Referential and Temporal Cohesion Dataset: 10 ChatGPT + 10 Grok Narrative Texts (500 Words) with Coh-Metrix Outputs and SPSS Analysis Files

This dataset contains 20 original 500 word narrative texts generated by ChatGPT (GTP-5 free tier) and Grok (Grok-4 free tier) using identical zero-shot prompts. the texts were generated to investigate referential and temporal cohesion differences between the two large language models. Also included are: raw Coh-Metrix 3.0 output tables (Referential indices CRFNO1-CRFCWOa and temporal indices CNCTemp, SMTEMP), cleaned .txt files and SPSS 2026 dataset.