Knowledge of the effectiveness and selectivity of research sampling techniques is imperative for developing efficient surveys that provide robust data, while minimising costs and impacts on habitats and non-target species. In this study, ten broad survey types (otter trawl, beach seine, benthic sled, bongo net, light trap, BRUV, acoustics, drone, crab trap and squid jigging) with multiple gear configurations were used to sample fishes and invertebrates of various life stages, sizes and habitat preferences (e.g. benthic vs pelagic, sand vs seagrass associated). Sampling occurred over two years (2021–2023) in Cockburn Sound, a sheltered marine embayment on the lower-west coast of Western Australia. The methods differed substantially in their efficiency (e.g. catch per unit of effort and cost/time to survey a given area) and community metrics (e.g. species richness and diversity). These trends were largely attributable to differences in the efficiency of various methods for recording larger and more mobile species (e.g. sparids and carangids) vs smaller and more cryptic species (e.g. syngnathids, monacanthids and gobiids). This was most apparent between ‘non-destructive’ techniques (e.g. BRUV, acoustics) and netting. Within species, efficiency and size selectivity was largely related to gear type and configuration (passive vs active gear, mesh size, hook size). There was also variation in each method’s effectiveness (catch rates and taxonomic resolution) due to environmental and other factors (e.g. water clarity, wind speed, operator ability). Recommendations for future surveys of fisheries species and ecological communities are presented.