Posted on 09/15/2012 9:17:57 AM PDT by Arthurio
Got an email from Romney OH campaign saying the WSJ/Gallup was badly off using 08 samples, and that right not it is tied. I suspect that this time GOP turnout will make the difference, but we thought that in 08. This time, I doubt tha ANY Republicans will be voting for Zero.
bttt
It is a standard technique in the social, behavioral and biological sciences called “stratified random sampling”: if you know a population is composed of several radically different sub-populations and you know (or think you know) the proportion falling into each, one gets better statistical models of the population as a whole by taking a random sample from each sub-population in proportion to their size (or not, but scale to weight the results according to the actual or surmised proportions).
The problem here is that the population you really want — folks who actually turn up to vote in November — doesn’t actually exist yet, so the proportions for the stratification *have to be* surmised. For realistic assessment of the likely outcome, as opposed to generating pro-Obama propaganda, using the proportions from 2008 is sheer folly. Averaging 2008 and 2010, or taking 2008 turn-out and shift it by the shift in voter-registration as a basis for the stratification would make much better sense.
Freepers, this is a conspiracy! The lame-stream has been putting out poll after poll after poll using 2008 voting percentages for the parties. Dims are being severely OVER-SAMPLED in these polls. Turnout is going to be closer to 2010 levels with Republicans even more fired up than they were then.
So WHY is the lame-stream doing this even though every poll exposes their corruptness when you click on the internals?
Answer: If they put enough polls out there, then we will look like the wackos if we claim there is a conspiracy. Repeat the lie that Zero is leading the polls enough times, and the uninformed will believe it.
Already I am seeing comments from Dims saying “ALL these polls can’t be wrong. You republicans just can’t see reality.”
This is what the lame-streams want to happen!
How do we fight it? Get onto EVERY site possible and EXPOSE every crap poll by showing what the party breakdown is.
PROOF THAT THESE ARE CRAP POLLS: Remember the Wisconsin Recall election? At 5pm on election day all the media outlets were quoting an exit poll that said the result was TOO CLOSE TO CALL. That exit poll over-sampled Dims JUST LIKE the polls that give Zero a lead over Romney. Do we remember the Wisconsin Recall result? Walker won 53-47.
Freepers, it is up to US to keep reminding people of how these crap polls were WRONG in Wisconsin and they will be WRONG for Zero. The lame-stream wants the uneducated to FORGET about the Wisconsin result and to IGNORE the over-sampling of Dims in these presidential polls.
DON’T LET THEM GET AWAY WITH IT! GET BUSY LEAVING COMMENTS!
It is great news for Romney-Ryan. Early voters are almost always Republican voters and they are highly motivated this year to vote.
Based on the breakdown of 42% Democratic/32% GOP/26% Indie (GOP -10) you end up with the following based upon historical turnout in Ohio according to CNN/Wiki over the past three election cycles:
2010: 37 (GOP), 36 (DEM), 27 (IND/OTHER) (GOP +1)
2008: 31 (GOP), 39 (DEM), 30 (IND/OTHER) (GOP -8)
2004: 40 (GOP), 35 (DEM), 25 (IND/OTHER) (GOP +5)
If you re-weight this poll just based strictly on party affiliation you get the following results based on prior races:
POLL: 47.12% Romney, 48.06% Obama
compared to:
2010: 51.98% Romney, 43.13% Obama
2008: 48.08% Romney, 46.82% Obama
2004: 53.70% Romney, 41.55% Obama
If you re-weight by party ID and apply an estimate that the “Other” vote will split 50/50 (taking from each candidate equally) and that Undecideds break 60/40 to Romney you get these results based on historical turnout:
2010: 54.09% Romney, 44.54% Obama
2008: 50.19% Romney, 48.23% Obama
2004: 55.81% Romney, 42.96% Obama
Husted is a Republican and he did not say which county. However, early voters are almost always Republican voters and they are high motivated to vote.
Does someone have a link to the party ID of this poll? I can’t find it at the link - http://americanresearchgroup.com/pres2012/OH12.html.
If it is indeed a DEM +10 poll this, like the other polls recently shows a lead even using the 2008 turnout results.
Nevermind - I see it now — sorry, just want to make sure that when I run these numbers I’m giving accurate info.
Some have asked “Why the oversampling of Dems?”
Biased reasons aside (and they’re a lot of those) many in the biz think the voting turnout will be similar to the 2008 models (jazzed up DEM base). They COMPLETELY ignore the 2010 results and current enthusiasm polls.
Anyone remember the PPP MO Senate poll weeks ago that showed Akin ahead of McCaskill by one point AFTER his major gaffe? Pew oversampled Republicans by nine points. Akin took the bait and stayed and whammo, Pew released an “update” shortly thereafter showing him down by ten.
If Romney is ONLY down by one, and they oversampled DEMS by +10, then you know it’s not good for Obama.
Either these pollsters see something we don’t or they’re tossing out BS like a farmer fertilizing the field.
http://www.video.theblaze.com/media/video.jsp?content_id=24739323&source=THEBLAZE
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.