Show Less

Text Complexity and Reading Comprehension Tests


Erik Castello

Based on the analysis of a specially compiled corpus of internationally recognized English as a foreign language (EFL) reading tests at different levels of proficiency, this volume explores the relation between the complexity of written texts and the difficulty of reading comprehension tests. It brings together linguistic investigations into the text-inherent complexity of the tests and a study of the data derived from their administration to groups of Italian university students. The study of text complexity draws on corpus linguistics, text linguistics and systemic functional linguistics. Both quantitative and qualitative analyses are carried out on the language used in the reading texts and in the related tasks that make up the corpus of tests. The assessment of test difficulty, on the other hand, is informed by research on language testing, and, in particular, by findings and methodologies of Classical Test Theory and Item Response Theory. Relevant aspects of these theories are used to analyze and interpret both the data obtained from the administration of the tests and the data collected by means of feedback questionnaires completed by test takers. The application of such diverse methodologies and the subsequent comparison of the results of the analyses has brought out interesting correlations between text-inherent complexity, perceived test difficulty and actual test difficulty.


Show Summary Details
Restricted access

4. Text Complexity, Participant Identification and Cohesion in the Corpus 123


4. Text Complexity, Participant Identification and Cohesion in the Corpus The investigation conducted in this chapter aims to explain how the choice and use of certain cohesive items and structures contribute to the complexity of some tests of the corpus. It also focuses on the bearing that these choices are likely to have on the successful completion of the tasks, that is, on test difficulty. The analyses carried out will mainly concern reference items, but in some cases other relevant cohesive structures and patterns, including thematic structure and lexical cohesion, will be taken into account. The framework for the analysis of reference is mainly based on Martin’s (1992) Participant Identification system, and the main working hypothesis underlying this part of the study is that, generally speaking, correctly retrieving the identity of the ‘participants’ in a given input text is fundamental to the successful completion of the accompanying reading comprehension tasks. The term ‘participant’ will be used here to indicate the ‘people’, ‘entities’, ‘places’, and ‘things’ that are introduced in a text and are made reference to in the text itself and in the related tasks. The present chapter is organised as follows: in section 4.1 the ‘tracking system’ network and the central Participant Identification system network will be introduced and the relevant literature on them reviewed; in order to clarify some relevant features of the linguistic model within which this part of the research project was conducted, in section 4.2 the systemic functional perspective on semantics, lexicogrammar, and the textual...

You are not authenticated to view the full text of this chapter or article.

This site requires a subscription or purchase to access the full text of books or journals.

Do you have any questions? Contact us.

Or login to access all content.