I pose to you that there is a different kind of voter suppression occurring, and it's conducted by the media. First, an outline: I often discuss the concept of rhetoric... or, as I like to put it, "rhetoricing." Rhetoricing is the process the media uses to sway people's minds. If the media constantly headlines pro-Obama or anti-Mitt Romney headlines/articles, people will be swayed. ABC, CBS, CNN, NBC News, USA Today, and major print publications should be reporting news and not opinion. But that's not the way it is. Just take a look at cnn.com's front page from September 18th. If you visited the site, you instantly encountered the five following headlines:
"Mitt Romney backing away from message caught on tape..."
"What's wrong with Mitt Romney?"
"Mitt Romney back on defense..."
"Conservatives split..."
"Bad timing for Romney."
None of these headlines paint a rosy picture for the former governor, and ALL use negatively connotative words to tell the "news" story. Of course, cnn.com did not balance this by providing five anti-Obama headlines, but this is now the M.O. for major news outlets.
So what do polls have to do with it? Well, this is the new form of rhetoricing. If you notice, most major polls are conducted by these same news outlets. They're released frequently, and realclearpolitics.com keeps track of them here: http://www.realclearpolitics.com/epolls/latest_polls/
The problem: Often times these polls claim "Obama is winning," or "Obama has the lead in," and so forth. The media headlines these "results" constantly, and it is making the case that the election is "unwinnable" or "out of reach" for Mitt Romney (in general, this is more pro-Obama, anti-Romney stuff, too). If an independent, on-the-fence Mitt Romney supporter was blasted daily with headlines making it clear the race was out of reach for the candidate he was potentially going to vote for, that voter would be encouraged to abstain (this tactic was used in Maryland by a rogue campaign manager for republican governor candidate Bob Ehrlich's campaign: http://www.foxnews.com/politics/2012/02/16/maryland-gop-campaign-operative-sentenced-for-robocalls/).
Voila! Voter suppression!
The truth is: The vast majority of these polls unrealistically over-sample democratic voters. For individual states like California, for example, sure, polls conducted here might comprise a more realistic spread of its voting populace. But nationwide, and especially in swing states, the mix of democrats and republicans is much more even. The media, again, the ones conducting the polls, heavily over-samples democrat voters so that the desired result of an Obama lead is produced. This same media can then put on the front page of www.leftleaningnewssite.com that "Obama is winning" (as we saw with a recent Washington Post poll), and the process is complete. The conservative-leaning Fox News has inexplicably done this too, and the fact remains that democrats are still over-sampled in their polls as well (maybe Fox truly is more balanced than people think).
If you don't believe me, fine... but let's take a look at some recent polls. I went ahead and balanced out the poll's voting split by party affiliation. I created a spreadsheet that uses actual voting demographics to show how the polls would really turn out if the sample of voters more evenly represented the country. You might ask, "well, how is the country really divided; what 'spread' did you use?" Don't worry; I have the answer for that. I used two different voting demographic spreads:
1. A 10-day rolling average measured by Gallup:
http://www.gallup.com/poll/15370/party-affiliation.aspx
2. A national average by Rasmussen: http://www.rasmussenreports.com/public_content/politics/mood_of_america/partisan_trends
Gallup shows a slight democratic lean for the country; Rasmussen shows a slight GOP lean. For my experiment, in order to paint a more accurate, balanced picture, I used both. I took 5 recent polls and put them through a calculation using these voter demographics. Additionally, it should be known that I can only use this spreadsheet/method if the poll provides the breakdown for how each party voted for each candidate. Here are the 5 polls used (the page numbers where the demographic breakdown can be found are in parenthesis in the chart):
1. A nationwide Pew Research poll claiming Obama has an 8-point lead: http://www.people-press.org/files/legacy-pdf/09-19-12%20Political%20release.pdf
2. A CBS/NYT/Quinnipiac poll showing Obama has a 4-point lead in Virginia (this was heavily broadcast by the media): http://www.cbsnews.com/8301-250_162-57515519/poll-obama-leads-in-virginia-wisconsin-tight-in-colorado/?pageNum=4&tag=contentMain;contentBody
3. A CNN/ORC poll showing Obama has an 8-point lead in Michigan: http://i2.cdn.turner.com/cnn/2012/images/09/19/topgenstate3mi.pdf
4. A Washington Post poll showing Obama has an 8-point lead in Virginia (this was heavily broadcast by the media): http://www.washingtonpost.com/local/dc-politics/post-poll-obama-up-8-points-over-romney-in-virginia/2012/09/18/ca691d9a-0193-11e2-9367-4e1bafb958db_story.html
5. A Morning Call poll showing Obama has a 9-point lead in Pennsylvania: http://www.mcall.com/news/nationworld/blog_pennsylvania_ave/mc-document-pa-presidential-poll-20120918,0,4409695.htmlstory
Please see the following results:
Click the image for an expanded view
As you can see, the original polls show a very large over-sampling of democratic voters. Of course, when CNN.com headlines "Obama is up by 4 in Virginia," it doesn't say "democrats out-sampled republicans 35% to 24%" (poll #2 in the chart). Using the Gallup voter spread, Obama would still lead in 4 of the 5 polls, but the margins are much thinner. His lead in Michigan, for example, drops from 8 percentage points to 1.4. Using the Rasmussen voter spread, Romney is winning handily.
So is this voter suppression? I think so. I do think the media does this intentionally. I think they over-sample intentionally, and I think they don't report that key bit of information intentionally. I think they headline these results intentionally, and I think they're trying to get people to give up on Mitt Romney.
What can we do? We can be skeptical. We can question. We can look into the make-up of these polls to see the truth. Since the media won't do it, we can do it. We can share this kind of information on Facebook, Twitter, emails, etc. I will do my best to publish the non-skewed poll results for major polls when they come out. I believe that people should see the truth, and I will do my best to put it on this blog.
What's your take? Do you agree with my analysis? Is my method off? Please feel free to check the numbers, and please share any opinions, comments, and questions below. Thank you.
Busy week, LME! :)
ReplyDeleteI'm attempting to assess the accuracy of your methods here, but I don't know the equations you're using in excel to come to your conclusions.
It appears that you're attempting to normalize the poll results across states, if an equivalent number of Dems/Repubs were sampled to reflect the national average distribution of Dems/Repubs (as taken from the Gallup/Rasmussen polls).
If that's the case, I don't think it would paint a very accurate picture. It’s an interesting and very well-done analysis, but I think that the most accurate picture would require the average distribution of Dems/Repubs for that given state (and not simply applying the national average over it).
Don't forget that our system easily allows for the situation where a President can poll ahead of another (and even win the popular vote), but still lose the election. The national average for the Dem/GOP split can’t be accurately fit to represent the Dem/GOP splits for every individual state; even if a swing state. Some states are overwhelmingly GOP, and some are overwhelming Dem, and then there’s everything in between.
Swing states in particular may be rather close to the national average, but by leaving that to assumption it hurts the accuracy/credibility of the final numbers.
I also found it a bit odd in particular that you mention the Gallup poll in this topic, as they’ve been under scrutiny for a good part of the year for under-sampling Democrats and on Democratic issues. On the other hand though, in comparing that to the Rasmussen, keep in mind that the Gallup polls sample literally twice as many people as the Rasmussen.
Don't get me wrong though, I like what you did here and found it to be a very interesting method of analysis.
RKen – good morning! And yes, it’s a very busy week :)
DeleteI will try to address each one of these points. Thank you for sharing them.
As far as the equations, my methods were (hang on tight here):
Adjusted Poll Results for Obama using the Gallup scenario for poll #1 (Pew), for example =
(LV for the poll X Gallup Poll Affiliation for D X how D voted for Obama in the poll) + (LV for the poll X Gallup Poll Affiliation for R X how R voted for Obama in the poll) + (LV for the poll X Gallup Poll Affiliation for I x how I voted for Obama in the poll) ALL divided by LV… for example:
((2192 x .305 x .93) + (2192 x .277 x .05) + (2192 x .415 x .45))/2192 = 622 D Obama voters + 30 R Obama voters + 409 I Obama voters = 1061 Total Obama voters / 2192 Total Likely voters = 48.4% Obama voters in the unskewed Pew poll using Gallup’s affiliation numbers.
For: “It appears that you're attempting to normalize… distribution of Dems/Repubs (as taken from the Gallup/Rasmussen polls).”
- I did address this. I stated: “The truth is: The vast majority of these polls unrealistically over-sample democratic voters. For individual states like California, for example, sure, polls conducted here might comprise a more realistic spread of its voting populace. But nationwide, and especially in swing states, the mix of democrats and republicans is much more even.“
I have accounted for this by focusing only on National polls (where the national Gallup and Rasmussen spreads would apply), and on swing states… states that, yes, while they don’t have an individual stat for party affiliation, it can be safe to assume they follow the national trend more closely. To account for this, I provided two samples… one with a democrat lean, one with a GOP lean to cover both scenarios. These are just scenarios of what the poll results SHOULD look like if there wasn't such an egregious over sampling. As always, there is a MOE… and I think, of course, my analysis has it, but again, these are swing states, and I definitely think it’s safe to use a national spread here.
I do acknowledge the popular vote vs. electoral vote parity, but that’s not my issue here. To me, the biggest issue is the constant and intentional oversampling of one side, both nationally and on the state level. For example, in the VA CBS poll, democrats had a sampling of 35% compared to republicans’ sampling of only 24%. In a “swing state” I think that’s pretty ridiculous. And, as always, this was not reported by any major news outlet when they reported the "lead" Obama had in this "crucial battleground state."
“Swing states in particular may be rather close to the national average, but by leaving that to assumption it hurts the accuracy/credibility of the final numbers.”
I don’t think it hurts it that much. And I do ask, what hurts more, or which is more accurate: having a massive oversample of one side, or getting to a more balanced result by using two different demographic spreads to assess the situation?
As far as the issue with Gallup, I haven’t seen anything about that at all. Do you have any information about that?
Busy day again for me today, but I definitely hope I can respond.
The equation looks solid, thanks for posting it.
DeleteI agree in that, the MOE is always going to be present regardless of the method of assessment (whether we look at polls themselves or draw different analysis from them). Which is really ‘worse’ though, is impossible to tell without all the numbers. If we assume that there was major under sampling, then yeah, your numbers would likely be more accurate. But while I don't disagree with the possibility, it's just impossible to tell how much under sampling really took place and for which side, etc.
Likewise, the same applies for without knowing exactly how accurate the national average is and using a method that depends wholly on that and how closely it matches swing states. Though you do attempt to represent both sides, it’s tough to say to what degree each may be ‘biased’ or ‘inaccurate.’
One of the other problematic aspects being that while one would think that the swing states should more closely resemble a nationally averaged distribution of Rs and Ds, the fact that every election they all turn different ways and in varying margins speaks to how inaccurate it would be to classify all of them as having the exact same distribution of Ds to Rs as the nation averages. Otherwise, if it was really that simple, every election would have every swing state go one way and by the same margin… but of course, that doesn’t happen.
All of that said though, I do think you did very well with a solid analysis, and don’t mean to take away from that. I just don’t think that it can be used as much other than at best a well-supported theory, as opposed to a factual representation (though I realize you didn't imply that).
But then again I feel the same way for the poll numbers themselves, and seems that you do too!
I can definitely see where you're coming from. I still think if I had to have the pendulum swing towards one side, it would be a "safer" bet to poll a more even sample of D and R than to have, as with the VA poll, for example, 38% more democrats than republicans (hard number, not difference in percentage). I'd argue that there is no state, not even California where that happens. :-)
DeleteI'm probably weird, but I don't pay any attention to 'polls'... One reason for this: I'm pushing sixty years old, have been a registered voter since I was eighteen, and have always owned a home, a telephone AND a driver's license...
ReplyDeleteIn all these years, I have NEVER been polled for ANY state OR national election. They don't have ANY problem finding me for jury duty - and due to the fact that they'll NEVER let me sit - I ALWAYS get called up TWICE within a cycle. I don't take it personally - I'm married to a (now retired) police officer; prosecutors love me... defense, not so much.
Another reason: I believe ANY poll can be skewed - even when one side or the other ISN'T oversampled. They do this with the WAY they word the questions. I work with words - and a tiny substitution, within a phrase can change the ENTIRE gist of the inquiry... sometimes it's so subtle that someone NOT paying close attention can MISinterpret and answer in a completely different manner to what they intended.
There's also a psychological factor here - particularly with the Liberals. I can't find the articles (many) right now, but those who lean to the left tend to be more (for lack of better words) - mob oriented, they engage more in a group think/go with the flow mentality. If they stray - and it's discovered - they can and WILL be chastised by their peers. We've actually seen this behavior in action with the Occupy people.
Hannity had a guy (don't know his name) on last night, who chanted, 'Romney's done - he's toast' over and over talking over Hannity, raising his voice and interrupting Monica Crowley when she tried to refute. I held my nose and took a peek at MSNBC to find Rachael Maddow doing the same thing, w/her mouth twisted in her best imitation of a Snydley Whiplash leer. Chris 'thrill up my leg' Matthews touted a couple of days ago how Romney was arrogant to even THINK he could run against Obama.
Back to weird. IMO the MSM and most polls no longer have any interest in the truth, they have one, and only one agenda: to get Obama re-elected... period. It is my belief that there are NO depths to which they will not sink to succeed with that agenda.
Say you've lost your footing on a treacherous trail; you've slipped and are dangling from a cliff, clutching at a wispy clump of grass. Below you someone shouts, 'Don't look down!' Why? If you look down - you lose your focus - and you'll FALL.
You know your chances of surviving are pretty slim - but you have faith in your OWN abilities, so you take their advice, cast your gaze upward and struggle for the safety of the ledge above.
As you near you goal, one leg up - the other still dangling - another LOUDER VOICE shouts 'LOOK DOWN'...
THAT voice belongs to the MSM!
NO MORE YEARS Romney/Ryan 2012
You make a good point Dara that I wanted to reemphasize.
DeleteThese polling institutions come under fire very often for how they 'phrase' their polls, because as you said it's incredibly easy to influence the numbers by simply changing some words around. And when these margins come down to a mere few percent, such differences can greatly influence the ultimate result and picture the poll paints.
Although, one minor tidbit I disagree on; I think that the ‘segregate outsides of the party’ idea is present in both parties. I agree in that I believe it's a problem, but if I had to say, I'd even wager that it was more prominent on the GOP side as evident by the coined 'RINO' term and the rush in 2010 to replace ‘moderate’ Republicans with Republicans that hold a more rigid set of ideals. Any GOP member that didn’t agree with the others or didn’t vote in unison, was signaled out and even put a target on to replace. Say what you will about the Democratic party’s problems, but they’ve yet to go that far in singling out the variations within the party.
Dara -
DeleteAs much of a shock as this may be, I agree with you 100% in regards to polling. The questions are skewed to seek a specific answer. Unlike you Dara, I have been called. Another thing they do is limit your answers. I can't tell you how many times the answer I wanted to give was not a choice. So the questions tend to guide one to an answer that the poller likely seeks. The best example of this that I always go to is the study or studies that were done to name marijuana as 'the gateway drug'. Essentially, they went to 500 heroin users (making up numbers and the drug of choice they went after) and asked if they smoked pot before doing heroin. That answer will likely almost always be 'yes' as I'd imagine people don't generally go from cocoa to heroin (unless it's refined into cocaine). They didn't ask about alcohol or cocoa or caffeine or nicotine... They asked solely about pot and now it's 'the gateway drug' that all parents must be warned about. It's really crappy but it's par for the course in the world we live in.
One thing though that I do struggle with, and it's been brought up frequently on this site, is the MSM being liberal. I'm not saying it's untrue but, quite frankly, it seems like a childish excuse (I used to blame my dog for things I used to do wrong as a kid - seems like the same thing). When I look at the highest rated political shows (tv and/or radio), all of the top spots are conservatives. How can that MSM be liberal when the highest rated shows are all conservative. I am having trouble reconciling that in my head (just as I have trouble reconciling LME's assertion that Dems win elections because they give stuff away and nobody would bite the hand that feeds them - in that case, how does the GOP win ANY elections? - sorry, I didn't get to that thread until late last night and wanted to bring that up). When referring to MSM, who/what is it that makes it liberal?
@RKen - In case you couldn't guess ; )... I'm a Constitutional/fiscal conservative - who leans more toward the libertarian side of the equation. As the R platform 'more closely' fits my views, I register that way for voting purposes.
DeleteIMO a 'moderate Republican' might just as well switch parties... I believe the RINO's are even more to blame for most of the financial problems we now face. Instead of pushing back against the most socialist and wasteful programs; standing up for fiscal conservatism, they worried MORE about their OWN hides (re-election chances) and 'went along to get along'.
When the new Tea Party people arrived in Washington, the RINO's immediately got busy trying to change them - trying to 'make them fit in' with the 'old guard.' Didn't work and the result: they have been demonized - by their peers on BOTH sides of the aisle and by the Pres... for doing EXACTLY what We The People sent them there to do.
I was - and continue to be - front and center for removing ALL remaining RINO's from office - posthaste. TERM LIMITS.
Note: I'm fully aware of Romney's RINO tendencies... he is, however, NOT a socialist; SHOULD he win - I'll be one who holds his feet to the fire.
Where you're wrong: I could care less what others in 'my' party say or feel about my views... I voice them loud and clear - and take whatever heat the 'moderates' dish out. I've lost some 'friends' - but have gained many more.
We have a president who cares not one whit about the fiscal cliff we (ALL Americans) are teetering on, who cares more about buying votes with 'free stuff' than working to figure out how to fix it... who thinks he can 'tax and spend our way out' of a $16T hole while partying all over h*ll and half of Georgia (on our dime)... who has shown, more than once politically, he 'rewards his friends - and punishes his enemies'... who boldly lies to us about what happened and continues to happen - 'over there'... who openly voices disdain for American values, principles, and exceptionalism...who bows to America's enemies - and poops on her friends.. IMO it's time for a change, before it's too late.
'Question with Boldness - The Truth has NO agenda.'
@Whatsamatta - Wow! We agree on something!
DeleteAs far as the MSM being liberal. Perhaps because they share your POV - you don't see it? The top spots ARE conservative due to ratings... but they are BY FAR outnumbered.
On t.v., we have Fox... The left has ALL of the alphabet networks plus Current and a couple of other very HARD left (Francis Fox Piven, Bill Ayers, Van Jones sort of stuff) channels, on cable.
Aside from Hannity and Meagan Kelly, many of those on Fox have taken to being 'more moderate'... or straight up leaning left. Their 'fair and balanced' approach once showed the differences between us... and where we agreed. Those lines have been blurred lately as Fox moved further left... (but, maybe it's me)
'...in that case, how does the GOP win ANY elections?' thanks for the chuckle :)
Honestly? I don't know... maybe there are many more people out there - like me - who are just too stubborn to give up?
The show's not over 'till the fat lady sings...lol
Thanks for the explanation Dara. I do agree that there are many more liberal outlets but their ratings are just nowhere near the conservatives - (http://www.huffingtonpost.com/2012/02/29/cable-news-ratings-top-30-february_n_1310969.html#s739933&title=1_The_OReilly). Again, it's just something that I don't understand. Perhaps you're right that I can't see the forest through the trees. Frankly, I don't watch much political tv anymore because I just think they're all full of it. I like the internet (thanks to Al Gore :)) for my news. I can read it instead of hear it and I can look for a counter point.
Delete