Polls of Only Self-Described "Definite" or "Certain" Voters Not Necessarily That Predictive

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Friday November 04, 2022 · 5:11 PM PDT

2022/11/04 · 17:11

We’re starting, once again this election cycle, to see some pollsters report results for self-described “definite” or “certain” voters, as well as for registered and, perhaps, likely voters.

maristpoll.marist.edu/…

Two years ago, Iowa pollster extraordinaire Ann Selzer reported in an interview with FiveThirtyEight she only includes definite voters (“they could have said they were probably going to vote… We don’t take that”).

fivethirtyeight.com/…

Monmouth University’s Patrick Murray appeared later on in the same episode. He noted that he had data from six states that included definiteness of voting and other possible markers of voting likelihood (e.g., voting history). As he reported, there was only one state in which using a definite-voter criterion improved the accuracy of the turnout model — Iowa.

In addition, some people who would not consider themselves definite voters may still vote. In fact, Obama was famous for his campaign organizations’ ability to bring out “unlikely voters.”

www.reuters.com/…

Definite-voter models may work occasionally, but I would treat them simply as one piece of data in the mix.

Biden meeting PA benchmarks (per Dave Wasserman)

ProfAlan

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Wednesday November 04, 2020 · 9:12 PM PST

2020/11/04 · 21:12

Within the past hour, the Cook Political Report’s Dave Wasserman has been issuing a series of tweets on how Biden is meeting his benchmarks across several key Pennsylvania counties (either winning them by more than he needs or losing them by less). The link is twitter.com/… Wasserman closes one of his tweets with the prediction: “At these rates, Biden should finish with a healthy lead in PA.”

New SurveyMonkey Poll: 53% Favor Impeachment Inquiry, 45% Favor Actual Impeachment

ProfAlan

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Thursday September 26, 2019 · 10:00 AM PDT

2019/09/26 · 10:00

According to a new Business Insider/SurveyMonkey Poll taken over the past two days, 53% of respondents supported an impeachment inquiry and 45% supported actual impeachment.

SurveyMonkey’s methods warrant some caution, however. The company builds a panel of opt-in respondents from the many surveys it conducts on various topics, draws a random sample of the panel, then applies post-hoc (or post-stratification) weights to bring the sample demographics into line with the population (Census).

I suspect a lot more polls will start coming out in the next few days, so we’ll have other results to compare these to.

Any insights on Nelson-Scott FL Senate race?

ProfAlan

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Monday August 27, 2018 · 10:37 AM PDT

2018/08/27 · 10:37

I neither live in Florida nor have an extensive knowledge of campaign dynamics in the Sunshine State. I’m thus interested to see what those more knowledgeable have to say about the U.S. Senate race between incumbent Democrat Bill Nelson and his challenger, Gov. Rick Scott. The impetus for my inquiry is today’s Politico article on whether Nelson is “choking” away the race (h/t Political Wire).

The article paints a pretty dire picture, claiming national Democrats might cut off funding for Nelson if it begins to look like he can’t win. Publicly released polling is scarce, as shown in this Real Clear Politics chart. The one August poll has Scott up 45-39, but most other polls have had Nelson between 44-49 (with Scott up 1.5% on average). Politico offers only the vague statement that, “It’s been a month since Nelson led a public poll. Private polling, even surveys conducted by Democrats, also show Nelson behind Scott.” The question is, by how much is Nelson behind in these private polls?

Given that Scott won both his gubernatorial terms by essentially 1 percent (1.2% in 2010 and 1.0% in 2014) in strong Republican years, one would think he’d have a much harder time in what appears to be a strong Democratic year. Also, despite Scott’s massive spending and advertising advantage thus far, he has not pulled away in the polls. My hope is that Nelson has been saving his money for the final two months and can gain some advertising traction (albeit probably not parity with Scott’s expenditures). Also, Nelson will need to start campaigning more vigorously and visibly.

Am I too optimistic?

National Polls Varying Widely in Hispanic Support for Clinton

ProfAlan

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Sunday October 30, 2016 · 6:52 PM PDT

2016/10/30 · 18:52

I was just diving into the internals of a couple of daily tracking polls that show the presidential race getting closer. The subgroup figures for Hispanics jumped out at me.

The Washington Post-ABC poll has Clinton pulling in 60% of the Hispanic vote.

Over at IBD-TIPP, things seem even more aberrant, with Clinton getting only 48% of the Hispanic vote.

Latino Decisions, which issues weekly tracking polls of Hispanic voters, has had Clinton’s support at 74%, 75%, 74% in its last three reports (Oct. 10, 17, and 24). It looks like it may be a registered-voter sample, but roughly 90% of respondents have been saying “almost certain” or “probably” regarding their voting likelihood. Also LD’s 500-person sample size (a new 250 weekly respondents, plus 250 from the previous week) would have a lower margin-of-error than the perhaps 100-150 Hispanics in a larger poll of 1,000 (a typical national sample size).

For comparison purposes, Obama won 71% of the Hispanic vote in 2012. Based on the kinds of things Trump has said during the campaign, one would not expect him to do better than Romney.

Keep an eye out for the next Latino Decisions poll (list by dates), coming out in the next day or so.

Survey-mode effects on Clinton-Trump polling

ProfAlan

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Friday October 28, 2016 · 9:15 AM PDT

2016/10/28 · 09:15

University of Montreal professor Claire Durand has broken out U.S. horserace presidential-election poll results by survey mode: telephone, IVR (“robo-poll”), and Internet. The modest tightening in the race appears attributable to IVR polls, according to Durand. Take a look:

ahlessondages.blogspot.ca/…

Bloomberg/Selzer OHIO Poll Deviates Heavily in Pro-GOP Composition Compared to 2012

ProfAlan

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Wednesday September 14, 2016 · 7:12 AM PDT

2016/09/14 · 07:12

The new Bloomberg/Selzer poll, showing Trump up 5 in Ohio (44-39), is bound to get a lot of attention. Ann Selzer is considered a genius at polling her home state of Iowa and many will assume her polls in other states are just as impressive.

We cannot know for sure, but it seems there’s an implausible swing in GOP composition in this Ohio poll. Forty-three percent of respondents considered themselves Republicans (33 + 10 leaners), compared to only 36% self-identified Democrats (29 + 7).

assets.bwbx.io/...

In the 2012 exit poll, there were 7 percentage points more Democrats than Republicans, 38-31.

www.cnn.com/…

I know we’re not supposed to second-guess each individual poll that comes along. Still, a 14-point swing (from D+7 in 2012 to R+7 this year) just doesn’t seem plausible. In fact, the 2012 figure is probably more reliable than the 2016 poll, as an exit poll, by definition, consists of people who actually showed up to vote.

Former Rep. Don Edwards (D-CA) Dead at 100

ProfAlan

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Friday October 02, 2015 · 3:57 PM PDT

2015/10/02 · 15:57

When I began following politics really closely as a college student in the early 1980s, one of my favorite members of Congress was Rep. Don Edwards (D-CA) from the San Jose area. I didn't live in his district, but rather, I saw Edwards regularly champion progressive causes in floor speeches and other appearances on C-SPAN.

Every so often, I wonder what became of former politicians who were in their 60s or 70s during the 1980s (of course, I would expect few of them to be alive today). Rep. Edwards, it turns out, lived to 100. I know that from the articles today announcing his passing. This obituary, from Edwards' home-city newspaper, provides a good overview of his career:

http://www.mercurynews.com/...

Sat. night polls mostly bad, but glimmer in Louisiana

ProfAlan

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Saturday November 01, 2014 · 5:36 PM PDT

2014/11/01 · 17:36

The shot heard around the polling world tonight is Joni Ernst's 7-point lead over Bruce Braley in the final Des Moines Register poll.

Though DMR pollster Ann Selzer has a record of accuracy in Iowa surveys, this poll is at odds with some other recent ones that have the race within 2 points either way.

Public Policy Polling (PPP) is on a weekend marathon of polls to close out this election cycle and its results are likewise disappointing for the Democrats.

Arkansas (Cotton 49-41) and Kentucky (McConnell 50-42) seem gone for the Democrats.

One glimmer of hope for the Democrats from PPP is that Landrieu leads Cassidy 43-35 in the November open primary, and she trails only 48-47 in a hypothetical run-off.

Is Alaska Senate race shaking up dramatically?

ProfAlan

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Friday October 24, 2014 · 3:24 PM PDT

2014/10/24 · 15:24

Earlier today, I saw that Sen. Mark Begich (D-Alaska) had pulled into a 44-44 tie with challenger Dan Sullivan (R), contrary to virtually all recent polls that had had Sullivan ahead (generally in the mid-single digits).

http://elections.huffingtonpost.com/...

However, this poll was by Harstad, a Democratic-affiliated outfit.

Now, via Political Wire, I see there's a new poll out showing Begich leading 48.6%-39.2%.

https://www.scribd.com/...

Needless to say, I'm skeptical. I know nothing about the pollster, Hellenthal & Associates. I highly doubt Begich is really up 10, but maybe the race is narrowing.

Could a major shift be going on?

Braley 48-47 in New PPP/LCV Poll

ProfAlan

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Friday October 17, 2014 · 9:37 AM PDT

2014/10/17 · 09:37

This new poll, in the field Oct 15-16, shows Braley taking a slim lead:

http://www.lcv.org/...

One dynamic that may be helping Braley is that he's holding his own among age 65+ voters, trailing by only two percentage points. Male and female voters don't seem to differ much, though, which seems odd. On the whole, the linked "internals" don't go into as much detail as usual.

This ain't good: Des Moines Register Poll Has it Ernst 44-38

ProfAlan

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Saturday September 27, 2014 · 8:00 PM PDT

2014/09/27 · 20:00

A new Des Moines Register poll has GOP Senate candidate Joni Ernst leading Democrat Bruce Braley, 44-38, in the race to replace retiring Sen. Tom Harkin (D).

The DMR poll is conducted by Ann Selzer, who is considered the expert of polling Iowa. In 2012, her final poll had Obama over Romney by 5 percentage points, and the president carried the state by 5.6.

I suppose the only good news for Braley is that such a high percentage of respondents in the poll were undecided or favoring a minor-party candidate.

I could not find a supplemental document of poll internals, but the main article and a sidebar provide some.

Braley is "suffering badly with rural voters. Only 15 percent support him compared with 58 percent for Ernst." (Main article)

"Ernst trails Braley among likely female voters by 13 percentage points. But Braley faces an even bigger gap than Ernst: He loses with men by 25 points." (Sidebar)

My main question after reading these articles is, who are the undecideds? Are they mainly people who voted for Obama (or have demographic characteristics suggesting a Democratic lean) or Romney (or have demographics suggesting a GOP lean) or is there no clear trend among the undecideds?

Newer
Older

RSS