Research Papers in Language Teaching and Learning (Feb 2012)
Re-examining text difficulty through automated textual analysis tools and readers’ beliefs: the case of the Greek State Certificate of English Language Proficiency exam
Abstract
This article reports on an exploratory study that aimed at describing and comparing a range of linguistic features that characterize the reading texts used at the B2 and C1 level of the Greek State Certificate of English Language Proficiency exam (KPG1). Its ultimate purpose was to explore the contribution of such features to perceived text difficulty while at the same time examining the relationship between strategy use and test-takers’ perceived level of reading comprehension difficulty reported in 7,250 questionnaires. Text analysis revealed significant differences between B2 and C1 reading texts for a specific number of text features such as word, paragraph and text length, readability indices, levels of word frequency and presence of words with rich conceptual content. A significant correlation was also found between B2 test-takers’ perception of reading module difficulty and specific text features i.e. lexical diversity, abstract words, positive additive connectives and anaphoric references between adjacent sentences. With regard to C1 test-takers, data analysis showed that two specific text variables i.e. positive logical connectives and argument overlap, correlated significantly with readers’ perception of reading module difficulty. Finally, problem-solving reading strategies such as rereading the text, guessing the meaning of unknown words and translating in mother tongue were found to correlate significantly with perceived text difficulty, whereas support-type reading strategies such as underlining or selectively reading parts of the text were less often employed regardless KPG test-takers’ perception of text difficulty. The findings of this study could help both EFL teachers and test designers gain valuable knowledge regarding EFL learners’ reading habits and also become more alert to the difficulty specific text features impose on the latter.
