Riley Poll is Actually Bad News for Smith (if it means anything at all).

Jon Isaacs

This is an excerpt from the Oregonian’s blog report on the recent released tracking poll by Riley Research:

Riley said he was most surprised by the level of bipartisan support for Smith, who faces re-election next year:

A total of 54 percent of the poll respondents rated Smith's job performance favorably. Among those giving him a satisfactory job rating were 57 percent of Republicans and 54 percent of Democrats.

"His support is almost dead even" between the two parties, which is a good sign for someone who has to run statewide. Riley said Smith's recent opposition to the Iraq war has helped him with Democrats and hasn't hurt him much among Republicans.

"He's played it very shrewdly from a political standpoint," Riley said. Anyone looking to challenge him, either in the primary or in the general election, might be dissuaded by such numbers, he said.

Say what?

Let’s be honest. The fact is this one question really doesn’t tell us anything about Smith’s potential vulnerability in 2008. For all intents and purposes, it means nothing.

But since Riley has decided to tout his findings as great news for Smith and the Oregonian is going to report that these numbers are going to dissuade potential challengers, than let’s point that out for what it is – bunk.

By any meaningful standard, the Riley poll is really BAD news for Smith. Here's why -

Smith’s 57% approval among Republicans is incredibly weak.

Riley’s quote in the third paragraph above is complete nonsense. Smith would much rather have his R support be in the low to mid 60’s even if that shot his #’s among D’s down a few points. If Smith’s approval rating among R’s is really this low he’s going to have a base problem next year.

For now, 54% Dem approval is somewhat impressive. But these numbers point out the ongoing dilemma facing Smith. He’s spent the past four months doing nothing but pander to liberals and moderates and he still only has a 54% approval rating among Democrats. What’s going to happen to that number once Smith starts trying to get his conservative base to come home?

Smith’s 54% approval rating is actually a drop of four points from a comparable poll done in January.

Survey USA’s January Oregon Senate approval rating poll, taken 1/12 – 1/14, had Smith’s numbers at 58% approval with 62% of D’s and R’s approving. If we take both of these polls at face value Smith has actually dropped in every category since January. His eight point drop with Democrats is a statistically significant.

In support of my point in bullet #1, take a look at Wyden’s approval ratings in the same January Survey USA poll. He received 52% approval from Republicans and a whopping 76% from Democrats. Now those are healthy numbers.

Smith has a huge problem with Independents/Non Affiliated Voters.

I made the point in my last post that the key to the 2006 Democratic wave was independents breaking for Democrats big time. For the time being, Republicans have completely lost this group of voters. Smith has a terrible 44% approval rating among independent voters. In the Survey USA poll it was 52%. Another eight point drop. This is a real problem for Smith.

If you take him at his word, all of Smith’s flip flopping, contradictory positions and explanations really show that he’s a thoughtful, independent decision maker. So far, independents aren’t going for it. In fact, if we take both of these polls at face value Smith’s numbers among Independents are dropping. Unfortunately, the only voters falling for his ruse at the moment are Democrats.

The bottom line is neither of these polls really tells us anything.

Here are some approval ratings for you from June 2005: George Allen – 53%, Conrad Burns – 50%. Those are results from Survey USA’s 2005 monthly US Senate approval rating polls. Both numbers are within the Riley poll 4.5% margin of error. We all know that these are ratings for former US Senators.

If Riley truly wanted to rate Smith’s vulnerability he would have done the following -

1. More accurate approval ratings. The poll should have given those taking the survey five options instead of three for rating Smith’s job performance – it should have asked strongly fav/unfav and somewhat fav/unfav. This would have given us some idea of the intensity and depth of Smith’s support. I suspect that much of Smith’s Democratic support is soft that will be easy for a well funded opponent to move to their column. Interestingly, Riley did provide five approval options in the question about Kulongoski’s healthy kids proposal.

2. Re-elect rating. Riley should have asked respondents whether they planned on voting to re-elect Smith or voting to replace him. I suspect this number would have been in the low to mid 40’s. This is an important rating because it gives us an idea of whether or not Smith’s job approval is translating into votes for re-election. Any US Senator with a re-elect rating below 50 is in trouble.

3. Head-to-Head match-ups. I’ll bet money that there are at least three head-to-head match-ups if tested today would show Smith trailing or tied - (candidates who have not unequivocally removed themselves from the race). Smith v. Defazio. Smith v. Blumenauer. Smith v. Bradbury. My hunch is that a potential upstart candidate such as Steve Novick, who admittedly would start with a severe name ID disadvantage, would start in the low 40’s v. Smith with plenty of room for growth. In my view, head to head questions would give us the biggest sense of Smith’s vulnerability. When voters are given a choice where do they go?

  • (Show?)

    Jon, while I think you analysis is on the whole pretty sound, I'd make a couple of clarifications:

    *Your points are matters of interpretation. There is, however, no reason to doubt his Riley's methodology. It's an important distinction, because often we see internet surveys and partisan surveys compared to independent telephone surveys, which are much more rigorous. Riley polled likely voters--those voting at least twice in the last four general and primary elections--and the poll was statewide. I sense in your post an indictment of bias, which I don't see. *The sub-sample of non-affiliated voters was just 75 people--substantially less than the 200 each for the Dem and GOP sub-samples. (Margin of error of about 12%, I think.)

    However, there's some additional data you didn't mention that supports your overall point.

    *The poll also mentions Kulongoski's numbers, which haven't moved much since last year's election. I take from this that the polls haven't moved much so early before the next election. This would tend to support your overall analysis that Smith is probably a lot weaker than polling reflects. On the other hand, we can't poll people about how they'll feel in a year.

    *The numbers on Smith also include a curious figure: 21% who are "neutral/unsure." If public opinion is turning against Smith, the trend may be reflected there (35% of non-affiliated voters were unsure, further evidence).

  • (Show?)

    You wanna know how crap that Riley poll is?

    On the presidential preference question for Democrats, he left out Joe Biden and Dennis Kucinich (whose support is certainly non-zero among Oregon Democrats.)

    But he included some crackpot named Randy Crow - who isn't filed with the FEC, but has a website where he rants about conspiracy theories.

    Seriously, Mike, what's the deal?

  • (Show?)

    Riley's likely voter screen is what's crap--that's a bad method that leaves out lots of voters who don't participate in primaries or midterms, but come out for Presidentials (which of course 2008 will be).

    Riley's track record of accuracy is not good, and historically shows Republican bias. I would take these numbers with a big grain of salt--but if anything, they're worse than they appear.

  • (Show?)

    Torrid, what are you talking about? His methodology included voters from the '06 primary and general (tat's a midterm AND a primary) and the '04 primary and general.

    Kari, fair enough on the answer categories, but "crap" suggests methodological bias. Do you see any?

  • (Show?)

    Spot on John! Of course Riley methodology is biased (my opinion). Just look up Republican Pollster. After observing Gordon Smith at the recent Town Hall with Senator Wyden in Bend it becomes quite clear that this guy is slicker than black ice. I hope we can come up with a strategy to use it against him.

  • (Show?)

    The poll seems suspect. When is Wyden going to stop hauling Smith around on his coat tails? It benefits Smith, not Wyden.

  • Betsy (unverified)

    Has anyone considered a Smith v. Kulongoski race?

    Coming off an eight-point win over a moderate Republican -- and a successful legislative session -- he might just make a better candidate than some of the alternatives.

    Am I crazy?

  • (Show?)

    "Torrid, what are you talking about? His methodology included voters from the '06 primary and general (tat's a midterm AND a primary) and the '04 primary and general."

    No, it didn't. He pulls people who have voted in two of the last four (06 gen, 06 primary, 04 gen, 04 primary). That means anybody who votes only in Presidential elections cannot answer the survey. Or if they just voted in the 06 general, also not counted.

    That's just a bad way to do LV screens. The best way is to ask if they voted in the last Presidential, or to rate their likelihood of voting on a scale of 1-10, and only taking the 9s and 10s. (or maybe the 8s as well).

  • BlueNote (unverified)

    In my opinion, every day that passes without the announcement of an "A List" opponent for Smith reduces the already marginal chances of beating Smith in 08.

    I have worked for two (non-Oregon) US Senate races and in each case, my candidate spent a full year (or more) doing groundwork preparation prior to the commencement of the primary season. There are hundreds of people to meet and cultivate both in Oregon and in DC. There are major political donors to cultivate. These days, it is also necessary to schmooze existing "safe seat" incumbents to convince them to move money your campaign. There are fund raising "beauty contests" to participate in. As one of my former bosses used to say, to win a senate election you need to spend a couple years kissing "A & B" (Asses and Babies) before you file for the primaries.

    Who and where is our candidate? Is he/she already out there kissing? It would be nice to know.

  • (Show?)

    BlueNote -- Yes, time is of the essence. But we're not in the redzone yet.

    Just because no one is public, doesn't mean that there isn't action happening... I don't have any special knowledge - but take note that our members of congress have just come into the majority; which means they have access to people, organizations, and money that they've never had access to before.

  • (Show?)

    Right on, John. Republicans don't trust the guy, and in the end Democrats will not vote for a Senator who voted to undermine Oregon's minimum wage and says that his greatest achievement in Congress has been passing a massive tax break for multinational corporations that stash their money overseas. Once the D's know his voting record, they will vote against him.

  • varner (unverified)

    Novick for Senate! Oregon's Smartest Choice.

  • Mike Riley (unverified)

    You have to admit, if I didn't post such a thorough explanation of my methods, you wouldn't have so much fun debating them, but a few points and clarifications are in order...

    torridjoe - you rarely get it wrong, but you said people who only vote in general elections (ie 04 & 06) couldn't be included, but with a 2 out of 4 selection, they certainly could be. I'm always considering methods that would recognize the most recent voters (particularly new voters in the 2006 general), as they may well best represent the current (and future) thinking. Trouble is, we saw a lot of the 2004 first-time voters disappear in the next primary. Also, most people are familiar with the 2/4 criteria and I'd risk being called a "cherry picker" if got too exotic with the selection methods. Just for kicks, I'll look to see if I can pull results for those who voted in just the 2006 general election to see how they answered, and if there's a big enough sample, I'll post those results next week.

    Kari - you're probably right, Biden & Kucinich might have gotten a point or two, and while it wouldn't have changed the fact that it's currently a two person race between Clinton and Obama, they should have been in there. Further, while I think the two current Democratic front runners are likely to be the same a year from now, I have much less confidence about the two current Republican front-runners (Giuliani and McCain).

    Jon - I couldn't agree more: a 5-point scale is a better indicator of the strength of support (or opposition), but I chose the favorable / unfavorable rating because I had three years of similar data for Governor Kulongoski and wanted to see how Smith's numbers compared to his. Thanks for the feedback.

  • (Show?)

    Mike, you misread TJ and what he said was not wrong. He did not say you missed people who only vote in generals. He said you didn't use people who only vote in presidentials. The fact that a lot of your 2004 first timers "disappeared" in 2006 only underscores his point.

  • Mike Riley (unverified)

    Fair enough, and by the way, I'm going to have to do additional interviews to reach those who voted only last November as I obviously have none in my current sammple. I would expect their responses to the questions to be different.

  • (Show?)

    Mike, thanks for taking the time to come in and explain your opinions and methodology to the people here at BO. My question is somewhat off-topic: Who is Randy Crow, and why did you include him in your presidential poll?

  • Hawthorne (unverified)

    At the risk of offending one of his supporters...Randy Crow's bio and website are pretty entertaining in a scary kind of way:

    In his own words, " I am the tenth generation or so grandson of Pocahontas through my mother and father. People seem to be fascinated by this fact. If I were to give my genealogy it would be very obnoxious because there is a herd of giants in it."

    Having looked at his website I have to say I'm perplexed that Riley included him in the poll. Mike, what were you thinking? This guy can't even win a seat on his own city council.

  • Mike Riley (unverified)

    Mea Culpa - Not thinking. I started out planning to list just the "major" candidates, then asked a staffer to do a Google search and create a comprehensive list. I failed to review the final list. Sorry - thinking big picture - forgetting to check the details.

  • (Show?)

    Interesting that you've got a staffer whose idea of a plausible presidential candidate includes someone who rants about "zionist conspiracies" -- but doesn't include a US Senator and a Congressman, each on their second runs for president.

    <h2>Next time, try a more authoritative source.</h2>

connect with blueoregon