Surname Sampling

Abstract

In 1984, Shin and Yu proposed that sampling Koreans by simply identifying those with the common surname Kim would yield a representative sample, as determined by geographic distribution. We extend the evidence that individuals with specific common surnames in Korea are representative of the whole population. We found that individuals with any of the five most common Korean surnames, not just Kim, were highly representative in Korea in terms of sociodemographic characteristics, attitudes, and health characteristics. In the United States, we found that Kim sampling produces a representative sample of Korean Americans among those who have a Korean surname. While Kim sampling generates a representative sample of Korean American men, it underidentifies Korean American women who grew up in the United States, many of whom do not have Korean surnames, thus potentially biasing samples of women.