The article summarizes the study of Reading Comprehension Tasks utilized in preparation for Unified (Russian) State Exam. The corpus of reading tasks was analyzed with the use of the classification algorithm developed by Weir and Urquhart (1998), and aimed at determining the level of engagement (local or global) and type of engagement (literal or interpretative) of a reader performing a Reading Task.

The analysis demonstrated manifestations of both levels and types of engagement: the majority of Reading Tasks of ‘Heading-Text Match’ type are found to be of ‘global-interpretative’ configuration, Tasks 2 in EGE Reading Section are typically of a ‘local-literal’ configuration, while Multiple Choice Tasks (Task 3) require both careful and expeditious (in rare cases) reading skills and two types of engagement—literal and interpretative.