Evaluating objectively teacher performance using variance partitioning analysis

Dissertation

College of Industrial Education Doctor of Education in Industrial Education Management Technological University of the Philippines 2021

The study tested the feasibility of a theory (algorithm) described as the variance partitioning analysis (VPA). This was done by partitioning out the non-teacher variance (variance of the control variables) from the total variance (1-R²) of student cognitive. Hence, what remains after partitioning is the variance attributable to the observable and non-observable teacher traits (variables), which is called ΘΘ under the Alicia's-VPA model. on the other hand, Dr. Sanders' " value-added is diluted by the effect of non-teacher variables while VPA removes such extraneous, Teacher effects. this study made use of the IQ, current grades in mathematics, science and English of the secondary student respondents (proxy variable for post-test) and their previous grades in mathematics, science and English (proxy variable for pre-test) as supplied by the Office of the registrar of the institution where the respondents came from. other demographics data were gathered using a questionnaire. the population consisted of private school teachers from Metro Manila. The samples comprised of eighteen (18) teachers from the Ateneo de Manila High School and their corresponding classes in mathematics, science and English. The necessary statistical analysis was done and the 1 of each teacher involved was computed. The teachers per department were ranked objectively on the basis of their 1 scores based on student cognitive achievement. The 1 represents the proportionate variance attributed to a teacher. The teacher who obtained the highest 1 score in his department is presumed to be the most effective. The VPA appears to be feasible as exemplified by the computed Θ1's of the teachers. to supplement the showing of the VPA, a teacher performance evaluation- Author's Abstract