Tag Archives: NBC-Marist

OK, let’s have some fun. NBC/Marist/WSJ polls are out and the race is OVER!

September 13, 2012 – 7:16 pm

You can only laugh at the professional irresponsibility of these polls.  We’ve busted Marist and NBC a bunch of times before  and now the Wall Street Journal sullies its name getting in bed with these jokesters.

** Now we can confirm thanks to Hot Air the party IDs are much like we thought –  very pro-Democrat replicating the 2008 turnout or every better for Obama. Sigh.

OHIO

I don’t have the cross-tabs yet to get exact party IDs but the Ohio survey has Romney leading Obama among independents by 3-points but LOSING the state by 7-points.  This tells you there is a HUGE over-sampling of Democrats.  The party ID on election day in 2008 had Democrats with an 8-point advantage in Ohio: Dem 39, Rep: 31, Ind 30.  The party ID in 2004 was a 5-point advantage for REPUBLICANS: Dem 35, Rep 40, Ind: 25.  The mid-term election in 2010 was split perfectly even between the two parties. Per usual, they are polling as if Obama can recreate the 2008 wave and probably even exceed his performance to achieve a poll where he’s leading.

Update: Now hearing this sample was D +10 compared to D +8 in 2008 and R +5 in 2004. Here it is: Dem 38, Rep 28, Ind 32 or D +10.  Confirmed wholly unrealistic sample.

FLORIDA

The Florida poll doesn’t have a  glaring issue yet, but Obama leads with Independents by 8-points but only 7 with the overall populations.  That doesn’t seem correct although Florida party ID shifts like Ohio. In 2008 the party ID was D +3: Dem 37, Rep 34, Ind 29. in 2004 the party ID was R+4, Dem 37, Rep 41, Ind 23. In the 2010 mid-term the party ID was split perfectly even.

Update: Party ID in the Florida survey is D +2, Dem 35, Rep 33, Ind 30 much like it was in 2008.

VIRGINIA

The Virginia poll has the same issue as Florida.  Romney leads by 2 among Independents but is down statewide by 5 points.  Elsewhere I’m see that party ID in Virginia was Dem 31, Rep 26, Ind 43 (thanks Ben).  In 2008 the advantage on election day was D +6: Dem 39, Rep 33, Ind 27. In 2004 the party ID was R +4: Dem 35, Rep 39, Ind 26. No figures available for 2010.

Update: The party ID above is correct  so the conclusions are all the same.  The press wishes to recreate the 2008 Obama wave and the pollsters reflect that despite the fact that no reasonable person believes that can be replicated.

BOTTOM LINE: Per usual the polling services continue to sample the electorate as if President Obama will have much the same turnout he had in his historic 2008 election.  Every single poll says much of his base is less enthusiastic (youth and Hispanics) while the GOP is unquestionably fired up in no small part due to antipathy towards Obama.  There are only two sane takeaways from these polls: Romney is doing decent with Independents in Virginia and Ohio while Obama is doing well with Independents in Florida.  After that, the only thing I can say is these polls have no relevance to November 2012 as they are representative samples based on Obama enthusiasm of 4 years ago which is non-existent today.

By keithbacker | Posted in Florida, Ohio, Virginia | Also tagged , , | Comments (11)

Romney and Obama Tied in New Hampshire, Obama +4 in Michigan — NBC/Marist

June 28, 2012 – 7:41 am

NBC/Marist released a few polls this morning and New Hampshire is a dead heat while President Obama holds a four point lead in Michigan 47 – 43. Included in the survey was North Carolina which Obama surprisingly led 46 – 44 (although based on party ID it is almost as if this was a poll of North Carolina DNC members). Party ID for all three after the jump.

NBC-Marist surveys conducted June 24-25

New Hampshire Michigan North Carolina
Barack Obama 45% 47% 46%
Mitt Romney 45% 43% 44%
Undecided 8% 10% 9%

Plenty of interesting take-aways underneath the headlines:

  • First, the President again fails to crack 50% in any Battleground state polls. This is really bad for an incumbent.
  • Two of the states in this survey — New Hampshire and Michigan — were state Obama won by double digits in 2008 and now they are toss-ups
  • President Obama’s approval also fails to crack 50%, and “approval” is a fairly good predictor of popular vote count on election day.
  • A majority of voters in each state say the country is headed in the wrong direction

Party  ID was a bit of a mess, especially in North Carolina.

Read More »

By keithbacker | Posted in Michigan, New Hampshire | Also tagged , , , | Comments (0)

Beware Funny #s: Ralston Dismantles Nevada “Dead Heat” Poll

June 2, 2012 – 11:09 pm

The dean of Nevada politics, and solid lefty, Jon Ralston opens up his column with a little chest thumping over his prescient 2010 call for Harry Reid to win his improbable Senate re-election. Normally this would be a bit unseemly, but Ralston’s call for Reid was consistent from start to finish and his analysis, in the face of mountains of contrary opinions, proved to be spot on. In his column deconstructing the recent NBC-Marist poll showing Obama up 2 but within the margin for error, he importantly identifies who and what gave him such successful insights into the Reid race:

[M]any polls erroneously created the impression that Angle was likely to defeat Reid. But there were problems with almost every one of those surveys, easily discovered by exploring the internals. And Reid’s pollster, Mark Mellman, turned out to have the only consistently correct numbers for one reason: His model of what the turnout was going to look like most closely approximated what it actually was.

The is the Rosetta stone for every poll and pollster.  Getting the party ID and turnout to most accurately reflect who will actually show up in the voting booth determines everything about the validity of any poll.  This is often as much art as science, but reputable pollsters get this more consistently right than wrong which is why I take great umbrage when Battleground state polling forecasts 2012 turnout to be a similarly Democrat year as 2008 (the D +8 spread we’ve seen multiple times) — no one thinks Democrats will be able to repeat such an advantage.  And Ralston was all over this in 2010 while GOP operatives fine tuned their Reid obituaries.

So the chest thumping is both warranted and a helpful reminder as Ralston digs into the recent NBC-Marist poll on Nevada:

[B]eing a poll junkie, I dived (sic) into the crosstabs, which can tell you something about the survey’s validity and also set the contours that will help determine the outcomes in November. The NBC/Marist poll makes some assumptions that might be instructive for November, but also might skew the current results. [T]he survey highlights the key factors that will determine the outcomes for the top-of-the-ticket races:

• Regional turnout: Nevada is three states – Southern, Northern and Rural. The greater the percentage of the vote in Clark County, the better for Democrats, who have a huge registration edge. The NBC/Marist poll turnout model indicates Clark County will make up 72 percent of the vote in November. That’s high by about 5 percentage points based on the past two elections, which shortchanges super-conservative rural Nevada by almost that much.

This reveals 3 things: the oversampling of the heavily Democrat region greatly increases Romney’s chances for victory in a normalized turnout. If Obama is able to achieve this level of Clark County mobilization (Ralston doubts it), he’ll likely win the state. But maybe most importantly, in this heavily Democrat region, Obama only leads by 5 points. This would spell doom for Obama in November according to Ralston.

• Hispanic turnout: The NBC/Marist poll has Hispanics as 19 percent of the electorate, which is something that has never happened in the state’s history. The highest was the past two cycles, in which 15 percent of voters were Latinos.

Read More »

By keithbacker | Posted in Nevada | Also tagged , , , , | Comments (6)

NBC-Marist Polls Shows Iowa, Colorado and Nevada are Dead Heats

May 31, 2012 – 10:17 am

New NBC-Marist polls released this morning show the Mountain West swath of Battleground states are anybody’s to win:

These three states are all battlegrounds that Obama carried in 2008, but George W. Bush won in 2004. “These are very, very competitive states,” says Lee Miringoff, director of the Marist College Institute for Public Opinion, which conducted these polls. “Everything is close.”

Bottom line: Fairly consistent results confirming most everyone’s suspicions, right now this race is wide open.

Normally I would have an issue with the party ID of the sampling like in the previous Ohio (D +9) and Florida (D +8) polls that oversampled Democrats.  But these polls are fairly balanced (D +2 in Nevada and R +1 in Iowa) with a slight oversampling of Republicans in Colorado (R +4).  Overall a fairly good sampling of registered voters.

Update: Great comment by Jon Ralston of the Las Vegas Sun on MSNBC. he points out that the Nevada poll was comprised of 19% Hispanics. Ralston thinks that is a meaningsul over-sampling of a Pro-Obama group meaning Obama is likely behind in that state. It is important to remember Ralston is a big ol’ lefty but also a good reporter. He was all over the hidden strength of Harry Reid during a difficult re-election in 2010 when outsiders had that race called wrong from start to finish.

Update II: The National Journal identifies some intriguing takeaways from these results:

  • Iowa: Obama’s 46% approval rating is little different than the 45% in January. 70% of voters say the economy is the #1 issue and favor Romney 46 – 41 in this area.
  • Colorado: In the all important “likely voter” category, Romney flips the one point deficit in Colorado and takes a one point advantage.
  • Nevada: Obama lags his 2008 performance in every race sub-category. In 2008 Obama won 43% of whites vs 38% today, 94% of blacks vs 89% today, and 76% of Hispanics vs 67% today.

Lastly, in the Colorado discussion, this observation stood out:

Just 45 percent of Colorado voters approve of the job Obama is doing as president, while 49 percent disapprove. A majority, 56 percent, say the country is on the wrong track, compared to only 38 percent who say the nation is headed in the right direction.

Those are horrible numbers for the President. If those figures don’t materially change by November, Barack Obama will lose the Presidency by a large margin.

By keithbacker | Posted in Colorado, Iowa, Nevada | Also tagged , , , , , , | Comments (1)

Obama Leading in Ohio, Virginia and Florida? That’s What the Headlines Say

May 24, 2012 – 10:02 am

New NBC-Marist polls were released this morning for three of the more important Battleground states (Ohio, Virginia and Florida) with President Obama leading in all three, albeit narrowly. Despite the headline worthy story that Obama leads in all three, a peak under the hood highlights the weakness in the incumbent and opportunity for the challenger:

[I]n each of these states, Obama’s share of the vote is below the 50 percent threshold usually considered safe haven for an incumbent president, and Romney has narrowed the margin in these three battlegrounds since earlier this year.

At first glance within the context that these states are admitted by all to be Battlegrounds, I noticed a curious thing about the party ID component — a key factor in weighing the accuracy of any poll as well as a way to create awfully funny numbers in polling results.

Let’s start with Ohio where Obama is leading by 6 (48 – 42) in this poll (notably down from +12 in March). The party ID here was Democrat 37, Republican 28, Independent 34 or D +9 for short (with also probably too high a sampling of Independents).  That is a large over-sampling of Democrats.  Nationally in 2008 — an overwhelming Democrat year — the party ID breakdown was D +8 (in 2004 it was an even split).  Is 2012 shaping up to being another “wave” election for the Democrats in Ohio, even larger than 2008?  Not a chance.  So while the President should be please to be leading this poll, it is highly unlikely to be representative of the actual electorate which will be closer to D+2 on election day by my guesstimate. This would push the Ohio poll to a practical dead heat with Romney picking up ~87% of the increased Republican sampling (87% of a 7 point ID swing is 6 points in favor of Romney based on the underlying data). At that point it’s a dead heat 48-48 although the subsequent adjustment for oversampled independents would likely shave a point off Romney’s total.  Clearly not the 6 point lead reported for the President in a state where Obama’s bailout of the auto industry would have the 2nd biggest impact (Michigan is #1).

The Virginia poll that has Obama leading by 4 (48 – 44) the party ID is much more reasonable at D +2 with again an unusually high sampling of Independents at 39%.  Virginia’s party ID will be key in November as the state has been trending Democrat over the last 15 years but swung back dramatically to the GOP in 2009 and 2010. Of note, a reduction of the Independents oversampling would hurt Obama since he carried this group 47 – 41 in the state poll.

The Florida poll with Obama leading by 4 (48 – 44)  falls into the same trap as the Ohio poll with a party ID of Democrat 43, Republican 35 and Independent 21 for D +8 with Independents being fairly represented. Again, no one expects Florida to be D +8 a reasonable adjustment (80% of registered Florida Republicans support Romney so 80% of 6 points equals a 4.8 point bump) would change Florida to Romney 49 and Obama 48 or Romney +1 — with the subsequent change in independents adding another point to his lead at +2. This is more closely consistent to yesterday’s Quinnipiac poll where Romney was +6 rather than down 4. Consistency in polling results like these are what any campaign is looking for to ensure attractive results are the norm and not the outlier.

Despite being behind in all three polls, the Romney camp can’t be too upset by what the underlying data of these polls are actually telling each campaign.

By keithbacker | Posted in Florida, Ohio, Virginia | Also tagged , , | Comments (5)
Follow

Get every new post delivered to your Inbox.

Join 217 other followers