Posted 8:18 AM by Robert

The Importance Of Being Bouncy

O, that post-Convention bounce. Every candidate wants it. Most get it; Kerry-Edwards didn't get much, maybe none, maybe negative.

The mainstream media - sometimes speaking through the avatar of Larry Sabato - is eager to tell us that the reason Kerry-Edwards got no meaningful bounce is that infamous dearth of persuadable voters, that consequence of the almost completely early-polarized electorate in which almost every voter has already made up his or her mind. And that's that. See, the media says, there are just no voters left for anyone to bounce off - and that means you, too, Mr. Bush. So don't get any ideas. And, of course, much of the media is also eager to tell us that post-Convention bounce all goes away, anyway. So it really doesn't matter that Kerry-Edwards got little ... or none .. or negative. Nor will the media let us forget that most undecided voters swing against the incumbent, since the first thing a voter does is to decide if he or she likes the incumbent, with whom the voter is already familiar, anyway.

Except that what the media is saying is all wrong. Worse for the Kerry-Edwards prospects, much of the media optimism for the Democratic ticket is based on exactly such errors.

Yes, post-convention bounce always goes away, except when it doesn't - as happened with Bill Clinton's 1992 huge and mostly permanent bounce. But even in election years less exceptional than 1992, bounce matters a lot. The importance of bounce is that it demonstrates how (or whether) a candidate can move voters under the best conditions, when almost all media attention is focused on one candidate and that candidate's message. If a candidate can't move more voters under such conditions, that's very bad news for that candidate's chances of moving voters under the less favorable conditions of the general campaign. Kerry-Edwards is there.

And, yes, most undecided voters swing against the incumbent - if they are still undecided towards the end of the campaign. But this campaign is just beginning - not ending - so the "swing voter" rule is simply not applicable at this point in the campaign.

To get an idea of how badly deluded the media coverage has been, consider Hispanic voters - arguably the most generally misunderstood group of voters in the country, if they can even be meaningfully called a "group." For example, the Washington Post recently reported on a poll that found that Kerry claimed support from 60 percent of all Latino registered voters in the 11 states surveyed while Bush had 30 percent. The Post chipped in this bit of analysis:

The findings suggest that, at this point in the campaign, Bush is falling short of his goal of notably improving on the 35 percent share of the Hispanic vote he received four years ago, although his advisers said they believe he is still on track to do so.

Are the President's advisers really this out of touch? Do the Post's findings suggest that the President really falling short on his goal of notably improving on his share of the Hispanic vote? Well, no - on all counts. One might start by pointing out that the Post is using the old "registered/likely" voter trick. The Post poll is among registered voters - and Hispanic voters have historically had a turnout rate of about 40%. And, historically, Hispanic voters who do vote tend to vote a good deal more Republican than what is seen in polls of registered Hispanic voters. The Post knows all that - but doesn't mention any of it.

But it gets worse. The Post article doesn't even mention post-Convention bounce, but the analysis implicitly assumes that Mr. Bush will receive much less post Convention bounce among Hispanic voters than he did in 2000. That's because the Post is comparing pre-Convention 2004 poll numbers with actual 2000 election results.

The picture changes quite a bit if we compare pre-2004-Convention poll numbers with pre-2000-Convention poll numbers. This Hispanic Trends Polling Report examined the effect of the 2000 Republican Convention on Hispanic voters. Hispanic Trends found that Mr. Bush entered his 2000 Convention with the support of just 24% of registered Hispanic voters and came out of it with a modest bounce among such voters to 28%. In other words, Mr. Bush enters his 2004 convention with six percentage points more support in the Post poll from registered Hispanic voters than what was recorded in 2000 Hispanic Trends poll. Even keeping in mind possibly differing poll methodologies, in light of the Hispanic Trends data, the Post survey suggests that the President has made lots of headway towards his goal of notably improving on his share of the Hispanic vote since 2000.

But the Hispanic story gets even worse for the Democrats: Al Gore's support fell 12% - from 58% to 46% - during the 2000 Republican convention. Here are some other effects Hispanic Trends found that the 2000 Republican Convention had on Hispanic voters:

The GOP's Convention succeeded in convincing a significant percentage of Hispanic registered voters nationally to give its presidential nominee, George W. Bush, a second look. 24% of all Latino voters told our interviewers that they are now undecided about the upcoming election. In contrast - 13% and 16% were undecided in our May and July (pre-convention) polls respectively. The large majority of Hispanic voters that moved from the Gore column to the undecided column are Mexican-Americans.

* The positive impact of the Philadelphia event on Hispanic voters becomes even more evident when we analyze those voters that watched at least some of the Republican Convention (65% of all Latino voters). Gore's lead over Bush among this group is only 8 points (41%-33%). In contrast, the Democratic nominee leads his Republican counterpart by 34 points (53%-19%) among those that did not watch the convention at all. There is little doubt that the Republican strategy of communicating messages of "compassionate conservatism" and of "minority inclusion" worked extremely well with the Latino electorate.

* George W. Bush made important gains among foreign-born Hispanics who became citizens and voters before 1995. The Republican nominee now leads among these voters - most of whom became politically active during the Reagan-Bush years- by 42% to 38%. Bush trailed among these voters in our May poll by 12 points. He also is continuing to increase his support among Cuban-American voters, among whom he leads 75% - 7% .

* Al Gore's most loyal constituencies in the Latino electorate are now Puerto Ricans, Dominicans, Central Americans and South Americans. He leads among these voters by over 40 points (61%-18%). But the Republican Convention hurt him significantly with the group that up until now had been his most loyal Hispanic constituency - "new immigrants". These voters - who became citizens in the aftermath of former California Governor Pete Wilson's anti-immigrant campaign - are one of the most important reasons why California was considered "not-in-play" before Philadelphia. In our pre-convention survey, Gore led among this group by 46 points (67%-21%). Our post-convention study indicates that his lead among "new immigrants" has been cut in more than half to 17 points (45%-28%). Is California "back-in-play"?

Do you think maybe the 2000 Convention effects are part of the reason why the President's advisers aren't agreeing with the Post? But, see, the media is telling us that none of that 2000 stuff matters now - because there are just no voters left for anyone to bounce off - and that means you, too, Mr. Bush. So don't get any ideas. And, of course, that post-Convention bounce all goes away, anyway. Except that Mr. Bush's 2000 Hispanic bounce didn't go away. It got bigger. In fact, Mr. Bush's support grew from 28% support among registered Hispanic voters right out of his 2000 Convention to 35% support among actual Hispanic voters on election day. The Post poll suggests that he will do considerably better than that in November.

MORE