ADHD is a heritable condition of childhood for which several risk alleles have been identified. However, observed effect sizes have been small and few replicated polymorphisms have been identified. There are many reasons for the lack of one-to-one correspondence between genotype and phenotype in ADHD. Endophenotypes are non-clinical markers of genetic risk which may facilitate gene discovery in complex disorders like ADHD. The most common endophenotypes under consideration in ADHD are neuropsychological measures of executive function, although a range of psychological, physiological and neuroanatomical endophenotypes have been proposed. If carefully chosen, endophenotypes have the potential to increase the power of genetic research to identify susceptibility genes. If not carefully selected, endophenotypes may generate false negative and false positive results. This paper reviews the theoretical rationale for endophenotypes and proposes a priori criteria by which ADHD endophenotypes should be selected and validated. The literature on motor response inhibition is reviewed to illustrate the validation process which is recommended in the selection of other candidate endophenotypes.