Nate Silver can't apply statistical analysis to isolated incidents because it is "almost impossible to come to meaningful conclusions about probability from a sample size of one." He suggests, instead, that we focus on threats:
If it turns out, for instance, that Democratic members of Congress are much more likely to receive such threats than Republican ones, that might tell us something meaningful. Likewise, if threats made against Mr. Obama routinely invoke his race, that could tell us something too. It might also be possible, with careful study, to see whether there is a correlation between the frequency of different types of political rhetoric and the number of such threats. ... [Such statistics] would probably provide us with considerably more insight into the risks posed by assassination, domestic terrorism, and other severe forms of political violence, than any degree of scrutiny of Mr. Loughner’s case could on its own.