So...What's with Hibbits?
Jeff Bull
I don't know that this is totally kosher, but, if possible, I'd like to turn something from this morning's paper into a teaching moment about polling and commenting on them.
Here's the big question: What do people know about Portland pollster Tim Hibbits?
I'm thinking about this as I read some polling results that ran in this morning's Oregonian on Portland Mayor Tom Potter's first days in office. I have no cause to dispute his findings generally, but Hibbits' quoted characterization of one aspect of the poll - a Portland City Council proposal on public financing for city campaigns – is either taken out of context or it's well over-stated.
By way of (quick and dirty) background, the proposal currently up for a vote would put public financing in place on a trial basis until 2010; Portland voters will pass judgment on the system at that time. And here's where Hibbits gets a bit weird. Asked to comment on that delay, he states:
"If it were placed before the public now, it would lose and lose badly."
But that doesn't mesh so well with the poll results as the Oregonian records them:
"Fifty-six percent of those polled said they oppose the campaign-finance plan."
OK, so far so good. But then you move to the polling sample size and margin of error:
"Hibbitts and his colleagues at Davis, Hibbitts & Midghall Inc. polled 300 registered voters over the weekend and on Monday. The margin of error in their survey is plus or minus 5.6 percentage points."
So, if we're looking at 56% opposed on a poll with a plus/minus 5.6% margin of error, how does that translate to "[losing] and [losing] badly?" The raw numbers suggest to me that it's anyone's ball game.
I dredged up some material on the 2004 elections that raised issues with how he characterized a withdrawn endorsement, as well as a national piece in Slate that contains the following:
"In the pollster Hibbits' disdainful tone about Portland's Democrats, I hear the resentment of the rest of the state that has led Oregon to cast more and more Republican ballots in each of the past three presidential elections. Portland's Democrats have become extreme 'latte leftists,' 'pathetic' rich wackos and oddballs, Hibbits tells me. His town is turning into San Francisco, he complains. 'San Francisco's politics no more represent America's than some rural county in Utah. Now [Portland is] heading off in the same direction.'
So, I suppose the question I'm asking is, how persistent a pattern is this? How much does his apparent "beef" with liberals color his commentary?
Anyone who knows more about this, feel free to chime in. It's not that I think affiliation with a given party and polling don't mix, but do suspect that it's possible that you can judge the poll and commentary on their separate merits. I'm also willing to entertain the possibility that the Oregonian didn't provide Hibbits space to expand on that comment; perhaps there's something in the follow-up sampling that points to firmer opposition than support. But based on raw numbers, I’m not seeing it.
 |
More Recent Posts | |
Albert Kaufman |
|
Guest Column |
|
Kari Chisholm |
|
Kari Chisholm |
Final pre-census estimate: Oregon's getting a sixth congressional seat |
Albert Kaufman |
Polluted by Money - How corporate cash corrupted one of the greenest states in America |
Guest Column |
|
Albert Kaufman |
Our Democrat Representatives in Action - What's on your wish list? |
Kari Chisholm |
|
Guest Column |
|
Kari Chisholm |
|
connect with blueoregon
1:06 p.m.
May 18, '05
I can't speak to Tim's current political views but I do know that before he and Adam Davis became independent pollsters they were the top Democratic party pollsters in the state. I think their last big political job was polling for Barbara Roberts in her gubernatorial race against Dave Frohnmayer.
If Tim is cranky with "latte leftists" it may be because he feels the Democratic Party he used to work for is being pulled away from the mainstream. (As a Republican, I know the feeling.)
May 18, '05
Thanks for that perspective...very helpful. And I can definitely relate to his frustration. But that only deepens the mystery. Absent some kind of political beef, what's with the slam?
May 18, '05
Glad you brought this up. My experience (as a Marion County resident I sure am not a "Portland Democrat", more like a mid-valley resident who leans independent)is that in recent years any Hibbits poll reporting beyond "The counties to look at on election night are Washington, Clackamas, Marion..." have been skewed.
I don't mean ideologically biased so much as "I'm Tim Hibbits and you are supposed to unquestioningly accept my poll numbers".
Too often I have been with friends, at work, at a gathering or whatever and heard a particular point of view ("this measure sounds interesting, no use for that candidate" etc.) which went against Hibbits polling. I seem to recall he was biased in favor of "Accept that Bill Sizemore is a power in Oregon" as were many in Portland media saying things in the mid-1990s like "Bill Sizemore, who will have a measure on the ballot this fall, said today..." as if that weren't an in-kind contribution to Sizemore and collecting signatures was only a formality in the days before signature gathering was closely scrutinized.
And as often as not the conversations I overheard were as close to the truth as Hibbits polling.
Oh, and one more thing--skepticism. I know people who refuse to answer polls, write down the questions and try to call someone they know on a campaign and say "are you aware of this poll?", say "they never call me", people who say margin of error isn't enough--they'd want to know age and geographic diversity as well as sample size, refusal rate, wording of questions before they would believe a poll.
Not to mention those who have skepticism for the "polls are Gospel" crowd, esp. those on election night with thin "of those who attend church once a week..." "demographics" variety. No,it was a "sample of people who answered the poll and said they attend church" without making clear that a Baptist, Methodist, Congregationalist, Presbyterian, Catholic, Greek Orthodox, Jew, Muslim might all answer differently.
Perhaps Hibbits is getting sour as he ages like Lars did (was a good TV reporter once upon a time).
And as far as the Oregonian, don't forget they fired their polling firm shortly after the 1992 primary.
Whose idiot idea it was to put polls on the front page of the Oregonian the day before a polling place primary has never been revealed. But they had these 4 polls on the front page, ALL 4 WERE WRONG, and one of those races ended in a statewide recount.
By and large I don't trust polls.
May 18, '05
I don't want to weigh in on the "what's with Hibbits" question, but regarding his "it would lose badly" comment, I think he's on pretty solid ground.
He reports the margin as 56 against, 32 for, with a margin of error of close to 6%. Take the 6 from "against" and give it to "for" and you're still at 50% no, 36% yes. Good idea or bad, Hibbits is right: it loses at the polls today.
May 18, '05
Ah, now we're cooking. I just re-read the Oregonian's article trying to find the 32% number Patrick Allen provided. I didn't see it (which isn't the same thing as saying it's not there).
This seems to switch the question from "What's with Hibbits?" to "what's with the Oregonian?" Provided that figure, Hibbits' comment doesn't look so out of place.
Then again, take a look at the post by Kari Chisolm further down the page about the effect of the question asked on polling. It's worth a think anyway....
May 18, '05
Pundits are pundits -- not academics. Pollsters often start as one (with academic rigor) and grow into the other (pundits with excess ego that overwhelms rigor).
They start to talk, then they draw conclusions that aren't solid. People keep hiring them, media like to call them, because they provide strong answers, not weak "unclear due to the data" opinions.
I'd say that at least half of what pundits/pollsters/campaign managers say is bs -- it's part of that culture.
2:12 p.m.
May 18, '05
Tim's been a little cranky lately, but of all the polls not commissioned by campaigns I've been involved with, Tim's have been the only ones I've ever really paid attention to or valued.
He's actually got a pretty consistent track record of accurately reading voter opinion and likely voter behavior. Having said that, some of his analysis has been a little off during the past cycle.
Specifically, on the day in which both Bush and Kerry held rallys here (Bush for a limited and closed event for supporters; Kerry with a packed, 50K+ mega-rally with a beautiful backdrop of the river), he opinined that he thought Bush got the better end of the news cycle. Not only was he off on the coverage, his analysis completely skips over the grassroots benefit of pumping up 50K supporters to get out there and work for you. Even from just a voter contact perspective, 50K is a huge amount of voter contact for a day.
Don't think its a conspiracy, just think he was wrong. And as we know, Kerry won Oregon by a reasonably healhhty margin.
May 18, '05
Hibbits might indeed be getting sour.
And I believe that Hibbits is a moderatish Republican. Adam Davis is a Dem, but I don't believe Hibbits leans that way.
And this idea of super-liberal Portland politicians is largely wrong. Most Dem candidates move to the middle -- where the money is -- well before they are ever elected.
The entire country has moved rightward (thanks the the purchase of the media and the investment of millions into right-wing leadership development and think tanks), and so anything liberal looks ultra-liberal by comparison. But what we have now in history is a moderate Democratic Party and an ultra-conservative (and successful) Republican Party.
--Taxes on the wealthy have gone down - not up --Environmental protections have been relaxed - not strengthened --Affirmative Action has receded not been strengthened. --Spending per pupil has gone down --We've voted for two armed conflicts in the last 15 years. (Much more than in the immediate wake of Viet Nam).
And Portland might have resisted these changes, but they haven't pushed many super-liberal things, unless you count money for schools and marriage equity.
As for public campaign money, it's good for the voters and for Democracy. If the voters don't like it, they'll be able to get rid of it. But the advantage of it is that elections are currently bought most of the time. We will never move far forward if we deal with that.
Potter is aces.
Time for me to shut up now.
2:15 p.m.
May 18, '05
Jeff,
You also have to understand how margins of error work. This is commonly misunderstood.
Suppose you have a poll with 56%, +- 5.6%. What that means is (not describing a bunch of statistical assumptions here, all based on assumed random sampling, unbiased questions, and normal distributions), that if you were to conduct the poll 100 times, we believe that 95 times out of 100, the figure will be between 61.5% and 50.5%.
Or put another way, we are 95% confident that the true number is somewhere between 50.5% and 61.5%.
Now, you are saying this means it is a toss up. No--there is actually a pretty small probability that the true number is down near 51%. We just can't say with 95% confidence that it isn't 51.
For the true figure to be 50.5%, you are in the far left hand tail of the distribution.
It is just as probable that it's 61, a blowout.
It is far more probable (again ... given a bunch of assumptions) that the true figure is somewhere in the 54-58% range.
And I suspect that Tim is also drawing on his insight into the polling results on the ITax, and the last Itax vote, and the failed vote to increase state taxes, and a whole series of indications that Portlanders have become significantly more skeptical of government spending in the past few years.
If that isn't clear, let me know, and I can post up some nice online examples of how sampling distributions operate.
May 18, '05
Hot damn! I got my teaching moment. I knew there was a reason I didn't post this one on my own blog; too few people read the thing to float a question and expect an answer.
May 18, '05
If the City Council doubts Hibbits' polling results they can disprove him by putting Voter Financed Elections up for an immediate public vote.
There is no organized opposition to it yet, and look at the headstart its backers would have on a campaign. The Council can even draft the ballot title.
What are they so afraid of?
The City Council's unwillingness to refer this to voters a validation of Hibbits' poll results.
May 18, '05
Jeff Bull wrote:
"I just re-read the Oregonian's article trying to find the 32% number Patrick Allen provided. I didn't see it (which isn't the same thing as saying it's not there)."
Well, in my copy of the paper it's right there in the chart next to the story in color and above the fold. Now, if you're getting your news via OregonLive, well, that's a whole 'nother thread, isn't it? :)
Also, slightly related to Kari's point, the chart is titled "Publicly Funded Elections."
May 18, '05
Oh my holy hell how I loathe Oregonlive. If there's a worse site for a media outfit of that size on the frickin' planet I haven't visited the thing and it should, by all that's right and good, be killed with due haste and extreme, Louisville-slugger-supplied prejudice. All other phrases I would use to express my outright hatred for that site would get me arrested.
May 19, '05
"To be honest with you, I'm a little suspect of his motives."
This Kulongoscopy is killing us. Randy Leonard for Governor.
9:49 a.m.
May 19, '05
To paraphrase Lloyd Bentsen: I know Tim Hibbitts. Tim Hibbitts is a friend of mine.
I think Tim is the best pollster in Oregon, by a fairly large margin. He used to do candidates, working for folks like Bud Clark and Neil Goldschmidt (back when he was believed to be a progressive reformer and not the evil SOB we know him to be). He has quit doing candidates and now has mostly corporate and media clients.
To the Tim that I know, reputation and integrity is everything. He (appropriately) takes pride in the fact that is correct in his calls more than anyone I know of. He is not about to slant questions to blow smoke to a client.
Polling is a tricky business as evidenced by Kari's other post on this topic. I believe that Tim is the best at that tricky business in this state.
Regarding the Slate quote, my rule of thumb on the accuracy of quotes is that if a reporter doesn't care enough to spell the quotee's name correctly, he or she is likely to get the quote wrong. The Slate "reporter" failed this test by putting too few t's in Hibbitts. The "'pathetic' rich wackos and oddballs" line just does not sound like Tim to me.
Finally, Tim is one of the funniest people I know, a great outdoorsman and an absolute fanatic college football fan. That's my two cents on my friend Tim Hibbitts.
May 19, '05
So, Jack Roberts is cranky [too] because the Democratic Party is "being pulled away from the mainstream." I imagine he is apoplectic over the radicalization of his own party.
while Portland Democrats are generally more progressive than the national average [bless their hearts], elected Democrats from Portland and environs are clearly more moderate than their constituents. Contrast this to the whigged out extremists controlling the Republican Party and legislative caucuses in Oregon, who are all decidedly farther right than the average Oregon Republican.
I wonder what specific "away from the mainstream" movements are causing Jack's crankiness. Might support for ending the corruption of city government by special interest campaign contributions be among them? How about the effort to join the other major west coast cities in supplying citizens the lower cost of publicly owned electric power? Maybe the desire to see no more gutting of public education galls Jack. Or maybe he is cranky about the idea of requiring health insurance plans to cover women's reproductive health needs. I bet the idea that gay people should have the same rights as straights really ruins his day.
If such things make Jack cranky I hope he has a really out of sorts future ahead of him.
11:58 a.m.
May 19, '05
I think it's important to keep in mind that 56+32 = 88, not 100. There are no undecided votes cast at the theoretical polls. Therefore, I think it's unwise to suggest a definitive outcome (as I think "losing badly" surely does), where a near-mathematical tie is within the realm of possibility.
Taking the extreme of the bias tolerances, it's true that a 50-38 result is possible although pretty unlikely. If you add the remaining 12% to the "yes" side, you get nearly 50-50. So while a "victory" is the better probability bet, depending on the definition of "badly" I think Hibbitts overstated, "badly."
In my personal estimation, based on a decade's worth of work at a survey research house and being a poll junkie in general, nobody should say anything stands to lose badly unless one side is polling at or above 60%--especially when undecideds are more than 10%, and where sample sizes are relatively small and the question is relatively new. How many people really have any idea what Council just passed? How it works, how much it will cost, who gets to use it, who has pledged not to? Asking an uninformed populace to give opinions on the subjects they're not fully informed about, is not necessarily a good idea--at least if you plan to draw future conclusions from them.
May 19, '05
"So, Jack Roberts is cranky [too] because the Democratic Party is "being pulled away from the mainstream." I imagine he is apoplectic over the radicalization of his own party."
Read for meaning, Mr. Civiletti. This is what he said.
Are your further comments specific to Jack Roberts, or is this a generic anti-Republican slime? Personally, I think it's a shame and a dishonor to smear Jack Roberts of all people this way.
Jun 2, '05
Tim is wrong. I am all for taxpayer funded campaigns. In fact, the first 1500 people to send me a $5 check will get a $100 check in return as soon as I am accepted as a mayoral candidate. If involved in a runoff, I will send another $100 dollar check to my first 1500 supporters. The other $150k will be spent on Kulongoski style bus to tool around the neighborhoods and mingle with the unwashed masses.
<hr/>