Riparian buffers, the vegetated region adjacent to streams and wetlands, are thought to be effective at intercepting and reducing nitrogen loads entering water bodies. Riparian buffer width is thought to be positively related to nitrogen removal effectiveness by influencing nitrogen retention or removal. We surveyed the scientific literature containing data on riparian buffers and nitrogen concentration in streams and groundwater to identify trends between nitrogen removal effectiveness and buffer width, hydrological flow path, and vegetative cover. Nitrogen removal effectiveness varied widely. Wide buffers (>50 m) more consistently removed significant portions of nitrogen entering a riparian zone than narrow buffers (0-25 m). Buffers of various vegetation types were equally effective at removing nitrogen but buffers composed of herbaceous and forest/herbaceous vegetation were more effective when wider. Subsurface removal of nitrogen was efficient, but did not appear to be related to buffer width, while surface removal of nitrogen was partly related to buffer width. The mass of nitrate nitrogen removed per unit length of buffer did not differ by buffer width, flow path, or buffer vegetation type. Our meta-analysis suggests that buffer width is an important consideration in managing nitrogen in watersheds. However, the inconsistent effects of buffer width and vegetation on nitrogen removal suggest that soil type, subsurface hydrology (e.g., soil saturation, groundwater flow paths), and subsurface biogeochemistry (organic carbon supply, nitrate inputs) also are important factors governing nitrogen removal in buffers.