ReadabilityWhat the talking heads don't tell you about the NBC/WSJ poll internals
There are a lot of numbers being thrown around about the NBC poll today. On Morning Joe they point out Obama’s is given a 49% approval, but there are some internal numbers in this poll that you need in order to understand that approval figure and where we actually are. So lets examine the internals.
We’ll start with the splits:
Strong Democrat .….….….…. 19
Not very strong Democrat .…. 13
Independent/lean Democrat …10
Strictly Independent .….….….14
Independent/lean Republican ..10
Not very strong Republican .…11
Strong Republican .….….….…13
Other (VOL) .….….….….….… 8
Not sure .….….….….….….…. 2
So this means that the skew of the poll to begin with is 42% dem vs 34% GOP. Gee do ya think an 8 point difference in party affiliation might have something to do with the numbers? I’d say polling nearly 25% more democrats than republicans might help the democrats bottom line figures a tad.
Now lets see how those people voted last time
Yes, Voted
Voted for Barack Obama .…. 42
Voted for John McCain .…… 32
Voted for someone else .…… 5
Not sure .….….….….….…… 3
No, Did Not Vote .….….…… 18
Not sure .….….….….….….… -
The spread of Obama voters vs McCain voters in this sample is 3 points greater than his actual 2008 election spread but lets look deeper:
Nearly 1 in 5 of the people who answer claimed they did not vote last time in an election that drew huge turn out, so lets extrapolate those figures without the 18% that claim they didn’t vote. What kind of spread does that leave us compared to the actual 2008 results (52.9 Obama vs 45.7 McCain)?
Voted for Barack Obama .…. 52
Voted for John McCain .…… 40
Voted for someone else .…… 6%
In the last election President Obama got 52.9% of the vote so the poll reflects that figure within 1% but John McCain got 45.7% of the vote. This means this poll under-represents McCain voters by over 11%
But if the McCain numbers were under-represented by almost six points how is it that Obama’s numbers are within 1% of his actual figure. Look at the “voted for someone else” figure 6%. In the election only 1.5% of the electorate voted for a non-major party candidate. How on earth does this poll manage to oversample that group by a factor of 4?
How is it possible? Well I’ve interviewed thousands of people over the past 18 months from all over the country. I have talked to many people who voted for McCain and were proud of it, but have not talked to a single person who voted for McCain who regretted it. I’ve talked to plenty of people who voted for Barack Obama as well, but I’ve also talked to many Obama voters who were very apologetic about that vote who now regret it.
I submit and suggest that quite a few of those who were “not sure” or “did not vote” or “voted for someone else” in this poll were actually Obama Voters who didn’t want to admit it. This would be consistent with both my interviews and the trends that we will discuss later on.
Now lets look at this set of numbers of registered voters and how they say they’ll vote in 2012:
Probably vote for President Obama .….….….…..45
Probably vote for Republican candidate .….….…40
Vote for other party (VOL) .….….….….….….. …2
Depends/Depends on who opponent is (VOL) .….11
Not sure .….….….….….….….….….….….….…. 2
In the internals of the polls you can see that the 45% figure (with one exception a 43%) has been steady since Feb. I submit that this is a ceiling for the president. Consider in a poll with 42% self identified Democrats only 45% will “probably” vote for President Obama.
Meanwhile in a poll where you had only 32% republicans 40% say they will probably vote for the GOP. This tells you where the independents are trending.
Finally lets tackle that 49% approval number.
There are two items that needs to be addressed. The first is the approval trends that I have alluded to before.
In this poll for approval the choices are not “Positive” “Negative” and “Neutral” they are “Very Positive” “Somewhat Positive” and “Somewhat Negative” and “Very Negative””
If you look at the tend numbers for the first 4 months of his term he was over 40% Very Positive (Max 47%), for the following three months he was over 30% Very positive, since then “Very positive” has been below 30% (27% this month) Meanwhile the “somewhat positive” has remained between 18 – 23% throughout his presidency while “Neutral” has ranged from 9 – 15% (14% this month) This means that the trend since inauguration has been a full 20% of support slipping into disapproval.
Finally consider these demographics:
White .….….….….75
Black .….….….…..11
Asian .….….….….. 2
Other .….….….….. 3
Hispanic (VOL) .….. 7
Not sure/refused … 2
It is not up to me to speak for the black community but I’m telling you that people of various nationalities tend to be less likely to critique on of their own publicly than privately. You will find this in families and in almost every ethnic group. You are even less likely to see an ethnic group publicly “disapprove” the first of their own in any particular job. The black community runs 90 – 95% Democratic and voted that way last election. If you think that more than one in 20 black Americans are going to tell a national pollster that they disapprove of the first Black President you are absolutely out of your minds.
One can spin it any way you want but if you look at the internals of this poll the bottom line is Barack Obama is hemorrhaging his popularity and barring some cataclysmic failure on the GOP side or incredible success on his side he is not going to win re-election.
Update: Just got some of the most crushing news I’ve had in years, An Instalanche doesn’t cure it but it certainly can’t hurt.
Update: 2 Got my first ever link from the weekly standard. That’s kind of cool. Still a tad down but hard let financial problems get to you when you are fundraising for people who have lost absolutely everything.
There are a lot of numbers being thrown around about the NBC poll today. On Morning Joe they point out Obama’s is given a 49% approval, but there are some internal numbers in this poll that you need in order to understand that approval figure and where we actually are. So lets examine the internals.
We’ll start with the splits:
Strong Democrat …………….. 19
Not very strong Democrat ….. 13
Independent/lean Democrat …10
Strictly Independent ………….14
Independent/lean Republican ..10
Not very strong Republican ….11
Strong Republican …………….13
Other (VOL) …………………… 8
Not sure ……………………….. 2
So this means that the skew of the poll to begin with is 42% dem vs 34% GOP. Gee do ya think an 8 point difference in party affiliation might have something to do with the numbers? I’d say polling nearly 25% more democrats than republicans might help the democrats bottom line figures a tad.
Now lets see how those people voted last time
Yes, Voted
Voted for Barack Obama ….. 42
Voted for John McCain ……. 32
Voted for someone else ……. 5
Not sure ……………………… 3
No, Did Not Vote …………… 18
Not sure ………………………. -
The spread of Obama voters vs McCain voters in this sample is 3 points greater than his actual 2008 election spread but lets look deeper:
Nearly 1 in 5 of the people who answer claimed they did not vote last time in an election that drew huge turn out, so lets extrapolate those figures without the 18% that claim they didn’t vote. What kind of spread does that leave us compared to the actual 2008 results (52.9 Obama vs 45.7 McCain)?
Voted for Barack Obama ….. 52
Voted for John McCain ……. 40
Voted for someone else ……. 6%
In the last election President Obama got 52.9% of the vote so the poll reflects that figure within 1% but John McCain got 45.7% of the vote. This means this poll under-represents McCain voters by over 11%
But if the McCain numbers were under-represented by almost six points how is it that Obama’s numbers are within 1% of his actual figure. Look at the “voted for someone else” figure 6%. In the election only 1.5% of the electorate voted for a non-major party candidate. How on earth does this poll manage to oversample that group by a factor of 4?
How is it possible? Well I’ve interviewed thousands of people over the past 18 months from all over the country. I have talked to many people who voted for McCain and were proud of it, but have not talked to a single person who voted for McCain who regretted it. I’ve talked to plenty of people who voted for Barack Obama as well, but I’ve also talked to many Obama voters who were very apologetic about that vote who now regret it.
I submit and suggest that quite a few of those who were “not sure” or “did not vote” or “voted for someone else” in this poll were actually Obama Voters who didn’t want to admit it. This would be consistent with both my interviews and the trends that we will discuss later on.
Now lets look at this set of numbers of registered voters and how they say they’ll vote in 2012:
Probably vote for President Obama ………………45
Probably vote for Republican candidate …………40
Vote for other party (VOL) …………………….. …2
Depends/Depends on who opponent is (VOL) …..11
Not sure …………………………………………….. 2
In the internals of the polls you can see that the 45% figure (with one exception a 43%) has been steady since Feb. I submit that this is a ceiling for the president. Consider in a poll with 42% self identified Democrats only 45% will “probably” vote for President Obama.
Meanwhile in a poll where you had only 32% republicans 40% say they will probably vote for the GOP. This tells you where the independents are trending.
Finally lets tackle that 49% approval number.
There are two items that needs to be addressed. The first is the approval trends that I have alluded to before.
In this poll for approval the choices are not “Positive” “Negative” and “Neutral” they are “Very Positive” “Somewhat Positive” and “Somewhat Negative” and “Very Negative””
If you look at the tend numbers for the first 4 months of his term he was over 40% Very Positive (Max 47%), for the following three months he was over 30% Very positive, since then “Very positive” has been below 30% (27% this month) Meanwhile the “somewhat positive” has remained between 18-23% throughout his presidency while “Neutral” has ranged from 9-15% (14% this month) This means that the trend since inauguration has been a full 20% of support slipping into disapproval.
Finally consider these demographics:
White ……………..75
Black ………………11
Asian ……………… 2
Other ……………… 3
Hispanic (VOL) …… 7
Not sure/refused … 2
It is not up to me to speak for the black community but I’m telling you that people of various nationalities tend to be less likely to critique on of their own publicly than privately. You will find this in families and in almost every ethnic group. You are even less likely to see an ethnic group publicly “disapprove” the first of their own in any particular job. The black community runs 90-95% Democratic and voted that way last election. If you think that more than one in 20 black Americans are going to tell a national pollster that they disapprove of the first Black President you are absolutely out of your minds.
One can spin it any way you want but if you look at the internals of this poll the bottom line is Barack Obama is hemorrhaging his popularity and barring some cataclysmic failure on the GOP side or incredible success on his side he is not going to win re-election.
Update: Just got some of the most crushing news I’ve had in years, An Instalanche doesn’t cure it but it certainly can’t hurt.
Update: 2 Got my first ever link from the weekly standard. That’s kind of cool. Still a tad down but hard let financial problems get to you when you are fundraising for people who have lost absolutely everything.
[…] already commented on how the black community is the foundation of the president’s polling: The black community runs 90-95% Democratic, and voted that way last election. If you think that […]
[…] already commented on how the black community is the cornerstone of the president’s polling: The black community runs 90-95% Democratic and voted that way last election. If you think that […]
[…] already commented on how the black community is the cornerstone of the president’s polling: The black community runs 90-95% Democratic and voted that way last election. If you think that […]
[…] every other pollster will ask people who claim to be “independent” about their “lean.” For example, in Massachusetts more than 51% of the registered voters are “unenrolled,” […]
[…] before it was unlikely that alone would make a difference for the first black president in the black community. It is not up to me to speak for the black community but I’m telling you that people of various […]
[…] The story as printed in the Journal had almost no answers to my methodology questions. But here’s the thing: the respondents to this poll voted for Obama over McCain by 10 points — maybe more. Obama only won by seven. Sample error! More from this guy. […]
[…] What the talking heads don’t tell you about the NBC/WSJ poll internals […]
I think there is a huge Wilder Effect. Wilder was the black dude who ran for gov. of VA and led all the polls up until election day where he lost by a substantial margin. later analysis showed that when polled people were afraid to admit they would not vote for a black guy when ask by a pollster for fear of being branded a racist by the hysterical cultural Marxist politically correct left. I think the same is happening in all Obama polls. In reality I think he will be destroyed in a landslide next election regardless of who the GOP puts against him.
I hope my ‘Republican’ friends who voted for this incompetent, unqualified assclown are happy with the destruction they have wrought upon the country.
Never before envisioned a scenario where I wished Hillary Clinton was the President.
…but have not talked to a single person who voted for McCain who regretted it.
That’s because he lost. I daresay that if we had President McCain in office today, there’d be more than a few who regretted it.
[…] all is not kosher. We’ll start with the splits: Strong Democrat …………….. 19 Not very strong Democrat […]
Great column, but that aside, your comment about “crushing news” is most unfortunate. Please know that you are in my prayers.
“he is not going to win re-election.”
from your lips to God’s ears.
My take is that I’m not sure if Charles is referring to the GOP-run House that has actually passed a budget or the Democrat-controlled Senate that hasn’t passed a budget in the past 750+ days.
But I’m sure Charles agrees with President Kerry’s assessment.
Sorry. The style sheet didn’t load correctly.
Barring an act of God, the repubs are going to wash the Left away in 2012.
Why then is the WSJ wasting money on these pollsters?
Change you font color. Your text is practically unreadable.
That was laid out real well. I have wondered about polls from liberal sources but just took a few points off in my own math for that. This made so much sense and it must have to many other readers especially Instapundit, or you would not have been linked again.
Common sense just tells me your post is spot on and I look forward to reading more from you as the next year rolls around.
I believe any positive posts from bloggers is important because Obama has honestly hurt America more than any president in my lifetime. I did not vote for him but I did hope our first black president would be a man the black community would be so very proud of and we would all see this, “Hope and Change” as a good thing.
I voted for McCain and I am sorry I only had that choice but I am not sorry I didn’t vote for Obama (his world view precluded him as a choice for me).
Don’t know if you have written on the weirdness of CNN internals yet. They seem to only poll Republicans over 50 from the South according to their own demographics shown in the poll internals.
Ghadaffy (? you know, that Libyan guy) got himself killed choking on a fish bone on Oct. 30, 2012, Obie claimed credits. MSM wet themselves on what-a-big-f*cking-deal his foreign policy was, enough hopey-changers would be impressed one more time. Frankly, the hopey-changers have some screws loosen up there, but they would plunge us into four more it’s-Bush’s-fault years.
It is a poll of adults and not registered voters or likely voters so it would skew left. But thanks for breaking it down to how much. And yes when discussing a poll, pundits on TV must — not should — explain the difference among adults, registered voters and likely voters.
[…] UPDATE: Da Tech Guy examines the numbers of those polled and finds it is skewed left. […]
Not quite convinced. People always over-report voting for the winner of the last presidential election. In 1964, I think the figure was that 66% of the American people remembered voting for JFK in 1960 (!). Smaller skews have showed up after every presidential victory. 52/40 Obama/McCain would be consistent with this history and NOT in and of itself evidence of a pro-Democratics bias in the sample.
The 42/34 Dem/GOP difference is more interesting. What is the evidence for current Dem/GOP affiliations in other current data? Dems traditionally have an edge here as well.
The 49% approval rating for Obama also looks reasonable: Obama is averaging 47-50% in the Rasmussen surveys over the last two weeks for likely voters (although Obama is down a bit today).
In November 2012, I hope the only vote Obama gets is own…but the NBC/WSJ poll looks pretty reasonable for right now.
I want to suggest another factor. In the last election, at least in Chicago, the Democrats used extensive and extremely aggressive and dishonest fake “poll” tactics. I would receive calls, the caller would ask me to participate in a political poll. The first question would ask if I preferred the Democratic or Republican candidate in one of the races. If I responded that I preferred the Democratic candidate, they would thank me and hang up. If I responded that I preferred the Republican candidate, then they would follow up the question with questions that stated some slanderous falsehood about the candidate and asked if hearing that made me more or less likely to vote for the candidate. At the end of the call, after resisting what was essentially a five minute one-on-one personal attack against the Republican candidate, the person on the other end of the line asked me, “One more question. Is this John I’m speaking to?”
I will no longer answer a telephone poll.
What they are doing may or may not be an effective campaign tactic to sway (or intimidate) undecided voters, but I think that they are definitely (and possibly deliberately) poisoning the polling well. One number I’d be interested in is what percentage of conservatives simply hang up on polling calls. I’ll bet that it’s non-trivial, and I’ll bet that it’s another factor skewing the polls towards liberals.
ah ha it was indeed my browser…sorry to bother.
Great analysis…thank you.
Not sure if my browser is broken or what but your blog came out with a brown background with light black font which was impossible to read. I ended up copying and pasting into notepad to check it out. Not being ungrateful just fyi.
I read the headline on Yahoo about this poll – it was something to the effect of “Most Americans Don’t Blame Obama for Bad Economy”. I immediately figured there’s gotta be some kind of manipulation going on there and didn’t bother reading the story.
My take is that an unpopular Congress is pulling down the Republicans, who should be taking nothing for granted.
http://www.leftbankofthecharles.com/2011/06/barack-obamas-popularity-rises-above.html