© 2016 Elsevier B.V. The need to address biodiversity information requirements has led to 'methods research' growing significantly in the field of ecology. Optimizing inference from survey data is dependent upon understanding the extent of sampling bias, which is inevitable when a subsample is taken to represent the entire population. In this study, angling is compared to Baited Remote Underwater Stereo-Video systems (stereo-BRUVs) as fisheries-independent sampling techniques for surveying temperate reef fishes in the Tsitsikamma Marine Protected Area, South Africa. The comparison highlighted a number of differences in the observed ichthyofauna, notably the ability of stereo-BRUVs to sample a wider range of species while also providing a more consistent index of relative abundance. Comparing relative abundance estimates indicated that both sampling methods were, in general, detecting similar environmentally-induced trends. Size distribution data from the two methods differed significantly amongst the predominant species, and angling produced consistently larger length-frequencies than stereo-BRUVs. However, when the hook-based size selectivity of angling was taken into account, there were no significant differences between the two methods. The ability to detect a species-specific population change over a period of five years was assessed using power analyses, and stereo-BRUVs consistently outperformed angling. The results identify stereo-BRUVs as the superior technique for monitoring the population and size-structure of temperate reef fishes.