|  |  | Multiple regression |
---|
 | Simple regression | Stepwise regression | Estimate | z-value | Residual deviation | P(>|χ|) |
---|
Null | 853.06 | (0) 855.07 | 2.535 | 25.310 | 853.07 | - |
Met
| 852.93 | (-) | -0.052 | -0.692 | 852.94 | - |
Cex
| 852.35 | (-) | 0.05 | 0.526 | 852.22 | - |
Igd
| 833.09 | (1) 837.09 | -0.312 | -4.172 | 832.48 | <0.0001 |
Let
| 853.02 | (-) | 0.044 | 0.453 | 832.33 | - |
Rec
| 850.4 | (-) | -0.084 | -0.935 | 830.95 | - |
Cre
| 845.2 | (2) 835.1 | -0.168 | -2.08 | 827.16 | <0.05 |
Pro
| 852.7 | (-) | -0.092 | -0.995 | 826.22 | - |
- The first column lists the seven predictors contributing to the generalized models and the corresponding null model. The second column shows residual deviance (equivalent to the residual sum of squares in ordinary regression analyses) of a model with a single determinant. The third column describes a stepwise forward regression according to the Akaike criterion with insertion order in parenthesis. The last four columns list the results of a multiple regression model (estimates and z-values) and the corresponding Anova with terms added sequentially from met to pro (residual and χ2 test).