Local extinctions are often non-randomly associated with range size, dispersal ability and habitat specificity, as well as body size, sexual dimorphism and phylogeny. We used a large data set of the Orthoptera species (bush crickets, crickets, grasshoppers) occurring in Germany and compared the number of occupied grid cells before 1980 to those occupied after 1980, corrected for monitoring intensity. The number of grid cells in which a species went extinct was non-linearly related to the number of occupied grid cells per species. Using generalized linear modelling we analysed extinction in relation to national distribution (the number of occupied grid cells before 1980), dispersal ability (derived from a large body of literature concerning wing development, colonization dynamics and within-habitat mobility), habitat specificity (moisture specialists versus generalists), potential reproduction (the number of ovarioles), the degree of sexual size dimorphism and phylogeny (twelve clades). Species with a large global range size also had a large national range size. Species with a large range experienced more total extinction events than species with smaller ranges but relatively fewer compared to range size. The latter relationship was largely shaped by the dispersal ability of the species: the interactions of range size×dispersal ability and range size×habitat specificity explained almost one third of the variation in the number of extinction events. Species with high dispersal ability went extinct in a similar number of grid cells irrespective of their range size. By contrast, species with low dispersal ability went extinct in proportion to their range size. Therefore, comparing the speed of extinction across species in the conventional way of extinction rates (that is the percentage of range contraction) might be flawed because it only applies to species with low dispersal ability. Sexual size dimorphism was not a significant predictor of extinction. Extinction was not concentrated on particular clades.