Table II.
DISCERN and readability values of ChatGPT 3.5 and 4.0 versions and inter-rater reliability analysis.
| ChatGPT 3.5 | ChatGPT 4.0 | |
|---|---|---|
| DISCERN (mean) | 61.6 | 61.6 |
| IRR | 0.94 | 0.96 |
| FRES (points) | 31.0 | 45.6 |
| FKGL (grade) | 12.6 | 9.10 |
| SMOG (grade) | 14.6 | 13.5 |
| Coleman Liau (grade) | 15.0 | 12.9 |
| Gunning-Fox Index (grade) | 16.6 | 15.0 |
| Grade Scale Mean (grade) | 14.7 | 12.6 |
FRES, Flesch-Kincaid Reading Ease Score; FKGL, Flesch-Kincaid Grade Level; SMOG, Simple Measure of Gobbledygook; ChatGPT, Chat Generative Pretrained Transfomer; IRR, inter-rater reliability.