SLIDE 8 8
A representative sample taking care of geographic
spread, partner diversity, financial weight of interventions, and multitude of components and interventions is a challenge.
Sample to be stratified to ensure adequate
representation of each cateogry.
Sample is expected to provide statistically valid
results at the project level.
Based on final stratificaton, 95% confidence level,
and ±3% error margin, a representative sample of settlements and households to be selected.
Based on simulation, an estimated 4,000 treatment
(and 4,000 control) households will be required.