The objective of this review is to clarify the role of matching in family genetic studies of autism as a way of defining endophenotypes for linkage analysis. The concept of a confounding variable is reviewed and the importance of considering these in family studies of three endophenotypes in autism are considered: cognitive/language impairments, psychiatric disorders, and autistic-like traits. The importance of matching in infant sibling studies of autism is also addressed. Matching as a way of dealing with confounding variables has an important impact on understanding the extent to which these phenotypes are associated with the genes that confer susceptibility to autism and to the early detection of the disorder. Matching continues to be an important issue in the planning and conduct of family-genetic studies of the autism spectrum disorders, particularly as the search for autism susceptibility genes enters the next generation of studies.