absorption) were positively related with teachers' confidence in the training courses. In fact, positive, medium, and statistically significant correlations were found between confidence in training and the three measures of work engagement. On the contrary, work engagement reported a negative and statistically significant correlation with the burnout measures (personal burnout, work-related burnout, and client-related burnout) meaning that the two latent variables were inversely related. Finally, teachers' burnout levels were negatively associated with confidence in training.

Mediation Analysis Results
According to the procedure of mediation testing, we modelled the hypothesized structural paths (see Figure 1) by estimating (1) the total effect (A) of work engagement on burnout; (2) the total effect (B) of burnout on confidence in training; and (3) both the total (C) and the direct (C') effects of work engagement on confidence in training.
The model showed an excellent fit to the data (χ2(15) = 48.7, p < 0.05, NFI = 0.980, CFI = 0.986, TLI = 0.974, and RMSEA = 0.069 (0.048-0.092)), and supported the adoption of the structural model (see Figure 3). The results showed that work engagement and burnout significantly predicted teachers' confidence in training, negatively and positively, respectively. In particular, work engagement reported a statistically significant, negative, and large standardized total effect on both burnout (β = −0.53, p = 0.005, and 90% CI = −0.425-−0.312) and confidence in training (β = 0.34, p = 0.010, and 90% CI = 0.201-0.601). Similarly, burnout levels revealed a statistically significant, negative, and large standardized effect on confidence in training (β = −0.27, p = 0.003, and 90% CI = −0.738-−0.242), meaning that the higher the burnout the lower the perceptions of confidence in training. Finally, the analysis of the direct (β = 0.19, p = 0.005, and 90% CI = 0.098-0.402) and indirect effects (β = 0.15, p = 0.003, 90% CI = 0.091-0.270) of engagement on confidence in training suggested that a mediating effect of burnout on the relation between work engagement and confidence in training was found. In particular, the ratio between the total and the direct effect indicated that approximately 45% of the positive effect of work engagement on confidence in training was attenuated by the burnout levels. As a consequence, the hypothesis of mediation was partially supported since data did not reveal a full mediation.

Discussion
The current study examined how teachers rate their own in-service training, which is mandatory in several education systems around the world, and how their confidence in the professional development courses, as well as the effects of work engagement and professional burnout, impact their scores on confidence towards these in-training courses. Firstly, we designed and validated the Index of Confidence in Training (ICT) to evaluate teachers' attitudes towards professional training. Our findings show a good reliability of the index for both latent factors, i.e., personal and professional growth in training, as well as for the unidimensional measure able to evaluates teachers' perceptions of training. Secondly, as expected, we found a negative association between teachers' work engagement and burnout, as well as a positive association between engagement and confidence in training levels. Similarly, as expected, our results show that teachers' burnout level is negatively associated with their confidence in training. Finally, the main core of the current study focused on the mediating role of teachers' burnout. Our findings show its partial mediation, affecting the relationship between work engagement and confidence in training. In the following sections, we discuss the implications of our findings for teacher professional development.

Between Teachers' Engagement and Confidence in Training: The Dramatic Impact of Burnout
Coherently with our hypothesis, in the current sample of Italian teachers, work engagement was negatively associated with burnout. It is an expected result in line with abundant international findings that have shown that the more workers are engaged, the less burnout symptoms they experience during their career (e.g., [26][27][28][43][44][45][46][47]). Furthermore, our findings have shown that engaged teachers express high confidence towards professional training. This result was partially supported by previous research that has shown that teachers with high work engagement are more likely to display positive attitudes towards their job in terms of persistence, productivity, job satisfaction, positive work attitudes, organizational commitment, and psychological well-being at work [31,[48][49][50][51]. More specifically, in the current study, we analyzed whether teachers' work engagement was associated with a specific attitude towards their professional career, i.e., in-service training. Our findings show that engaged teachers are more likely to be confident in their professional training. Effectively, when they are asked to take stock of the different training moments in their professional career, engaged teachers strongly agree with ICT items like "they enabled me to discover new topics; they enabled me to try out new things in the classroom; they increased my confidence in my own competencies".
Similarly, as expected, teachers with high levels of burnout show lower confidence in professional training than their colleagues. More specifically, they strongly agree with ICT statements like "they didn't help me to change my attitudes and styles of teaching; I thought that I would never attend such demanding courses again; I found it difficult to apply the course contents with my students." Effectively, burned-out teachers are characterized by detachment from their work (e.g., [28][29][30][31]) and cynicism (e.g., [52][53][54]) regarding their relationships with colleagues and students. In this regard, our results seem to confirm that it might be more difficult for teachers who are experiencing exhaustion and demotivation towards their work to be involved in professional learning experiences.
The main result of our study is that teachers' burnout has a significant impact on the positive relation between engagement and confidence in training, specifically by having a detrimental effect on this positive relation. More specifically, while teachers' engagement is, as expected, directly predictive of their confidence in training, it also has an indirect effect through the mediating role of burnout. In other words, almost half of the positive effect of engagement on confidence in training is weakened by their burnout level. Nevertheless, our results add some new perspectives regarding the burnout dimension focusing on teachers' attitudes towards professional development. The findings are in line with previous research that has shown the detrimental effects of burnout on teachers' commitment to their job [52,[55][56][57][58][59][60]]. Yet, results shed a light on their demotivation to adopt novelty in their work, coming from professional training as well as from their colleagues' suggestions [61]. Conversely, our findings further support this dimension by showing how burned-out teachers have less motivation to positively advance themselves through professional training (on both dimensions of personal and professional growth). In our opinion, policies on teacher professional development needs to take into account the teachers' attitudes towards in-service training, which, in turn, may significantly impact on the effectiveness of the training (e.g., [31]).

Conclusions
Overall, negative attitudes towards their job hinders teachers from actively pursuing job resources (e.g., professional training opportunities). Principals, practitioners, and education policy makers should bear in mind that teachers' attitudes towards professional learning experiences are influenced by their level of burnout during training.
One of the key challenges in improving education today is optimizing teachers' professional development. It has been documented that countries with high-performing education systems implement professional development policies [4]. Furthermore, previous international studies have highlighted the important role played by teacher professional development in improving teachers' beliefs and fostering their relations with students [22,31]. Despite this, there is research evidence that both novice and experienced teachers tend to not feel sufficiently prepared to face the daily challenges at work (e.g., [62,63]). This suggests that teachers perceive their ongoing professional training as ineffective or not effective enough.
Policies providing for teacher professional development should take into account the principle that effective training requires a positive attitude and confidence on the part of the attending teachers. Scholars (e.g., [30]) have found that effective training can lead to a change in teachers' educational practices, provided that teachers have a positive attitude towards their own learning activities. In our opinion, training can positively impact teachers' professional development when it focuses on their professional well-being in the first instance [64]. Negative emotions towards one's job can undermine the impact of any professional learning opportunity.
As proposed by Korthagen [65], a key component of teacher development should be enhancing teachers' ability to recognize their core personal strengths (see [66]), and to draw on them in their everyday educational practice. Research assessing a strengths-based approach to professional learning in teachers has shown that an increased ability to identify their core strengths in their daily work at school helps teachers to feel more autonomous, competent, and effective as professionals [67]. Overall, it seems of great interest to develop a multidimensional approach to teacher development and learning that integrates the professional and personal aspects of teaching (see [65]).
Finally, in the case of other empirical studies, a number of limitations should be addressed in considering the results. First, the participants were recruited as a convenience sample, meaning that the results should be considered more in terms of transferability rather than generalizability [68]. We tried to compensate for this issue by recruiting a large sample of teachers from different school grades following the principle of maximizing sampling heterogeneity. Further studies are needed to comprehend whether and to what extent work engagement and burnout affect confidence in training in the general population of teachers. A second limitation is linked to the use of quantitative self-reports as the principal method for data gathering. Mono-method quantitative studies are vulnerable to common method variance (CMV; [69]), meaning that associations between variables can be distorted. Future studies examining the pattern of association from a mixed-method research design are consequently welcome. Moreover, the present study was cross-sectional and consequently the measures were gathered at one single point in time. Further research is needed to observe variables longitudinally by perhaps including the direct measures of training effectiveness for individual teachers. Finally, it could be interesting to analyze how teachers change their attitudes towards professional training by further taking into account their previous training experiences.
Funding: This research received no external funding.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A
Instructions "In terms of providing an overall assessment of the training courses you have been offered during your teaching career, could you please indicate to what extent you agree with the following statements?"