Threatening stimuli have been shown to preferentially capture attention using a range of tasks and measures. However, attentional bias to threat has not typically been found in unselected individuals using behavioral measures in the dot-probe task, one of the most common ways of examining attention to threat. The present study leveraged event-related potentials (ERPs) in conjunction with behavioral measures in the dot-probe task to examine whether more direct measures of attention might reveal an attentional bias to threat in unselected individuals. As in previous dot-probe studies, we found no evidence of an attentional bias to threat using reaction time; additionally, this measure exhibited poor internal reliability. In contrast, ERPs revealed an initial shift of attention to threat-related stimuli, reflected by the N2pc, which showed moderate internal reliability. However, there was no evidence of sustained engagement with the threat-related stimuli, as measured by the late positive potential (LPP). Together, these results demonstrate that unselected individuals do initially allocate attention to threat in the dot-probe task, and further, that this bias is better characterized by neural measures of attention than traditional behavioral measures. These results have implications for the study of attention to threat in both unselected and anxious populations.