PMA2020 uses a two-stage cluster design with typically urban-rural and major regions as the strata. A representative sample of enumeration areas (EAs) are drawn from a master sampling frame covered, usually provided by the national statistical agency in each country. Ahead of data collection, households, and key landmarks in each EA are listed and mapped by resident data collectors. Within each EA, a random sample of households are selected. The survey aims to include a sample size that would allow analysts to caluclate a national estimate for all indicators, including calculating the modern contraceptive prevalence rate (mCPR) with a margin of error of ±3 percentage points. The target sample assumes an expected number of eligible women per household and accounts for non-response rates. All resident eligible females are contacted and consented for interviews. Up to three private SDPs within the EA are also selected for interviews along with the public health posts, district hospitals and regionl hospitals serving the EA.
Data collection is conducted bi-annually in the first two years in each country and annually thereafter.
SURVEY PROTOCOL DOCUMENTS
PMA2020 General Sampling Strategy Memo >> A summary of PMA2020's general survey sampling strategy including survey sampling assumptions and selection. Available in English and French
PMA2020 Service Delivery Point Sampling Memo >> A summary of PMA2020's general survey sampling strategy including survey sampling assumptions and selection. Available in English and French
PMA2020/Nigeria Sampling Memo >> Summarizes the overall survey design and sample size calculation method of last section provides methods regarding post-stratification weights to calculate national-level estimates, unique for Nigeria PMA2020. Available in English
PMA2020 Segmentation and Supplementation Memo >> General guidelines on when segmentation and supplementation should be used. Available in English
PMA2020 Weights Memo >> A summary of the creation of PMA2020 household and female weights. Available in English