Teacher made test scores as indicators of national public examination scores in Kenya schools: a case study of national public secondary schools.

Abstract

In this research project the relationship between schools based mock examination scores and final KCSE scores have been investigated and the corresponding predictive validity determined. Consequently how well mock scores evaluate the curriculum and whether or not they can be used as surrogates to KCSE scores is discussed. Ultimately the present and the future role of teacher based examinations such as mocks have been put on focus. The process involves a case study on the public National secondary schools in Kenya, which best represents the Kenya society as intake in those institutions is done on quarter allocation basis and where facilities are most comparable. Data capture was done by random sampling techniques mainly on old records from schools and KNEC archives. The Design basically involved determination of correlations, the use of both descriptive and inferential statistical analysis to depict the nature of sample data distribution, as well as association or differences between variables. The process has applied graphical instruments, sequential plots and error plots to depict differences. Association was measured using Pearson's correlation coefficient, Pearson's regression analysis, Spearman's correlation for strength and the student's t-test. The SPSS computer application software was applied in data analysis. The duration of the study lasted about eight months starting from December, 2007 to August, 2008. The relationship between mock (trial) scores and KCSE scores was found to be Linear, positive and significant at 95 % confidence level.