When an animal has learned a discrimination between two stimuli on the same dimension, exposing the animal to intermediate stimuli will produce a generalisation gradient. Events that that shift responding from this baseline are said to bias responding. Shifts in gradient which show more responding to intermediate stimuli are called positive judgement biases; while shifts in gradient showing reduced responding on intermediate stimuli are called negative judgement biases. While initial studies showed that exposure to poor conditions prior to testing produced negative biases, not all of the literature has supported this. This study aimed to clarify inconsistencies in the literature, using a within-subjects design and a short-term aversive event. Chickens were trained under two-component multiple schedules with each component associated with a stimulus location. Phase 1 involved 15 min exposure to white noise at 100dB, followed immediately by judgement bias testing. Phase 2 involved exposure to white noise at the same dB level, but for the duration of the test. Phase 3 involved judgment bias testing, interrupted by 15 min white noise at 100dB, following which the test resumed. According to early studies, chickens would show a negative bias in Phase 1 and in the second half of Phase 3. However, if release from stressful conditions produces a positive judgement bias, then a negative judgement bias would be expected in Phase 2 only. Results showed negative biases for some birds in Phase 1, and for all birds in Phase 2. No statistically significant bias was observed in Phase 3. The author argues that idiosyncratic differences may have influenced the direction of bias observed in Phase 1, and suggests further investigations be conducted.