Evaluation of Kenyan primary education performance using canonical correlation analysis

Abstract

Education is very important in the life of every person and is rated the best inheritance a parent can give to his child. Many studies have been carried out on the Kenyan education system but none regarding the interrelationships between the subjects taught at primal)' school level. Therefore, it would suffice to find out the importance and relevance of each subject as well as how performance in one subject impacts on another. This thesis briefly analyzes the relationship between eight core subjects examined during the year 2004 Kenya Certificate of Primary Education and goes ahead to deduce the most important subjects offered at the said level. The method of Canonical Correlation Analysis (CCA) has been used since it is best suited to study relationships between two sets of variables. One set consists mathematics and languages while the other consists Science, GHC and Religion. CCA describes the internal structure of the relationship within and between the sets i.e. intreset and interset correlations and suggests a linear predictive formula for the dependent set given the observed variables of the independent set.