How are respondents selected in PAPI?
PAPI classifies provinces of Vietnam into three groups: large (Ha Noi and Ho Chi Minh City) with population over 5 millions, medium (Thanh Hoa, Nghe An, An Giang and Tien Giang) with population between 2 and 5 millions, and small (57 remaining provinces) with population under 2 millions
Selection of Districts: To ensure comparability across all provinces, the capital district of each province is forced into the selection. Additionally, two other districts (5 in medium and large provinces) were selected based on the PPS method (proportion to measures of size).
Selection of Communes: In each of these selected districts, the commune being district’s administrative seat was purposely included in the sample, along with another commune selected using PPS.
Selection of Villages: At the village level, the village being commune’s seat was purposely selected, together with another village identified using PPS. This sampling design ensured a full range of possible geographical and socio-economic situations, from dense urban residential areas in the proximity of provincial institutions to remote villages. Thank to the use of PPS, units with significantly different sizes had the same probability of being included into the sample.
Selection of Respondents: In each of the identified villages, 20 citizens of age between 18 and 70, with another 10 for replacement (30 and 15 respectively for large provinces) were randomly selected with the target to reach a quota of 16 respondents (24 for large provinces). Again, random selection ensured that each person in the given age bracket had the same probability of being selected into the survey.
For more detailed information see PAPI Methodology