Introduction

In this chapter I discuss how different research methodologies can inform the design and validation of vocabulary assessment in oral proficiency interviews. The rationale is twofold. Firstly, despite the profuse literature on vocabulary testing, the assessment of spoken vocabulary is under-researched and a number of key questions thus remain unanswered, such as: What role do the various dimensions of vocabulary knowledge play in speaking tests contexts? What is the relationship between candidate production of different types of vocabulary (for example, (in)frequent words and colloquialisms) and examiner ratings for vocabulary? Secondly, in oral proficiency interviews vocabulary is assessed as part of a larger construct, namely the ability of candidates to interact with an examiner. Consequently, the validity of this assessment is affected by factors that go beyond candidates’ lexical output per se. As this chapter will show, these factors can be best ascertained and understood by applying different research methodologies to their study.

Oral proficiency interviews vary considerably. One manifestation of this is whether the roles of interlocutor and rater are performed by the same person or by two different people. For simplification, this chapter uses the term examiner throughout, except when it is important for the discussion to make a distinction between the two roles.

Although the chapter's emphasis is mainly methodological, I analyse illustrative extracts and report findings from empirical work on the assessment of spoken vocabulary. This work draws extensively on a corpus of oral proficiency interviews of Spanish as a second language (L2) collected over a four-year period (2000–2004).