Models were tested separately for mothers and fathers. The baseline structural model for mothers included dummy-coded variables for group status (father alcoholic vs. not), marital satisfaction, warmth/sensitivity during parent–infant interactions and depression. In order to examine the longitudinal associations, marital satisfaction and parental warmth/sensitivity were included at each of the three time points (12, 24, and 36 months) to control for concurrent associations among the variables. Given the temporal stability of paternal alcoholism and depression over the three time points these variables were included in the analyses only at Time 1. The baseline model included covariances between exogenous predictors (alcohol diagnosis, marital satisfaction, warm/sensitive parenting and depression). Direct paths were also included from alcohol group status, marital satisfaction and depression at 12 months to marital satisfaction and warmth/sensitivity at 24 months and paths from marital satisfaction and warmth/sensitivityat 24 months to warmth/sensitivityat 36 months. A direct path from depression at 12 months to parenting at 36 months was also included (see Fig. 1).
