论文部分内容阅读
学校规模的英语口语课程考试一直面临着评分一致性不高的难题。为此,本研究开发了计算机辅助评分系统(CARS),力图通过评分标准实例化、评分任务随机分布处理、单项任务评分、在线评分培训和一致性验证、利用波形图谱实现可视化评分和反馈等手段来提高评分一致性。由6位英语教师先后进行传统评分和机助评分试验,并在评分结束后围绕提纲进行座谈。通过对两次评分结果和座谈记录进行定量和定性分析表明:该系统能够帮助评分员把握评分标准、增强责任心和集中注意力,评分员内部一致性和评分员之间的一致性都比传统评分方法有显著提高。最后为进一步开展评分信度研究提出建议和启示。
The school-based Oral English exams have always faced the challenge of not having the same score. To this end, this study developed a computer-aided scoring system (CARS), trying to instantiate grading standards, random distribution of scoring tasks, single-task scores, online scoring training and consistency verification, visualization scoring and feedback using waveform profiling To improve scoring consistency. Six English teachers conducted traditional grading and mechanized grading tests successively, and held discussions around the outline after the grading. Through the quantitative and qualitative analysis of the two rating results and the discussion record, it shows that the system can help the scorer to grasp the standard of scoring, enhance the sense of responsibility and focus, the internal consistency of the scoring staff and the consistency between the scorers Scoring methods have significantly improved. Finally, it puts forward suggestions and enlightenments for further research on credit rating.