Background and Objective: The exact design of the residency exams (standards-based) is very important. The purpose of this study was Quantitative and qualitative indicators evaluation of residency exams in Qazvin University of Medical Sciences and study of the relationships between taxonomic levels by quantitative indicators.

Materials and Methods:  All questions of residency exams from 7 Medical Specialties of Qazvin University of Medical Sciences in 2012 and 2013 were studied. Taxonomy of questions in three levels: remembered, understanding and judges were determined. Discriminating power was studied in four groups as good, fair, poor, and negative. Difficulty was classified into 5 levels of easy, medium, hard, very difficult and extremely difficult. SPSS v.19 software was used for the calculation of the T-test and ANOVA.

Results: A total of 2100 questions, 38 percent of the questions reminded taxonomy, 24.1 percent of the level of understanding and 37.9 percent level were judges. 24.9 percent of questions had negative discriminating power, 34.1 percent had poor, 22.5 percent had average and 18.5 percent had good discriminating power. For the degree of difficulty, 22.1 percent simple questions, 22.6 percent average, 24.5 percent difficult, 17.5 percent very difficult and 13.2 percent was extremely difficult. A significant association was found between taxonomic questions and differentiation factor, taxonomy remembered had less differentiation index and questions, taxonomic understanding and application  had differentiation index higher(p<0.0001).

Conclusion: The results showed increased levels of taxonomic questions can enhance the discriminating power test questions. Since questions designed the taxonomy of two and three were lower than the standard (80%), teachers were recommended design more questions higher taxonomy in residency exams.