The majority of Iranian universities which offer doctoral programs, have started to accept only those candidates who have a good command of English proficiency, which is usually measured either by a general English test administered by the same university, or by tests like MCHE, TOLIMO, or MOHMET which are administered by the related ministries" The results of these tests are used to make very important decisions on the fate of candidates; hence, these tests should be carefully investigated in terms of reliability, validity, and the other characteristics of the test. The English Examination of Tarbiat Modarres University (TMU) is a TOEFL-like English language proficiency test used as a prerequisite for the Ph.D. Entrance Examination at Tarbiat Modarres University. This year, 6000 male and female postgraduates from different universities and different fields took part in this exam. This study attempts to investigate its reliability, test difficulty, and speededness. To achieve the purpose of the study, the data obtained from more than 2000 male and female Ph.D. applicants were analysed through different types of reliability estimates. The reliability coefficients of the total test and its subparts were estimated through K-R21, K-R20, split half, and cronbach DC. Also the Standard Error of Measurement (SEM) was computed. Then the reliability coefficients of different types of reliability were compared for significant differences. In addition, the test difficulty as a general characteristic of the test, which may influence the size of reliability estimates, was also investigated. Related to this issue, the difficulty values of the test items are calculated. Based on the results, the relation between test difficulty and different types of reliability were discussed. This study also investigated speededness as a source of measurement error. Here, we determined if section 2 and 3 of the TOEFL-like TMU English Examination were truly speeded according to established criteria. Two exploratory techniques employing regression analyses were used in an attempt to account for the possibility Quarterly Journal of Humanities, Al-Zahra University / 57 of random 'or patterned responses at the end of each section. One technique provided an estimate of the degree to which all examinees truly reached 75 percent on the sections, and the second determined if 80 percent of the examinees truly completed the last set of items. The results revealed that although the TMU English Examination was to some extent reliable, the reliability coefficients did not meet the levels which would be required for a high-stake test. Also it was found that based on norm-referencing criteria, the TMU exam did not have an acceptable level of difficulty that is, it was too difficult for the examinees. Further, the results showed that the test was slightly speeded as far as the ETS criteria are considered. Results of this study may be of benefit to test developers in general and language test designers in particular. More specifically, policy makers and test developers of TMU, the Ministry of Science, Research and Technology, the Ministry of Health and Medical Education and other such institutes may get some benefit from this study.