Table 1. Results of multilevel Poisson regression with nest ID as a random factor.
Variable | Incidence rate ratio (95% C.I.) | chi2 | z | p |
nest base | 5.56 | 0.06 | ||
power pole vs platform | 0.69 (0.51–0.94) | −2.35 | 0.019 | |
wild vs. platform | 0.98 (0.75–1.29) | −0.15 | 0.88 | |
power pole vs. wild | 0.708 (0.484–1.036) | 1.78 | 0.075 | |
distance (km) | 0.920 (0.88–0.96) | −3.60 | <0.001 | |
preci April (mm) | 0.997 (0.995–0.999) | −3.46 | <0.001 | |
preci May (mm) | 0.996 (0.994–0.997) | −5.42 | <0.0001 | |
preci June (mm) | 0.997 (0.995–0.999) | −3.19 | <0.001 |
Rate ratios (95% confidence intervals) are shown with chi2 or z test statistics and p values. An observation–level random effect was included to model possible overdispersion. Variance of i) random effects “nest ID”: 0.076 (95% C.I.: 0.035–0.17), p<0.05 and ii) observation-level random effect: 0 (exact 1.86e-23), n.s. The overall model explained 28.7% of the variance of which the random factor nest ID explained 13.2% and the fixed effects nest base, distance and precipitation in April, May and June 15.5%.