The Retention Rates of Students In Public Secondary Schools Using The Cox Proportional Hazard Model: A Case of Kisumu County, Kenya

Abstract

The study sought to propose a statistical model for the public Secondary School students' retention rates for Kisumu County. We used survival regression analysis in which students were grouped according to the their  performance in KCSE, mean school fee payment, school category, sex, mean age and teacher - student ratio using a desirable survival function. The model was of interest because the study sought to address a life testing experience mostly restricted to survival models and most of the existing models on student retention addressed Universities and tertiary colleges' trend and were particularly developed outside the country. This was in spite of the fact that Kenya suffers from high dropout rates at the Secondary level. The annual Secondary School students data traced from a cohort in form one in the year 2010  to form four in the year 2013 were obtained from the Kisumu County Ministry of Education headquarters and analyzed using survival regression. The variables which were insignificant were dropped to get the desirable function. Survival rates of students in the Secondary Schools were traced at the end of each and every level. It was found out that dropping out of students was influenced by; the category of the school, average fee payment, performance in KCSE and sex, with more girls dropping out than boys and the dropping out mostly rampant at form two. The model will be of relevance to the concerned Secondary education stakeholders in improving the quality of education for it will inform the planning of necessary educational interventions to ensure enhanced retention rate and transition.

Keywords: Retention, Cox Regression, Drop-out rate, Probability Density function, Kaplan Meier.