President (To Win Colorado)
See Full Big Line
(D) Kamala Harris
(R) Donald Trump
80%↑
20%
CO-01 (Denver)
See Full Big Line
(D) Diana DeGette*
(R) V. Archuleta
98%
2%
CO-02 (Boulder-ish)
See Full Big Line
(D) Joe Neguse*
(R) Marshall Dawson
95%
5%
CO-03 (West & Southern CO)
See Full Big Line
(R) Jeff Hurd
(D) Adam Frisch
50%
50%
CO-04 (Northeast-ish Colorado)
See Full Big Line
(R) Lauren Boebert
(D) Trisha Calvarese
90%
10%
CO-05 (Colorado Springs)
See Full Big Line
(R) Jeff Crank
(D) River Gassen
80%
20%
CO-06 (Aurora)
See Full Big Line
(D) Jason Crow*
(R) John Fabbricatore
90%
10%
CO-07 (Jefferson County)
See Full Big Line
(D) B. Pettersen
(R) Sergei Matveyuk
90%
10%
CO-08 (Northern Colo.)
See Full Big Line
(D) Yadira Caraveo
(R) Gabe Evans
52%↑
48%↓
State Senate Majority
See Full Big Line
DEMOCRATS
REPUBLICANS
80%
20%
State House Majority
See Full Big Line
DEMOCRATS
REPUBLICANS
95%
5%
June 22, 2010 11:04 PM UTC
31 Comments
Bennett is still up and will clean the clocks of either….oh how I love Republican infighting
if the poll that showed Buck up is invalid, it’s the same poll that showed Bennet up…which questions the validity of that as well.
Also, these don’t address the general. Earlier polling that did have the 4 potential general election matchups showed Bennet getting beat by Norton or Buck (and Romanoff as the stronger dem candidate).
That said, I love seeing the Republicans fight too 🙂
is against copying entire NEWS articles. An except and a link to the original is the better plan. That way the source gets credit for the hit (page view) on its blog.
That’s a matter of journalistic courtesy and, in the extreme, copyright law.
In your case, you aren’;t copying a news article, you’re copying a news release. I.E., the campaign wrote it, sent it out without copyright, and prays that the whole world will reprint it.
We wouldn’t normally post a full news release if it was felt to be self-serving but I think your quotation of it in this context is fine.
I wasn’t sure so I thought I’d throw in the comment.
That said, what do you think about all this Mr. Newly-minted Front Page Editor? Was there some massive voter shift overnight, is one of the polsters that bad, or who is cooking the voter file books to get the response they want?
Here’s the real point polling on primaries is very volatile, as is issue polling. In the real world, whereever that is, people don’t follow politics the way Those of Us Without A Life do. Primaries pit people most people have never heard of against each other.
In a general election, typical state, about 40 percent would vote for a dead dog with a Republican label and 40 percent for any Democrat with no recent felonies. Goldwater got 40 percent, McGovern got 40 percent. General elections are fought for the one third of voters who might be persuadable. (That’s higher than the 20 percent I imply in the Goldwater/McGovern case but not all of the persuadable third or so ends in the same camp. Goldwater picked up Southern conservatives as Johnson picked up liberal republicans, etc.
In a primary, there’s no such anchor. Buck/Norton, Romanoff/Bennet agree on 95 percent or so of the issues. If Jane says something dumb, it’s easy to shift to Buck and vice versa. Likewise, a clever add may pull support for one Republican at the expense of another.
Issues — do we kill all highway funds etc. — are even more volatile.
So I routinely expect wide volatility in primary polls. It’s not the methodology of the poll itself so much as the total volatility of what is being measured. Absent the gravity of party loyalty, the opinions being measured are just very shallowly based and capable of wide swings.
was badly skewed.
In the 506 likely democratic voters polled for Bennet vs Romanoff,
the poll sampled only 20% of the 65 and older voters with the majority of the percentage from 25-64 (65%)
If you know politics, the MOST active and likely voters – especially in a primary,
are the retired 65 and older activists.
The fact that they only represented 20% of the poll indicates that the most likely to vote was under represented.
and especiailly true for an off-year election. But an all-mail (well, 54 out of 64 counties) primary should have higher turnout across the board, not just in the demographics that always vote.
Referring to the mail in balloting…
Translation: voters are screened for their likelihood to vote for Jane Norton. Plus their “expanded universe” contains more voters who have never voted in a primary. A lot of these people won’t vote in a primary this time, and the only ones that will are Tea Party supporters who, despite Norton’s claims, are solidly behind Buck.
Then there’s this little paragraph at the end:
I think it’s pretty clear who this poll is pulling for. It makes sense that an establishment pollster would be for the establishment candidate.
Finally, the small sample size of only 400 voters leaves room for a large margin of error at 5%. According to this margin of error, it is conceivable that Buck is beating Norton 38-34 in this very same poll. If there was ever a flawed poll, this is it.
…and I think it’s clear I’m not a Norton supporter — I just find this stuff really interesting.
1) Isn’t there some validity to the idea that because of mail in voting more people will vote? That has been tested and proven to have an effect.
2) You say that the only new voters are the teaparty folks and that they support Buck (the 2nd point wouldn’t surprise me in the slightest). Why would only teaparty people be the new voters?
3) As for who the poll is “pulling for” — well unless I read it wrong Norton did pay for it — that’s no secret. That said, the company is reputable and polsters don’t want a reputation of being really wrong all the time/willing to fake polls for candidates. It sinks them big time.
4) Margin of error. I’m pretty sure that a 6 point lead with 5% margin of error means that at worst she has a 1 point lead and at best an 11 point lead. In other words, there is 5% wiggle room.
As to 1) and 2) there is validity to a potentially expanded universe of voters, but remember both the Magellan Poll and the Denver Post Poll had Buck leading every demographic, age, sex, race etc.
3) forget who paid for it, we don’t know how they got to the 400 people, friends and family? This was not a poll done with no purpose in mind. It was conducted the same day the Denver Post poll came out.
4) I don’t think there was a margin of error. I think the error was planned.
If they release the poll we will know. (They won’t) Until then, in my view, it is the Penry poll. He just keeps makin’ stuff up.
That doesn’t matter if it’s qualifying likely voters based on a different model. I’m not saying Norton’s poll is a better model, but if it is, whether Buck led among any demographic under a flawed model would still lead to flawed results.
“a 6 point lead with 5% margin of error”
The 5% margin of error captures 95% of all possible outcomes. The 6-point lead has the maximum likelihood of being correct.
As you get move away from the max likelihood and closer to the tails of the distribution (either a 1% lead or an 11% lead), the likelihood decreases significantly.
In other words, while there may be 5% “wiggle room”, it’s much more likely that she has a 6% lead than a 1% lead.
Thus endeth our statistics lesson for the day.
you explained that in a way that makes a lot of sense. So it’s more likely it’s a 6 point lead and possible but extremely unlikely that it’s either an 11 or 1 point lead (and a 7 point lead is more likely than an 8 point).
Makes sense, and you used cool math-speak.
I’m a math major. Still, people refer to polls as “statistically tied” if it’s within the margin of error. I guess maybe the best possible outcome within the margin of error would be Norton ahead by 1% (with their flawed method). And the sample size is so small that it does lead to a somewhat fuzzy result. I wouldn’t be surprised if they did 3 polls with 400 people and then picked the best one or something.
I would expect it to be higher.
ANd I would expect bigger turnout to help Buck and Bennet.
A lot of primary voters grab their ballot, pull up Google, do some searches, and then vote. That means these polls come before they have read the content that will determine their vote.
as a keyboard macro, David?
It is not up on RCP. It was commissioned by a candidate the same day that they got creamed in a real poll and does not disclose how the 400 people were selected, what questions were asked, etc.
No poll is going to show Norton leading Buck with the Tea Party crowd. I don’t care what they say.
Seems like more Penry disinformation that is likely going to come back to bite him in the butt.
Seems to me the Denver Post, with the integrity of their work questioned, might look into it.
Norton strategy, just make it.
I would be really surprised if Buck wasn’t leading with that crowd. Yes — I would also be interested to read the actual questions, etc.
Was the other any more credible though? These robopolls allow anyone to respond — not just the intended person (a live caller will end the call if the person on their list isn’t available). Anyone who has ever volunteered for a campaign can tell you that a decent chunk of those phone numbers are just wrong too. Add that to older people who just press “1” thinking that it will end the call or whatever, and robopolls are pretty laughable.
So we are left with the question of if either of these can actually be viewed as valuable in any way. And if we can’t look at polling, what can we look at? Candidates who raise more money will say money. Candidates who win caucuses will say caucuses. What is a fair way of looking at how the campaigns are doing?
After the DP poll came out Penry recognized the game was over. He concluded he needed to come out with a poll that showed Norton was leading. I believe the facts will show that the 400 people were not arbitrarily selected, but purposefully skewed to obtain that result.
The Poll will never be released and Norton will still lose by 20%.
Polls are the best predictor and yes there are inaccuracies incident to robopolling but the inaccuracies are random and work out statistically.
The Magellan poll with Buck holding a 10% lead understated his real lead. He does better with men and it had a 50-50 split. The real turnout in a republican primary is about 53 or 54 to 46-47 Male vs Female. Magellan also had Buck’s voters as more enthused and therefore more likely to vote. The Denver Post poll had a 57 to 43 male vs female split so you might look at him leading by 15 instead of 16, but neither poll adjust for enthusiasm. With that factored in, it would suggest a 20% victory in August.
The Penry poll can’t be explained other than by a purposeful skewing of the 400 people contacted to obtain a desired result. Given Penry’s history, that is the only thing that explains it.
Now the problem for Penry is he can’t publicize this poll without it coming back to bite him and it sure won’t be released to the people that track these things.
I guess it allows him to continue to lie to his campaign and pretend they have a chance when the race is already over.
I’d say this is pretty standard. If you have internals that show you ahead, you release them. Especially when the last poll showed you getting your ass kicked.
Here are some differences between the Survey USA poll and this one, though:
1. Much smaller sample size. This poll asked 400 likely Republican primary voters, Survey USA asked 747 likely Republican primary voters. For the Democratic race, the number was smaller than the GOP but still more than this poll at 506.
2. Wider margin of error. I’m not a polling expert (probably better than a lay person, but no expert by any means) but I would say this is due to the smaller sample size. The MoE was +/-3.7% for Buck/Norton and +/-4.4% for Bennet/Romanoff in the Survey USA poll, as opposed to 4.9% for this poll.
3. Money. Survey USA was an independent pollster along with 9News and the Denver Post, whereas Public Opinion Strategies is on Norton’s payroll. Still, they have Buck losing to Norton in other key areas. I think that their methodology of asking the “expanded universe” of primary voters, led to the different results.
But like Voyageur astutely pointed out in the open thread, primary polls are volatile. The only thing you can do is watch them as you get closer to the election. If you spend all your time (and, by extension, money) on polls, then you’re going to get distracted. Norton can’t afford to get distracted right now.
No matter what this poll’s results say, Norton is in serious trouble in this race.
I tend to agree that campaigns will release internals in cases like this (unless the advantage is not a straight up number but rather some demographic data they don’t want out — X type of voters can be persuaded if we say Y and that will give us the margin to win, etc).
Your 1 and 2 are related of course — larger sample size means you have a smaller margin (ask everyone on election day and you know the truth of course). 4-6% doesn’t make up the difference between these though. If both were really tight and had a similar margin, sure; I wouldn’t think much of it. In this case though, the respective polls are their candidates outside their respective margins.
3) This is really interesting and gets at a much larger question — who are “likely” voters. Campaigns will spend a lot of time trying to figure out just that, because they can’t target everyone (well I guess it’s not really targeting at that point).
So if Norton thinks she can get a lot of votes from these marginal voters (phrase I heard during the Obama campaign to mean voters who vote less often), then it makes sense that she sees them as likely — and they will go after them to turn in their ballots.
Additionally, if the age demographics are skewed, it could have really different results.
I agree that Norton has a serious fight, and I think Buck might pull it off; by no means am I saying he won’t. I just find the huge difference in numbers pretty interesting.
the info like who their likely voters are? Wouldn’t that be like releasing their targeting and give their opponents an advantage knowing that they are doing?
as I said above, if it’s something they need to keep under wraps, they won’t (or will be vague). If not, then they might — especially when they need to change the conversation away from a bad poll that just came out.
Between “likely voters” and “persuadable voters.” You have to poll “likely voters” to get reliable results, but identifying those people isn’t really a trade secret. The list of “persuadable voters” is different — that’s the targeting data you are probably asking about. Campaigns won’t usually target all “likely voters.”
I believe “likely voter” is based on a registered voters record of voting in primaries. I know they use this data to construct “calling lists” for campaigns. I believe this same data is used by pollsters. So a likely voter is (for example) a registered voter having voted in the last two primaries, or perhaps 2 out of the last 3 primaries.
Just my two cents worth.
is we need a PPP or Rasmussen poll to come out to get a more firm result. However with Norton’s own pollster saying she’s up a little bit, Buck’s pollster saying he’s up a lot, and the independent Denver Post poll saying Buck is up more; I think it gives us a pretty clear picture. Buck is solidly in the lead.
Agree Buck is in the lead. Disagree it’s necessarily by a lot. Think the Denver Post poll is missing a critical dynamic in his race, which is tea party support from independents identifying as republicans but whom aren’t actually registered. Suspect a lot of people will make that mistake this year.
Also, Rasmussen? Really? Those polls are literally worth nothing. PPP is spot on usually, and I’d welcome another poll from them. Suspect it will show a closer race, as it did last time, and a huge lead for Bennet.
It seems like the fight over Bennet / Romanoff has really been missing the boat. Seems like all the action is really on the Republican side.
The tea party types that are not in the poll are unaffiliated. I don’t see this intention to deceive going on. I do agree with you about the enthusiasm dynamic.
The PPP poll, done pre Assembly had Buck picking up 12 point from their March poll. He went from down 17 to down 5. If they did one today in my view it would show him in the +10 to +15 range. But considering the enthusiasm of his core voters, this will likely translate to a 20 point win.