Another potential limitation is the representativeness of the sample. Selection of families through a proband pair of alcohol dependent women, one of whom was in treatment at the time of identification, by study design, results in a sample that may not be representative of families of alcohol dependent individuals in the general population. However, the strength of this approach is that a significant number of offspring have the outcome variable of interest to relate to child/adolescent precursors. Population-based samples provide the distinct advantage of drawing on large and often representative samples. However, the base rate of alcohol dependence and other substance use disorders especially among mothers is often reported to be quite low limiting the number of variables that can be assessed at one time.
