Quantcast

Categorized | News

2014 Senate Polls are All Skewed Up

There are a lot of things going on politically that are working against the Democrats this mid term election. And trying to take a poll and “weight” it to reflect the 2014 electorate is a difficult if not impossible task.

And to top it off, some of these polls may be less concerned with being accurate than they are at making things look better for Democrats – who are often paying the pollsters. And we all know how Democrats like love to drive the narrative.  Especially when the expectations this year are for the Democrats to suffer a bloodbath at the polls.

The annual “tightening of the polls” is underway and things are starting to move. But watching the polls over the years I have noted several times that the races are not getting tighter as much as the polling demographics are changing. Polls that were polling +9 or +15 Democrats in spite of historic data showing that the party affiliation is actually closer suddenly start using models more founded in reality. So the polls in many instances are not “tightening” but the models start changing in order to try to get as close as possible in order to maintain credibility in future polls.

Or in other words, they stop trying to control the narrative and start working towards accuracy. The closest polls get bragging rights which they use to drum up business (and money) next year.

But it looks like some are getting smart. They realize that intentionally skewed polls are called out so they look for ways to “hide” the bias. And it should also be noted that Democrat polling firms are far more likely to use questionable weighting models than Republican or non-partisan pollsters.

So sometimes you need to look beyond the percentages of Republicans, Democrats and Independents and dive deeper into the numbers. And many don’t give you the numbers behind the numbers, also called internals.

And there are also some questions about which models pollsters are using. Obviously the ones using 2012 Presidential Election numbers are going to be off. Far fewer people turn out and while blacks turned out in large numbers in 2008 and 2012 to elect Obama, they largely stayed home in the 2010 mid term election.

And this is a mid term election.

And we must also consider several other factors.

When I see polls that fail to make the real numbers available I really don’t give them much credibility because there is no way to verify their results. And it seems that the more out of line with the other polls they are the less likely they are to show their data.

In looking at some of the current polls, these guys are getting creative with the numbers. Common sense tells us that 2014 will be more like 2010 than either 2008 or 2012. Obama is not on the ballot so the black vote will be considerably less than in the presidential election years.

Lets look at two examples of polls that use weighting to show a narrative. We will look at the latest poll from Arkansas which plays with the weighted averages to arrive at a result that shows arace closer than it really is and a Virginia poll that is using a pretty good model now, but used more Democrat rich models in their past 2 polls that caused the news stories to report a wider margin than most likely existed. With a very wide margin, the Party is less likely to throw a lot of money into a race. So polls are much more than just “feel good” instruments for one party. They actually effect the ground game. And Democrats play this game well.

In looking at the latest poll in Arkansas by NBC/Marist is it reasonable to believe that blacks will turn out there in greater numbers than in 2008? NBC believes so. In 2008 83% of the Arkansas electorate were white and 12% were black. In 2010 83% were white and blacks were down to 11%. And now, NBC would have us believe that this year, without Obama on the ballot that blacks will be 15% of the electorate and whites will be just 79%. And the statistics show that blacks vote for the Democrat most of the time.

In the same poll, age was skewed towards younger voters that also vote Democrat more often than older voters. In fact the poll admits Respondents in the household were selected by asking for the youngest male.

In the 2010 exit polls, 27% of the voters were under 45 years of age and 74% were 45 or older. While in 2008 44% were under 45 with just 56% older. This new poll predicts 2014 will see more under 45 year old voters (35%) than the 2010 mid terms with 65% older than 44. I am not sure how they justify that logic.

So I believe this poll which shows Republican Cotton up by only 2% has overestimated the blacks and young people that will show up to vote. But what this poll did to the Real Clear Politics average was to take the Arkansas race from “Leans GOP” to “toss up”.

Amazing how that works.

And if we take a look at the CBS/NY Times polling for the just released Virginia US Senate poll we find that the Democrat Mark Warner is leading Republican Ed Gillespie by 10 points, 49% to 39%. This poll does provide the raw numbers but only gives the “weighted” averages, which is only as accurate as the model they use.

So how does the CBS/NY Times model compare to exit polls from previous elections?

Weighted polling shows that 34% were Democrats, 28% Republicans and 33% were Independents. (Just to note the raw data shows that they actually polled 30% Democrats, 30% Republicans and 39% Independents.) So the “weighted” polling model shows Democrats +6.

By sex, the weighted poll numbers show 48% male, 52% female.

By race, 74% white, 18% black.

By age, 36% under 45, 64% over 45.

In 2012, Democrats were 39% of the electorate and Republicans were 32%.

By Sex 47% male to 53% female.

By Race – 70% white, 20% black.

By age – 46% under 45, 55% over.

And let’s look at exit polls for last year’s Governor’s race in Virginia:

By Party – 37% Democrats, 32% Republicans 31% Independent.

By sex – 49% male, 51% female.

By race – 72% white, 20% black.

By age – 36% under 45, 64% over.

So the CBS/ NY Times poll model looks to be pretty close to the 2013 governor’s race in Virginia. In 2012 voting by women, blacks and younger people was up and last year, absent a presidential race, the demographics showed a drop in participation by Obama voters. Democrats still won, but the overall turnout was down from 2012.

I think the 2013 model will be pretty close to the 2014 mid terms and it looks as if CBS/NY Times put out a realistic model.

So while the Arkansas poll looks to be modeled after unrealistic (for 2014) demographics, the Virginia CBS/NY Times poll fits my expectations. And while this is not such good news for Ed Gillespie and Virginia Republicans, it is probably good news for Republican Cotton in Arkansas.

Warner by 10 points in Virginia passes my sanity check, Cotton by 2 is a bit low. It is probably Cotton by 5 or 6.

And if we look back and examine the CBS/NY Times polls from September and August it looks like the polls are “tightening”. But the fact is, the only real significant change is the fact that the 2 polls prior to the latest one sampled Democrats at +10 and this poll samples Dems at +6. And the 2 older polls showed Democrats up by 14 points, the current poll Dems are up by only 10. There’s the 4 point difference.

So as we get closer to election day it looks like there is some movement towards the Republican, but the polls have remained steady. The pollsters are starting to go with more realistic poll weighting.

September CBS/NY Times poll.

By Party – 36% Democrats, 26% Republicans 35% Independent.

By sex – 50% male, 50% female.

By race – 75% white, 16% black.

By age – 21% under 45, 79% over.

About Tom White

Tom is a US Navy Veteran, owns an Insurance Agency and is currently an IT Manager for a Virginia Distributor. He has been published in American Thinker, currently writes for the Richmond Examiner as well as Virginia Right! Blog.Tom lives in Hanover County, Va and is involved in politics at every level and is a Recovering Republican who has finally had enough of the War on Conservatives in progress with the Leadership of the GOP on a National Level.

13 Responses to “2014 Senate Polls are All Skewed Up”

  1. souvenir pernikahan says:

    I am genuinely grateful to the owner of this web site who has shared this fantastic
    piece of writing at at this time.
    souvenir pernikahan recently posted…souvenir pernikahanMy Profile

  2. twink porn says:

    I think this is one of the most vital information for me.
    And i am glad reading your article. But want to remark on few general things, The
    website style is great, the articles is really nice :
    D. Good job, cheers
    twink porn recently posted…twink pornMy Profile

Trackbacks/Pingbacks

  1. […] VA Right! – 2014 Senate Polls are All Skewed Up […]

  2. […] VA Right! – 2014 Senate Polls are All Skewed Up […]

  3. […] Fifth place with 2/3 vote – VA Right! – 2014 Senate Polls are All Skewed Up […]

  4. […] Fifth place with 2/3 vote – VA Right! – 2014 Senate Polls are All Skewed Up […]

  5. […] Fifth place with 2/3 vote – VA Right! – 2014 Senate Polls are All Skewed Up […]

  6. […] Fifth place with 2/3 vote ? VA Right! – 2014 Senate Polls are All Skewed Up […]

  7. […] VA Right! – 2014 Senate Polls are All Skewed Up […]

  8. […] Fifth place with 2/3 vote – VA Right! – 2014 Senate Polls are All Skewed Up […]

  9. […] Fifth place with 2/3 vote ? VA Right! – 2014 Senate Polls are All Skewed Up […]

  10. […] Fifth place with 2/3 vote – VA Right! – 2014 Senate Polls are All Skewed Up […]


Leave a Reply

Your email address will not be published. Required fields are marked *

CommentLuv badge

Tom White Says:

Nothing is more conservative than a republican wanting to get their majority back. And nothing is more liberal than a republican WITH a majority.

Check out NewsMax!

Sign up for Virginia Right Once Daily Email Digest

No Spam - ever! We send a daily email with the posts of the previous day. Unsubscribe at any time.
* = required field

Submit a Blog Post!

Submit a Blog Post for our 'Boots on the Ground' feature

Click Here for Instructions on How to Submit a Post

Google Ad

Google Ad

Follow Us Anywhere!

Google Ad

Archives

Facebook Auto Publish Powered By : XYZScripts.com
%d bloggers like this: