Abstract
This paper provides a number of reflections on the difficulties involved in measuring monitored knowledge of second language grammar, and describes the construction and validation of a prototype test for measuring this knowl edge. It is argued that a fill-in-the-blanks format is to be preferred over multi ple choice, grammaticality judgement or error correction tasks. Problems in defining the structures to be tested are discussed, and a pragmatic solution is proposed. The psychometric characteristics of three versions of the prototype test are described, based on the results obtained for 62 learners of French as a second language in their last year of high school in Dutch-speaking Belgium.
Get full access to this article
View all access options for this article.
