Table 2 Regression results for rodent abundance for crude and adjusted multiple regression models (n = 585). Results are shown as regression coefficients for key variables, with standard errors reported in parentheses, and the overall model log likelihood.
Crude poisson | Crude negative binomial | Adjusted negative binomial–random effects | Adjusted negative binomial–random effects + a priori variables | Fully adjusted negative binomial | |
|---|---|---|---|---|---|
Constant | − 1.2 (0.22) | − 1.2 (0.25) | − 1.47 (0.25) | − 1.81 (0.32) | − 2.4 (0.41) |
> 8 holes | 0.58 (0.24)** | 0.58 (0.28)** | 0.45 (0.28) | 0.54 (0.30)* | 0.58 (0.30)* |
Presence of burrows | 0.66 (0.16)*** | 0.66 (0.18) *** | 0.73 (0.18)*** | 0.85 (0.18) *** | 0.84 (0.18)*** |
Presence of food | 0.28 (0.18) | 0.28 (0.21) | 0.32 (0.20)* | 0.31 (0.20) | 0.29 (0.20) |
Building type | − | − | − | 0.57 (0.20)*** | 0.67 (0.21)*** |
Log Likelihood | − 631.9 | − 619.8 | − 612.7 | − 597.7 | − 592.9 |
