Vertical And Horizontal Analysis Of Test Data And Its Significance On Educational Evaluation

Educational evaluation is of vital importance to the new curriculum reform, and scientific test data analysis is considered the key to educational evaluation. Taking a reflective study of the current test data analysis, we find it has some negative impacts on education:students'initiatives are suppressed by mere score-ranking, evaluation of teaching abilities is affected by simple data analysis, let alone management of schools and educational administrations. The underlying reasons lie in the long-lasting mechanism for talent selection, lack of talents doing educational statistic work, and the social disapproval of the existing statistical methods, etc.. Test data analysis should never be the impediment to educational development, which is advancing swiftly and vigorously.Different dimensions need to be taken into test data analysis. This thesis attempts to have both vertical and horizontal analyses of students'test data in order to ensure scientific evaluation of learning, teaching and effectiveness of teaching value, with the ultimate aim to promote and facilitate teaching and learning using an overall holistic and developmental approach.The first chapter of the thesis introduces background, significance, method, problems and relative concepts of the study.The second chapter is a literature review, which includes information on obtainment and use of test data, requirement on educational evaluation from the new curriculum reform, and relative study of overseas data analysis, especially on PISA testing program from international Organization for Economic Cooperation and Development.The third chapter is a detailed analysis of a bank of test data, which is collected from a high school enrollment examination in one of the districts in Shanghai. Based on traditional analysis, it makes both vertical and horizontal analyses using graphics and standard score, and analyzes test data at different periods from different aspects of schools and subjects in order to have constructive analysis of teaching value for schools and subject studies.The fourth chapter summarizes and draws a conclusion. It sets up a vertical and horizontal analyzing framework and states some limitations of the study. Further research assumptions are made, such as to enhance 4D study on test writing, to enhance vertical and horizontal analysis of micro-data, to switch partial distribution into normal distribution through nonlinear transformation first before vertical and horizontal data analysis while using standard score to analyze test data.