Assessing reading comprehension in CET-4: a cross-sectional case study

Testing reading has always been indispensable in language tests. The present study aims to evaluate three parallel reading tests from the National College English Test Band Four (CET-4) in June 2021 in terms of reading ability orientations, text readability and its correlations with reading performances. The quantitative analysis reveals: (1) in Banked Cloze (BC), lexical items designed and examined were largely confined to an intrasentential level. Meanwhile, an either-or situation was found in each test on adverbs; (2) Matching prioritized the skill of scanning for explicit details; (3) Reading in Depth (RD) focused on comprehension of facts and details while seldom involved testing gists at textual levels; (4) text readability in the same question type among the three target tests was short of an expected internal consistency and (5) the correlation between test results and text readability in CET-4 reading comprehension (RC) was not absolute but dependent on question types. Based on findings derived, this study would hopefully offer some vital clues to a more scientific design and assessment of CET reading tests in both theory and practice. And it is also expected to shed some light on evaluating text readability from new perspectives.