Research on intimate partner violence (IPV) has led to the development of more than a dozen specialized risk assessment instruments. The present study evaluates the Danger Assessment (DA; Campbell, 1986; Campbell et al., 2003), which was designed to assess risk for lethal IPV based on victim information. We completed DA ratings for 100 male offenders convicted of IPV related crimes. Follow-up data on the presence of lethal or near-lethal IPV were obtained on average 5.19 years after assessment. Findings indicated that DA ratings varied as a function of the sources of information relied upon; ratings were lower when based on victim reports alone, compared with victim reports and other information. The DA also appeared to yield unrealistically high estimates of risk for lethal IPV. Although more than half of the sample achieved the highest DA ratings, none of the offenders was convicted of lethal or near-lethal IPV during the follow-up period. Findings call into question the utility of the DA.