Baited underwater video techniques are increasingly being utilised for assessing and monitoring demersal fishes because they are: 1) non extractive, 2) can be used to sample across multiple habitats and depths, 3) are cost effective, 4) sample a broader range of species than many other techniques, 5) and with greater statistical power. However, an examination of the literature demonstrates that a range of different bait types are being used. The use of different types of bait can create an additional source of variability in sampling programs. Coral reef fish assemblages at the Houtman Abrolhos Islands, Western Australia, were sampled using baited remote underwater stereo-video systems. One-hour stereo-video recordings were collected for four different bait treatments (pilchards, cat food, falafel mix and no bait (control)) from sites inside and outside a targeted fishery closure (TFC). In total, 5209 individuals from 132 fish species belonging to 41 families were recorded. There were significant differences in the fish assemblage structure and composition between baited and non-baited treatments (P<0.001), while no difference was observed with species richness. Samples baited with cat food and pilchards contained similar ingredients and were found to record similar components of the fish assemblage. There were no significant differences in the fish assemblages in areas open or closed to fishing, regardless of the bait used. Investigation of five targeted species indicated that the response to different types of bait was species-specific. For example, the relative abundance of Pagrus auratus was found to increase in areas protected from fishing, but only in samples baited with pilchards and cat food. The results indicate that the use of bait in conjunction with stereo-BRUVs is advantageous. On balance, the use of pilchards as a standardised bait for stereo-BRUVs deployments is justified for use along the mid-west coast of Western Australia.