In a straight alley containing a cat odor stimulus rats show high rates of risk assessment, including flat back approach and stretch attend behaviors oriented toward the threat stimulus and contact with the stimulus. In this situation, diazepam (2.0 and 4.0 mg/kg) significantly reduced risk assessment measures (flat back approach + stretch attend), while not reliably altering control locomotion (curved back). In combination with earlier findings that the same doses of diazepam reliably increased risk assessment from a movement arrest baseline, these results strongly support a behavioral model of anxiety involving predictable nonmonotonic changes in risk assessment as a function of anxiety reduction. In comparison to diazepam, ethanol had less pronounced effects on the Cat Odor Test, as in earlier tasks of reactivity to potential threat. However, reliable dose x time interactions for risk assessment measures suggest ethanol effects similar to those of diazepam but most pronounced in initial stages in the test session.