Testing discontinuity in the foreign language classroom anxiety–proficiency link

This study re-examines the well-established negative relationship between foreign language classroom anxiety (FLCA) and foreign language proficiency by testing whether this linear association remains stable across the proficiency continuum. Drawing on cross-sectional data from 413 English Foreign Language learners in Mexico, we used piecewise regression to examine whether the strength of the FLCA–proficiency link changes after a data-driven proficiency threshold. Model comparisons indicated that a segmented model provided a better fit than a standard linear model, identifying a breakpoint at a LEXTALE score of 66.25, roughly corresponding to the B2 level of the Common European Framework of Reference. Below this threshold, the FLCA–proficiency relationship was weak and non-significant after correction; above it, the association was significantly stronger. These findings suggest that FLCA may become more closely tied to proficiency after learners reach certain proficiency thresholds, such that more proficient learners feel less anxiety (rather than less proficient learners feeling more anxiety). Methodologically, the study demonstrates the value of piecewise regression for testing theoretically motivated discontinuities in otherwise linear emotion–proficiency relationships.