The family history method (inquiry only) was directly compared to the gold standard of the family study method (family members interviewed directly) in a large pedigree segregating a stuttering phenotype. Fifty-three individuals from one branch of a large family were asked whether they had a family history of stuttering. Additional data were collected concerning the recognition of stuttering in family members based on proximity of affected family members, affection status of the informant, gender of the informant, and the number of stutterers in the immediate family. Sensitivity was reduced for both sexes and for affected and non-affected cohorts. Affected family members had consistently higher sensitivities than non-affected family members. Gender differences were also found. Non-affected individuals' recognition of their family history was influenced by the proximity of affected family members. When the question of family history was limited to first-degree relatives, specificity was found to be high, whereas sensitivity decreased. This study confirms previous findings that the sensitivity of the family history method is reduced compared to the gold standard of the family study (Andreason et al., 1986; Thompson et al., 1982; Mendlewicz et al., 1975; Orvaschel et al., 1982). This study also replicates recent findings that affected family members are more accurate informants than nonaffected family members (Kendler et al., 1991). In addition, when informants are asked to focus on first-degree relatives only (vs, the family as a whole) the sensitivity actually decreases, suggesting that familiarity may play some role in the informants failure to identify affected family members.
