Abstract: Sample-selection bias (SSB) arises when the process that selects observations into a sample involves a correlation with the question of interest. SSB affects many quantitative historical sources and has analogues in other forms of historical inquiry. Three extensive literatures in quantitative history raise important issues of SSB: the anthropometric literature that uses height as a proxy for material welfare; family-reconstitution studies that derive demographic rates from parish registers; and the more recent "linking" literature that creates longitudinal data sets from cross-section sources such as censuses. The linking literature, unlike the first two, has yet to come to grips with SSB.