Deaf intermarriage has limited effect on the prevalence of recessive deafness and no effect on underlying allelic frequency.

The idea that deaf intermarriage increases the prevalence of deafness was forcefully pushed in the late 19th century by Alexander Graham Bell, in proceedings published by the National Academy of Science. Bell's hypothesis was not supported by a 19th century study by Edward Allen Fay, which was funded by Bell's own organization, the Volta Bureau. The Fay study showed through an analysis of 4,471 deaf marriages that the chances of having deaf children did not increase significantly when both parents were deaf. In light of an apparent increase in non-complementary pairings when a modern dataset of Gallaudet alumni was compared with the 19th century Fay dataset, Bell's argument has been resurrected. This hypothesis is that residential schools for the deaf, which concentrate signing deaf individuals together, have promoted assortative mating, which in turn has increased the prevalence of recessive deafness and also the commonest underlying deafness allele. Because this hypothesis persists, even though it contradicts classical models of assortative mating, it is critically important that it be thoroughly investigated. In this study, we used an established forward-time genetics simulator with parameters and measurements collected from the published literature. Compared to mathematical equations, simulations allowed for more complex modeling, operated without assumptions of parametricity, and captured ending distributions and variances. Our simulation results affirm predictions from classical equations and show that intense assortative mating only modestly increases the prevalence of deafness, with this effect mostly completed by the third generation. More importantly, our data show that even intense assortative mating does not affect the frequency of the underlying alleles under reported conditions. These results are not locus-specific and are generalizable to other forms of recessive deafness. We explain the higher rate of non-complementary pairings measured in the contemporary Gallaudet alumni sample as compared to the Fay dataset.