Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)

Graduate Group

Epidemiology & Biostatistics

First Advisor

Jeane Ann Grisso


Witnessing domestic violence during childhood has been associated with various negative health outcomes. In this work, we conducted a series of analyses to further explore the effects of witnessing that persist into adolescence, adulthood, and, taking a novel approach, to the next generation of children.

First, we compared witnessing status and subsequent experiences with different types (physical, sexual, emotional, or multiple types) of adolescent relationship violence by comparing standardized marginal effects from multiply-imputed data. Witnessing was associated with all forms of violence, and additive effect modification indicated that female witnesses were more likely than male witnesses to experience victimization-related outcomes. In contrast, male witnesses were more likely than female witnesses to experience perpetration-related outcomes, except for physical perpetration.

Second, using standardized multinomial regression, we compared whether witnessing same-gender, opposite-gender, or bidirectional domestic violence perpetration was associated with different adolescent relationship violence outcomes for boys and girls. Compared to non-witnesses, boys who witnessed adult males perpetrate had increased adolescent perpetration. In contrast, girls had increased adolescent victimization if they witnessed males and females perpetrating together and had a tendency toward higher victimization when witnessing males perpetrate alone. Notably, boys and girls who witnessed females perpetrate, alone or together with a male, had increased risk for combined victimization/perpetration compared to non-witnesses.

Finally, perhaps the greatest challenge in this field is having to rely on observational studies, particularly because witnessing co-occurs with many confounding experiences. To overcome this limitation, we used propensity score weighting and applied a new approach to understand the impact that intergenerational violence has on health. Comparing the effects of witnessing domestic violence on first- and second-generation health outcomes using parent-child pairs from a population-based study, we found no effect of witnessing on general health of adults who witnessed violence during childhood. However, children whose parents witnessed domestic violence had worse health compared to children with non-witnessing parents.

This work further supports the theory of intergenerational violence transmission and provides a springboard for future studies by offering a novel approach to studying multi-generational effects of witnessing and promoting more rigorous methods to remove the effects of commonly confounding exposures.

Available for download on Thursday, July 23, 2020