09 October 2012

About that Rasmussen poll from Ohio from 10/05....





As we call all tell, there is contra-indicative polling coming in post-Debate no. 1 between President Obama and Governor Romney.

Scott Rasmussen, of Rasmussen Research, put out a poll of the all-important state of OHIO on 10/05. The poll, according to his website, was conducted all on 10/04 - completely after the first debate - from the usual 500 LV and a MoE of +/-4.5.

The results:

Obama 50
Romney 49
Margin: Obama +1


Up until 10/05, Obama had a margin average of +5.55 in the Buckeye State, which is just outside of the Battleground zone +5 boundary. After that RAS poll, the average fell to Obama +4.93 and right now, it is Obama +4.52.  At this exact time in 2008 (October 6th, 2008 29 days before the election), it was Obama +4.20 average over McCain.

So, people see a 1-point spread in a poll and either they start to panic (some Democrats) or they start to cheer (some Republicans), but please remember, CONTEXT IS EVERYTHING.

Here is a complete listing of all the Rasmussen polls for Ohio for 2012 to-date (many will find this very, very interesting):

Nr.OHIO – Rasmussen polls onlyDateSampleMoEObamaRomneyUndMar.

AVERAGEN/AN/AN/A












49Rasmussen05.10.12500 LV+/-4.5504911
36Rasmussen13.09.12500 LV+/-4.5474671
26Rasmussen14.08.12500 LV+/-4.54545100
22Rasmussen19.07.12500 LV+/-4.5474582
16Rasmussen31.05.12500 LV+/-4.5444610-2
10Rasmussen20.04.12500 LV+/-4.54642124
8Rasmussen28.03.12500 LV+/-4.54840128
2Rasmussen09.02.12500 LV+/-4.54541144



There have been exactly 8 Rasmussen polls of the Buckeye State for all of 2012 thus far (there were 12 RAS polls of Ohio in 2008 at this exact corresponding time: 10/06/2008). Please notice that the last RAS poll was also "only" Obama +1 and the one before it was an absolute tie. The one before that was Obama +2 and the one before that was Romney +2. In fact, you have to go all the way back to April to find a RAS poll of Ohio  with a margin more than +2.

Like I said, context is everything. This means that the RAS poll results have barely contributed to the circa Obama +4 average that we have been seeing in Ohio for many months now. Remembering that Nate Silver at 538 has mathematically proven not just once, but at least three times, that Rasmussen has a mathematical bias on the order of +4 to the right (see HERE and HERE), we can also safely assume that there is still mathematical bias in a RAS poll today, so a +1 for Obama from RAS could just as well be a +4 or +5 in real life...

And you don't have to go farther than the last RAS poll of Ohio in 2008 to see this:


Ohio 2008: 17 Rasmussen polls


Pollster
Date
Obama
McCain
Other
Und.
Margin
FINAL AVG:

48.99
46.68
0.91
3.09
+2.31
11/03
49
49
1
1
0
10/27
49
45
1
4
4
10/20
47
49
1
2
2
10/16
49
49
1
1
0
10/13
49
47
1
3
2
10/06
47
48
--
4
1
09/29
47
48
--
4
1
09/25
46
47
3
3
1
09/22
46
50
2
2
4
09/15
45
48
1
6
3
09/08
44
51
1
3
7
08/20
43
48
2
6
5
07/22
40
46
7
7
6
06/19
43
44
7
7
1
05/19
44
45
5
2
1
04/10
40
47


7
03/14
40
46


6



Yes, the last RAS poll for Ohio in 2008 showed an absolute tie: Obama 49 / McCain 49. Actual result from 2008: Obama +4.58. This means that Rasmussen was off on the margin in Ohio by 4.58 points, or 4.6 points, rounded. (Note: these links are probably defunct. I used the Rasmussen paywall in 2008. But these numbers are also corroborated here at RCP).



Go back and look at that top number for 2012, for the poll from 10/06:

Obama 50.

Now, look through all the other RAS polls for 2012 and 2008.

See any other "50" for Obama anywhere?

That's right. Nope, there are no other "50"s for Obama. In 2008, he hit "49" by RAS four times, but never "50". And thus far, in 2012, the second highest topline number for Obama was one "48".

BTW, there were thirteen 2012 GE polls taken for Ohio in 2011, but Rasmussen did not poll Ohio even one single time in that off-year.

Let's go back to 2004, the first presidential election that Rasmussen polled (he started up in 2003):




Ohio 2004 Polls – 11 Rasmussen pollsSend This Page to a Friend
Presidential Race





2004 Electoral Votes: 20 | 2000 Results: Bush 50.0, Gore 46.5, Nader 2.5 (Bush +3.5) 
Poll | Date
Sample
MoE
Bush
Kerry
Nader
Spread
RCP Average | 10/25-11/1
-
-
48.8
46.7
Off Ballot
Bush +2.1
600 LV
4.0
50
46
-
Bush +4
537 LV
4.0
50
46
-
Bush +4
537 LV
4.0
47
47
-
TIE
564 LV
4.0
49
47
-
Bush +2
597 LV
4.0
48
47
-
Bush +1
668 LV
4.0
48
45
-
Bush +3
LV
4.0
46
48
-
Kerry +2
LV
4.0
45
46
1.0
Kerry +1
n/a
4.0
46
42
-
Bush +4
n/a
4.0
46
44
-
Bush +2
500 LV
4.5
41
45
-
Kerry +4

Rasmussen polled the Buckeye State 11 times in 2004. Kerry's highest topline by Rasmussen was 48, one time. Rasmussen predicted a Bush +4 right before election day. Actual result: Bush +2.11. In 2004, RAS was off by +1.89 in margin - to the Right.

Facit: what we have just seen on 10/05 from Rasmussen was a history-making first: the first time that a Democratic Presidential candidate has hit the "50" mark in a Rasmussen poll of Ohio. Ever.

No comments:

Post a Comment

Constructive comments and critique are always welcome. Please keep it polite and respectful.