Wisconsin Walker Recall True Vote Model
The version of the browser you are using is no longer supported. Please upgrade to a supported browser.Dismiss

Still loading...
ABCDEFGHIJKLMNOPQRST
1
Richard Charnin
2
3
7/24/2012
4
http://www.nytimes.com/interactive/2012/06/05/us/politics/wisconsin-recall-exit-polls.html?_r=0
5
This document contains all the Wisconsin True Vote Models for recalls and recounts in separate worksheets from 2011-2012.
6
7
Walker won in 2010 by 124,638 votes with a 52.3% share.
8
Iin 2012, he won the recall by 171,105 votes with a 53.1% share.
9
10
The following DEFAULT assumptions can be overidden on the INPUT sheet:
11
1) Equal turnout rates for returning Obama and McCain voters.
12
2) Barrett's share of returning Obama voters is estimated as an increment applied to his base case 90% Wisconsin share.
13
3) Shares of New voters are set to the estimated 2008 vote shares.
14
4) Barrett won 5% of returning McCain voters, a conservative estimate. According to the WI 2010 Exit Poll, Barrett had 7%.
15
16
I just added the option of letting the user set Barrett shares of returning Obama and McCain voters as defaults on the Input sheet (they were originally hard coded as 90% and 5%). In the 2010 (adjusted ) Wisconsin Governor exit poll, Barrett had just 83% of Obama voters (which I believe low-balled his actual share) and 7% of McCain voters. If Barrett's share of McCain voters is set to 7%, the required Walker share of Obama voters increases from 24% to 25%.
17
18
The model calculates Walkerâ€™s share of Obama returning voters that were required to match the state/county recorded vote.
19
The "2010-2012" worksheet provides a comparative analysis of the 2010 and 2012 elections
20
21
KEY RESULTS
22
The very conservative assumption is that Obama's recorded 56.2% share was his True share. But he did better than that. He had 63.3% in the unadjusted exit poll, therefore Barrett probably did better than his True Vote shown below.
23
24
Wisconsin
25
Obama had 56.2%.
26
Barrett had 47.1% and lost by 171,000 votes.
27
Walker needed 24% of returning Obama voters.
28
Barrett True Vote: 53.6%, 193k margin.
29
30
Waukesha
31
Obama had 36.4%.
32
Walker had 72.4% and won by 96,000 votes.
33
Walker True Vote: 65.1%, a 65k vote margin.
34
Walker needed 36% of returning Obama voters.
35
36
37
Milwaukee
38
Obama had 67.3%.
39
Barrett had 63.2% and won by 107,000 votes.
40
Barrett True Vote: 65.1%, 122k margin.
41
Walker needed 10% of returning Obama voters.
42
43
Dane
44
Obama had 72.83%.
45
Barrett had 69.1% and won by 99k votes.
46
Barrett True Vote: 71.2%, 110k margin.
47
Walker needed 8% of returning Obama voters.
48
49
50
The "Quick TVM" worksheet
51
52
Barrett's Quick County True Vote is a "first-cut" eyeball estimate for ALL 72 counties. It assumes 1) Obama's 56.2% WI recorded vote share (a CONSERVATIVE assumption since he had 63.3% in the WI exit poll, 2) 85% TURNOUT of Obama and McCain voters in 2012, 3) Barret had 90% of RETURNING Obama and 5% of McCain voters. Note: the Quick County True Vote is usually a good approximation to the county TVM. However, county turnout and vote share estimates will differ from the Quick calculator fixed estimates. For instance, the assumption that Barrett won 90% of returning Obama voters in DANE COUNTY is most likely TOO LOW. Therefore, Barrett's True Dane share will be understated. His required share of returning Obama is auomatically calculated (see the Input sheet) as 95%. Note that Barrett's 53.1% WEIGHTED AVERAGE Wisconsin share in the Quick Calculator is very close to his 53.6% in the "WalkerRecall" TVM. Note: the Quick Calculator does NOT factor in NEW voters.
53
54
The Walker Recall County/Muni Database
55
56
Wisconsin 2012 recall vote data is available for downloading from GAB as an Excel worksheet.
57
http://gab.wi.gov/elections-voting/results/2012/recall-election
58
59
The file consists of 3500 Unit/Ward vote records. By itself, it is not very useful since the viewer must scroll through the records to get to the desired county â€“ a time-consuming process. A data filter function is required to quickly view the ward totals for a given county. In addition, vote shares needed to be calculated and added to the data record.
60
61
This Google Spreadsheet has a filter capability. County Ward vote shares are automatically sorted
62
https://docs.google.com/spreadsheet/ccc?key=0AjAk1JUWDMyRdHdWYWpBemZYZHZ3S08xMUhvQTRNSUE#gid=1
63
64
My Wisconsin blog postings are here:
65
http://richardcharnin.wordpress.com/category/2011-wisconsin-supreme-court-recall-elections/
66
67
68
69
70
71
72
73
74
75
76
Loading...
 IntroSheet1InputTVMCopySumSheet40WalkerRecallTrueVoteDevBarrett True v. RecordedCountyMachinesMachinesObamaBarrettChart2010-2012ExitPollBarrettVsObamaWtdBarrettDiff2008BarrettVsObamaBarrettVsObamaChartWalkerVsJkloWalkerJkloGraphGovernorMoEProbBarrett vs ObamaSampleManual InputRecallFinalCompareBackup