Joe posted a recent Gallup poll showing a significant tightening of the Congressional vote (from heavily pro-Dem to even among likely voters) and a sizable jump in President Bush’s job approval numbers after the 9/11 anniversary.
But in a poll conducted after the Gallup poll, Bush’s numbers appear to have dropped again to a point virtually unchanged from a month ago.
According to the NYT/CBS Poll, Bush’s numbers have jumped a whopping one point between mid-August and today, from 36% to 37%. Views on Congress have gone significantly downhill since August: from 29% approval to 25% approval. In fact, the 25% approval rating for Congress as a whole is one of the lowest it has ever reached. Approval of one’s own Congressman is at 53%, which sounds quite high. Except when you consider that that is the lowest approval number for “your own Congressman” since July 1992, it appears more significant.
But when the question is posed: “Do you think your Congressman deserves re-election?”, 48% say “NO.” For comparison’s sake, the number of people willing to elect a “new person” is the highest since, you guessed it, November 1, 1994 (when 53% wanted somebody new).
Comparing polling firms’ result is a dangerous game. Each firm uses different sampling methods (on party ID especially). In this poll, Party ID is actually fairly even, with 32D, 30R and 32I – this is much more Republican of a sample than August when Dems were 35 and Republicans were 30. Ideologically, this poll sample includes much fewer self-described liberals than August (24% versus 17% today). So the voting sample cannot be cited as the reason for such a pro-Democratic poll result.
As for “likely voters,” each firm uses a different “likely voter” model. In fact, this poll does not even look for “likely voters”, relying instead on “registered voters.” While “likely voters” is a more important result in the last week before the election, at this point in the cycle any likely voter model is going to be driven more by 2002 performance than by actual intentions to vote this time. And when you’re comparing registered voters, even the Gallup poll numbers are not far off from this one.
On the generic Congressional ballot, 50% choose a Democrat and 35% choose a Republican. This number is of limited use, but it does serve as an indication of overall movement. This 15-point spread is the largest in the cycle, though it matches the same spread in August. The difference is that the “don’t knows” moved equally into each camp.
On the top issues, Republican advantages are at a low point.
Voters trust Democrats over Republicans on Iraq by 5 points (though this is actually narrower than in April when Democrats led by 11). On terrorism, the Republican advantage is only 5 points, which is the smallest lead the GOP has ever had (though tied with May 2006). Significantly, in August Republicans had an 8 point advantage. The difference may be that the August poll was taken shortly after the British bomb plot was discovered, so voters may have been more approving of the GOP on terrorism then than they are now.
What does all this mean?
It means we have a long race ahead, where both parties will ebb and flow. Bush’s numbers may have actually jumped after the 9/11 anniversary, and then dropped again a few days later. Major speeches tend to have that effect. Terrorism has, indeed, jumped as an issue of concern to voters.
But that hasn’t necessarily redounded to the Republicans’ benefit. This is why the Republican divide over detainee policy is so damaging to the GOP on this issue. Unity breeds confidence; disunity yields demoralization. For once, the Dems are united on how to approach terrorism and the GOP is not. And GOP efforts to paint Democrats as “soft” or “unserious” about terrorism seem not to have resonated with the public.
More importantly, Iraq is getting noticeably worse, with Bush Administration officials voicing frustration at Maliki’s inability to clamp down on death squads. Without objective improvements in Iraq, the GOP is left to warning about premature withdrawal – a sizable difference in tone from early predictions of “victory.” My guess is the recent jump in support for Bush on Iraq, from a paltry 30% to a less-paltry 36%, is more a result of the 9/11-campaigning than anything. But if nothing radically changes, I see this number dropping again. Voters might not know if the Democrats can handle Iraq better. But they’re pretty sure that Bush and the GOP have screwed up nonetheless.
The cautionary tale may be that this is just registered voters. Republican GOTV efforts put the Dems to shame. And when the campaign seasons really heats up, which will be in a few weeks, the numbers could really change. But as of today, September 20, the Democrats are in very strong position for the election.