The sample for Options 1 and 2 of the NCD Mobile Phone Survey should be selected using a multi-phase, stratified design to ensure the respondent sample closely matches relevant demographics of the entire target population. In summary, this sample selection process should proceed as follows.
The first phases of the design involve randomly selecting MPNs. This selection process is highly dependent on how the MPN assignments and recordkeeping are handled in the particular country that is fielding the survey.
The selected sample for each of the age and sex strata defined for sampling in the second phase of the design for Options 1 and 2 should be allocated so that the final respondent sample is proportional to the age-by-sex distribution of the country’s target population. The age-by-sex distribution of the country’s target population can be constructed from a recent census, administrative records, or a recent large and reputable demographic survey and, in general, should equal the total eligible population, 18 years of age and older. If NCD Mobile Phone Survey target population counts are not available, corresponding total population counts can be used as stratum-size measures.
MPNs should continue to be collected until all strata sample sizes are achieved or until the data collection has expired.