The perception of temporal order can help infer the causal structure of the world. By investigating the perceptual signatures of audiovisual temporal order in rats, we demonstrate the importance of the protocol design for reliable order processing. Rats trained with both reinforced audiovisual trials and non-reinforced unisensory trials (two consecutive tones or flashes) learned the task surprisingly faster than rats trained with reinforced multisensory trials only. They also displayed signatures of temporal order perception, such as individual biases and sequential effects that are well described in humans, and impaired in clinical populations. We conclude that an experimental protocol requiring individuals to process all stimuli in a sequence is compulsory to ensure temporal order processing.