Hello everyone,
We want to conduct a KAP survey of mothers in nutrition in a department that has 51 villages. In these 51 villages, we set up FARN (Learning and Nutritional Rehabilitation Centre). We want to use 2 stage cluster sampling: the first stage selecting the villages and the 2nd stage households.
Should we include all the villages? otherwise how do you make the selection?
We have the population of each village and also that of children from 6 to 59 months. The size of the population differs per village. And so, we used the probability proportional to the size to have the places of the clusters in all the villages.
My concern here is whether we have to use the total population or the population in the age range of 6 to 59 months to calculate the sampling interval?
What is the point of having the sample size if we proceed by households?
Finally, is it possible to limit our investigation just to our FARNs (Learning and Nutritional Rehabilitation Centers) since they are implanted in all villages (51)? I fear a bias because mothers who come to these homes are sensitized daily. This may affect their answers to some questions.
I remain available for more precision.
Dear Zola:
Your questions are quite basic to survey design and require substantial discussion to answer. My first recommendation would be that you consult someone with experience and/or training in survey sampling and methodology to assist you in making the many decisions required as a part of survey planning.
One of the first steps in survey planning is to clarify exactly what questions you are trying to answer and what indicators you will measure with your survey. These objectives and indicators will determine many aspects of survey design. Without knowing this information it is difficult to provide useful answers to questions about methodology.
For example, you ask whether to use the total population or the number of children 6-59 months of age as the measure of village size for PPS sampling in the first stage. The measure of size of the primary samplings units used to carry out PPS sampling should be as close as possible to the unit of analysis for the indicators measured by the survey. Therefore, your question cannot be answered without knowing what data you will collect from whom.
I understand this is an unsatisfying response to your questions, but I hope you find the local expert assistance you need.
Answered:
6 years ago