Table 3.
Prediction bias (mean of predictions minus mean of observed scores) as well as Pearson correlation between predictions and observed scores for each model, all on the PROPr scale.
| Model | Prediction bias | Correlation |
|---|---|---|
| Linear model | 0.0176 | 0.932 |
| Bayesian neural network | 0.0175 | 0.933 |
| Floor | −0.231 | 0.930 |
| Ceiling | 0.0672 | 0.930 |
| Zero | 0.0251 | 0.930 |