Understanding animals' aversion is important to improving their welfare. Aversion is often assessed using an approach-avoidance (AA) test in which animals have to forfeit a reward if they want to avoid an event or environment presented in the same place. However, sometimes the event/environment suspected to be aversive may physically impair the animal's ability to withdraw from that place (i.e. its ability to express aversion), leading to incorrect interpretations. Combining AA with a Conditioned-Stimulus that predicts the event/environment may overcome this problem by allowing animals to demonstrate aversion without exposure to the stimulus. We aimed to validate this paradigm for testing aversion in chickens. Seven Hyline-Brown chickens were trained to obtain a food reward from a coloured bowl located in the test chamber (TC) of a two-chambered box; the reward was presented in a green bowl with an inactivated air canister or a red bowl with the canister activated to deliver an air puff. Two 5-minute tests were conducted, one with each bowl colour and both with the canister inactivated. All chickens entered TC with the green bowl. With the red bowl, two chickens entered on their first attempt, one fully entered after a partial entry (3/7 fully entered), two made only partial entries and two made no attempts to enter. Chickens spent less time in the TC with the red bowl (median 31s, IQR 7-252) compared to the green bowl (293s, IQR 290-294; p = 0.008). The higher ratio of partial to full entries, failure to enter the TC and less time spent in TC reflected chickens' aversion to the air puff, signalled by the red bowl. The paradigm allowed chickens to demonstrate aversion without exposure to the aversive stimulus during testing.