Combining Teacher Assessment Scores with External Examination Scores for Certification: Comparative Study of Four Statistical Models
This study investigated four statistical models for combining teacher assessment scores with external examination scores for certifying secondary school graduates in Ghana in terms of validity of the composite scores. The models studied were (1) external examination moderation with nominal weights, (2) reference school moderation using nominal weights and square-mean-root combination, (3) reference school moderation using achieved weights and square-mean-root combination, and (4) external examination moderation using nominal weights and square-mean-root combination.
The population of study was 56,924 graduates of the 455 public senior secondary schools in Ghana for the 1996 academic year. A sample of 2,378 graduates from 11 randomly selected public secondary schools was chosen for the study.
The data collected for the study were teacher assessment mathematics scores for the three years and external examination mathematics scores. University entrance examination scores in mathematics were obtained for a subsample of 115 graduates who took that examination.
Although the results show that two of the models have a potential to serve as alternatives to the external examination moderation with nominal weights, their improvement in this study over that model is not substantial. Further research with subject matters other than mathematics is needed.
IFE PsychologIA Vol 9, No1 2001, pp. 12-34