Replication Data for: Participant Accuracy Data for AI-Assisted Phonological Drills on English Plural Inflections by Mandarin

This dataset contains demographic and performance data from 185 intermediate-to-advanced Mandarin-speaking undergraduate EFL learners participating in a quasi-experimental study on the effects of AI-assisted phonological drills (using the SpeakGuru platform) for overcoming L1 phonological interference in spontaneous production of English plural inflections. Data includes participant ID (experimental or control group), age, gender (1=male, 2=female), age of onset of English learning, years of English learning, English proficiency level (1=intermediate, 2=upper-intermediate, 3=advanced), and accuracy percentages for pre-test, post-test, and one-month delayed post-test (assessed via semi-structured interviews eliciting ~20 obligatory plural contexts). The experimental group (n=95) showed significant gains (pre: ~50%, post: ~75%, delayed: ~73%), while the control group (n=90) showed minimal changes, supporting the Prosodic Transfer Hypothesis and behaviorist principles. Data is in tabular format for statistical analysis (e.g., mixed-effects models). Ethical standards were followed, including informed consent.