タスクに基づくライティングテスト自動評価採点システムにおける客観的評価指標の検討 Examination of Objective Rating Indices to an Automated Essay-Scoring System for Task-Based Writing Tests

Access this Article

Author(s)

Abstract

This study focuses on the validity of objective rating indices to an automated essay-scoring system for task-based writing tests. The relationship between the holistic ratings of <i>Accuracy</i> and <i>Communicability</i> tasks and their objective indices was examined. Independent variables to measure the holistic ratings were chosen from existing objective indices based on the constructs of each task: organizational skills, linguistic accuracy, communicative quality and effects. As a result of the correlation analyses, a moderately high correlation was found to exist between the holistic ratings of the <i>Accuracy</i> task and its indices (tokens and type levels). A moderately high correlation was also found between the holistic ratings of the <i>Communicability</i> task and its indices (type levels and relevant ideas). Through multiple regression analysis, the scores of each task could be predicted with moderately high dimensional accuracy. These findings were discussed with regard to the usefulness and improvement of the automated essay-scoring system.

Journal

  • ARELE: Annual Review of English Language Education in Japan

    ARELE: Annual Review of English Language Education in Japan 27(0), 17-32, 2016

    The Japan Society of English Language Education

Codes

  • NII Article ID (NAID)
    130005549241
  • Text Lang
    ENG
  • ISSN
    1344-8560
  • Data Source
    J-STAGE 
Page Top