Over the past fifteen years or so examiner input in oral proficiency interviews has provided a constant topic of discussion within language assessment quarters. Despite the attention received, the precise nature of such input remains largely elusive, as does its relationship to test-taker performance and its assessment. The aim of this paper is to shed some further light on this matter by focussing on the relationship between examiner support and test-taker vocabulary. An analysis of 30 Spanish oral examinations reveals that test-takers' lexical richness and diversity, on the one hand, and examiner's support strategies, on the other, discriminate between grades awarded for vocabulary in these examinations. Qualitative analysis of the transcribed corpus and of examiner retrospective verbal protocols, however, reveals that vocabulary output on its own does not adequately explain some of the grades awarded. Nor does frequency of examiner support entirely explain the relative impact of accommodation strategies and vocabulary on grades. In view of this, an integrative approach that combines statistical and qualitative analysis is suggested as an optimal framework within which to develop current work on oral proficiency interviews.