This study investigates the effect of increasing sampling resolution on our ability to detect the effects of large-scale processes on marine benthic communities. The sampling resolutions used are common to many marine benthic studies (i.e., in order of increasing resolution: video transects; grab sampling; small-core sampling). We investigate whether: (1) loss of information about the finer scale inhibits our ability to detect relationships between large-scale (e.g., environmental) processes and communities at coarser resolutions; or (2) information collected at the finer scale is noisy and obscures detection of large-scale relationships; or (3) the same information is available from all 3 resolutions. We detected fewest relationships between environmental variables and communities at our coarsest resolution; this has important implications to study designs utilising video methods. The results of this study suggest that finding an 'appropriate' scale of sampling for impact assessment may not be a question that should be concerning us. Rather, multiscale modelling and sampling may provide us with an increased ability to detect and interpret large-scale relationships.