You could quite easily test that, by comparing bit-flipped domains to keyboard-neighbour domains. Or alternatively, just by looking at the distribution of hits

You could quite easily test that, by comparing bit-flipped domains to keyboard-neighbour domains. Or alternatively, just by looking at the distribution of hits to bit-flipped domains (which should be broadly uniform; typos will be non-uniform).