Our understanding of the effect of scale on runoff and sediment transfers within catchments is currently limited by a lack of available data. A multi-scale dataset of 17 rainfall events collected simultaneously at four spatial scales within a small agricultural catchment in 2005–2006 is presented. Analysis using exploratory techniques and a two-step, zero-inflated lognormal mixed-effects regression model, has demonstrated that event responses, and event response characteristics representing runoff and sediment peaks and area-normalized yields, are scale dependent, and hence cannot be transferred directly between scales. Runoff and sediment yields increase as scale increases, and it is proposed that this effect, which differs from that observed in the few other studies of scale effects undertaken, is due to increasing connectivity within the catchment, and the dominance of preferential flow pathways including through macropores and field drains. The processes contributing to scale dependence in the data, and the possibility that certain processes dominate at particular scales, are discussed. The data presented here help to improve our spatial understanding of runoff and sediment transport in small agricultural catchments, and provide examples of the type of spatial dataset and the type of analysis that are essential if we are to develop models which are able to predict runoff and soil erosion accurately, and allow us to manage runoff and sediment transport effectively across scales.