An association study can be used to investigate how individuals with unique genetic variants respond to a drug treatment. In an association study, individuals may come from different ethnic groups or an admixed population. The heterogeneity of genetic backgrounds among individuals in association studies may lead to false-positive or false-negative results. Confounding caused by population structure and recent admixture may be one major factor that contributes to the lack of replication of association study results. Confounding can be detected and adjusted. Major methods that adjust for population stratification are described and explained in this chapter. Their advantages and disadvantages are discussed.