Table 2.
Regression model examining whether bipolar subtype moderates agreement between Bayesian estimates and LEAD ratings (N=562)
| Predictor | B | SE | p | Part r |
|---|---|---|---|---|
| Bayesian Estimate | 25.4 | 4.8 | .000 | .14 |
| LEAD Diagnostic Status | ||||
| BP-I Yes or No | 78.9 | 7.3 | .000 | .29 |
| BP-II Yes or No | 82.6 | 10.7 | .000 | .21 |
| Cyclothymic Disorder Yes or No | 78.1 | 6.0 | .000 | .34 |
| Interaction Terms (Diagnosis * Actuarial Bayesian Estimate) | ||||
| BP-I | -11.4 | 23.2 | .622 | -.01 |
| BP-II | -27.4 | 32.5 | .399 | -.02 |
| Cyclothymic Disorder | -34.8 | 18.5 | .061 | -.05 |
| Constant | 2.2 | 0.9 | .015 |
Note. The significance of the interaction terms was the direct test of the hypothesis; none were significant, p > .05. Overall model R2 = .61, adjusted R2 = .61, p < .0005.