Speaker's Abstract

Title:
"From Test Scores to Learning Trajectories: Rethinking Academic Language Assessment through Ecological Languaging and Bayesian Updating"
English Medium Education (EME) institutions often rely on single-shot proficiency measures such as IELTS, Duolingo, or C-tests as proxies for students’ academic English readiness. These instruments assume that language ability is a stable, internal trait that can be captured through decontextualized testing. However, academic performance in EMI classrooms is dynamic, relational, and deeply context-sensitive.
This presentation introduces Ecological Languaging Competencies (ELC) (The New Territories Group, 2026) as an alternative framework for understanding English for Academic Purposes (EAP). Rather than viewing language as a hidden ability inside the learner, ELC conceptualizes competencies as emergent coordination potential: the capacity to mobilize linguistic, cognitive, social, and material resources to successfully manage real academic tasks.
Building on the ELC perspective, I present a Bayesian assessment approach that tracks students’ EAP development over time. Instead of assigning fixed levels, the system updates teacher-informed expectations using evidence from learning-oriented assessment (LoA), ecologically authentic academic tasks. The result is not a single score, but a probabilistic developmental profile that reflects uncertainty, growth, and contextual variation.
In contrast to snapshot testing, this approach models development as a trajectory, which is continuously refined through classroom participation. I will demonstrate how Bayesian updating (using our teacher-friendly app) can formalize professional judgment without replacing it, support formative decision-making, and provide a more developmentally sensitive alternative to proxy EAP tests.
The talk concludes by considering implications for EMI policy, placement practices, and the future of academic language assessment.