Monday, April 30, 2007

Smith Slides to Pre-Iraq Flip Levels--A Closer Look

[How bad is it to get scooped by my own partner?!? Since I'm the numbers geek of the pair, you lucky readers will get two versions of the same story: Carla's Headline News version for people who just want the results; and my wonkier search for hints and clues to the dynamics of Smith's rating...abbondanza!]

I don't know of a more regular independent survey of US politicians than Survey USA's monthly look at the President, Senators and Governors by state. Whatever you think about computer-voice political surveying, replicating the same method month after month gives the poll its greatest natural strength--the ability to conduct sound trend analysis. The key utility for polling is never prediction, but rather a point-in-time snapshot and the ability to compare this month to last month (and the month before that, etc.).

The latest numbers for Gordon Smith and the rest of the Senate are in for April, reflecting interviews done mid-month. With a hat-tip to Senate Guru for reminding us that new numbers were in, let's take a look, shall we? (click to enlarge, sorry)


To get a little perspective on what's happening here, recall that Smith's Pearl Harbor attack on Bush's "criminal" war in Iraq took place December 7th, and was followed up by an appearance on Sunday talk shows December 10th. Clearly, both the December and January 2007 ratings for Smith reflect those events, as approval in mid-January spiked to 58%, easily his best showing since spring of 2005.

It's less clear whether any particular set of events caused him to drop right back to stasis the following month--was it the series of "clarifications" that pointed out that Smith really didn't have a clear position on Iraq at all? Was it general anger at Republicans for backing the surge newly being implemented at that time? Or was it just the natural ebb of a dead-cat bounce, akin to Saddam's capture causing a spike for the President before almost-immediate retrenchment?

Also notable in the February rating is a return to relatively high "don't know" responses, at 14%, reminiscient of a year previous. Since that time, Smith's approval AND disapproval numbers have gone up, reflecting firmer opinions on the part of the electorate, with only 8% failing to give a response one way or the other. Otherwise, it would seem that his rating has remained relatively flat over the last three months, with approval hovering right around the sentinel 50% mark--above it and you can feel pretty good about your re-election chances; below it and you should start to be concerned, at least.

But the apparent moderation of the trend line masks some definite changes among subgroups, and SUSA deserves major kudos for allowing the reader to pore over them. Under the heading "Track Points" in the chart, find the dropdown list of subgroups--gender, age, political affiliation, political ideology, race, geography...it's a gold mine, and it's where we find some curious recent shifts in the way Oregonians are evaluating Smith.


Clearly independents are a wildly shifting part of the electorate, and Smith's volatility with them has to be of major concern for his campaign. If he catches them in a good month, he'll sail to victory--but if he gets unlucky and November 2008 is one of the bad months, he's completely fucked. File April 2007 under "completely fucked;" in one month independents' view of him dropped by a whopping 21 points, from 56/35 to a dead-even 44/44.

Note however that as with the electorate at large, his ratings among NAV's spiked immediately after the Iraq speech, but has since returned to previous levels. Without making any predictions about what level will be required to put Smith in serious danger, what seems clear is that his performance on the Iraq issue has ceased to be a positive driver of his poll rating, and he's basically back where he was before the speech.

You want to see something a little shocking? Check out the Democratic subsample:


Yes, that's right--after the Iraq speech, Smith's approval among Democrats spiked at 62% in January, with an equally surprising low of 31% disapprovals. That's higher than his ratings both with the general electorate and independent voters. Had Smith not appeared to lurch back into his old patterns and begun hemming and hawing about just what he meant, we might be talking right now about who else to go after in the Senate next year--because there simply is no way he loses with 60% of Democrats on board the Smith train. It just goes to show you how absolutely desperate Democrats are to hear unfiltered language about the giant clusterfuck that is our foreign policy right now--they'll stand up and cheer a Republican who talks the talk, too.

But like everyone else, Democrats have figured out there's less than meets the eye to Gordo's conversion, as his numbers have floated back down to pre-speech levels. One smidgen of concern here is that those figures are still somewhat above ratings from last winter, when he was barely pulling a third of Democrats into his favor. On the other hand, it's unlikely those numbers will stay where they are as the opposition machine picks up the pace in the run-up to the election; Dems will obviously be most easily convinced by the attempt to persuade folks of his unsuitability.

Now, a look at Republican sentiment:


Now this is interesting. Note the lack of any significant movement at all among Republicans during the two post-speech "spike" months for Democrats and Independents--but in the following two months, as with the other groups, Smith's approval fell among Republicans by several points. And unlike the other two groups, our junior Senator has seen a resurgence in April--although to a lesser point than his 70+ days in the fall of 2005, and in the same pattern showing a return to pre-speech levels.

So no matter what party affiliation you claim, the net effect of all Smith's posturing and speechifying the last six months has been zero--but you got there in different ways. Democrats and Independents are still unhappy with Smith, but a touch less so recently, while Republicans are still happy overall, but somewhat less so compared to a year or more ago.

One final pair of charts that have special significance in Oregon--Portland Metro vs. the rest of the state. First Portland and environs:

Here we see roughly the same trend as for Democrats and Independents--a spike, followed by general retrenchment to earlier levels (although there's a bit of an uptick in April that the other two subgroups don't share). Somewhat like independents, the Portland area (including its suburbs) represents the barometer of the swing vote in the state--stay right around 50%, and you've got a decent shot, but under that and you're going to struggle unless the rest of the state just loves you...and right now, that's really not the case:


Once again we have a pattern of improved feelings right after the speech, and a current return to previous levels--but with a strange twist: a moderated improvement in December and January, with a bit steeper drop in February than we saw elsewhere, but a spike in March and then back down for this month for an overall return to stasis. Was it the attention paid to a potential DeFazio run by Southern Oregonians, complete with partisan words about Smith's service, that pulled him down? Can't say, but it's clear that the rest of the state is not yet in a position to rescue Smith if his Portland numbers atrophy as the election looms.

What's the overall analysis? Essentially, that we're back where we were around the 2006 elections: with a vulnerable incumbent who nonetheless still maintains a solid-enough appeal among Independents, Democrats and Portland-area voters to keep him in a relatively favorable position. The Iraq speech has neither hurt him nor helped him--and in the context that he surely hoped it would drive moderate opinion his way, that has to be a big disappointment for Smith. What other "moderate" arrows does he have left in his quiver? But by the same token, his flip on Iraq doesn't seem to have done any serious damage to his Republican base (perhaps because a significant chunk of them are sick and tired of the war, too?)