Until recently, domestic violence was considered a social problem that primarily affected adults. In recent years, however, increased attention has been given to how domestic violence adversely impacts children. Findings that children's exposure to domestic violence can adversely impact their emotional, behavioral, and physical health, combined with changing ideas about what constitutes a good childhood, have led some family advocates and policymakers to conclude that child exposure to domestic violence (CEDV) is a form of child maltreatment warranting public intervention. However, little is known about how child welfare agencies and their workers respond to allegations of CEDV, or if these agencies construe CEDV as a type of child maltreatment.

While past studies have documented the co-occurrence of domestic violence and child maltreatment among child welfare involved families, this dissertation considers the possibility that CEDV itself might prompt an initial child welfare investigation and may, at times, be construed as a type of child maltreatment. Through analysis of case records and administrative data, this dissertation examines one child welfare agency's response to investigated referrals that allege domestic violence.

A random sample (n=322) of case records for all households referred to and investigated by the agency between July 1, 2011 and June 30, 2012 were reviewed for indications of active and historic domestic violence. Data generated from this review were linked with administrative child welfare data to examine: 1) the prevalence of identified historic and active domestic violence among households referred to and investigated by a public child welfare agency; 2) the demographic and categorical differences between households experiencing active domestic violence at the time of investigation, and those households not experiencing active domestic violence at the time of investigation; 3) how workers in this public child welfare agency construe active domestic violence as causing harm or risk of harm to children in their investigative documents; and 4) how these constructions shape service decisions and impact families.

Examination of these data finds that both active and historic domestic violence are common experiences among households referred to and investigated by this agency. Sixteen percent (n=52) of investigated households in the sample indicated active domestic violence at the time of investigation. Fourteen (n=45) percent indicated prior domestic violence at the time of investigation, but no domestic violence in the 12 months preceding it. Significant demographic and categorical differences were found between those households that were experiencing active domestic violence (active households) at the time of the investigation, and those households that were not (non-active households). On average, caregivers and children in active households were much younger than caregivers and children in non-active households. Approximately one-fourth (n=14) of active households had a child under the age of one year residing in the home, suggesting that the children in active households were more vulnerable than children in non-active households. While active households were more likely to have their referrals promoted to case status than were non-active households, active domestic violence itself was not associated with promotion.

Closer examination of the referrals for active households finds that 60% (n=31), or 10% of households in the sample, were referred to the Agency for domestic violence alone, that is, domestic violence was the reason for referral. How the agency responded to these households offers the clearest picture of how the agency responded to allegations of CEDV. Among households that were referred to the Agency for domestic violence, 94% (n=29) were assigned an allegation of emotional abuse at the time of referral; for more than half of these referrals, emotional abuse was the most severe form of maltreatment assigned. After controlling for other households risk factors and referral characteristics, households referred to the Agency for domestic violence alone were 34 times more likely to be assigned an allegation of emotional abuse than were households for which domestic violence was not the reason, or not the sole reason, for referral. Lacking a specific CEDV maltreatment type, these data strongly suggest that agency workers frequently used the allegation type, "emotional abuse," as a proxy for CEDV.

While review of these data finds that workers construed domestic violence as a type of maltreatment at the time of referral, data indicates that workers construed less than one-third of domestic violence referrals as a type of maltreatment at the time of disposition, that is, the allegation was not substantiated. Examination of investigation outcomes finds that of the 31 active households referred to the agency for domestic violence alone, seven (22.6%) had their referrals promoted to case status. Of these referrals, only one, or 3% of the sample, resulted in a child being removed from their home. Instead, most households that were promoted to case status received in-home or family maintenance services.

While a relatively small proportion of households that were referred to the agency for domestic violence alone had their referrals promoted to case status, many of these households (81%) did receive, or were referred to, child welfare or community-based domestic violence services.

Overall, this dissertation finds that workers at this agency did construe some allegations of domestic violence as a type of maltreatment at the time of referral, but fewer domestic violence allegations were construed as a type of maltreatment at the time of disposition. In addition, this study finds that while the agency did investigate households for alleged domestic violence, most investigations did not result in a child being placed outside the home. Instead, most households were referred to community-based domestic violence organizations and some received in-home family maintenance services.