Two experiments tested whether innocent victims threaten observers' belief
in a just world. In both experiments, participants viewed an innocent victi
m then performed a modified Stroop task in which they identified the color
of several words presented for brief exposures (followed by a mask) on a co
mputer screen. When the threat to justice beliefs was presumably highest, c
olor-identification latencies were greater for justice-related words than f
or neutral words. In Experiment 2, under conditions of high threat, justice
-related interference predicted participants' tendency to disassociate them
selves from and derogate the victim. These findings suggest that innocent v
ictims do threaten justice beliefs and responses to these victims may, at t
imes, be attempts to reduce this threat. The methodology presented here may
be applied to future investigations of defensive, counternormative process
es reflecting people's concern with justice.