A Parametric Modeling of Factors Associated With Teachers? Exit from Service in Kenya

The Survival analysis methods have been widely applied in studying factors associated with duration in employment. This study implemented an accelerated failure time parametric Weibull model to determine the factors associated with teachers exit from service in Kenya, using data from the Teachers Service Commission of 178, 063 teachers who were still working in public schools and institutions up to October 2014 and 28, 403 teachers who had already exited from the profession. The accelerated failure time Weibull and log-logistic models were fitted to study data. To find the best fitting model we used graphical methods and the Akaikes Information Criterion (AIC). The results indicated that the accelerated failure time Weiibull model was the best fitting model to the study data, with the least AIC value. Results of the Weibull model revealed that age, gender, Job group, province or region where the teacher was working and type of exit were the factors that have an effect on the survival time of teachers in service. Therefore accelerated failure time models can be applied to a survival data with even better and more accurate estimates.