Environmental Engineering Reference
In-Depth Information
train research assistants, calculate sample sizes and develop a wealth ranking
system, as well as being a pilot for the survey in a more conventional sense.
●
Households were chosen for the main study by systematic sampling of every
fifth household along footpaths.
●
Data were collected with the help of local research assistants, who were trained
during the pilot study.
●
The study period of 16 months was chosen to ensure that months of agricul-
tural scarcity and plenty were equally represented.
The
data collection methods
used were:
●
For assessing wild food consumption: 24-h recall by the person in each house-
hold who prepared the food. The respondent was asked to give a detailed
account of all food and drinks consumed by the household during the previous
24 h, and where they came from (bought, collected from the wild, gifts). The
questions asked were simple, and the recall period was as short as possible.
Because the households were revisited numerous times, random variation in
consumption patterns on particular days was evened out.
●
Participatory assessment of wealth: A group of four key informants visited
each household with the first author during the pilot phase. While walking
between houses, they informally discussed the wealth characteristics of the
house they had just visited. Subsequently each informant individually placed
the households into groups of similar wealth. The informants then met and
discussed their wealth groupings. Any discrepancies between the informants
in where they placed households were discussed, and consensus was reached.
Then a list of attributes of each wealth group was drawn up. This list was used
in the main study as a template for assigning households to wealth groups.
●
Quantitative assessment of wealth: One of the strengths of this study is that
they assessed wealth (a key variable for their research questions) in two differ-
ent ways, and then were able to cross-validate between their research methods.
This is particularly important because the participatory wealth groupings
were not based on the main sample, but instead were developed using the pilot
sample. For the quantitative assessment they collected data for each house-
hold on key measures that the participatory assessment highlighted as impor-
tant wealth indicators (field size, disposable income, non-monetary income,
expenditure, capital assets, food reserves). Direct observation was used for
field size and capital assets, while income and expenditure came from the
recall survey. Capital assets were measured using a formal questionnaire
survey on presence of items such as a bicycle, radio, shotgun.
The
data analysis methods
used were:
●
A
cluster analysis
was used to check that the quantitative and qualitative
wealth rankings were consistent. The authors used the four continuous quan-
titative wealth measures (incomes, field size, expenditure, assets) as the basis
for their cluster analysis. The method of k-means clustering allows you to
