摘要
基于某资格类考试考后数据,对包含单选题和多选题的复合选择题的雷同答卷进行分析,提出单选甄别、多选甄别和合并甄别3种甄别方案,设计不同测验长度、难度、被抄袭考生水平、题目抄袭比率、作弊考生比率及显著性水平的实验样本。研究结果显示:合并甄别和单选甄别都表现出较好的甄别性能,多选甄别由于多选题数量方面的劣势,甄别率低且Ⅰ型错误率高。据此,提出甄别雷同答卷的建议:如果对甄别出的雷同答卷考生进行违纪违规处理,可将单选甄别和合并甄别分别甄别出的结果取交集作为最后处理依据;如果不对甄别结果进行处理,可将2种方案分别甄别出的结果取并集,以此来对考区或考点的考风考纪进行监管。
Based on the post-test data of one qualification examination,this paper explores a way of copy detection by analyzing the similar-response of compound choice item with multiple-choice single-key item and multiple-choice multiple-key item,proposing three detection schemes including multiple-choice singlekey item detection,multiple-choice multiple-key item detection,joint detection,as well as by designing the test with different length,difficulty,level of plagiarized candidates,plagiarized rate of test items,cheating candidate rate and the significance level of experimental samples.The results show that both joint detection and multiple-choice single-key item detection have performed better in copy detecting.Due to the disadvantage of the number of multiple-choice multiple-key item,multiple-choice multiple-key item detection has lower detection rate and higher error rate of typeⅠ.Accordingly,the following suggestions are made for similar response detection:If the detected candidates with the similar response are dealt with in violation of disciplines and regulations,the intersection of the results of multiple-choice single-key item detection and joint detection can be used as the final treatment basis.If the similar response detection results are not taken as the basis for treatment,the results of the two detection schemes can be combined to supervise the examination discipline of the examination areas or locations.
作者
孔祥
常颖昊
KONG Xiang;CHANG Yinghao(Beijing Language and Culture University,Beijing 100083,China;Beijing Normal University,Beijing 100875,China)
出处
《中国考试》
CSSCI
北大核心
2021年第2期63-68,共6页
journal of China Examinations
基金
2019年北京语言大学院级项目“基于言语指征的面试作伪识别研究”(19YJ140009)。
关键词
考试作弊甄别
雷同答卷甄别
单选甄别
多选甄别
合并甄别
考试安全
cheating detection
similar-response detection
multiple-choice single-key item detection
multiple-choice multiple-key item detection
joint detection
test security