This study examined the influence of the number of orthographically similar candidates, neighborhood size, on the word and pseudoword naming performance of normal, dyslexic, and beginning readers. Participants were 23 Dutch dyslexic fourth-graders matched to 23 fourth-grade chronological age controls and 17 second-grade reading age controls. Unexpectedly, neighborhood size had similar effects in all groups: It did not affect word naming and facilitated the naming of pseudowords. However, the presence of a high-frequency neighbor had different effects. In contrast to normal readers, words with a high-frequency neighbor were named more slowly by beginning and dyslexic readers. These findings suggest a dissociation between global and specific effects of neighbor words. Nevertheless, both findings seem to be compatible with the view that orthographic representations of beginning and dyslexic children are not (yet) sufficiently specified.