A New Poll on the Eve of the Election

Public Policy Polling has released the results (pdf) of a poll for the statewide primary races conducted from May 17-18th. For President, the poll has Barack Obama with a large lead over Hillary Clinton:

Barack Obama is likely to win an overwhelming victory in Oregon, according to Public Policy Polling’s final survey in the state. He leads Hillary Clinton 58-39.

Obama has likely pretty much won the election already. Oregon votes by mail, and he is polling at 59% with the 74% of poll respondents who reported already having cast their ballots.

US Senate:

The hotly contested primary for the US Senate is going right down to the wire. Steve Novick has a 38-33 lead over Jeff Merkley, a lead within the poll’s margin of error. Candy Neville is at 6%, while none of the rest of the candidates exceed 1%. 19% remain undecided even on the eve of the primary.

Attorney General and Secretary of State:

There is more clarity in the state’s races for Attorney General and Secretary of State. John Kroger leads Greg MacPherson 42-31 for AG, while Kate Brown is the frontrunner for SoS, polling at 42% followed by Rick Metsger and Vicki Walker at 19% and 17% respectively.

Read the rest (pdf). The margin of error is 2.7%.

Discuss.

  • Randle McMurphy (unverified)
    (Show?)

    Two words: automated poll.

    Face it folks, with the exception of the presidential race, we don't know what is going to happen tomorrow night.

  • (Show?)

    two more words:

    "so what?"

    Automated polls are just as good if not better than human polls, we're finding. SUSA and Rasmussen have done very solid jobs in the 2006 cycle and so far this primary. PPP has also been decent as well.

    This has to worry Merkley. First Hibbits, now this? If the SUSA result that is likely to come out tonight shows Novick up too, that's going to be a hard spin, isn't it? "Hibbits is wrong, PPP is wrong--and so is the polling outfit we said was so awesome a week ago!" (Or maybe it will have Jeff moving further ahead on the strength of all that momentum they claim they have; guess we'll see).

    And what chilling words--Novick's vote is tied to Obama's. If they called folks yesterday, about 75,000 Obama voters weren't home. :)

  • Other Info (unverified)
    (Show?)

    The poll is not a scientific poll.

    It is just a random sample.

    They don't have cross tabs via region or voting record.

    You can take this poll with a grain of salt.

    Also when you realize that this automated poll came out with all of these Robo-Calls you can see why most people would hang up on these calls.

    Personally I think this is really just a poll of activist. They are the only folks who would respond to these kind of polls this close to the election.

  • (Show?)

    "The poll is not a scientific poll.

    It is just a random sample. "

    You do realize you just said: "This is not a scientific poll. This is a scientific poll," right? A random sample goes a long way towards what makes a poll "scientific."

  • (Show?)

    I guess I'm just going to be skeptical of any poll that was done over a weekend that included a) the first hot weekend of the year, and b) 75,000 people at a Barack Obama rally.

    I'm not sure what the effect would be, but I'm guessing that some demographics would be more affected than others by the hot weather (parents with kids outdoors? seniors taking shelter at the mall?).

    And 72,000 Portlanders at an Obama rally? It's gotta suppress the Obama vote in the poll - and who knows what the effect would be on the rest of the races.

  • (Show?)

    Here are the two other polls released today:

    Survey USA

    55 Obama, 42 Clinton, 2 other and 1 undecided.

    Suffolk University

    45 Obama, 41 Clinton, 8 undecided and 6 refused.

  • James X. (unverified)
    (Show?)

    It's ridiculous that the Senate candidates are still polling in the 30s, according to this poll. If this poll is accurate, and I'm not taking that as a given, that would mean that either none of the candidates' advertising really made that much of an impact, or people like both Jeff and Steve, or don't like either, or ... something. And if there really are that many undecideds, I think the last-minute factors will be the voter's guide, newspaper endorsements that can be found online, and the last ads and mailers a voter is exposed to. I think Merkley wins with the voter's guide — the first words on Novick's page are "Occupation: Consultant", which is only marginally better than "Occupation: Lobbyist," and you don't do any favors following it up with "Occupational Background: Attorney." In comparison with "Occupation: Speaker of the Oregon House of Representatives" and "Occupational Background: Executive Director, Portland Habitat for Humanity; National Security Analyst, Pentagon and Congressional Budget Office; President, Oregon World Affairs Council," Merkley's page is just more impressive. His photo was printed with terrible brightness and contrast, though, so he looks like he has no lips. In newspaper endorsements, Novick has all the good ones. Novick came out with a last-minute negative mailer on Merkley. I don't know what Merkley's been mailing out lately, but he's got the TV ad with negative Novick quotes and the positive ad with his daughter.

    I just don't know what will be the deciding factor for most people.

  • Other Info (unverified)
    (Show?)

    TJ,

    Where did you get your degree in statistics?

    No really. Where? Please tell us where you gained all this knowledge about scientific polling from.

    Please Define: Validity, Optimality, Invariance.

    Please tell us how this poll did or did not achieve these criteria which you need to meet inorder to be a scientific poll (HINT: the poll only achieved one of the three).

    So come on smart guy. You like to write and apparently you like to put words in other peoples' mouth.

    I'm not going to do the same thing to you rather I will just give you an opportunity to prove to us if you are as smart as you say you are.

    I look forward to reading your response.

  • (Show?)

    "And 72,000 Portlanders at an Obama rally? It's gotta suppress the Obama vote in the poll - and who knows what the effect would be on the rest of the races."

    According to the pollster, the effect on the Senate race would be that Novick's vote is being suppressed as well.

  • 18yearoldwithanopinion (unverified)
    (Show?)

    This poll is quite different from all other polls, with the expection of the presidental and senate races. Looks like Obama will win and the senate race will be down to the wire. Did the new Kroger ad really help him and the latest Greg Mac ad piss of people? I would have to say the Secetary of State race polling probably is heavily influenced by the huge ad buys on both radio and TV by kate brown. I can't wait till tomorrow.

  • (Show?)

    "Where did you get your degree in statistics?"

    Degrees are in Sociology, actually--Bachelor's and Master's both from William and Mary. My Master's thesis was based on survey research. I spent a decade working for VCU's Survey Research Lab, the first part of that under the tutelage of Scott Keeter, now Asst. Director at Pew Research.

    As part of the curriculum I had to do a stats course for both degrees. But my degrees are in social science, strictly speaking.

    When you say validity, do you mean internal or external? :)

  • trishka (unverified)
    (Show?)

    i think i got called for this poll. some automated poll or another called me on saturday afternoon w/ "press 1 for hillary clinton, 2 for barack obama" &c.

    i was more than happy to skew the results novick's direction.

    heh.

  • 18yearoldwithanopinion (unverified)
    (Show?)

    Other info,

    anyone who reads blueoregon.com regulary knows that I almost never agree with Torridjoe. But on this poll I have to agree with him. While I wont say its a legit poll without seeing the crossbars and how it was conducted I wont say its not a legit poll. A poll in order to be legit needs to have a random sample or its not legit. You toss around a bunch of big words but in reality a poll needs to reduce response and voluntary response bias to the highest degree, be a random sample, have a fair method of asking the questions, and have a decent sized sample. Reading the pdf I have no reason to believe that it doesnt meet those guidelines. Reading the pdf I know that its a random sample of a decent size.

    PS- You dont need a degree in stats to read a poll and know whether its legit or not. Any decent college states course or even AP stats course will teach you what you need to both read and conduct polls.

  • Other Info (unverified)
    (Show?)

    TJ,

    So you actually do know something about stats.

    So you're not a blow hard rather you are simply just disingenuous.

    Feel free to keep digging a deeper hole for yourself.

  • Eric Parker (unverified)
    (Show?)

    I hung up on this turkey. My family dinner was rudely interrupted.

  • (Show?)

    The only poll that matters, blah, blah, blah.

    Having said that, as my friend TJ always points out, it's better to be up than down.

    And if a lot of people were out at the Obama event, and Obama trends with Novick, then that suggests that the Obama / Novick numbers might have been understated if lots of people were, say, out at the largest political event in the history of Oregon. %^>

    Tomorrow: bring it on, I say. Let's get this show on the road.

  • Jack Murray (unverified)
    (Show?)

    Since we have a few polls coming out at the same time, I thought I'd show this chart that ranks the performance of the pollster firms in relation to the actual result (I think that's what this is).

    1. SurveyUSA
    2. Suffolk
    3. PPP
    4. Rasmussen

    To see the whole chart, click here.

  • (Show?)

    "So you're not a blow hard rather you are simply just disingenuous.

    Feel free to keep digging a deeper hole for yourself."

    Care to explain what the heck you're talking about?

    Re: the chart of accuracy--note that most of those pollsters have only done a couple of races. SUSA has done best overall, but they've had some headscratchers.

    But if I understand you right Jack, we should take SUSA as the most likely example of reality? Is that what you're saying?

  • Jack Murray (unverified)
    (Show?)

    TJ: Not necessarily, because this chart is based on the Presidential Primary.

    I just thought it was handy to have a way to interpret all of the different polls that are coming out. Or, having at least one source that isn't couched in either party's camp.

  • Lou (unverified)
    (Show?)

    I think that the numbers of undecided voters left in the Senate Race is troubling. I think we are giving these voters too much credit if we assume that they will be using the voter's guide or any last minute mailers to make a decision. I am starting to think that an increasing portion of them may not even vote on the Senate race. These folks may be so hyperfocused on the presidential race that they are leaving the Senate race blank. I will be interested to see how much the presidential race totals dwarf the senate totals in the end.

  • Logan Gilles (unverified)
    (Show?)

    I think these poll results make a lot of sense, including the SoS race. I can't speak to the validity of the process, but the results seem right.

  • trishka (unverified)
    (Show?)

    hey, did anyone else notice that on the PPP poll, women were polled at 54% and men at 46%

    and obama and novick still came out ahead.

    for two candidates who supposedly need to work on connecting with women, well...

    could the actual numbers be even higher? or is 54/46 representative of the oregon demographic?

  • TOTAL BS (unverified)
    (Show?)

    Great point from Ben Rivers on Politickeor.com:

    "I am surprised most by the Sec of State results...This is a huge point swing from the Tribune poll. I am convinced no one knows what is going on. Fun times in Oregon. We will know the answer to all of these questions in a mere 30 hours...

    BTW, these guys predicted a 4 point win by Obama in Pennsylvania...soooo

    http://www.publicpolicypolling.com/pdf/PPP_Release_042108.pdf"

  • (Show?)

    "I am surprised most by the Sec of State results...This is a huge point swing from the Tribune poll. "

    Doesn't look so strange if the number of undecided have truly dropped from 43%. Also, it's two entirely different polls, conducted in entirely different ways. Comparing them in a trend fashion is very dangerous.

  • backbeat, woman (unverified)
    (Show?)

    So many times I've felt that Senator Clinton finally jumped the shark, only to be shocked once again. Today she said that KKKarl Rove thinks she should stay in the race. Wow. Dear Senator Clinton, if you mention his name, it should be in a sentence with the word "prison", not one admiring his great polling skills. Disgusting, but not surprising.

  • Gary (unverified)
    (Show?)

    Sorry folks, but the only poll that really matters isn't released until 8pm tomorrow.

    BTW- TJ, you should really stick to what you do best, destroying perfectly good blogs.

  • KJ (unverified)
    (Show?)

    The latest in Multnomah county re turnout: 38% of ballots returned as of this afternoon. Lots of last minute deciders, I think. So far this is trailing behind 2004 when there was not a contested presidential primary.

  • (Show?)

    btw Gary, do you have a fucking point or would you simply prefer to be oblique about it? And anyway, aren't I taking your advice right now? :)

  • trishka (unverified)
    (Show?)

    what i want to know is - what happened to merkley's SURGE?!?!?!?!?!!!

    kidding!

    ::cue dazzling dissertation from chris lowe on margin of error::

in the news

connect with blueoregon