Mark Blumental expects Obama's bounce for killing bin Laden to follow historical trends:
The reaction to the arrest of Hussein in 2004 is particularly noteworthy, since it presents the closest parallel to the killing of bin Laden. The chart below shows the trends in all polls compiled by University of Wisconsin Professor Charles Franklin. His trend estimate shows that the capture of Hussein produced a bounce of roughly 5 percentage points in his summary trend line, which estimates the underlying trend represented by all polls (including the aforementioned Gallup numbers). The chart also shows that the increase in Bush's rating quickly dissipated.
Is killing Bin Laden akin to the capture of Saddam Hussein, which caused a relatively small and short-lived jump in George W. Bush’s approval ratings? That probably understates the case: Osama bin Laden did orders-of-magnitude more damage to America than Saddam did. But is it tantamount to achieving victory in a war? It’s not that either, exactly: Bin Laden is just a single (very) high-value target in what United States policymakers describe as a global campaign against terrorism.
Jonathan Bernstein believes the bump will be short-lived:
It’s going to be hard to run against Obama .?.?. for the next few weeks. Which doesn’t matter much electorally, since he won’t be on a ballot for the next few weeks. After that, the most likely result is that, assuming no other events intervene, things will return more or less to normal. We have 50 years of data on rally-round-the-flag effects, and what they tell us is that the half-life on these things isn’t very long.