Defining target population and sampling

[agoxtinhoParticipant]

Hi to all!

I have a question regarding a very specific situation.

We’re starting a new project and we’re going to use SMART as one of our tool for baseline.

We’ve already defined that we’re going to work in 2 different municipalities (out of 9 in the province) in villages that were targeted with a CLTS program and that are now ODF (Open Defecation Free).

We’re going to cooperate in a 3rd municipality with the Ministry of Health, but our purpose is just to run the SMART survey, not to implement the project within this municipality.

Clearly we’re going to a Cluster sampling.

My questions are:

1. Which is the population we should consider?
a. our population is the population of the 3 municipalities
b.run separate SMART survey for each municipality, considering the population, the whole population of the municipality
a. our population is the population of ODF villages from the 3 municipalities
b.run separate SMART survey for each municipality, considering the population, the population of ODF villages in each municipality

Looking forward for your replies!

Hugo

[SMARTKeymaster]

A) Villages in two municipalities that are now ODF: You have indicated that these villages have been targeted for your project; therefore, the total population of each of these villages should be included in your sampling frame for one SMART survey. In your report it is important to emphasize that the sampling frame used for your baseline survey was the villages in the two municipalities that were ODF and NOT all the villages (population) of the two municipalities.

B) The third municipality should be a second SMART survey. Since your program will not be operating in this municipality there is no need to include it in your baseline survey of the ODF villages in municipalities 1 and 2. If the third municipality contains some villages that are ODF and others that are not, separate surveys should be performed (or only target ODF/ or non ODF villages depending on objectives) because CLTS intervention has a significant impact on nutrition status. Also, the proposed survey for the third municipality should only take place if an action is likely to take place in a reasonable amount of time (if results indicate actions are required), otherwise, the survey would be unnecessary at this time.

[agoxtinhoParticipant]

Hi!

Thanks a lot for your reply!

Another question.

We’re just going collect anthropometric data. Can we do a shortlist of households with children u5 and do a random selection within these households?

Best,

Hugo

[SMARTKeymaster]

Yes, if only anthropometric data from 6-59 month children is being collected, a shortlist of households with 6-59 month children can be used for random selection. However, it is VERY important that the household list is up to date. If the household list is even 6 months old it can be considered out of date. It is recommended to create an under 5 updated household list and begin data collection as soon as possible after the list has been updated.

[agoxtinhoParticipant]

So I’ve created my survey planning on ENA.

Cluster sampling, correction for smalll population size, 30 clusters…and what about the samples»

I go to 680 Households and I measure all children from 6-59 months

I stop my survey when I measure 697 children from 6-59?

🙂

Thank you

Hugo

• This reply was modified 11 years, 1 month ago by  agoxtinho.

[SMARTKeymaster]

You should continue to go to all clusters and households after the minimum sample size (697 children) is reached to maintain a representative sample. The following resources found on the website describe in detail how to calculate sample and sampling methods:

Survey Planning Tools/ Sampling for SMART: Sampling Methods & Sample size calculation for the SMART methodology: https://smartmethodology.org/survey-planning-tools/smart-methodology/

Survey Planning Tools/ SMART Capacity Building Toolbox: Survey Manager Module 3 Sampling: https://smartmethodology.org/survey-planning-tools/smart-capacity-building-toolbox/

[Author]

Posts