Question 1: Given this information, what would be the overall sample size?

The overall sample size calculated based on the information will be 360,000

The sample size calculated based on;

A: Number of strata (districts) A=100

B: Effective sample size (ESS) based on the previous survey data (95% coverage of ; will be: (with IC95%, Penta3=85%, Precision=5%) B=265

C: Design effect (DEFF) will be : (with ICC=1/3, and m=10) C=4

D: Number of households to find an eligible respondent D=5

E: Inflation factor for non response per cluster (with 15% of non response) E=1.18

Target number of completes responses per cluster m=10

Question #2: How many households would have to be visited?

Using the same informations as above, the total of households to visite would be 621,177

Question #3: What do you think of this estimated sample size?

This sample size is very large. It will lead to obtain results for decisons and operational actions. The coverage of national and province levels can be obtained by aggregation.

But, this sample size can also introduce more difficulties or errors in the survey. It will need more ressources in terms of times, finances, personnel. The challenges will be its affordability (financing++) and the team capacity to conduct a high quality work simultaneous in all strata on short time.

Question #4: How feasible will it be to conduct this survey?

This survey would be feasible in case there is sufficient ressources (budget, timeline, personnel, logistics, informations)

Question #5: What are the trade-off in terms of time, money, and quality of survey implementation?

Compromise will be to respond to the challenges of ressources and quality of the work. The survey team must conciliate the need of granular results and the challenges by making the survey on province level.

For the province level, the sample size, calculed with the same informations provided above, is 10,560 children. It's the tenth of size of the earlier size. And only 62,118 households would be visited. This can reduce more the need of financing and personnel, and contribute to improve the quality of the implementaion and final results.

Perhaps, if the available ressources is superior to the need for province level, the team can combine two levels (province and district) by selecting randomly some districts per province and estimate their coverage. This coverage can be aggregated to obtain the province and national coverage.