Monday, August 29, 2011

CNN has a New MISLEADING National Poll out taken ENTIRELY in the South

Here is the article about the poll:

New CNN Poll: Perry sits atop GOP field
A new national survey is further proof that Texas Gov. Rick Perry's entrance earlier this month into the race for the White House has dramatically altered the battle for the Republican presidential nomination.

According to a CNN/ORC International Poll of Republicans and independent voters who lean towards the GOP, Perry now sits atop the list of Republican presidential candidates, with strong support from most demographic groups.

The survey, released Monday, indicates that 27 percent of Republicans nationwide support Perry for their party's nomination.....
(emphasis mine)
Surprise...Surprise....Rick Perry is out front. I can see the media outlets lead in now, "Perry winning another National Poll".

Oh Really?.....Not so fast. National Poll you say?.....

Of course, unless you are interested in WHERE the poll was taken, you wouldn't bother checking the crosstabs, Here;s What I found:

Click chart to enlarge

Notice the columns in the chart above as to what part of the country the folks surveyed were from:

Northeast = NA (Not Applicable)
Mid West = NA (Not Applicable)
West = NA (Not Applicable)
South = Add up the column....about 100%

Now I've read CNN's entire write-up and NO WHERE does it mention that this was a Southern Poll. So is it any surprise to anyone that Perry is leading in the south?

The problem is, that's not the message that they are portraying here. They're making it sound as if Perry is leading Nationwide with these Southern skewed results.

I'll let you draw your own conclusions

Crosstabs can be viewed HERE.

Please check us out on Facebook and If you like what you see, please "Like" us. You can find us here.


Ben said...


Just Wow!

What a find this is. Do you think they want America to believe Perry is out front?

craigs said...

Hi Bos
Good post, but I think they've done this before. I also note the variability in the sampling error, up to + / - 7 %.
This is a really strange poll, but probably accurate in the south today.


Aron said...

This poll was conducted nationwide. When CNN doesn't publish a result in their crosstabs, it doesn't mean that the northeast, midwest and western regions of the country weren't surveyed, or that they only polled people in the suburbs, but that there simply wasn't a large enough sample in a given region to render statistically relevant data. The South was the only area large enough to yield results worth reporting, albeit with a relatively high MoE of 7%.

Right Wingnut said...

I'd be inclined to agree, but this poll tracks fairly close to PPP and Gallup. Unfortunately, it's probably accurate.

Ohio JOE said...

Not to be a bugger, but by your logic, they did not poll any rural or urban voter, just Southern Suburbanites?

Right Wingnut said...

It also tracks Rasmussen fairly close as well

BOSMAN said...


The bottom line is, it is a HEAVILY SKEWED SOUTHERN POLL if "there simply wasn't a large enough sample in a given region to render statistically relevant data" in those other areas

Nate G. said...

My two cents:

I don't think this is correct BOS. The cross-tabs are only published for sub-groups that a had large enough number of respondents to be statistically significant. BUT that doesn't mean that their answers were excluded from the total results. Proof of this is that the total results were 27-14 for Perry over Romney. But, the results for the south are a greater gap at 31-14. They should be the same if they only polled that south.

It's likely that the south was the only region with a large enough sampling to be statistically significant because they are the largest region of GOP supporters.

It is very likely thought that CNN could have over sampled the south, but we'll never know because they don't provide the percentages of respondents that came from each region.

Nate G. said...

Oh, looks like I had my comment field open for a while and Aron had already posted what I was thinking.

The sampling for each region would have to be 300-600 people in order for it to be significant. That would greatly increase the number for the whole poll. Plus they'd have to do it for each age group and ethnicity. Pretty soon it just isn't feasible. The headline they seek it just a national poll result, and they can get that with 1000 respondents instead of doing tens of thousands to make sure there is a large sampling for each sub-group.

Ohio JOE said...

"The bottom line is, it is a HEAVILY SKEWED SOUTHERN POLL if "there simply wasn't a large enough sample in a given region to render statistically relevant data" in those other areas" Wrong! How many freaking sub-groups to you want them to poll.

BOSMAN said...

Suppose they had this Poll with Romney in the lead:

"South = NA" and the other area showing data.

What would you say?

Anonymous said...

The GOP runs Perry, Obama will get re-elected.

Perry is already being referee to aka W, jr.

craigs said...

Wouldn't the regional results outside the south HAVE to be pretty significant to impact the overall polling results so much ?? The MOE DROPS from 7 % in the south to 4.5 % overall and Perry drops 4 % overall .So, why wouldn't CNN publish numbers that have such a big impact ???


Terry said...

Why can't CNN try harder to poll a more even amount of Republicans from each area.

It doesn't matter for the primaries that there are more Reps from the south.

Each state has delegates regardless of how many Reps reside there.

craigs said...

I noticed a recent poll in Florida, dated June 22, polling a sample of 3200 voters. The poll showed a 3:1 negative for the Tea Party among Independents and a 2:1 Negative among all voters. Want to lose Florida ???
Here's a primer. Run a candidate with full Tea Party support. A full 64 % of Florida Moderate Republicans are opposed to the Tea party. Now, go look at the latest Florida poll with Romney leading Perry 28-21 % and check the cross tabs to see where Perry's support is.Finally, get Perry to campaign on his social security comments...." a monstrous rip off" ," Unconstitutional", " a Bernie Madoff Ponzi scheme" where everybody should be in jail.
This is like having a blind brain surgeon work on your head. The results of this kind of analysis are blindingly obvious and the Florida story is as clear as a bell.


Aron said...

Looking at all of the crosstabs and perusing through previous polls, it appears CNN's policy is to report the polling data for every sub-group, provided the margin of error is no more than +/- 8.5 percentage points. If it's +/-9% or higher, you'll see an N/A.

Right Wingnut said...

Guys, we now have four polls in the past week that show similar numbers. It is what it is.

I wonder what they will look like after Palin gets in?

CF said...

"Guys, we now have four polls in the past week that show similar numbers. It is what it is."

That is irrelevant. The vetting has begun, common sense says Perry's poll numbers SHOULD start dropping not holding steady.

There are now thousands of people more informed on Perry's poor record, you aren't going to get a poll identical to the ones in the first week unless something sneaky is going on.

Right Wingnut said...


What vetting? GOP primary voters watch Fox News, and listen to talk radio. Who's doing this vetting that you speak of?

CF said...

Right Wingnut -

Independents and moderates don't only listen to Fox News and Talk Radio. Perry's first polls showed that he was picking up a ton of indys. Logic says that these same people are going to start peeling off, even in small amounts, because they get their news from more places.

Anonymous said...

This post's title is entirely false. Are you too dumb to read the crosstabs that you yourself posted? Look at that JPG image. Total Sampling Error: 4.5% Sampling Error in the South: 7.0% That alone proves that it's not entirely, or even mostly, from the South. The South has a plurality of Republican voters, but not a majority.

Matt Y.

Concerned Patriot said...

Its CNN so its not very hard to believe. Anderson Cooper was probably giggling like a little girl about the poll.

Mike Sage said...

I got the inspiration for my article at America Needs Mitt from Right Speak, but just thought someone should go a little more in-depth on it. Nate does have a valid point.. which is, just because a column on the cross-tabs says NA, doesn't mean the poll "was conducted entirely in the south." It could mean that the numbers for other regions were so low as to be unreportable. Even so, I don't buy it. The south doesn't have THAT much of a monopoly on Republicans, sorry.