Access to maps can be challenging and they are not always readily available. In the occasions that they are available, more often than not they are not usable because they have very few details or are in small scale (if scaled at all). This tutorial will show how a stage 1 spatial sample can be obtained without requiring maps but instead using villages lists.

 

Step 1: Creating a village list

A list of all (or nearly all) the villages in the survey area is essential for this stage 1 sampling approach. Most of the time, these lists are already available through official government bodies (e.g., census / statistics bureau) either at national or local level. UN organisations and/or NGOs may also have these lists for use in their programming. In the rare occasions when such lists are unavailable or are available but outdated and unusable, a village list can be created through a listing of known villages in the survey area verified and triangulated through various sources such as different key informants, villages names in patient records found in the health clinic/s covering the survey area.

An important feature of the villages list is that it should be organised or stratified using the most relevant spatial segregator. This can be based on the administrative structure of the country / area to be surveyed, programme catchment areas, health clinic catchment areas or a combination of these categories (i.e., administrative levels and then clinic catchment areas). It is therefore important to recognise and understand the administrative structure and organisation of the survey area so as to be able apply the appropriate groupings for the villages in the list. Figure 1 gives an example of the administrative structure in Sierra Leone. Table 1 shows an example of a village list from Sierra Leone based on the administrative structure described in Figure 1.

 

Figure 1: Administrative structure of rural districts of Sierra Leone
adminStructure

 

Table 1: List of villages in Bombali district, Sierra Leone organised by chiefdom

District Chiefdom Village
Bombali Biriwa Katadumgbu
Bombali Biriwa Kamathoro
Bombali Biriwa Kamalankan
Bombali Biriwa Kamafutay Gay
Bombali Gbanti Kamaranka Bonkoh
Bombali Gbanti Kamaranka Rothenkeh
Bombali Gbanti Kamaranka Rochain-Yanka
Bombali Gbanti Kamaranka Kamasondo
Bombali Gbanti Kamaranka Gbenkfay
Bombali Gbanti Kamaranka Misra
Bombali Gbanti Kamaranka Royana

 

Step 2: Selecting a systematic sample from the list

Using the list produced from Step 1, a systematic sample is drawn. First, a sampling interval is determined using the following equation:

 
$$ \mbox{sampling interval} ~=~ \left\lfloor\frac{\mbox{Number of villages on list}}{\mbox{Number of villages to sample}}\right\rfloor $$

 
If the total number of villages on the list is 325 and the target sample size is 40, the sampling interval is calculated as:

 
$$ \mbox{sampling interval} ~=~ \left\lfloor\frac{325}{40}\right\rfloor ~=~ \left\lfloor{8.13}\right\rfloor ~\approx~ 8 $$

 
Second, we need to determine a random starting position from the top of the list. This can be done by selecting a random number between 1 and the length of the sampling interval which is 8. Random number selection can be done through a number of different techniques such as lottery method or generating random numbers using a spreadsheet. The random number selected is used to select the first village from the top of the list. Subsequent selections are made by repeated addition of the sampling interval. This process (see Figure 2) ensured that the sample was distributed over the entire survey area.

 

Figure 2: List of villages in Bombali district, Sierra Leone sorted by section and chiefdom showing systematic sampling with random start = 16 and sampling interval = 20
systematicSample

 
 

Tagged on: