Presentations

ABSTRACT

RATIONALE. It is generally agreed that vocabulary knowledge is a multicomponential construct (Nation, 2013) and that the individual components are highly interrelated (González-Fernández, 2022). Research further shows that vocabulary knowledge and its components predict L2 proficiency in listening and reading (Zhang & Zhang, 2022). The ability to recognise words quickly, automatically, and accurately is also important for these two skills (Schmitt, 2019). However, we have a limited understanding of the extent to which ‘fluency’ measures (e.g., speed and automaticity of word recognition) relate to ‘traditional’ measures of vocabulary knowledge (such as size or depth) and the relative impact of these two measures on listening and reading comprehension. One approach has conceptualised fluency as a component of vocabulary knowledge together with size and depth (Daller et al., 2007; Godfroid, 2019). These proposals warrant further validation, particularly among low proficiency learners when lexical representations are being established. 

RESEARCH QUESTIONS. Our study explores (a) the extent to which fluency measures are psychometrically distinct from traditional measures of vocabulary knowledge, and (b) the extent to which these measures predict listening and reading comprehension.

METHOD. 113 16-year-old learners of French who had recently taken their (high-stakes) GCSE exams after approximately 400–450 hours of instruction completed three vocabulary tasks: a form–meaning recognition, a form recall (English-to-French translation), and a yes–no lexical decision, each containing the same 50 high-frequency words from their GCSE curriculum. These tasks elicited three traditional measures (recognition, recall, and yes–no accuracy scores) and two fluency measures (response times indexing lexical speed and coefficients of variation indexing automaticity). Measures of proficiency were the listening and reading components of the standardised Diplôme d'études en langue française and GCSE listening and reading scores. Data were analysed using Spearman’s correlations, confirmatory factor analyses, and structural equation modelling. 

RESULTS. We found that the traditional measures were highly correlated with each other, but not with either fluency measure, and that each traditional measure significantly loaded onto a global ‘vocabulary’ latent variable, but the fluency measures did not. Furthermore, the traditional measures (individually and when loaded onto a traditional latent variable) significantly predicted L2 proficiency in listening and reading, regardless of measure, whereas the fluency measures did not. Taken together, we interpret these findings as preliminary evidence that traditional and fluency measures are psychometrically distinct and that vocabulary knowledge plays a more instrumental role in predicting L2 listening and reading comprehension than fluency, at least among these beginner-to-low intermediate learners. At this level of proficiency, vocabulary knowledge is perhaps yet to be fully automatised, weakening observed relations between fluency and knowledge or proficiency. We discuss implications for foreign language learning, teaching, and testing.  

ABSTRACT

Knowledge of vocabulary, and in particular high-frequency vocabulary, is widely reported to be one of the strongest predictors of second language (L2) listening and reading (Zhang & Zhang, 2022). And yet, the introduction of a compulsory frequency-driven word list for GCSE French, German, and Spanish has created considerable controversy, with some critics warning that GCSE languages are being dumbed down. But what do we really know about the vocabulary knowledge of GCSE learners? To date, most research into vocabulary knowledge and its components and correlates has involved adult learners of English in universities rather than adolescent learners of languages other than English in low-exposure instructed contexts. As such, the language education community in England, including teachers, curriculum designers, test developers, policy makers, and researchers, have had to operate with only a partial understanding of the learner lexicon when designing materials for teaching and high-stakes assessments. 

How we have begun to address these gaps constitutes the focus of our talk. First, we present evidence that underpins aspects of the recent French, German, and Spanish GCSE reforms relating to lexis (DfE, 2022). This includes sobering data on (a) the number and frequency of words that have been used in exams to date and (b) the coverage that current word lists, published by awarding organisations, provide of these exams relative to new frequency-informed word lists (Dudley & Marsden, under review; Finlayson, Marsden, & Hawkes, under review; Marsden, Dudley, & Hawkes, 2023). In the second part of the talk, we present findings about the components of vocabulary knowledge among learners of French GCSE, including size, form–meaning recognition and recall, and speed and automaticity of word recognition (Dudley et al., under review a & b). We examine the extent to which these components predict L2 listening and reading and compare this to the predictive power of knowledge of grammar and grapheme-phoneme correspondences. 

These findings have the potential to improve our understanding of L2 pedagogy, assessment, and the nature of vocabulary knowledge in low-exposure classroom contexts.