Table 6. AICc multimodel selection results. Akaike weights and coefficients for five logistic regression models predicting CHANNEL distribution. The last two columns on the right provide model weighted average coefficients and relative variable importance measures for each variable (Burnham and Anderson 2002). The inequality model captured the largest Aikaike weight, and best explains the distribution of irrigation channels.
| |
| |
|
|
Inequality |
Individual |
All |
Base |
Village |
Model weighted
average |
Relative variable
importance |
| |
|
|
weight |
0.97 |
0.01 |
0.01 |
0.00 |
0.00 |
|
|
|
AICc |
258.2 |
266.6 |
266.8 |
318.5 |
319.4 |
|
|
|
df |
6 |
8 |
12 |
2 |
6 |
|
|
| |
|
|
LNPOP |
- |
- |
0.13 |
- |
0.01 |
0.00 |
0.01 |
|
DISTANCE |
- |
- |
0.02 |
- |
0.01 |
0.00 |
0.01 |
|
CASTES |
- |
- |
0.01 |
- |
0.09 |
0.00 |
0.01 |
|
WEALTHGINI |
-0.08 |
- |
-0.07 |
- |
-0.03 |
-0.08 |
0.99 |
|
AGE |
- |
0.01 |
0.02 |
- |
- |
0.00 |
0.03 |
|
EDUC |
- |
0.06 |
0.07 |
- |
- |
0.00 |
0.03 |
|
HHSIZE |
- |
-0.07 |
-0.07 |
- |
- |
0.00 |
0.03 |
|
LNWEALTH |
0.53 |
0.54 |
0.52 |
- |
- |
0.53 |
1.00 |
|
DALIT |
-1.50 |
-1.28 |
-1.24 |
- |
- |
-1.49 |
1.00 |
|
FRAC |
1.39 |
0.73 |
1.65 |
- |
- |
1.39 |
1.00 |
|
| |
|