Human and AI Judgments on Arabic–English Code-Switching

This dataset contains experimental materials and anonymized response data from a two-alternative forced-choice (2AFC) study investigating determiner–noun switching in Arabic–English code-switching. The study tested predictions from the Matrix Language Frame (MLF) model using 48 test items and 36 fillers, presented under English and Arabic matrix language (ML) conditions. Data include human judgments from 51 Emirati Arabic–English bilinguals (Ex 1) and AI-generated judgments from ChatGPT-4 (Ex 2), along with corresponding stimuli, condition labels, and summary statistics. All participant data are fully anonymized and reported in aggregate.