日本テスト学会誌 Vol.2 No.1 Abstract
トップ>学会誌>既刊号一覧>既刊号（2-1）>Abstract
JART Vol.2 No.1
| ▶ General research | |
|---|---|
| International Comparisons of Mathematical Thinking Skills in the Timss-r | |
| Kikumi Tatsuoka, Anabelle Guerrero, James E. Corter, Curtis Tatsuoka, Tomoko Yamada, Tao Xin, Enis Dogan, Michael Dean, Seongah Im | |
| Department of Human Development Teachers College, Columbia University | |
| The TIMSS studies have generated a large amount of research aimed at understanding differences among countries in mathematics and science teaching and curriculum. However, few studies have attempted to investigate differences among countries at the micro-level of student performance using a cognitive diagnostic approach. The present study applies the Rule Space Model to diagnose eighth graders' latent micro-level skills (termed "attributes") and knowledge states for twenty countries. Comparisons of the countries show that students in the six top-ranking countries produced good perfbrmance on the TIMSS mathematics test via distinct patterns on these latent achievement variables and knowledge states. Singapore, Korea, Belgium (Flemish), Czech Republic and The Netherlands had almost the same profiles in their learning progress. Japanese and Finnish students tended to learn mathematical thinking skills earlier than mastering content knowledge.On the other hand, Hong Kong students acquired solid content knowledge first and then stronger mathematical thinking skills.U.S. students achieved fairly well, and produced attribute profiles similar to students in Australia, Canada, England, and Russia. In general, countries located in the same geographic vicinity tended to have very similar profiles of learning paths. For example, Israel, Italy and Jordan are similar, as are Indonesia and the Philippines, and Finland, Russia and Netherlands performed similarly although students from The Netherlands had somewhat higher mean scores. | |
| Keywords: mathematical thinking skills, mathematics problem-solving, TIMSS, international comparisons, cognitive diagnosis | |
| ▶ General research | |
| An investigation on the official trial of a new English listening comprehension test with IC players | |
| Uchida Teruhisa, Otsu Tatsuo, Shima Kunmiko, Hayashi Atsuhiro, Ito Kei, Shojima Kojiro, Sugisawa Taketoshi | |
| Research Division, the National Center for University Entrance Examinations | |
| The National Center Test will include an English listening comprehension test in January 2oo6. In the exam, every examinee will individually operate an IC player and listen to English scripts via earphones. Prior to the implementation of the new test, three trial tests were held lbom 2002 to 2005.Of these tests, 'pre'trial' and 'trial' tests were conducted on high schoo1 students, and Post'trial test' was on university freshmen. This paper reperts overview of the third test, 'post'trial test' including the results of the reading test and the listening comprehension test. In addition, effects of artificially-created aircraft noise were examined in terms of the examinees' performance. Finally, we discuss the importance of the coooperation between fundamental and practical researeh, and the needs of the study on social influence upon the implementation of the high-stake test with new elements and formats. | |
| Keywords: the National Center Test,1istening comprehnension test, IC Players, aircraft noise | |
| ▶ General research | |
| Detection and Classification of DIF by using log-linear model | |
| Sato Hiroyuki | |
| Tohoku University Graduate School of Educational Informatics | |
| Differential Item Functioning(DIF), Mantel-Haenszel statistic, log-linear model,Chinese character reading test.Detection and Classification of DIF by using log-linear modelSato HiroyukiTohoku University Graduate School of Educational Informatics The present study examined detection and classfication methods for Differential Item Functioning (DIF) using Mantel-Haenszel statistic analysis and the log-lmear model analysis (log-lmer method). A set of 100 items in Chinese character reading subscale, which is a part of The Broad'Range Japanese Fundamental Language Ski11s Test (BRJFLST, Taira et.al, 1992; 1993; 1995; 1998) for Japanese high schoo1 students, was used for this examination. In this study, responses were sampled from year '91-'92 participants data and year '03 participants data. As a result, 96 items were classified into the same DIF pattern categories by both methods. It means that log-linear method is almost identical to Mantel-Haenszel method in terms of DIF detection. In addition, log-linear method can directly show more detailed DIF features by referring to item interaction parameters. We may make use of log-linear method in order to obtain more detailed information usefu1 for interpreting DIF phenomena. | |
| ▶ General research | |
| Estimation of Difficulty Parameters in a One-Parameter Logistic Model from Incomplete Data | |
| Kyotaka Yoshida1, Takuya Watanabe2, Manabu Iwasaki1 | |
| 1Faculty of Science and Teckmology, Seikei University, 2Graduate School of Engineering, Seikei University | |
| We consider a problern arising in estimation of difficulty parameters of a one-parameter logistic model fiom test scores reported from the test examinees.We deal with the cases that not all the test scores are available.Some examinees do not report his/her scores with some reason such as the test scores were lower than expected. If we analyze the reported data only, it is quite likely to obtain biased estirnates and reach a wrong conclusicm. ln this paper,we first show that incomplete data provide biased estirnates. ln order to obtain unbiased or at least less biased estimates, we prepose a new estimation procedure of difficulty parameters from incorrrplete data by applying multiple irnputations. It is shown by a simulation stady that our procedure can provide reasonable estirnates.Keywords:difficulty parameter, incomplete data, item response theory, multiple imputation one-parameter logistic model | |
| ▶ General research | |
| Scholastic Achievement Structure of The National Center Test 2005 using Self-Organizing Map and Full Information Maximum Likelihood Method | |
| Kojiro Shojima, Tatsuo Otsu, Masaaki Tagumi | |
| The National Center for University Entrance Examinations | |
| Scholastic Achievement Structure of The National Center Test 2005using Self-Organizing Map and Full lnformation Maximum Likelihood MethodKojiro Shojina, Tatsuo Otsu, Masaaki TaguriThe National Center for University Entrance Examinations The scholastic achievement structure in the data of the National Center Test 2005 was exploratorily analyzed using a self-organizing map (SOM). Next, the mean, the variance-covariance, and the correlation structure for the sixteen principal subjects were computed using the full-information maximum likelihood (FIML) method. 'The target population of the analysis was able-bodied high schoo1 graduands. Extraction of nine scholastic achievement layers using SOM, clarified that the subject-selection patterns and favorite subjects were different in each layer. FIML analysis also suggested that two factors which were abilities for liberal arts and sciences were assumed in the scholastic achievement structure. | |
| Keywords: academic achievement, National Center Test, self-organizing map, full-information ML | |
| ▶ General research | |
| Refining of Kanji Placement Test Using the Rasch Model | |
| Eri Bamno1, Mieko Sakai2 | |
| 1Okayama Unmverslty, 2Oita Universty | |
| This study examines and refines a kanji placement test using the Rasch model. The participants were intemational students who were studying Japanese at a college in Japan. One hundred eighty seven students took a non-revised multiple-choice kanji placement test and 176 students took a revised placement test. The results of the analysis revealed some items anddistracters which did not fi nction well in the original version, therefore the following revisionswere made:1)revision or elimination of malfunctioned items,2)elimination of a distracter from each questioq and 3)elimination of an item which had the same item difficulty as another. Afterrevisions were carried out twice on the test, the finial version was found to be more efficient thanthe original one, since it had less items and distracters while keeping the same reliability level asbefore. Iterns that can measure the candidates'abilitay properly also increased in the final version. | |
| Keywords: Rasch mode1, placement test, Kanji, item response theory, classical test theory | |
| ▶ General research | |
| Performance Assessment System that Considers the Uncertainty of State Recognition | |
| Ogata Hiroyuki1, Yamamoto Saeko2 | |
| 1Faculty of Science and Technology, Seikei University, 2Graduate School of Engineering, Seikei University | |
| Authors are developing a performance assessment system that can quantify examinees' skill by observing and recognizing their task gestures in the real environment using sensors such as cameras. In works of Ogata and Igarashi, a prototype system was developed to show the feasibility of automating a certain type of performance assessment. In the process of development, it became apparent that providing solution to the uncertainty of gesture recognition is an important point of this theme. The main purpose of this paper is to deal with this problem by taking account of the uncertainty in quantifying examinees' skil1. Method to improve the robustness of gesture recognition is also discussed. | |
| Keywords: performance assessment, human skil1, task recognition, nearest-neighbor method, Bayesian inference | |
| ▶ General research | |
| Proposing a Visual Imagery Test: ameasurement of visual imagery ability | |
| Tamayu Fukumachi1, Masanorl Nakagawa2 | |
| 1Labour College, the Japan Institute for Labour Policy and Training, 2Graduate School of Decision Science and Technology, Tokyo Institute of Technology, Japan | |
| In increasingly common Internet on-line shopping situations, customers have no opportunity to inspect a real object and must generate the mental image of the goods they wish to purchase. Under these circumstances, the customers'feeling of satisfaction depends on their ability to generate a visual image of the object. This paper proposed a new test to measure human visual imagery ability to examined its reliability and validity through 3 studies. In Studies l and 2, the Visual Imagery Tbst(VIT)was proposed and its standardization with test-retest reliability was established. In Study 3, an attempt was made to demonstrate the relationship between a specific inner structure measured by the VIT and a daily-life imagery-related situation in order to examine the evidence of construct validity. Results showed that high test・retest reliability and some evidence of inner structural validity were attained sucoessfully. In the future, further evidence of VIT validity should be investigated for effective use. | |
| Keywords: visual imagery ability, factor analysis, testing, reliability, validity | |
| ▶ General research | |
| Rule Space Methodology: Apsychometric approach to diagnostic testing | |
| Kikumi Tatsuoka1, Naoki Kuramoto2 | |
| 1Teachers College, Colunmbia University, 2Center for the Advancement of Higher Education, Tohoku University | |
| Combining cognitive science and psychometrics has been one of the most difficult tasks. This paper introduces the Rule Space Methodology(RSM), which diagnoses test takers'cognitive errors. The basic idea of developing the RSM comes from statistical pattem recognition and classification problems. A studentgs response pattern will be classified into one of the predetermined knowledgestates that are logical1y determined from domain experts'hypothesis. The hypotheses are expressed in an item times cognitive skills matrix Q. Rule space is aCartesian product of IRTθand, person-fit statistics, which are designed to map similar patterns into neighboring points. For each student's response, Bayes decision rule determines its most plausible knowledgestates. Once the classification of an examinee is completed,one can indicate which attributes will need remediation. The diagnosed results from the RSM have been proved to be useful through the various studies. | |
| Keywords: cognitive diagnosis, incidence matrix,knowledge,knowledgestate, Rule Space Methodology | |