Tuesday, July 12, 2016

More than 50% of respondents in latest Reuters/Ipsos general election poll are Democrats

It may seem like I'm beating a dead horse here, but I'm not. The night mare is very much alive.

Today Reuters/Ipsos formally released the results of its latest five-day tracking poll headlined "Clinton extends lead over Trump to 13 points". The national poll of likely general election voters, which was administered entirely online, was made up of a sample that was 50.7% Democrat, 36.3% Republican, and 13.0% independent or third-party.

That's wildly out of line not only with other measures of partisan affiliation in the US--in late June Gallup found a 31%-28%-39% distribution, respectively--but also with the polling samples of Reuters/Ipsos' polls just a couple of months ago, when Democrats made up around 45% of those surveyed and Republicans around 42% (and when Trump briefly led).

What's particularly remarkable is that there is no mention of the partisan distribution of the sample anywhere in the write up. It can be found at the source but not in the media release.

The change in the political makeup of the polling sample accounts for virtually all of the gains Hillary (appears) to have made against Trump since May. Hillary's support among independents and third-party voters has actually declined over the last couple of months while her support among Republicans has stayed constant at around 7%, even as her overall support in the R/I poll has increased.

Reuters/Ipsos is an outlier in the current RCP average. While it shows Hillary leading Trump by 13 points, her average lead across the other seven polls currently displayed is just 3.6 points. The most recent polls released in Iowa (today) and Florida (yesterday) actually show Trump ahead in both, and these are states Obama won in 2012.

Something feels fishy. To reiterate, it could be a consequence of Republican recalcitrance in the face of Trump's refusal to morph into John Kasich, the tendency for people to identify with whatever party the candidate they intend to vote for is from, a problem with Reuters-Ipsos' sampling methodology (either inadvertently, or, more sinisterly, intentionally), some combination of these things, or something else entirely.

Then again, maybe I'm not being cynical enough. Perhaps Heartiste, who is nothing if not perspicacious, offers a better explanation:


pithom said...

When you exclude the obvious BS results (Bloomberg, Rasmussen, Reuters, Zogby, Raba), you get Trump losing in the HuffPo polling average by two points. I suspect that's optimistic for Trump, though:


As shown in 2012, to be sure Trump is winning, he has to be doing well in all the reliable polls.

Audacious Epigone said...


The Morning Consult polls consistently sample in the high 30%s Dems and low 30%s Republicans. Does R/I not look at this at presume something is awry? These figures are more representative of 2012 (which was 38% Dem, 32% Rep). It seems unlikely that Democrat turnout will surpass that of '12 even with the steady accretion in partisan affinity that has accrued over another four years of American demographic transitioning.

IHTG said...

Hey AE, how do you feel about this: https://youtu.be/brli1aAYxSs?t=180

Cicatrizatic said...

Quinnipiac polls out today show Trump leading in FL, OH, and PA. They also confirm that adding Johnson and Stein helps Trump, not Clinton.

Mil-Tech Bard said...


Instapundit on the same POLL --

POLITICO: Swing-state stunner: Trump has edge in key states. “New swing-state polls released Wednesday by Quinnipiac University show Trump leading Clinton in Florida and Pennsylvania — and tied in the critical battleground state of Ohio. In three of the states that matter most in November, the surveys point to a race much closer than the national polls, which have Clinton pegged to a significant, mid-single-digit advantage over Trump, suggest.” Trump’s not a typical GOP candidate, which means that traditional election models won’t work very well.

21 Posted at 8:44 am by Glenn Reynolds

Mil-Tech Bard said...

Audacious Epigone,

This late May 2016 Washington Post/ABC poll (link at bottom) shows Trump winning against Hillary with an Obama coalition turn out model (D+8).

From the poll Trump leads:

Men: Trump 56-34%
Independents: Trump 48-35%
No degree: Trump 52-38%
Whites: Trump 57-33%

To achieve this —

Men: Trump 56-34%
Whites: Trump 57-33%

Trump has to be up at least 35% over Hillary with white males over all and much higher than that with the working class white males (WCWM) for those numbers above to work.

Demographically, white males alone outnumber the total African American population in America by seven to one. WCWM represents 60% of the total adult white male population and by themselves outnumber all African Americans by a little over four to one.

Since 1980, WCWM presidential voter participation has dropped from 75% to 50% in 2012.

The Obama coalition model assumes white working class male voters will only turn out at 50% as they have for the last four Presidential election cycles.

The newest media polls assume higher than +8% Democratic turn out advantage.

Neither scenario will be the case with Trump in the General Election, as he has demonstrated in the GOP Primaries higher than the last four GOP Presidential election cycle WCWM turn out voting for him.

Disproportionate working class white male turnout will not only bury Hillary in the Rust Belt, the down ballot effects promise to make for yet another GOP wave election in the Federal Senate and a lot of state/local offices. (Assuming the GOP Nevertrumpkins in the Senate don't eff it up)


Trump Leads Clinton In ABC/Washington Post Poll, Even With D+8 Poll Sample…


Frank Gappa said...

What is the strategy behind fudging poll numbers? Is it to discourage certain voters from participating? I would think that if a candidate was down 5-10 points that this could have the potential to motivate voters into voting for his candidate more so. Or, conversely if my candidate was ahead by the same margin perhaps I would stay home feeling confident my candidate would win anyway. I just don't follow how manipulating polls for psychological effect is something as a pollster they can control with certitude. My own take on these polls is about confirmation bias and putting like minded voters at ease in the short term.

Audacious Epigone said...


Not good. His instincts usually are, but it's pretty clear that all he did was look at the video clips of the incidents without doing any of the necessary homework. He's trying to keep the citizens-vs-outsiders dynamic going and police shootings don't fit into that very well. He at least allowed for the explanation that these were "bad people" and he's had nothing good to say about BLM that I'm aware of.


Underscoring how wildly at odds the polls are, NBC/WSJ did many of the same states as Quinnipiac and got much different results. NBC/WSJ polls have consistently been among the worst for Trump, not just now but also during the primaries (as opposed to Reuters-Ipsos, which was actually one of the best for Trump through the primaries and early in the general but has since shifted dramatically).

Mil-Tech Bard,

R/I showed the same thing in mid-May, with +4 Dem Trump was slightly ahead.

There aren't seven times as many white men as there are blacks though, in total or electorally. It's closer to half that. Did you mean 7x as many whites as blacks?

Audacious Epigone said...


There is "momentum" for one thing. It probably also influences campaign contributions. If it doesn't look like your guy has much of a chance it may not dissuade you from ultimately voting for him but it could keep you from taking out your checkbook. It also discourages #NeverTrump cucks from falling in line and at least offering formal tepid support for Trump. To the extent that it's intentional, maybe the intent is to create a discordant convention?

Cicatrizatic said...

I think there are some voters who are swayed by the fact that polls show one candidate leading by a large margin. Where the margin is greater than +5, some voters who are leaning toward the loser candidate end up staying home. You see this with primaries where the top candidate leads heavily in polls going into primary day, and the second and third place candidates end up under-performing. I expect that is because many of the supporters of those second and third place candidates stay home since their candidate appears to have no shot.

If you can show one candidate leading by 5-10 percentage points for 3 months straight, you create psychological conditions under which a portion of voters will give up and assume the result has been determined.

Mil-Tech Bard said...

Audacious Epigone,

I went back and checked the census and voting records, and yep. You are correct. It is six to one over all White total versus Black total and 3-to-1 for white males vesus all blacks.

On the subject of the polls, see this --


Now 3rd party candidates are hurting Hillary more than Trump.

We have the markers of a preference cascade in progress.

A preference cascade that is -away- from Hillary.

Cicatrizatic said...

Today's Rasmussen shows Trump +7. Romney's largest Rasmussen lead was +4 in June of 2012.

CBS News poll has them tied, both with the one-on-one poll, and with third party candidates included.

CBS News appears to weight their sample based on latest census data, so it ends up being 39% I / 33% D / 27% R.

Both Rasmussen and CBS News reflect that Trump wins Independents by about the same margin.

Mil-Tech Bard said...


Gatewaypundit on the same Rasmussen poll --


Mil-Tech Bard said...

Audacious Epigone

TCTH members in NY report Hillary ads on local TV and radio channels in up-state NY and in NJ.

Other TCTH commenters there say Hillary for President ads are on-air in Massachusetts OF ALL PLACES.

It looks like Dallas is for the General election what San Bernardino Muslim Terrorist attack was for the GOP Primaries...the pro-Trump inflection point that defined the campaign.

IMO, we are looking at a range of Trump landslide somewhere between Reagan 1980 and Reagan 1984.

Mil-Tech Bard said...

This Scott Adams comment sounds right --


...To my eyes, the biggest change is that Clinton’s team just became the cop-killing side. At least that’s how it looks to our irrational minds. Your brain thinks cops are probably Trump supporters (true or not) while you probably see cop-killers as Clinton supporters (true or not). And that means the recent slaughter of five policemen in Dallas changed your mental equation.

Now it seems – to our irrational minds – that we no longer have a contest between crooked and racist. Now we have a contest between cop-killers and racists. And in that contest, the racists win.

Before 7-7-2016 Hillary had successfully labeled Trump a racist.

On 7-7-2016, the blood splatter from Micha Johnson, and the Black Lives Matter movement that inspired him, labeled Hillary as a cop-killers.

Nothing she does, or does not do, makes a difference anymore.

Audacious Epigone said...

Mil-Tech Bard,

That Hillary is spending ad money in New York, New Jersey, and Massachusetts is astounding. Do we have confirmation or just reports at this point?

Less than a week out from Dallas and another mass slaughter in France. The driver of the truck that smashed through a throng of people has yet to be publicly identified. That he has yet to be virtually guarantees that he is not a native.

Mil-Tech Bard said...


Just reports so far about the Hillary ads.

As for the Nice' terrorists, see this --

“INFO NICE MORNING. A Niçois of Tunisian origin at the wheel of the truck,” translated from Nice-Matin, July 14, 2016 (thanks to Bill):

A 31-year-old Niçois of Tunisian origin was driving the truck that claimed over 70 people on the Prom in Nice on Thursday night.


Mil-Tech Bard said...

Via instapundit --

NICE ATTACKER IDENTIFIED: “The attacker has been named as Mohamed Lahouaiej Bouhlel. He was 31 years old and born in Tunisia. He was known to police as a career criminal, but not known to intelligence services. He reportedly carried grenades in his truck, suggesting he was planning a even more elaborate attack.”

Plus: “A Saudi-funded mosque opened in Nice just two weeks before last night’s attack.”

22 Posted at 8:13 am by Glenn Reynolds

Audacious Epigone said...

Mil-Tech Bard,

My brother is currently in Bavaria taking foreign exchange law school classes. He says the cities in Germany and Belgium (Nuremberg and Brussels) are like quasi-police states, police armed with AR-15s out in force. The West has done this to itself.

Joshua Sinistar said...

One of the things that creamed corn-fed Little Jebbie's corniness was all the voters who don't usually participate coming out and voting. In Truth the Left is actually a Lunatic Fringe of around 18-19% of the Total Population. They have a larger weight because most people right of a far left center have sat it out because they hate the GOPe candidates and know the votes are rigged. One thing the system fears, and why they really fear Trump is that these people who usually tune out and don't vote could actually turn out for Trump and shit all over their myth of a 50-50 split country. Actually its really more like 66 to 18 with the rest on the fence. If the disenchanted dropouts come back, the whole game collapses and the myth of 50-50 split electorates decided by small insignificant groups like homosexuals and other leftist projects gets blown up.

Audacious Epigone said...