Communicative approach to instructions in national-level ESL testing in Sri Lanka

Abstract

This study focused on issues in the language used in instructions in the A/L General English examination. It attempted to identify the extent to which language issues occur in test-item-instructions and the degree to which these test-item-instructions adhere to the CLT theories. The format of the examination and instructions have remained the same over the years. For primary data, past papers between 2008 and 2016 and course-book vocabulary were analysed, while questionnaires were administered to stakeholder groups for opinions/attitudes. Secondary data were obtained from the Department of Education. Instructions were analysed through grammar and mechanics of English language. The issues were identified as, ambiguous instructions, grammar, confusing/literary language, punctuation/mechanics, vocabulary, complexity/difficultly and ethicality. Part A of both papers, carrying an average 33% marks contained form-focused instructions. The five factors needed for a communicative test (Rea Dickins, 1991) seemed compromised due to grammatically-explicit information which reduce communicative-purpose. Eight test-item-types revealed that most issues were in the use of vocabulary, grammar, language and expressions, while some instructions were ambiguous due to complex grammatical structures. Words tested extra to the text-book were increasing over the years and vocabulary-related issues were 13% of the total. Vocabulary-testing items were not aligned with the recommended CA theoretical norms, although they carried around 35% of the marks. Thus, by analysing approximately 160 test-items, the study found that use of language caused 64% of the issues. The GE examination was primarily testing students for grammatical form rather than meaning. For example, they use unitary sentences instead of the contextbased ones. The study concluded that language-related issues occurred to a considerably high degree and to a significant extent, instructions were not based on CA norms. The recommendations are to better align the test with CLT norms, rather than reproducing old formats, by using contextual and integrative tests able to evaluate a variety of knowledge/competencies simultaneously.