Soil Quality Measurement Imputation Regressions (n = 234)
| Variable | Carbon | Nitrogen | Potassium | pH | Clay | Silt | Sand |
|---|---|---|---|---|---|---|---|
| Principal component 1 | -21.866 | -1.61 | -0.809 | 77.237 | -17.844 | 0.411 | -62.696 |
| (3.630) | (0.308) | (0.708) | (27.731) | (34.451) | (1.809) | (40.802) | |
| Principal component 2 | 34.244 | 3.2 | -0.058 | -147.601 | 92.245 | 12.092 | 55.525 |
| (6.167) | (0.523) | (1.218) | (47.572) | (59.100) | (3.103) | (69.996) | |
| Number of soil samples | 234 | 234 | 234 | 234 | 234 | 234 | 234 |
| Village fixed effects | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Adjusted R2 | 0.88 | 0.89 | 0.57 | 0.94 | 0.90 | 0.99 | 0.37 |
Source: Adapted from Barrett, Bellemare, and Hou 2010.
Note: Standard errors are in parentheses. The results of these regressions are used to impute the values of the dependent variables (i.e., carbon, nitrogen, potassium, pH, clay, silt, and sand) for the entire sample of 473 plots. Each column regresses a soil quality measurement obtained by wet chemistry on the principal components obtained from spectral analysis. The estimated coefficients are then used to predict each dependent variable for the whole sample.