Sampling Frame

The Korea Health Panel Survey uses 90% of the 2015 Population and Housing Census data as its sampling frame in order to maintain national representativeness. "Resident type," "Gender," and "Home ownership" were used as variables for population analyses, reflecting reality and conducting a survey only of general enumeration districts and apartment districts, with islands, special facilities, and dormitories being excluded.

Sample Stratification

Sample households were chosen through the following processes: 1) Select sample districts (cluster); 2) Choose sample households in enumeration districts, using a probability proportionate and stratified cluster sampling method. In other words, in the order of enumeration district size by region, dong, eup, and myeon, households were arranged before selecting sample districts by using a systematic random? sampling. Then, sample households in sample districts were systematically chosen in the order being listed in the household directories.

Sample Size

With the total number of enumeration districts being 350, the initial study participants included approximately 8,000 households across the nation and their family members residing together. Rationale behind selecting such number of cases was to account for the attrition rates across time and to secure appropriate number of cases for the analysis. The minimum sample size was calculated based a maximum error tolerance of 4.5% at the 95% confidence level, resulting in 16 cities/provinces with 494 households.

Population by City/Province and Korea Health Panel Survey Table