Table 7. Penalized logistic regression prediction results for burnout, exhaustion and disengagement (n = 57).
| Burnout | Exhaustion | Disengagement | |
|---|---|---|---|
| True Positive | 8 | 8 | 7 |
| True Negative | 46 | 45 | 37 |
| False Positive | 3 | 3 | 11 |
| False Negative | 0 | 1 | 2 |
| Recall | 100% | 89% | 78% |
| Precision | 73% | 73% | 39% |
| F1 Score | 84% | 80% | 52% |
AUC statistics are .96 for burnout, .96 for exhaustion and .84 for disengagement. The Hosmer-Lemeshow GOF test is non-significant for all models. The prediction threshold is .3 for burnout and for disengagement and .4 for exhaustion.