ABSTRACT
The purpose of this chapter is to discuss tasks as the means by which we can elicit a sample of language that can be scored. Tasks need to be designed or selected for use that will strengthen the inferences that we can make from scores to constructs. For example, it is unlikely that we would wish to make a claim about a learner’s ability to take part in a discussion from a task that requires the learner to read aloud; reading aloud a passage would not pro vide any information about a test taker’s ability to use turn-taking strategies, open or close topics, or many of the other key constructs that we associate with taking part in a discussion.
