Posted on 09/18/2006 7:21:00 PM PDT by dvwjr
WASHINGTON (AP) - President Bush and Republicans want to convince voters the unpopular Iraq war is central in the anti-terror fight. Democrats argue they can win control of Congress if voters view Iraq - and the continued bloodshed there - on its own.
The latest Associated Press-Ipsos poll found Republicans haven't made their case despite a sustained effort to link the conflicts; a majority of the public views the two as distinct.
(Excerpt) Read more at apnews.myway.com ...
Talk about burying the lead... The AP-Ipsos Bush Presidential job approval numbers for September were up six points from 33% to 39% - the largest single monthly change (in either direction) since the AP polls before the 2004 election. AP pollitorials usually trumpet 3 point monthly drops, not so on the upside. This sharp jump might not have occurred but for the fact that these September AP/Ipsos poll samples had Democrats out-numbering Republicans by a mere 8 percentage points. Last month in August that poll sample had Democrats outnumbering Republicans by 14 percentage points, which probably masked the Bush August uptick which was detected by other polls. So all the statistical chickens came home to roost this month instead of being spread over two months.
Now on to the central thesis of the AP article about the control of the US House of Representatives come November 2006. Let's take a look at the composition of the AP/Ipsos poll sample for September and the poll results for certain questions which resulted in the "Poll Shows GOP Not Making Its Case" AP headline.
Poll questions 2a, 2b, 2c, 2d were about party ID affiliation and how strong is such affiliation.
Poll questions 1a and 1b were about which party's House candidate would get your vote and how sure are you that you would vote that way in November.
Holy correlation, Batman! The results of the 'which party gets your vote' questions agree with the Registered Voter party ID sample composition of the AP poll. Who'da thunk it? Note the stunning congruence between the 17% 'Strongly Republican' poll party ID in the polling sample and the seemingly un-related fact that 18% of all the registered voters polled who intend to 'Definitely Vote Republican'. It boggles the mind... Look at the Democrats - those who are 'Strongly Democrat" represent 24% of the poll's party ID sample. Now contrast this with the suspicious fact that 27% of the registered voters in this poll intend to 'Definitely Vote Democrat' come November. Now there are co-incidences and then there are co-incidences. One might even draw the conclusion that the composition of the party ID sample of this AP/Ipsos poll has something to do with the results of the very poll questions about which party's candidate would get the nod from the registered respondents.
What changed from last month's AP/Ipsos poll in August of 2006? Now their September poll sample composition of Republicans and Democrats are now separated by 8 percentage points vs last month's sample where the difference was 14 percentage points... In a defensive move, AP-Ipsos reveals in this month's polling results that their initial screen of poll respondents consisted of 26%R, 35%D, 27%I, 11%None, 1%NotSure. (Hmmm, even more Independents than Republicans?) Compare this to last month's 27%R, 35%D, 24%I, 13%None, 1%NotSure. Now maybe it's just me, but if someone does not self-identify as a Republican or Democrat - they are by these poll definitions an Independent (it does not matter if they are communist, socialist, vegan, green, etc). So what AP-Ipsos has done is to start with a sample that was 26%R, 35%D, 39%I - then they push 'Independents' for leaning Republican or Democrat, if they don't lean the person is classified as a hard-core 'Independent'. Looks like AP has salted the sample again by hiding Democrats in the overly large 'Independent' category.
Just to add to the media polling bias, here is the recent history of the "ever-reliable" AP/Ipsos polls (January 2005 - September 2006) and their political Party ID sample composition for each poll. The AP/Ipsos polls push Independents to identify with either Republicans or Democrats, only the 'hard-core' remain as self-identified 'Independents'...
This is an mixed ADULTS, Registered and Likely Voters poll, not just ALL Adults. The new "Likely Voters" category is not compared so that all previous comparisons will be consistent. Note the persistent gap in the party ID between Democrats and Republicans, strange given that the 2004 Presidential Election exit polls identified 37% Republicans, 37% Democrats and 26% Independents among the voting American public. The Republicans have achieved voter self-identification parity with the Democrats for the first time since the late 1920s.
This series of AP/Ipsos poll result on Presidential job approval and the American public views on the 'direction of the country' was as always based on a skewed demographic sample that has 'averaged' 40% Republican, 50% Democrat and 10% Independent composition since January 2005. I think that when the AP poll fails to get the results they desire, they 'play' with both the registered voters party self-identification composition sample and vary the un-registered voter portion of the poll sample. Since the Registered voter portion of the poll averages around 79% and the 'breathing' voter averages around 21% you can see they have quite a bit of "statistical slack" to play with to get the results they require...
Note: The AP/Ipsos polls are published in PDF format and become unavailable after one month. You must be a 'premium' member to access those PDF files on their web site. No consistent links are provided for that reason.
Source: AP-Ipsos Poll, September 11-13, 2006 Project #81-5681-09
Hope this helps,
dvwjr
Ping. This month's AP/Ipsos poll data.
dvwjr
What pathetic spin from the editors. Should be titled: Bush starting to bitch slap the democrats, they are starting to squeal.
So if in one month 6% really changed their party affiliation as AP claims then the logical explanation could only be related to gas prices because that's the only real drastic changing factor.
THANKS and BUMP!!!
Too bad the MSM is not subject to the truth in advertising laws.
Imagine if your analysis had to be disclosed with EVERY singlie poll report.
Bush running for anything? Republicans biggewst problem is they all sound like McCain - and spend like him too.
I often wonder why there isn't a Fox News/NY Post poll to counter this leftist propaganda, it is the last monopoly the socialists enjoy.
Fox is tued to opinion dynamics.
Fox is tied to opinion dynamics.
The threat to the pope is going to make a lot of people really start to wonder if they sincerely want the demorats protecting us.I think that should get Bush another 6 points or more.
Poll Shows GOP Not Making Its Case -- NOT
Poll Shows GOP Gains Ground In Battle For Congress
There, fix it!!
All those numbers/charts/graphs establishes clearly that things are undecided.
And will not be decided until election day.
Mr Glover is a promiment reporter/dem who does a reasonable job of minimizing his personal bias. One evening he did note that his daughter was working on the campaign of dempcrap Blowhard--which at least is noting your investment choices upfront.
Karl Rove orchestrated it. That and the London airplane bomb plot. He is just EEEEEEEVIL
Thanks for the info. :0)
Note that the best poll of all, with real money down, is the Iowa Electronic Markets.
The current quote is "Republicans Hold House" 3.5 % above "Republicans Lose House".
Great work. I know that analyzing this data takes time, and I'll bet that AP is counting on most folks not taking the time you took to get into the subtleties of the raw data. I'll also bet that AP/IPSOS isn't the only polling outfit manipulating the data. In fact, I think that polls are a tool of political warfare for the left. The polling before the last presidential election didn't indicate that Bush would win as decisively as he did. Since then, the political debate has largely been driven by poll results, quite possibly skewed like this one. Again, great work.
crAP sucks!
Is it possible that their bias is the result of wishful thinking rather than willful manipulation? Or maybe they just tweak the parameters until they get the result that they "know" must be true?
Oh it gets better. On the morning the poll was released, the AP changed to title of their article to hide the positive Bush news.
http://www.freerepublic.com/focus/f-news/1703043/posts
The original title was "GOP Gains Ground In Battle For Congress" and a few hours later it miraculously became "Poll Shows GOP Not Making Its Case"
AP - scumbags
I have always liked the term- bleat to describe what democrats do when they condescend, accuse, pontificate, and bloviate...
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.