Intimate partner violence (IPV) and psychopathology pose risk to families, and parenting has emerged as a key outcome variable and intergenerational conduit for maladaptive paths. Little research, however, has examined maternal communication in IPV-exposed populations, and observational data on this construct are particularly lacking. The current dissertation aims to integrate observational data to assess risks (e.g., IPV subtypes, psychopathology) for mothers' communication in dyadic interactions, using variable-centered and person-centered analyses in two studies. Participants included a community sample of (N = 70) women and their children (N = 112) from a study of IPV in families.Mothers and children participated in dyadic play, and interactions were coded to measure maternal communication. In study 1, observed communication data were used in variable-centered analyses to identify the potential emergence of macro-factors of communication and analyze possible risks for maternal communication. No macro-factors of communication emerged, but results found that a subtype of psychological and physical IPV was inversely related to mothers' praise of children. Additionally, psychopathology did not moderate IPV subtypes. In person-centered analyses, latent profile analysis was implemented in study 2 to explore the emergence of maternal communication profiles, and analyses revealed a five-profile model of maternal communication, including Reserved, Harsh Directive, Warm Directive, Inquisitive Guidance, and Interactive typologies. IPV subtypes and psychopathology were not related to profiles of maternal communication in regression analyses, but models found significant roles of child age, maternal age, and mothers' minority status. Results from the current dissertation underscore the importance of understanding parenting from granular, interactive, and contextual perspectives.