Department of Nursing Studies, Faculty of Medicine, University of Hong Kong, 4/F, William M.W. Mong Block, 21 Sassoon Road, Hong Kong.
Nurse Educ Pract. 2006 Dec;6(6):354-63. doi: 10.1016/j.nepr.2006.07.002. Epub 2006 Nov 13.
Multiple-choice questions are a common assessment method in nursing examinations. Few nurse educators, however, have formal preparation in constructing multiple-choice questions. Consequently, questions used in baccalaureate nursing assessments often contain item-writing flaws, or violations to accepted item-writing guidelines. In one nursing department, 2770 MCQs were collected from tests and examinations administered over a five-year period from 2001 to 2005. Questions were evaluated for 19 frequently occurring item-writing flaws, for cognitive level, for question source, and for the distribution of correct answers. Results show that almost half (46.2%) of the questions contained violations of item-writing guidelines and over 90% were written at low cognitive levels. Only a small proportion of questions were teacher generated (14.1%), while 36.2% were taken from testbanks and almost half (49.4%) had no source identified. MCQs written at a lower cognitive level were significantly more likely to contain item-writing flaws. While there was no relationship between the source of the question and item-writing flaws, teachergenerated questions were more likely to be written at higher cognitive levels (p<0.001). Correct answers were evenly distributed across all four options and no bias was noted in the placement of correct options. Further training in item-writing is recommended for all faculty members who are responsible for developing tests. Pre-test review and quality assessment is also recommended to reduce the occurrence of item-writing flaws and to improve the quality of test questions.
选择题是护理考试中常见的评估方法。然而,很少有护士教育者接受过构建选择题的正规培训。因此,学士护理评估中使用的问题往往存在项目编写缺陷,或违反了公认的项目编写准则。在一个护理系,从 2001 年到 2005 年的五年期间,从考试中收集了 2770 道多项选择题。对 19 个常见的项目编写缺陷、认知水平、问题来源以及正确答案的分布进行了评估。结果表明,近一半(46.2%)的问题违反了项目编写准则,超过 90%的问题编写的认知水平较低。只有一小部分问题是教师编写的(14.1%),而 36.2%的问题取自题库,近一半(49.4%)的问题没有来源。认知水平较低的选择题编写缺陷明显较多。虽然问题的来源与项目编写缺陷之间没有关系,但教师编写的问题更有可能编写在较高的认知水平上(p<0.001)。正确答案均匀分布在所有四个选项中,正确选项的位置没有偏差。建议所有负责开发测试的教师成员都接受项目编写方面的进一步培训。还建议进行预测试审查和质量评估,以减少项目编写缺陷的发生,并提高测试问题的质量。