Abstract
Highly significant differences in mean age, blood pressure and phenotype frequency distributions between the non‐migrants and “emigrants” of a total unselected community sample were discovered. Use of the mean of BP scores collected from epidemiologic surveys over a period of time as an individual score allows sample attrition to produce both a genetically and demographically biased sample of a population intended to represent an unselected community of people. Multiple regression analyses estimated the contribution of an individual's age, genotype and mobility out of the sample to predicting blood pressure variation. Variation in blood pressure means among certain marker phenotype classes was greater in those who leave than in those who stay, but only the upper portion of the pressure distribution contributed to this relationship. A genetic‐environment interaction is suggested.