Table 4 Results of various statistical tests examining the convergence between laypeople and linguists, based on a re-analysis of human data from the ME task in Sprouse et al. (2013).
From: Grammaticality representation in ChatGPT as compared to linguists and laypeople
One-tailed | Two-tailed | LME | Bayes factor | |
|---|---|---|---|---|
Significant in the opposite direction | - | 2 | 2 | 2 |
Marginal in the opposite direction | - | 0 | 0 | 0 |
Non-significant in the opposite direction | - | 0 | 1 | 0 |
Non-significant in the predicted direction | 11 | 10 | 16 | 14 |
Marginal in the predicted direction | 1 | 4 | 2 | 2 |
Significant in the predicted direction | 136 | 132 | 127 | 130 |