Bush Job Rating and 2006 Turnout

While MP and his family were enjoying time off at their undisclosed location, Mickey Kaus raised several questions about the Bush job rating, particularly as measured by the Rasmussen Reports automated tracking survey.  The most intriguing was this one:

According to the Rasmussen robo-poll, 43% of Americans approve of how Bush is doing his job, while 55% disapprove. But what’s really striking is that the disapprovers disapprove much more vehemently than the approvers approve–41% of those surveyed "strongly disapproved" of Bush, while only 21% "strongly approve." Doesn’t this imbalance of fervor mean something in low-turnout elections, such as the upcoming 2006 mid-terms? Specifically, doesn’t it mean the anti-Bush forces should do very well in 2006, in a mirror-reversal of the 1994 mid-terms?

[Note: as of 8/31, Rasmussen shows 34% strongly disapproving Bush’s performance and 25% strongly approving].

The answer is a bit murky.  The Democrats may ultimately gain a turnout advantage in 2006, but MP is not sure this result supports that argument. 

There are two issues here.  The first is the long term decline in President Bush’s overall job rating since January, evident in surveys conducted by every public pollster except, ironically, Rasmussen.**  MP commented on this trend earlier in the month.  Since then, Gallup, Harris, ARG and now ABC/Washington Post have reported their lowest ever job ratings for the president (links via RCP). 

This trend is certainly not welcome news for the Republicans, as it represents a significant decline in support for the president since the election.  Obviously, a shift of that magnitude in vote preference last November would have put John Kerry in the White House.  If this trend persists, it does not portend well for the Republicans. 

The issue Kaus raises is a bit different but more intriguing:  Does the Rasmussen result, which continues to show more strong disapproval of the president than strong approval, suggest a potential turnout advantage for Democrats in 2006? 

The answer, unfortunately, is cloudy.  To get it, we need to check two things: 

First, do other surveys confirm the Rasmussen result?  The answer appears to be yes, but the answer was not as easy to find as MP assumed.  While pollsters typically use follow-up probes to measure intensity of opinion ("do you feel strongly about that?"), very few probe intensity on the standard presidential job approval rating question.  Most simply ask whether respondents "approve" or "disapprove" and leave it at that. 

There are a few exceptions.  Other than Rasmussen, the exceptions that MP can find include the Westhill/Hotline poll, Harris and Zogby:

• The Westhill/Hotline poll asks a question similar to Rasmussen’s.  In July, they found 23% strong approval and 34% strong disapproval of the President Bush.
• Zogby asks the job approval question with different answer categories:  excellent, good, fair or poor.  In survey conducted at the end of July, they reported 19% who gave President Bush an "excellent" rating vs. 28% who rated the president’s rating as "poor." 
• Harris also uses the excellent, good, fair or poor scale.  In their most recent August telephone survey sponsored by the Wall Street Journal, they reported 13% of Americans rating the president’s performance as excellent, 34% rating it poor.  Their June survey showed similar results.

Thus, at least two other surveys confirm the Rasmussen finding that more Americans express a strongly negative view of his job performance than a strongly positive view.

The second issue: Were these results different just before the elections last year?  That one is harder to answer.  While Bush’s ratings are certainly lower, the strongly negative opinions appear to have been more frequent than strongly positive opinions even in the midst of the 2004 elections: 

• Rasmussen’s free website provides only the current result for the four-category job rating.  I have requested pre-election results from Rasmussen directly but have not yet received a response. 

• According to a recent release, Zogby found 18% of "likely voters" giving Bush an excellent job rating in September 2004, while 32% gave Bush a rating of "poor."  In mid November, 24% rated Bush excellent, 31% poor.

• The Westhill/Hotline poll was started in January 2005, although at that time they reported that 29% strongly approving Bush’s performance and 34% strongly disapproving. 

These suggest that even back in October 2004, more Americans strongly disapproved of Bush’s job performance than strongly approved.   How can that be?  Didn’t George Bush do a better job turning out his supporters than John Kerry? 

The problem is that perceptions of George Bush alone do not tell the full story.  The vote was not just a referendum on Bush (despite the predictions of a certain officious blogger), it was a choice between Bush and Kerry.  Questions that measured intensity of support in the Bush-Kerry showed something different.  Consider:

• On a survey fielded October 28-30, 2004, the CBS/New York Times poll showed 67% of Bush voters saying their "strongly favored" their candidate compared to 49% of Kerry supporters.   Thus, roughly 34% strongly favored Bush, 23% strongly favored Kerry.

• On a survey fielded October 27-30, 2004, the Pew Research center showed 34% supporting Bush "strongly" vs 29% supporting Kerry strongly.

• On a survey fielded October 16-18, 2004, the Hart/McInturff NBC/WSJ poll ($) showed that 91% of Bush voters would "definitely support" their candidate compared to 84% of Kerry voters.   Thus, roughly 44% were "definitely" supporting Bush, 39% were "definitely" supporting Kerry.

The point here is relatively simple.  One may have some reservations about Bush’s job performance — enough to give him a "good" rating rather than "excellent" rating or "somewhat" approve rather than "strongly" approve — but still strongly prefer Bush to Kerry.  This point is related to the observation made recently by a kausfiles reader:

Alert kf reader G.S. suggests leaderless Democrats take another look at that Amazing Dr. Pollkatz Polling Graphic. The only time Bush’s steady polling decline stopped was in 2004, when he actually had some identifiable Democratic champions (Dean, then Kerry) to be set off against.

So what does all this tell us about the potential for turnout and the political landscape in the 2006 elections?

1) Again, obviously, Bush’s falling job rating — should it remain low — is bad news for Republican candidates in 2006.  Anecdotal evidence from the recent special elections suggests that the Democratic base is now more energized than Republicans. 

2) However, neither Bush nor Kerry will be on the ballot in 2006, so the ultimate impact of the Bush rating on turnout in the 2006 results is unclear and debatable. 

3) If we really want to check intensity of opinion regarding the Senate and Congressional elections, we need to measure it directly.  The best gauge on a national survey is the so-called "generic" Congressional vote with a follow-up question about strength of support.  In other words, we need to ask respondents whether they "plan to vote for the Republican candidate or the Democratic candidate in your district" and then whether they support that candidate "strongly" or not.   

Unfortunately, MP has not seen a strength of support follow-up yet regarding the 2006 congressional elections, but they will be coming soon. 

**On August 17, when Kaus posted these numbers, the Rasmussen survey had shown a sudden decline in the Bush job rating. However, over the last two weeks, Bush’s numbers on the Rasmussen survey have risen steadily. The result posted this morning — 49% approve, 50% disapprove — is roughly the same as Rasmussen has been reporting all year.