Abstract
In this paper the author argues that issues of construct validity should be given more prominence in the validation of lexical test items. One way of determining the construct validity of vocabulary items is to interview subjects directly after taking the items to ascertain what is actually known about the target words in question. This approach was combined with the framework of lexical competency proposed by Nation (1990) in an exploratory study which investigated the behaviour of lexical items on TOEFL. In individual interviews, six TOEFL vocabulary items were given to 30 pre-university international students who were then questioned about their knowledge of the target words’ associations, grammatical properties, collocations and various meaning senses. The results suggest that the type of item currently employed in TOEFL does not adequately reflect association, grammatical and collocational knowledge, and that even meaning knowledge is not captured as well as might be hoped. This indicates that the field could benefit from deeper exploration of what vocabulary test items are actually measuring.
Get full access to this article
View all access options for this article.
