Identify with precision the limits of assigned sub-cell

User Emmanuel has asked the following question in the context of conducting listing of an area determined by grid-based spatial sampling:

Q: How each enumerator can identify with precision the limits of his sub-cell or quadrant?

A: In the grid sampling we sample geographical areas, typically (though not necessarily) from a rectangular grid for convenience aligned with meridians and parallels. So a result of such sampling is an area bound by 4 numbers: N,S,E,W for extreme values (latitude between S and N, and longitude between E and W).

For practical applications, the whole sampled area may not be always covered by a single interviewer, and can be subdivided into sub-cells (quadrants) where each sub-cell may be assigned separately to different interviewers. The supervisor or the person making the assignments should inform the interviewer which sub-cell he/she is responsible for.

For interviewers to visualize the boundaries of assigned areas a shapefile must be generated for the sampled grid cells, and depict all sub-cells of that cell (if it was subdivided).

Here is an illustrative image:

The interviewer should pay attention to his current location shown with the blue pointer-circle, and should visit only the houses located inside the assigned boundary, meaning - no visits should be made to X-marked houses.

In this example, the sampled cell is subdivided into two sub-cells (numbered 1 and 2). The supervisor should inform the interviewer whether he is responsible for the sub-cell 1, sub-cell 2, or both sub-cells.

Assuming that the interviewer is responsible for sub-cell 2, he should be submitting the data from the green-marked houses, and not from the red-marked houses.

In practice the reality is typically not as nicely looking as the illustrative picture above:

  1. The boundaries of the cell/sub-cell may be going through the building, leaving it up to the interviewer to decide whether to include it or not. Here a rule is needed to formalize this decision, such as “The building belongs to the area where the most of the roof at the picture belongs” If such a rule is not formalized, then there is a risk that none of the interviewers (neither 1 nor 2 in the above example) will include such a house (omission) or both will (duplication).
  2. There may be new buildings that are not depicted in the satellite imagery, which still need to be enumerated. Or depicted buildings may no longer exist.
  3. GPS accuracy may cause the current position marker to wobble, and prevent the interviewer from understanding how far or how close to the boundary he or she is, or even whether they are still in the sampled area.
  4. Quality of the imagery (resolution, cloud coverage, coverage with vegetation or snow) may prevent the interviewer from being able to distinguish closely positioned buildings.

Hope this helps.

Best, Sergiy