You are viewing an obsolete version of the DU website which is no longer supported by the Administrators. Visit The New DU.
Democratic Underground Latest Greatest Lobby Journals Search Options Help Login
Google

Reply #133: Mitovsky's "Exuberant Kerry survey responder" identified [View All]

Printer-friendly format Printer-friendly format
Printer-friendly format Email this thread to a friend
Printer-friendly format Bookmark this thread
This topic is archived.
Home » Discuss » Topic Forums » Election Reform Donate to DU
Blue22 Donating Member (42 posts) Send PM | Profile | Ignore Fri Mar-18-05 11:21 PM
Response to Original message
133. Mitovsky's "Exuberant Kerry survey responder" identified
Minvis,  As promised I have created some aggregations of NEP
data by precinct with some of the key demographics and vote
results that we will need to compare with reported votes and
census characteristics for the matched NEP precincts.  Below
is an example of the Ohio exit poll data but aggregated
statewide.  The first two columns show the characteristics of
the 2042 survey respondents before weighting and after
weighting.  

Mitovsky applied one weight to each response to try to match
Bush's vote total in the precinct but this obviously changed
the population characteristics as well.  For instance, Kerry
received 53.5 percent of the vote in the unweighted survey and
46.5 percent after Mitovsky had diminished the value of some
of the respondents who voted for Kerry.  This brought Kerry's
total closer to the reported statewide vote total and turned
an 8 point raw survey advantage into a 2 point loss post
weighting, quite a large adjustment.

Mitovsky's weighting also changed the characteristics of the
overall sample.  For instance, the number of female voters (a
Kerry majority group) fell from 54.2% to 52.7%.  To do this
Mitovsky had to change the equivalent of 30 women into 30 men.
Yikes! Further, Kerry's percent of the weighted female vote
fell even further from 54.5% to 49.5% suggesting that Mitovsky
also diminished the value of remaining female respondents who
voted for Kerry compared with those who voted for Bush.  In
order to get the thing to match Bush's reported vote Mitovsky
made the entire survey group more male, older, more white,
less poor, more protestant, less educated, more married, more
Republican and more religious.  In other words Mitovsky's
exuberant Kerry responder had a very specific profile.  She
was young, non-white, unpoor, catholic/jewish/agnostic,
single, not pious and a Democrat.  Sheesh.  Isn't this a
stretch?  I feel like I could go to Ohio and pick her out of a
crowd on the street.

While Mitovsky's exuberant Kerry respondent thesis continues
to look ludicrous, his weighting to match the Presidential
reported vote had the side effect of making the survey also
more closely match the reported results for the Senate race
and Ohio Issue #1 (look way down toward the bottom on the
table).  This is consistent with several theories: 1) that
there were too many liberals in the survey, 2) that general
vote suppression was more effectively targetted than believed
or 3) that the manipulation occurred beyond the Presidential
level.  The latter is somewhat illogical because why would
anyone try to jack up the vote count in a race that was a
landslide?

I'll report on the weighting for our matched NEP precincts
over the weekend.  As always, anyone who wants a copy of the
Excel formatted version of the NEP Ohio survey, just send me a
PM.


Effect of Mitofsky reweighting  on vote and demographics in
Ohio Exit Poll				
				
	      Raw     Wgt'd  Kerry%Raw	Kerry%Wgt'd
				
Number	     2,042.0   2,040.9   1092	950
				
Demographics of sample				
				
Total	       100.0%	100.0%	53.5%	46.5%
				
Male	        44.9%	46.1%	52.2%	46.8%
Female	        54.2%	52.7%	54.5%	49.5%
				
Age18-24	13.0%	11.2%	65.7%	61.6%
Age25-29	9.9%	9.9%	54.2%	49.9%
Age30-39	19.4%	18.0%	54.7%	50.0%
Age40-44	12.0%	11.5%	46.1%	41.7%
Age45-49	11.8%	11.2%	52.1%	45.9%
Age50-59	19.0%	17.5%	55.3%	48.8%
Age60-64	6.0%	8.1%	49.2%	44.7%
Age65-74	6.3%	9.1%	46.5%	45.3%
Age75+	        2.2%	3.1%	31.1%	29.4%
				
Race-White	83.9%	85.0%	48.4%	43.5%
Race-Black	10.8%	10.1%	85.5%	80.9%
Race-Hisp	1.8%	1.6%	77.8%	72.6%
Race-Asian	1.0%	0.9%	70.0%	65.4%
Race-Oth	1.1%	0.9%	68.2%	63.8%
				
Inc<15000	7.4%	6.8%	71.5%	68.0%
Inc15k-30k	14.0%	14.2%	68.2%	62.7%
Inc30k-50k	22.5%	22.6%	54.8%	49.3%
Inc50k-75k	19.5%	19.8%	47.6%	40.9%
Inc75k-100k	13.8%	13.4%	49.1%	45.4%
Inc100k-150k	8.1%	7.8%	46.4%	41.2%
Inc150k-200k	3.1%	3.3%	42.2%	36.5%
Inc200k+	1.8%	1.8%	35.1%	34.1%
				
RelProt	        48.5%	50.2%	47.2%	43.1%
RelCath  	23.6%	23.3%	48.2%	43.9%
RelJew	         0.8%	0.8%	64.7%	57.8%
RelOth  	6.0%	5.2%	79.7%	72.2%
RelNone 	10.4%	9.1%	74.1%	67.9%
				
EdNoHS  	2.9%	3.1%	64.4%	57.6%
EdHS    	25.6%	26.0%	54.2%	50.0%
EdSoCo   	26.3%	25.1%	52.9%	47.2%
EdColl  	22.1%	21.8%	48.5%	44.3%
EdCol+  	12.4%	11.7%	56.7%	48.7%
				
MarYes  	56.8%	58.4%	45.3%	40.3%
				
SzPl500k+	4.8%	6.1%	48.5%	42.5%
SzPl50-500k	19.7%	19.2%	67.9%	61.8%
SzPlSubur	49.8%	49.5%	53.8%	48.3%
SzPl10-50k	7.1%	6.3%	51.4%	49.1%
SzPl<10k	18.8%	18.9%	39.4%	35.7%
				
BrnAgnYes	27.6%	28.9%	36.7%	31.9%
				
Atnd1+wk	11.6%	12.4%	36.4%	31.0%
Atnd1wk 	22.9%	23.8%	39.6%	36.1%
Atnd1+mo	14.1%	13.7%	53.7%	49.0%
Atnd1+yr	27.2%	25.5%	63.8%	59.7%
AtndZero	13.3%	12.3%	68.6%	62.5%
AtndOmit	2.4%	2.8%	48.0%	41.6%
				
GeoCuy  	12.4%	12.0%	67.3%	63.5%
GeoNE   	25.1%	26.5%	54.4%	49.6%
GeoCent 	19.8%	20.3%	45.4%	40.1%
GeoNW   	16.2%	14.4%	57.1%	51.6%
GeoSW   	26.4%	26.7%	49.9%	44.2%
				
The Vote				
				
Dem     	37.3%	33.4%	90.7%	88.9%
Rep     	33.2%	37.2%	5.8%	6.3%
Ind     	23.6%	23.3%	63.1%	57.8%
Oth     	5.9%	6.2%	48.8%	45.0%
				
Raw Kerry	53.5%	48.2%		
Raw Bush	45.2%	50.2%		
Raw Other	0.2%	0.4%		
Raw No vote	1.1%	1.1%		
				
SRawFing	39.9%	34.1%	94.5%	93.1%
SRawVoin	54.7%	60.1%	24.0%	22.8%
SRawOth  	0.0%	0.0%		
SRawNoVt	0.0%	0.0%		
				
Iss1Yes  	54.8%	59.4%	34.9%	31.8%
Iss1No  	41.2%	36.6%	77.6%	73.9%
Iss1NoVt	0.0%	0.0%		
Printer Friendly | Permalink |  | Top
 

Home » Discuss » Topic Forums » Election Reform Donate to DU

Powered by DCForum+ Version 1.1 Copyright 1997-2002 DCScripts.com
Software has been extensively modified by the DU administrators


Important Notices: By participating on this discussion board, visitors agree to abide by the rules outlined on our Rules page. Messages posted on the Democratic Underground Discussion Forums are the opinions of the individuals who post them, and do not necessarily represent the opinions of Democratic Underground, LLC.

Home  |  Discussion Forums  |  Journals |  Store  |  Donate

About DU  |  Contact Us  |  Privacy Policy

Got a message for Democratic Underground? Click here to send us a message.

© 2001 - 2011 Democratic Underground, LLC