What do willingness and involvement ratings actually measure? Evidence from L2 speech and writing assessment

Language assessment rubrics sometimes ask raters to judge learners’ willingness and involvement alongside language and content quality. However, it is unclear whether these ratings capture distinct learner characteristics or simply reflect overall performance quality. This study examined 22,400 ratings of 280 English speeches and essays from the International Corpus Network of Asian Learners of English Global Rating Archives. It also linked learner performances to independent survey measures of motivation, willingness to use English, proficiency, and language-learning background. Willingness and involvement ratings were extremely closely associated with ratings of fluency, accuracy, intelligibility, and content quality. This overlap remained after accounting for differences among raters and performances. Moreover, learners’ self-reported motivation and willingness provided little additional explanation once general performance quality and background variables were controlled. The findings indicate that willingness and involvement ratings did not function as clearly distinct assessment dimensions in this dataset. Rather, they appeared to reflect a broader evaluative impression of performance. The study highlights the need to define engagement-related criteria through observable behaviour and cautions against inferring internal motivation from a single spoken or written performance. For learner-corpus applications, attitude subscores should therefore not be treated as independent metadata variables without modelling their shared variance with general performance.