Indeks Kompleksitas Sintaktis dalam Teks Bahasa Kedua Mahasiswa vs Teks yang Dihasilkan oleh AI = Syntactic Complexity Indices in Students Second Language Writings vs AI-Generated Texts

This study investigates the relationship between syntactic complexity indices and writing proficiency, as well as provides an effective indices-based recommendation for second language learners. This study employs a quantitative approach, employing ANOVA and effect size analysis. The data consist of writings samples from second language users and two types of AI-generated text, basic AI texts, and AI prompted to imitate second language writings and in addition, the syntactic complexity indices score data. Writing samples from second language users is labeled as SLW, they were collected from second semester students of Bacharuddin Jusuf Habibie Institute of Technology. Basic AI texts is labeled as AI and were collected from the students who uses random AI models. AI prompted to imitate second language writings samples is labeled as AI Human Mimic, collected from ChatGPT using SLW samples as a model for imitation. Syntactic complexity indices score data were collected by processing the three types of writing samples into TAASSC. This study findings shows that some indices relationship challenges previous findings (MLS, CT/T, VP/T), aligns with previous findings (MLC, CP/C, CP/T, CN/C, DC/C, T/S), challenges or support indices with multiple previous findings (C/S, DC/T, C/T), and unable to determine some indices relationship (MLT, CN/T). Furthermore, this study found indices with the largest practical significance (MLC, MLS, C/S, C/T, CT/T, DC/C, DC/T, CP/C, CN/C), and with moderate practical significance (MLT, CP/T, T/S, CN/T, VP/T).

Keyword : Syntactic Complexity, Second Language, AI, writing proficiency, TAASSC.