| dc.description.abstract | Analysing Kenya Certificate of Secondary Education (KCSE) exam results is vital to
inform the design of interventions to improve teaching and learning of Mathematics in Kenyan
schools. Here, further analysis of the 2011 Kenya Certificate of Secondary Education results
explores patterns in the results and factors that affected the grades scored in Mathematics by
students. KCSE results for 699 schools with a total of 50,584 candidates in Western Kenya
region were collected and further analysis was done. Descriptive analysis of the data involved
calculation and presentation of summary statistics using tables and graphs produced using R.
This study demonstrates the application of two approaches to analyse the clustered data to
compare mathematics scores between the different school types and between genders. The first
approach, Analysis of Variance (ANOY A), assumes independence of observations while the
second approach, Residual Maximum likelihood (REML) recognises dependency of
observations. Analysis of the data revealed that Mathematics was generally poorly performed by
students. The results indicated that students in single gender schools performed better in
Mathematics than those in mixed schools while boys performed better than girls. The type of
school and gender of the student were found to have a significant effect on the candidate's
Mathematics score. The parameter estimates are observed to be lower when the random term is
included in the Linear mixed model that uses Residual Maximum Likelihood. The interpretation
and significance of fixed and random effects in the model is discussed and areas for further
research are highlighted. | en_US |
