By eery coincidence, this week I read Malcolm Gladwell's The Tipping Point (which I neglected to read when he first wrote it). The book describes the phenomenon of social epidemics, including suicide and murder-suicide epidemics. Gladwell cites compelling evidence that high-profile media coverage of suicides leads to a spike in the suicide rate and that high-profile media coverage of murder-suicides causes murder-suicides to spike. (Most creepily, even automobile fatalities spike following this pattern, suggesting that some people commit suicide and murder-suicide by auto.) In light of this evidence, it's a near-certainty that people will be killed by acts in imitation of the Virginia Tech shooter---and indeed, it appears that there has already been a wave of scares. (It's not always easy to tell which "successful" murder-suicides are attributable to copycatting, which is why the large-n study looks at the change in rate.)
As Gladwell explains, prominent murder-suicides create a "script" that others can then follow so as to get out their "message." Accordingly, any substantial news coverage of a murder-suicide is potentially lethal, but it seems especially problematic to show the killer's videorecorded statement because it makes the script that much easier to follow.
It may be tempting to think that news organizations can show the video on the theory that with its widespread availability on YouTube and elsewhere, the damage has already been done. But this is almost certainly wrong. Some, probably most, people watching the news will not have seen the video before, and each additional viewer raises the likelihood of further violence. That's not an argument for censorship. It may not even warrant self-censorship: A news organization could make a judgment that the newsworthiness of the story outweighs the marginal unwitting encouragement to murder-suicide that coverage provides (although I wouldn't reach that judgment). But at the very least, one would hope that responsible news organizations consider the very real cost. The fact that the networks have now said they will limit re-airing the video in response to claims of offense by victims' family members suggests that they don't appreciate the more substantial harm to which they have already likely contributed.