My bounce is bigger than yours – or is it? And does it matter?

Earlier I posted a brief look at the slightly smaller than average bounce, historically speaking, that Obama received from the Democratic convention. Today I want to look at the Republican bounce: just how much did McCain gain from the Republican convention? How does that compare to previous Republican conventions? Again, if we use the same measures that I did for Obama (the change in the RealClearPolitics average of polls), then McCain received almost an identical bounce from the Republican convention as did Obama from the Democratic convention – about 5.6%. This is slightly less than 1% point greater than Obama’s bounce of of about 4.8%, which – given the noise in these polls – is essentially a draw. Historically, it is also slightly smaller than the average Republican bounce, dating back to the 1960 convention, which is about 6.4%. (Again, historical comparisons are dicey, because I’m using a slightly different measure to calculate today’s bounce from the measures used to calculate previous convention bounces.) Note that the size of the bounce will obviously vary depending on what polls one consults. Using the RealClear politics poll averages, McCain has used that bounce to pull ahead of Obama in the national surveys, 46.7 to 45.7%. Again, as I’ve pointed out many times, RealClearPolitics averages good polls with not so good polls. Also, their average is a rolling average, so that it mixes slightly older polls with newer polls. Thus, today’s USAToday/Gallup poll has McCain up 10%, and the daily Gallup tracking poll has McCain up 3%. This might be construed by some as a better measure of the convention bounce than is an average of polls some of which are three days old. But to be consistent with my earlier post, I will stick to the RealClear average of the polls as the measure of convention bounce. And it seems to indicate that, as we begin the general election campaign, this race is a dead heat.

Several of you have commented on the smaller than typical Obama “bounce” and wondered how much had to do with the fact that the Republican convention began almost as soon as the Democratic convention ended. As Chris Abbot put it, perhaps Obama’s “bounce” hit the Republican convention “ceiling”. In response, I’ve checked the dates of every Democratic and Republican convention dating back to 1960, and this year’s were held closer in time than any previous conventions during the last 48 years. So it is certainly possible that Obama’s bounce was truncated because of the Republican counter bounce. At the same time, McCain’s bounce may have been diminished by the continuing reverberations from the Obama campaign. In short, they may have cancelled each other out to a greater degree in this election cycle than in previous years.

The more important thing to remember, however, is that the impact of the convention will tend to fade in the next few weeks as the more fundamental influences that normally shape the vote begin to reassert themselves in the polling measures. In the next week, I will take you through a brief discussion of these fundamental forces as a lead up to making my quadrennial prediction of the popular vote totals in the presidential election. As I’ve told many of you, although political scientists aren’t very good at predicting the nominating race outcome (too many candidates, too many separate contests) we are somewhat more successful at predicting the general election outcome soon after the two conventions have ended and the tickets are established. So, what do the forecasts models say? We will turn to that topic next.

P.S. Several of you who prefer to view these posts at the Presidential Power website have complained about the bleedover in the text. I am aware of it and have asked Middlebury’s tech support to take a look.  With luck it will be corrected very soon.

Leave a Reply

Your email address will not be published. Required fields are marked *