An assay in which groups perform at both similar and significantly different levels within the same test session may have experimental advantages both in understanding underlying behavioral-cognitive processes and in helping to resolve neuroimaging issues regarding functional significance vs. performance confounds. Here we report behavioral data from a response inhibition (Go/No-Go) task with two levels of No-Go difficulty (easy, hard). The sample included individuals with current cocaine dependence (N=18) and controls (N=15). Using signal detection methodology (d' and Beta), significant main effects were observed for group and trial type on d'. Post-hoc analyses revealed the cocaine-dependent individuals performed significantly worse than controls on difficult, but not easy, trials. Differences on d' but not Beta, and slower reaction times in cocaine subjects, suggest that response inhibition deficits were related to disruption in visual information processing rather than inhibition of motor activity.