Is the RNC/Palin bump over? Almost. But it's happening far too soon for them -- soon enough that suggests there may be significant movement toward Obama underneath the bump.
What does a bump look like? According to some nice work done by Poblano (aka Nate Silver) of fivethirtyeight.com a standard convention bounce looks like this:
Day 1 is the first day of the convention.
Now, the back-to-back conventions were unique, and there's no precedent on how the bounces would interact. I took a shot at it. My thought was that the second (RNC) bounce would halve the effect of the first bounce once its magnitude was bigger. So I charted that (blue line) and the change in the combined daily tracking polls* (red line). Here's where we were Friday:
My guess looked pretty good. But look at that same chart as of today:
The RNC bounce is very nearly gone -- I mark it at 0.9 points. If this trend continues, it very likely is showing some underlying bounce. Keep it up.
*The baseline on the bounce poll is a Gallup/Rasmussen composite of Obama +1.5 which he'd held with little variation for about a week before the DNC bounce started. Once the Hotline On Call/Diageo and DK/R2k tracking polls began, I normalized them based on the first three releases versus the composite of the already included (and normalized) polls from the same period. Note that Gallup and Rasmussen have a combined house effect of over 2 points +R, so the baseline for the bounce tracking should be roughly equivalent to Obama+4 on a poll with no house effect.