The generality of ecological patterns depends inextricably on the scale at which they are examined. We investigated patterns of crab distribution and the relationship between crabs and vegetation in salt marshes at multiple scales. By using consistent monitoring protocols across 15 U.S. National Estuarine Research Reserves, we were able to synthesize patterns from the scale of quadrats to the entire marsh landscape to regional and national scales. Some generalities emerged across marshes from our overall models, and these are useful for informing broad coastal management policy. We found that crab burrow distribution within a marsh could be predicted by marsh elevation, distance to creek and soil compressibility. While these physical factors also affected marsh vegetation cover, we did not find a strong or consistent overall effect of crabs at a broad scale in our multivariate model, though regressions conducted separately for each site revealed that crab burrows were negatively correlated with vegetation cover at 4 out of 15 sites. This contrasts with recent smaller-scale studies and meta-analyses synthesizing such studies that detected strong negative effects of crabs on marshes, likely because we sampled across the entire marsh landscape, while targeted studies are typically limited to low-lying areas near creeks, where crab burrow densities are highest. Our results suggest that sea-level rise generally poses a bigger threat to marshes than crabs, but there will likely be interactions between these physical and biological factors. Beyond these generalities across marshes, we detected some regional differences in crab community composition, richness, and abundance. However, we found striking differences among sites within regions, and within sites, in terms of crab abundance and relationships to marsh integrity. Although generalities are broadly useful, our findings indicate that local managers cannot rely on data from other nearby systems, but rather need local information for developing salt marsh management strategies.