Previous eye movement studies of attentional bias in spider fear reported inconsistent results with respect to early attentional capture, suggesting that overt attentional capture only reliably occurs under specific circumstances. In addition, none of these studies explored covert attention. The present study examined attentional bias in spider phobia using a change detection paradigm that was expected to provide good conditions for documenting attentional capture. In contrast to our expectations, eye movement data showed that all participants' first fixations were fastest on general negative targets, whereas participants' first fixations on spider targets were slower in the spider fearful than in the nonfearful group. In addition, spider fearful participants made more nontarget fixations before fixating on a spider target than did nonfearful participants. Thus, we found that participants' overt attention was more quickly focused on general negative targets, whereas covert attentional processes enabled initial avoidance of fear-relevant (i.e. spider) stimuli. The present findings have important implications for research on attention and fear as they indicate that fearful individuals are not characterized by static attentional orienting toward threat but, under certain conditions, may avert attention from threat automatically.