Hey Bob,

I read a book once too. I made no claim regarding the statistics. I posted them, linked back to the entire article, and made an assertion regarding Smith's motivation.

The overall trend of multiple polls for this race is kinda short on.....polls, and hence has a very large margin of error, and I did mention that it's not a large difference between the two Dems.

If you find yourself unable to comprehend simple English, perhaps you are, to quote an eminenent writer, just a partisan shill, but I won't make that judgement about you 'cause I don't know your motivation from that of Adam's Off Ox (TM Bill Clinton).

Hey, Pat, why not just cop to being a partisan shill and have done with it? I mean, it's pretty obvious. Takes one to know one and all that.

Disclaimer: (on the advice of a local attorney).

In the above post I took swipes at McCain, Novick, Clinton, Bush, Smith, the Republican Party, and SurveyUSA. I remained neutral on Rasmussen and praised Obama and Merkley. I also voted for the latter two and my ballot is already in Gladstone.

Of course the above sentence seems to be redundant to me in light of the actual text of the post, but what do I know?

Yeah Stephanie V, talk about the pot calling the kettle black. Pretty obvious.

Paddy my boy, another thoughtful post despite the Novick hit squad moving in to protect their guy. I'm guessing Novick will take about 38% of the vote.

Since February, Merkley has gained 15 points on Smith. Since March, he has gained 10 points on Smith.

Since February, Novick has gained 7 points on Smith. Since March he has gained 5 points on Smith.

Disbelieve the Rassmussen data if you want, but it clearly shows Merkley's strength against Smith is surging at twice the rate that Novick's has. Further, it establishes a trend with three separate data points which is always going to reveal significantly more than two data points.

everything he says at the end about a Merkley/Obama tag-team holds for a Novick/Obama pairing, and i can say that without having to bad-mouth Pat. i'm still pulling for Steve, who i think will run a more outspoken campaign, but both will depend, and win, based on the national Obama movement.

Barack's coattails will rule the day!

I recall that when Novick was leading in the polls, talk of "margin of error" was verboten by certain partisans. Now I see the math does matter.

The very serious finding here is in the text from Rasmussen--an incumbent below 50% is in big trouble.

And there is a sign of strength here for Merkley v Novick. The trendline for Novick has only gone up six since February, but for Merkley it's been on a steady trajectory.

For Novickians, though, the news is still pretty good--for a candidate who's never held elective office to be within six of a two-term incumbent is encouraging.

[Official Notice: since some readers may have the expectation of perfect neutrality from commenters on blog threads, I hereby disabuse said readers from any such notions.]

I will not argue that it is unreasonable for Merkley's supporters to read these data in the ways that they do. They are plausible readings.

But they are not the only way to read them. First, Jeff's trend line and Steve's trend line from March to May are similar. Jeff was behind Steve in January, caught up in March, and both have advanced in a similar way since then.

Second, neither is likely to gain much more in the way of trending up unless & until Smith's support starts eroding. While it is true that Smith has been under 50%, it hasn't declined a whole lot. The match-up with Jeff in this particular sample shows him (apparently) throwing more people from Smith into "undecided" than Steve -- but his own actual support is only 1% above Steve's.

T.A. has a good point about Obama and we can hope that helps -- but you can bet that Smith will spinning himself along lines of "I'm the one best able to work with whichever presidential candidate wins -- McCain's in my own party, but I have a long record of bipartisan work with Ron Wyden that I could extend to Obama too, if he reaches out to "end the bickering" as he says he will."

Struggling for rhetorical dominance for Jeff or Steve on BlueOregon seems really pointless to me now.

The overall message I take from this poll is 1) that there is real reason to think Smith may be beatable this year, and 2) both Steve and Jeff have raised their recognition significantly enough to be above 40% support, which is good, and 3) that insofar as their support reflects longer-term non-supporters of Smith who were previously "undecided", it appears that the voters agree with the common proposition here that both Jeff and Steve are progressives worth supporting in preference to Smith, to a very similar degree,

BUT, neither Steve nor Jeff has got much beyond having already existing Smith non-supporters being willing to support each of them, into significantly eroding Smith's support.

How to do that in the face of Smith's money is going to be the real problem for the general. The answer is bound to be a bit different for our two potential candidates.

Can I persuade anyone with column-writing privileges to pose a challenge along these lines: What's the best way to go after Smith -- answers for Novick to be provided by Merkley supporters, and answers for Merkley to be provided by Novick supporters (we all know who we are, pretty much), undecideds pick one or the other or do both.

The point of this exercise would be that whoever wins the primary, a large proportion of those who choose to put energy into defeating Smith will be working for their second choice in the primary. We probably have more developed ideas about the "beat Smith" problem for the primary candidate we favor. So let's build up the muscles on the other side, see if there are approaches common to either, and see if we gain anything in understanding Smith from looking at him through a less familiar lens.

To be very clear, this is not a claim that anyone must support the primary winner, but an invitation to those of us who are inclined that way to begin refocusing on Smith.

P.S. To clarify, by "apparently" above I mean that it looks like the poll was Smith-Merkley and Smith-Novick, so that the missing % apparently is undecided, and not say Frohnmeyer. I wasn't trying to suggest that Smith's 45% vs. Jeff as opposed to his 47% vs. Steve is somehow not real for this sample -- it is.

And then there's the interesting table that Pat left out: (sorry can't figure out how to make it come out in clean tabular form)

------ Favorable Ratings for Senate Candidates in Oregon Smith Merkley Novick Very Favorable 18% 9% 13% Somewhat Favorable 37% 33% 33% Somewhat Unfavorable 23% 22% 16% Very Unfavorable 16% 14% 14% Not Sure 7% 22% 24% -------

Good for Smith, bad for Ds: Smith has a reservoir of good will into which he might be able to tap. His "very favorable" level is twice Jeff's and 1.5 times Steve's.

Bad for Smith: He's losing 8-10% of those favorable to him, apparently.

Good for Jeff (See also "Bad for Steve"): His vote in sample vs. Smith is at the same level as his favorables (42%).

Bad for Jeff: His unfavorables are 6% higher than Steve's and only 3% less than Smith's. His favorables trail Smith by 13%.

Good for Steve: His favorables are 4% higher than Jeff's and trail Smith's by only 9% compared to Jeff's 13%. His unfavorables are 6% lower than Jeff's and 9% lower than Smith's. He has a small reservoir of good will he might tap.

Bad for Steve: Despite having a 4% favorable advantage over Jeff, his vote in sample vs. Smith is 1% lower than Jeff's.

Inferences: 1) On favorability ratings, Merkley in this sample clearly is the weaker of the two Democratic candidates. But that doesn't translate into weaker in votes vs. Smith.

2) The argument that Steve's style will make voters dislike him more than Jeff is not borne out by the data in this sample, though both still have a large % not sure. N.B. Jeff would have to have his unsures break considerably more positively than Steve's to end up with favorable parity.

3) Steve is having some trouble converting favorable views of him into votes vs. Smith. Possible reasons: Score one for LT, i.e. some people like his ads but don't find them a reason to vote for him? Disability sympathy or desire to avoid appearing unsympathetic? Some people like both Smith and Novick but like Smith more?

4) This combination might be read to provide some evidence for the Merkley thesis of Steve having "electability" troubles, but if so, it's not for the reasons Merkley supporters hypothesize (acerbic = higher dislike), at least not yet.

5) I haven't had time to try to figure out exactly how this relates to timing of ads, but Jeff's higher negatives could reflect effects of Smith attacks, backlash against negative ads against Steve, or both. Steve's negatives could reflect effect of Merkley & Smith attacks, anti-disability feelings, or dislike of his ads. Equal level of "very" unfavorables may mean = "very unfavorable to any D."

Bob is right, Kevin and Pat are wrong. As Merkley partisans have happily ignored all weekend despite having it explained over and over, BOTH CANDIDATES ARE IN TIES WITH SMITH. THE SMITH VARIATION IS MoE NOISE. THE SIGNIFICANT DIFFERENCE BETWEEN NOVICK AND MERKLEY IN THE POLL IS NONEXISTENT.

And when you are farther behind, you have more ground you can make up. Spinning the poll as if making up more ground represents "momentum" rather than "catching up," when the opponent is ALSO moving higher, is silly.

Jeff Alworth, who are you referring to about the "math suddenly mattering?" If it's me, you're off base. In fact, perhaps you might ask Kevin Kamberg why general election polls suddenly matter in a primary race, when for months he was saying they don't, and now suddenly a mathematically insignificant difference in a general trial heat becomes evidence Merkley is surging.

I recall that when Novick was leading in the polls, talk of "margin of error" was verboten by certain partisans. Now I see the math does matter.

Last time I checked, Novick has led Merkley in every head-to-head poll that we know about. It has always been the Merkley people who claimed the polls had methodology issues, etc. Now that the head-to-head polls show Merkley within two points, suddenly the methodology is no longer an issue even though it hasn't changed.

Can I persuade anyone with column-writing privileges to pose a challenge along these lines: What's the best way to go after Smith..........

Chris, that would be you. Submit a column. I think most of us that participate in the comments have developed some respect for your POV and your writing, and I'm sure that the Blue Crew would love to put it up........

Come on TJ, Kevin isn't saying that the current difference represents a surge. That's clearly a reference to Jeff having closed the gap vis a vis Steve since February.

Where he and Jeff A. are spinning, at least a little, is in refusing to acknowledge that Jeff had closed that gap by March, and that since March, in the public polls, Steve and Jeff have both gained at similar rates.

Not following you, Chris. Kevin is saying the SUSA and Rasmussen polls represent surges from before. Previously the argument was that it was simply a matter of name ID. Well, when you are further behind and it's due to name ID, you have more points to make up than your opponent--and you're making up points based on familiarity. Novick's earlier numbers were dismissed as the effects of name ID. Why is this result not subject to the same principles, given that Merkley was the only one running ads for most of the interval?

Where he and Jeff A. are spinning, at least a little, is in refusing to acknowledge that Jeff had closed that gap by March, and that since March, in the public polls, Steve and Jeff have both gained at similar rates.

How so?

Merkley has gained 10 points on Smith since March.

Novick has gained 5 points on Smith since March.

Where's the spin?

TJ, I'm not disagreeing entirely with you. I'm just less interested in inconsistencies in pro-Merkley interpretation and trying to stick with a consistent view that poll numbers are poll numbers.

I'm trying to take on the "trend line" / "average rate of gain since February" argument more directly. In the Rasmussen polls, the rate of change in positive voting since March is the same for both candidates.

Between February and March Jeff essentially catches up with Steve. Since then, in terms of gains in positive willingness to vote for a candidate, both have advanced at essentially the same rate. If Jeff is "surging," so is Steve. In both cases it's probably due to increased recognition, name and otherwise, in the context of Smith's relatively weak support, for an incumbent.

For Jeff, March to May is a decline in rate compared to his Feb to March rate, but that's essentially an artifact of his starting so low, just as is his higher average rate over the whole period, as you have repeatedly and correctly said.

For Steve, his gaining the same amount as Jeff from March to May, after being flat from Feb to March, is an increase in his rate of gain.

Insofar as a surge means an increase in rate, if anyone's been surging since March, it's been Steve.

But I don't think the concept of "surge" usefully describes the situation vis a vis Smith in any way.

So I don't want to claim it for Steve. I don't see either of them as surging. I see them both has having made steady but not yet sufficient gains of the same proportions since March, i.e. in the period when things have gotten more serious.

Possible surges in head-to-head numbers involving Merkley, Neville and Novick is an entirely different argument, about which I have no reliable information (if SUSA addresses this, that's because I haven't looked at it).

Novick's earlier numbers were dismissed as the effects of name ID. Why is this result not subject to the same principles, given that Merkley was the only one running ads for most of the interval?

Not sure where the dismissal is occuring TJ. Novick was out there opening beers with his hook, talking like a pirate, and flinging bon (and not so bon) mots for months beofre Merkley started airing his ads.

This gave Steve greater name familiarity. Jeff started running his ads and that gave him greater name familiarity than he had previously enjoyed.

Now, if you believe your lying eyes, Jeff is in a slightly better position than Novick vis a vis Smith. If the information from the polls is now somehow irrelevant or misleading, then there must be a reason that they weren't irrelevant or misleading for the past several months when they were being spun by your crew as favorable to your guy.

Just sayin'........

You seem to be fully on the right track to me, Chris. If there was a disagreement I didn't see it, or you don't quite see what I was saying--but it doesn't seem like we're saying different things. You might be taking a different angle, no less valid.

The bottom line:

"Electability" should not a be a major factor in deciding the primary; both candidates show strong promise against Smith. The extent of differences has remained marginal, whether Novick or Merkley was numerically "ahead."

"Now, if you believe your lying eyes, Jeff is in a slightly better position than Novick vis a vis Smith. If the information from the polls is now somehow irrelevant or misleading, then there must be a reason that they weren't irrelevant or misleading for the past several months when they were being spun by your crew as favorable to your guy."

I don't have a crew, but I know I wasn't doing any such spinning with a single poll. When three or four general election polls in a row show the same phenomenon, then one gains confidence in the result. But I was never foolish enough to suggest that Novick had a statistical advantage in any of those polls.

It's always better to finish up than down. But the Merkley campaign's dishonest parsing of this one poll--particularly after dismissing pretty much any poll that didn't show him ahead--stretches what the science allows them to claim with confidence.

Don't think there's base disagreement TJ.

Pat, what do you make of the Favorable Ratings table on which you did not comment? (Read here if my TypePad formatting problems make my version above too hard.)

Or of my alternative reading of the slight differences that I do agree exist, in this sample? (A sample is only one sample, but it is a sample.)

Pat, I will try, but my recent experience with guest column submissions has not been encouraging. Probably just the press of higher priorities with the elections upon us.

Per Kevin:

Where he and Jeff A. are spinning, at least a little, is in refusing to acknowledge that Jeff had closed that gap by March, and that since March, in the public polls, Steve and Jeff have both gained at similar rates. How so? Merkley has gained 10 points on Smith since March. Novick has gained 5 points on Smith since March. Where's the spin?

Kevin, first, please note that I specified that this argument is reasonable, but is not the only reasonable way to try to understand the poll results.

My argument above refers to absolute gain in voting support at the time of the poll. Merkley has gained 8 points, Novick 6 points. The were separated by one point in March and remain separated by one point in May, though the advantage of that one point has shifted to Merkley.

It appears that Merkley has thrown a little bit of Smith's support into the undecided/other column in May, but I don't think that's the same thing as positive voting support. I'd listen to arguments about why that's wrong.

The spinning comes in two parts. One is only considering the most favorable way to read the evidence, both in these numbers and also the exclusion of the Favorable Ratings table which pretty clearly favors Novick though with complexities noted above. I.e. no effort to consider interpretations against interest, as I did in looking at the Favorable Ratings.

The other spin is in the tendency of both you and Jeff A. to construct exclusively an overall rate of gain / trend line using February as a baseline, whereas Jeff had a slower rate of gain from March to May than from February to March, while Steve's rate of gain increased in comparing the two periods. Not entirely unreasonable, but arguably not the most reasonable interpretation and certainly not the only reasonable one.

Both Jeff's slowing rate of gain (spin bad) and the steeper average rate of gain / trend line (spin good) are artifacts of his very low February starting point. Merging the two periods elides the changing character of the campaign as the election approaches. The dynamic of the first period was Jeff catching up to Steve in basic name recognition, and Steve's support staying flat. The dynamic of the second period was both candidates building up new support more slowly than their first periods of rapid gain (Steve's earlier due to earlier advertising) at comparable rates.

This isn't the only reasonable way to read the data, but it is a reasonable way to do so, & not acknowledging it is spin.

Yes: Chris Lowe, please submit guest column. Yes, yes, yes.

The dynamic of the first period was Jeff catching up to Steve in basic name recognition, and Steve's support staying flat.

So you are saying that without even running TV ads, Jeff's name recognition was narrowing the gap? Interesting.

I think you've just made the same point that Merkley supporters were making back then about the relative effectiveness of Steve's ads. In fact, the exact opposite occurred which one would expect to occur where only one had any media buys - the one who didn't buy any gained ground faster for the period than the one who did.

This isn't the only reasonable way to read the data, but it is a reasonable way to do so, & not acknowledging it is spin.

I gotta be honest here, Chris. I find your assertion there both quixotic and intriguing. Am I to understand you as asserting that the failure to tout a competing explanation, even if it's "arguably not the most reasonable interpretation," as you phrased it, is itself an act of spin?

Kevin,

Yes (compensating for your rhetorical spin*) I am saying that failure to acknowledge the existence of competing reasonable explanations is an act of spin.

This is partly based on a long-held view that arguments are more persuasive if they take on strong versions of competing arguments or claims.

The question is entirely what kind of discussion you, or we, want to have. Do we want to understand the poll as well as we can, since it has a bearing on what happens after the primary? Do we want to try to persuade one another? Is whatever we're doing here in some sense collaborative, about co-thinking and improving one another's thinking & articulation?

Or is it really just about trying to establish some kind of rhetorical dominance? Get the last word, and you win?

I guess one question for me is whether you think the overall BlueOregon audience is less decided at this point than the frequent commenters. If you really think that I can sort of see a point to putting out one-sided arguments, or strong advocacy if you prefer, though it's also the advocacy context that makes it spin.

My own feeling is that we're mostly past that point.

Anyway, I really am interested in thinking about what the polls mean & how you'd answer (as opposed to rhetorically deflecting back) alternative views.

Also, with apologies for my convoluted writing, what I was saying was that it was your "straight line from February" approach was "arguably not the most reasonable approach," though not entirely unreasonable.

*On "touting": you're spinning again when you turn "failure to acknowledge" into "failure to tout." You're a good enough writer of English that I know you know the difference between acknowledge and tout. I expect you chose tout exactly because it has negative connotations of overstatement, and in more specialized contexts (e.g. "labor touts") perhaps deception. So you shade a reasonable point over into unreasonable territory & then argue against the straw man.

I.e., you're spinning like the end of a rinse cycle -- which can be almost as hard on straw men as flying monkeys. :->

[Disclaimer: Although at times I may seem to preach like a pompous and sententious twit, that's actually my evil-twin demon possessing me, and he speaks only for himself.]

[Disclaimer: Don Quixote speaks for me.]