The Lexical Content of High-Stakes National Exams in French, German, and Spanish in England

Abstract

Surprisingly little is known about the number and frequency of words that beginner-to-low intermediate 16-year-old learners of French, German, and Spanish are expected to know when taking high-stakes national exams in England. This study presents exploratory analyses of the lexical content of the listening and reading tests of these exams, a corpus totalling 116,645 running words. Specifically, it seeks to understand the number and frequency of words that (a) this demographic seems to be expected to know and (b) could be needed for awarding organisations to create exams year on year. Key findings include that the proportion of low(er) frequency words in the corpus of exam papers seemed large, given the stage of the learners and the purpose of the assessments. Critically, these low(er)-frequency words changed at a high rate between papers, likely incurring a heavy reliance on the lexical inferencing abilities of these relatively inexperienced language learners.

Use of word lists in a high‐stakes, low‐exposure context: Topic‐driven or frequency‐informed

“…Very little research, however, has examined the specific lexical demands of the GCSE exams. In our recent analysis (Dudley & Marsden, 2023), we observed extensive use of unpredictable, low(er) frequency, and likely unknown words in four sets of GCSE exams and suggested that this may be contributing to the perceived and actual difficulty of foreign languages.…”

“…This insufficient coverage could be a consequence of organizations not actively using their lists when creating the exams or the lists not containing a high enough proportion of high-frequency words, given that high-frequency words would normally represent over 80% of any written or spoken text (Dang & Webb, 2014;Nation, 2006b;Webb & Nation, 2017;Webb & Rodgers, 2009a, 2009b. Corroborating the latter explanation are Dudley and Marsden's (2023) findings that 44% of the flemmas used in the corpus of exams at foundation and 45% at higher are indeed low(er)-frequency (defined in that study as beyond the 2,000 most frequent words). Relatedly, it may be that the lower coverage observed in our study is in part due to the high proportion (49% at foundation and 52% at higher) of low(er)-frequency flemmas on the AQA lists (see Online Supporting Information B for more information).…”