Table 4—
Cost Comparison: Full Network versus ARD Surveys
| Traditional network survey | ARD survey | |
|---|---|---|
| Panel A. Assumptions | ||
| Project duration (months) | 8.2 | 3.2 |
| Number of villages | 120 | 120 |
| Census sampling rate (percent) | 100 | 100 |
| Fully enumerated census | Yes | No |
| Network/ARD survey sampling rate (percent) | 100 | 30 |
| Traditional network survey | ARD survey | |||
|---|---|---|---|---|
| Total cost($) | Per village cost($) | Total cost($) | Per village cost($) | |
| Panel B. Costs | ||||
| Variable | ||||
| Census | 29,904 | 249 | 12,816 | 107 |
| Networks survey | 84,954 | 708 | 4,486 | 37 |
| Data entry and matching | 14,284 | 119 | — | 0 |
| Tablet rentals | 8,584 | 72 | 1,026 | 9 |
| Fixed | ||||
| Project staff salaries | 20,185 | 168 | 7,959 | 66 |
| Travel | 1,617 | 13 | 638 | 5 |
| J-PAL training/staff meetings | 1,916 | 16 | 1,886 | 16 |
| Office expenses | 3,047 | 25 | 1,201 | 10 |
| OH | ||||
| J-PAL IFMR OH (15 percent) | 24,674 | 206 | 4,502 | 38 |
| Total cost | 189,164 | 1,576 | 34,512 | 288 |
Note: This cost comparison was prepared by J-PAL South Asia, the organization that implemented the network surveys for Banerjee et al. (2013) in Karnataka, India.