أثر إضافة بديل "لا أعرف" في أسئلة الاختيار من متعدد على تقدير قدرة الفرد وخفض بارامترات التخمين وتحسين كفاءة الاختبار في ضوء نظرية الاستجابة للمفردة
Guessing behavior is an important topic on multiple choice tests, due to greater the potential for systematic error or bias which that inflates observed test scores. The purpose of this research is to determine the effect of adding an alternative "I don't know" in multiple-choice questions on reducing guessing parameters, improving item parameters, efficiency of exam, and estimating of ability, based on item response theory. The study sample consisted of 97 male students, their ages ranged from 21 to 22 years with an average age of 21.26, and a standard deviation 0.44, they were divided into two groups, one of them is a control group 38 students, and the other is an experimental group 59 students. The equivalence of the two groups in achievement ability was verified. The instrument is an achievement test consisted of 20 multiple-choice items, with 4 alternatives each. Each item was scored as correct or incorrect. A fifth alternative "I don't know" has been added, for the exam of the experimental group. It is given to whoever chooses it (1/4) a score, to be added to his total score to verify the reduction of the guessing effect. The results showed that there was no significant difference for estimation ability, t(95)=0.44 p= 0.66 despite the control group (M= 12.58, SD= 3.69) attaining higher score than the experimental group (M= 12.22, SD= 4.09). The results also showed that adding an “I don’t know” alternative reduced the guessing parameters for the test items. The experimental group who used an “I don’t know” alternative (M = 0.19, SD = 0.03) was significantly higher than the control group (M= 0.21, SD= 0.02). "t" (95) = 3.08, p= 0.001. The results also revealed that adding the “I don’t know” alternative to the experimental group and using it in the test improved the discrimination parameters. The experimental group (M= 1.29, SD= 0.55) was significantly higher than the control group (M= 1.20, SD= 0.28) in the mean of discrimination, "t" (95) = 2.21, p= 0.03. The results showed that the efficiency of the exam for the experimental group (M= 3.79, SD= 1.16) was significantly higher than the control group (M = 3.12, SD= 0.08) in the average amount of test information, "t" (95) = 24.32, p= 0.001. The results also revealed that the experimental group (M = 0.50, SD = 0.13) was significantly less than the control group (M= 0.59, SD= 0.17), in the mean standard error of estimating ability, "t" (95) = 3.13, p= 0.002. The study concluded with some educational implications, and suggested related research. (Published abstract)