Statistics Random Sampling
Survey Methodology (Statistics Random Sampling)
To better understand public opinion on citizen’s preferences, PACE surveys citizens of Myanmar who were 18 and above at the time of the survey every year. To capture the opinions across Myanmar, PACE conducts the survey in all states and regions. All surveys are conducted according to internationally recognized methods of random statistical sampling as detailed below.
Step 1: Stratification by township
Using data from the population projection by Myanmar Population and Housing Census, PACE calculates the proportion of adult population in each township and allocates the same proportion of survey locations in that township.
Step 2: Stratification by urban and rural
Using population information described above, PACE calculates the proportion of urban and rural population within each township. Based on the proportion within each township, PACE allocates the same proportion of survey locations between urban wards and rural villages.
Step 3: Random sample of villages and wards
Based on allocations for each township and allocations for urban and rural locations, PACE selects wards and villages using simple random sampling. PACE uses a list of wards and villages in each township compiled by the Myanmar Information Management Unit (MIMU) as a sampling frame.
Survey Methodology (Statistics Random Sampling)
Step 4: Random household selection
Trained enumerators travel to survey locations where they randomly selected households using a random walk sampling method beginning in a randomly selected starting point. Enumerators select every 10th residence in rural locations (villages) and every 20th residence in urban locations (wards).
Step 5: Random respondent selection
Once a household has been selected, PACE enumerators randomly select a resident (male and female alternatively) of that household who is over 18 and a citizen of Myanmar. Respondents are selected using the “lucky draw” method. In total, each PACE enumerator is tasked to interview the required number of respondents in each village/ ward location.
See the random respondent selection instruction video below. (The video is in Burmese language.)
Survey Methodology (Statistics Random Sampling)
Step 6: Analysis
Following data collection, the data is weighted by non-response in state/region and rural/urban to bring the realized sample in line with the actual distribution in Myanmar. There may be slight variation between numbers presented due to rounding where the difference is never greater than one percent.