Scale properties.

Grounded in flow theory and the person-artefact-task (PAT) model, this study examines the impact of flow on EFL learners’ excessive use of translation software and identifies the antecedents of flow. This study also investigates whether students’ academic background (English vs. non-English) moderates the relationship between flow and excessive use. Data from 575 Chinese university students were analyzed using Partial Least Squares (PLS) path modeling, revealing that flow significantly predicts excessive use, especially among non-English major student. In terms of the antecedents, social norms exhibit the strongest predictive power on flow, followed by task-technology fit and perceived translation quality, while task perception does not significantly influence flow. In addition, the explanatory power of the model was significant, evidenced by an R² value of 0.577 for flow and 0.272 for excessive use of translation software. These findings underscore the importance of flow in understanding excessive use behaviors and inform educational strategies that promote balanced technology use.