教育・研究におけるデータ解析 : マークセンス方式によるテストのシステム分析

概要

論文の詳細を見る
Japanese teachers have been slow to utilize machine scorable tests. Too often the proponents of multiple choice testing have given the impression that this is the only way to test and teachers are reluctant to give up the tried and true (even if time-consuming to grade) essay type of test for something new. A bigger problem is the widely felt sense of venturing into an unknown world of test theory, statistics, and computers when utilizing such machine scorable tests. There is also the related fear that without understanding multiple choice test item construction, answer sheet design and computer programing you are bound to make a mess of this new form of testing. To try to calm these fears and make machine scorable testing a practical alternative to teachers who are not test theory specialists, statisticians, or computer programers this brief manual and the computer program explained in it have been prepared. We try to give a brief background of testing and the role of multiple choice tests, a number of examples (both bad and good) of multiple choice items, a brief discussion of basic test statistics that will help the test writing teacher evaluate both his students and the tests he writes, an explanation of item analysis which we hope will allow novice test writers (and experienced ones too) to improve their test writing by learning about the weaknesses in earlier attempts, and introduce ASTEM, the computer program developed to allow teachers with no knowledge of programing to have their multiple choice tests scored and analysed. ASTEM (Analytical System for Testing by Mark-Sensing) is a three part program. Part one inputs test data from an optical mark reader (OMR), scores the test, computes basic statistical characteristics of the test population (mean, standard deviation, range, kurtosis, etc.) and, if necessary, stores the results for later computations. Storage of raw (input) data and/or scored test results (subtotals and totals for each student) is possible. Part two is for the modification (combining, subgrouping, etc.) of the stored data. Part three allows the user to do computational work (correlations, item analysis, etc.) with the stored data. This part also provides an interface with Statistical Package for the Social Sciences (SPSS) . It is the hope of the authors that this program and manual will allow many teachers to improve their evaluation of their students through the regular testing that the multiple choice format permits. This project was supported by the Institute for Integrated Communication Research and Development and the Information Processing Research Center of Kwansei Gakuin University.
関西学院大学の論文

教育・研究におけるデータ解析 : マークセンス方式によるテストのシステム分析

スポンサーリンク

概要

著者

関連論文

スポンサーリンク