Mahmoud Ekrami; Negar Elhamian; Somayeh Rajabzadeh
Volume 3, Issue 10 , December 2015, , Pages 49-58
Abstract
Present study, titled “Estimating the parameters of multiple-choice questions in the evaluation of the educational systems virtual" was performed on failed test to prevent the repetition of tests. For parameters estimation of the questions, responses of 49 final test of failed students was investigated ...
Read More
Present study, titled “Estimating the parameters of multiple-choice questions in the evaluation of the educational systems virtual" was performed on failed test to prevent the repetition of tests. For parameters estimation of the questions, responses of 49 final test of failed students was investigated in an academic semester. In the classic model, the degree of difficulty of each question (p) the proportion of correct responses in all subjects, and judgment (Rpbis) double-stranded correlation with the test point considered, the calculations showed: (1) In-89.7% of tests there wasn’t even a very easy and reliable question. (2) In 63.26% of tests wasn’t observed even a very difficult and reliable question. (3) Only 2.04% of tests is without unreliable question. After deleting the unreliable questions, the students’ marks were determined again and about 15.11% of students passed the test. Thus, by deletion unreliable questions, percent of failed tests reduced remarkably, the waste of resources is prevented, the criteria of students’ academic achievement became real more and provided better basis for decision-making in university.