User Guide


IPUMS-PMA is the harmonized version of the multinational survey Performance Monitoring and Accountability 2020 (PMA2020). IPUMS-PMA enables researchers to browse the contents of the survey series and craft customized microdata files they download for analysis.

How to access the data

Register an Account

In order to download data, users must register for a free account. Apply for access here. Your application will be reviewed within two working days.

Upon approval, you will be granted access to download data for the specific countries you indicated in your application. You may subsequently request access to additional countries, if your research question has changed to require the expanded geographic scope. Apply for access to additional countries' data here.

Browse the contents of the PMA survey

IPUMS is an integrated data discovery and extraction system. Video tutorials describe how to use the dissemination system.

Design a customized data file

Specify a dataset containing only the samples and variables you need for your analysis. IPUMS-PMA microdata is accessed through our online data extract system.

How to use the data

Published figures using PMA2020 data

For examples of key indicators that can be calculated using PMA2020 data, see the briefs and other reports produced by PMA2020 for each survey round. They can be found, organized by topic, at PMA2020's website.

Understanding the data

Documentation regarding sample design and methodology is available on the source documents page.

IPUMS-PMA uses consistent codes to specify different types of missing data: NIU (not in universe), missing, and other non-valid response codes. The following is an example of codes for a 4-digit variable.

Logical edit - missing
Not interviewed (female questionnaire)
Not interviewed (household questionnaire)
Don't know
No response or missing
NIU (not in universe)

If a sampled respondent for the household questionnaire or an eligible respondent for the female questionnaire for whatever reason did not take the survey, they are included in the data file but their responses are all missing. IPUMS assigns these cases distinct codes (*95 or *96) to distinguish these cases from other kinds of non-responses. These cases are identified in the variables CONSENTHQ or CONSENTFQ. The data extract system by default will filter out these non-interviewed cases, but users can opt to include them.

Calculating representative estimates (weights)

For information about weights and weighting procedures, see our weighting guide, and the FAQ discussion.

Other questions?

Refer to the FAQ or to other user resources at the help page.