STATISTICAL MODELING OF TEACHERS’ SURVIVAL TIME IN SERVICE IN PUBLIC SCHOOLS AND INSTITUTIONS IN KENYA
Abstract
Teachers in public schools in Kenya have been leaving the classroom for greener pastures. Teacher attrition has continued to reduce the survival time of teachers in service. According to research most of this attrition has been due to voluntary resignation and to a smaller extent to retirement of teachers who reach the mandatory retirement age. The voluntary exit of teachers disrupts learning and compromises with the quality of learning. It is also very costly for the government in terms of replacement, training, and induction costs. This research therefore sought to identify and analyse factors associated with teachers’ exit from service and to identify geographical areas with high risks to exit of teachers for better intervention strategies. In this research, survival analysis methods were used namely the Kaplan-Meier and Log-rank test, the Cox Proportional hazards model, the Accelerated Failure Time models and the hierarchical Bayesian Cox spatial modeling and mapping under the Integrated Nested Laplace Approximation. In this research, data was obtained from the Teachers’ Service Commission. This data consisted of teachers employed by the Teachers’ Service Commission in public schools and institutions in Kenya up to October 2014 including those who had exited the service. In this research, the explanatory variables included age, gender, job group, salary, province and county where the teacher was working and type of exit while the response variable was the survival time of a teacher in service. R software was used to analyse the data in this research. Results shows that age, gender, job group, salary, province and county where the teacher was working and type of exit were found to be statistically significantly associated with the survival time of teachers in service. Further results from spatial mapping indicate that, North-Eastern and part of Coast region have higher risks to exit of teachers compared to other regions. These results will assist in the understanding of the causes and patterns of exit of teachers and identification of high-risk geographical areas. From these results, education administrators and planners will be able to formulate policies and appropriate strategies to counter the exits.