To examine whether these measures might account for any of the environmental variance associated with remission, twin status on each measure was regressed on co-twin status on the same measure, twin remission status, and their interaction in same-sex twin pairs, adjusting for gender. A significant interaction would indicate that twin pair similarity on the environmental measure varied as a function of twin remission status. None of the interaction terms were significant. Finally, adjustment of the final genetic model for marriage and for childhood abuse did not substantially change the estimates of genetic and environmental influences on remission, suggesting that these variables did not account for much of the environmental variance associated with remission.
