A large number of Large-billed Crows Corvus macrorhynchos and Carrion Crows C. corone are culled annually either by shooting or trapping in Japan, but the effects of such culling on crow behaviour have never been evaluated. I hypothesized that in an area where crows are shot they would become more sensitive to human disturbance than those in an area where they are trapped. I compared alert and flight initiation distances of crows in two nearby, but separate, areas in Iwate Prefecture, Japan, in March 2004. In 2003 all 1,025 crows culled in the Daito area were shot, whereas 95.3% of 1,586 crows culled in the Tono area were trapped. Topography, land use, and human population densities were similar in the two areas. Analyses using generalized linear models showed that the area (and hence the method of culling) was the sole explanatory variable in the best model and included in each of the top five models for both distance measures. The median alert distance and flight initiation distance were greater where crows were shot than where they were trapped (60.0 m vs. 20.5 m and 46.0 m vs. 18.0 m, respectively). Other factors, such as species, flock size, behaviour, and habitat type, were far less important. These results show that shooting has a greater non-lethal effect on crows than trapping. While trapping may be a more efficient way of removing crows than shooting, shooting not only reduces the number of nuisance animals directly, but also affects their behaviour and habitat use so that damage is reduced indirectly. Shooting is thus the more effective choice for culling in order to reduce damage caused by crows.