investigating the reliability of the reading module of the iranian ministry of science, research, and technology’s english proficiency test

As a national high-stakes test of english proficiency, msrt needs further scrutiny of reliability. thus, the present study aimed to investigate different sources of variation that may impact the msrt test-takers’ reading comprehension. accordingly, some factors including the reading topics, item types, and participants’ general proficiency were delved into based on the scores obtained from 60 msrt prospective candidates. upon administering a sample of the reading subtest taken from a recent version of msrt, the collected data was dichotomously scored and then analyzed in terms of internal consistency, inter-correlations, and causal patterns. the yielded results showed an overall reliability of 0.86 for the reading module, while a moderate interrelationship was obtained amongst the passages (r = 0.47) as well as the item types (r = 0.44). furthermore, the mixed anova results demonstrated that topic and item type significantly affected the reading performance, whereas the proficiency factor did not play a conspicuous role in distinguishing the participants’ reading accomplishment. both theoretically and operationally, the results reported through this study may stress the need for reconsidering the influence of such test-method facets as topic and item type in the msrt reading subtest to upgrade the test’s unidimensionality and fairness.