Determination of the Parameters of Six Multiple Choice Tests of Mashhad University of Medical Sciences (1389-90) based on Item-Response Theory (IRT)

According to the industrialization of countries and development of societies, tests and methods are required to employ people in industries and organizations and make the best selection in getting workforce. Interviews, Written tests, and multiple choice tests are common methods used in employing people. Among these methods, multiple choice tests is the easiest one because of the facility of correction and getting results. In this study, general and technical tests which were held in 1389 and 1390 in Mashhad University of Medical Sciences were studied based on Item Response Theory(IRT) for determining the quality of multiple choice tests and the effective factors in their results. In this applied study, the participants included male and female general and special medicine students in the fields of emergency medicine and euro logy of Mashhad University of Medical Sciences. As in educational assessment sampling is not allowed therefore all available data was judged. The study was conducted on 251 answer sheets of 6 tests (2 tests of atagers,1 of internship, and tests of residency) based on IRT. The comparison of the difficulty level, diagnosis capability, and line slope of each question based on IRT showed that in the first test which was related to residency questions number 1,5,and 27 had better omitted because their line slope was out of the interval of 0.5 to 205.Questions of 3 and 26 because of being bigger than the degree of freedom had meaningful qui squares and should be omitted from the test. Questions number 14 and 9 should be omitted their scores ignored because of low. After the determination of the three parameters of slope 12ag' type="#_x0000_t75">, difficulty 12bg' type="#_x0000_t75">,and success by chance 12cg' type="#_x0000_t75">,for each of the 284 items, and omitting the incorrect irems,ᶿ ability of the participants was determined,and therefore valid items were identified and differentiated in this way:out of 251 studied questions based on line slope, item difficulty, and guess in the six tests it is considered that 7 questions with line slope of out of the interval of 0.5 to 2.5,10 questions were bigger than three times of degree of freedom and had meaningful qui square,9 questions with guess parameter of higher than 0.25 were invalid and should be omitted. In a 31 item test which was related to residency 7 items were omitted because of being invalid, therefore it is worth to plan such questions by which examinees are assessed properly and the intervention of other factors in getting their real score is prevented.