I do have one comment, however, on an Ezra Klein post on the subject that I mostly agree with. Klein says, sensibly:
My hunch is that these moments only become “campaign defining” if the campaign was already defined that way in the first place, and in that case, they don’t have much of an impact. If that’s right, the election won’t be any different at all because of the Etch a Sketch comment. Which isn’t to say that it won’t be defined by it.I think that gets it exactly right...for the general election. But Klein also discusses Rick Perry's "oops" moment in the same terms, and I'm not convinced we want to think about it the same way.
Remember, for the November election we have lots of very strong influences on our vote: party, of course, is the strongest, but there's also our view of the economy, and of the president in general. The candidates will have very different positions on matters of public policy. And we have several years, lots of events, and tons and tons of information to use to make those decisions. The chance that a gaffe or a debate moment or an ad will change anything is vanishingly small in that context.
But in primary elections, every voter is up for grabs, and most voters have very little information to use to differentiate the candidates. They do it; we know, because for example the exit polls tell us that (most) social conservatives figure out that they're supposed to vote for Rick Santorum. But especially early in the process, it's a pretty difficult task. Why Santorum and not Bachmann? Why Romney and not Pawlenty?
Because voters have little to go on, high-visibility and trusted party actors become important. So do ads (which of course are funded or not funded by party actors). And so too is information in the mass media. When something such as "oops" blows up, there's very little to prevent it from damaging a candidate.
Now, Klein is probably correct that this particular campaign gaffe mattered more to Rick Perry than it would have to Romney because Perry had already acquired a terrible reputation from previous debates. So, yes, the single moment has to be seen in the context of other moments. Remember too that this all happens (or doesn't happen) at the elite level, too. Party actors who had a lot at stake were paying close attention to the debates and other campaign events, and may well have made their choices in part based on candidate performance there. But in the general election, all those party actors are not only committed to their candidate, but for the most part committed to at least pretending in public that their candidate is totally dominating everything.
In other words: yes, it makes lots of sense that there can be "game change" moments in primary elections, but that they'll be few and far between in the general election.