Acting on science and technology teachers' professional burnout variables forecasting with artificial neural network (The sampling of Zonguldak)

The teaching profession is job that has got important responsibilities and high expectation level in terms of its tasks. This situation can lead to it that teachers are under stress and pressure while they are practicing their profession. Circumstance that these teachers live burnout over time will be inevitable. Thanks to this study, professional burnout variables that play a role in development of teachers, in being motivated to work and in affecting the quality of education prescribed by artificial neural network, examined and revealed the relationships between these variables. The research was carried out by 51 science and technology teachers in Zonguldak. In this study, mixed method research used. Respectively survey model and the case study were used in the quantitative and qualitative aspects of research. As data collection tools in this study, "Maslach Burnout Inventory" that was developed by Maslach (1981) and adapted to Turkish by Çam (1991, 1992); "the Self-Efficacy Scale for Science and Technology Education" that was developed by Kaya, Polat ve Karamüftüoğlu (2012); "Ohio Teacher Efficacy Scale" that was developed by Tschannen-Moran and Woolfolk-Hoy (2001) and adapted to Turkish by Baloğlu and Karadağ (2008) and "Personal Information Form" is used. "Professional Burnout Interview Form for Science and Technology Teachers (κ = 0.68)" and "Science and Technology Teachers' Self-Efficacy Interview Form for Science Course (κ = 0.59)" that was developed by researcher to collect qualitative data and was done the content validity work with Fleiss Kappa coefficient (κ) is used. SPSS and MATLAB package programs were used in the analysis of quantitative data. The quantitative data were analyzed through parametric t-test, one-factor variance analysis (ANOVA), simple correlation and multiple linear regression. Artificial neural network is applied to data to find degree that self-efficacy, competence variables and the dimensions of burnout explain burnout and to make predictions about the subject of research. MATLAB package program was used for artificial neural network analysis. Content analysis was done to qualitative data in order to examine deeply questions in burnout and self-efficacy inventories and get the views of teachers. According to research results,sex, having children status, marital status, satisfaction status of the work environment, course load (hours per week) weight, support status of colleagues, the adequacy situation of the income of the science and technology teachers has no effect on professional burnout, self-efficacy and competence. Professional burnout (emotional exhaustion subscale) mean scores of teachers who want to retire is significantly higher than the teachers's who don't want to retire but the confidence to performance mean scores of their were significantly lower. Professional burnout (emotional exhaustion subscale) mean scores of teachers who look forward to weekend vacation were significantly higher than the teachers's who don't look forward to weekend vacation. Professional burnout (emotional exhaustion, depersonalization and personal accomplishment subscale) and the confidence to laboratory knowledge mean score of teachers who are satisfied with the current education system is significantly lower than teachers's who aren't satisfied with the current education system. Professional burnout (emotional exhaustion subscale) mean score of teachers who are optimistic about the professional future is significantly lower than teachers's who are pessimistic about professional future. Science and technology teachers's thoughts about professional future and seniority have no effect on professional burnout. But causes why choose the teaching profession has a effect on burnout. There were significant correlations between the confidence to performance, competence and occupational burnout. There was not a significant relationship between the confidence to laboratory information, the confidence to area knowledge and professional burnout. Four estimation studies such as "emotional exhaustion", "personal success", "depersonalization" and "competence" were conducted with artificial neural network and with the acceptable error rate, some predictions produced in response to the inputs of artificial neural network was closer to the real output value. When results of regression analysis and artificial neural network compared to; approximately 40% network performance and approximately 30% regression performance for "emotional exhaustion"; approximately 20% network performance and approximately 38% regression performance for "depersonalization"; approximately 50% network performance and approximately 29% regression performance "personal success"; approximately 80% network performance and approximately 58% regression performance for "competence" is seen to be. According to the results of the study, artificial neural network can be used safely to determine the variables that influence the burnout and to predict burnout. Thanks to this study, burnout which is subjects including especially teachers, students, school, education system and the whole society was examined its effect on self-efficacy and teacher competence variables. Also professional burnout issue that be included in the human factor was tested with artificial neural network and was reached the conclusion that it can be used safely. A new perspective to the analysis of research data in the education has been gained in this context.