Table 3.
Comparison of the ten best beta-binomial models of Rift Valley fever serologic incidence measured in small ruminants, Barkedji area, Senegal, 2003 rainy season
| Model | Deviance | Parameter (nb) | AICc | ΔAICc | |
|---|---|---|---|---|---|
|
1 |
VDI_500m |
2.76 |
3 |
25.4 |
0 |
|
2 |
VDI _100m |
3.71 |
3 |
26.3 |
0.95 |
| 3 |
VDI _1 000m |
7.43 |
3 |
30.1 |
4.67 |
| 4 |
VDI _500m + PDI |
1.81 |
4 |
33.8 |
8.39 |
| 5 |
VDI _500m + Parea |
2.41 |
4 |
34.4 |
8.98 |
| 6 |
VDI _500m + Ferlo |
3.09 |
4 |
35.1 |
9.66 |
| 7 |
Ferlo + WVI |
3.29 |
4 |
35.3 |
9.87 |
| 8 |
VDI _100m + PDI |
3.85 |
4 |
35.8 |
10.43 |
| 9 |
VDI _100m + Parea |
4.04 |
4 |
36 |
10.61 |
| 10 | VDI _1 000m + PDI | 4.17 | 4 | 36.1 | 10.75 |
Models are ordered from best to worst among a set of 24 candidate models. The two first models can be considered having substantial support (ΔAICc <2) (bold text).