Exit Poll Response Optimization: Confirms the USCV Simulation in Debunking the rBr Hypothesis

 

Richard Charnin (TruthIsAll)

 

Updated: Feb. 10, 2012

 

Further Confirmation of a Kerry Landslide

 

In January, 2005, exit pollsters Edison-Mitofsky hypothesized that the 6.5% average exit poll discrepancy was due to differential exit poll response rates of Kerry/Bush voters (56/50). It was dubbed “reluctant Bush responder” (rBr) hypothesis. But the exit pollsters did not provide any evidence to back up the theory. In fact, independent simulation, optimization and regression analysis indicated that Bush voters were in fact slightly more eager to be interviewed – meaning that Kerry actually did better than the exit polls indicated.

 

In Sept. 2005, US Count Votes provided a simulation analysis which effectively debunked rBr. USCV noted “However, our analysis below shows that even if the “E/M hypothetical,” is interpreted as referring to average, rather than constant average, partisan exit poll response rates, it is inconsistent with the reported WPE data. There is no configuration of partisan response rates, however varying across precinct partisanship categories, with overall averages of K=0.56 and B=0.5 that  can produce the reported values of the actual E/M exit poll data for mean WPE and median WPE”. See below.

 

The following set of three exit poll response optimization models confirms the USCV simulation.

 

1. Precinct Partisanship Exit Poll Response Optimizer

The pollsters summarized the exit poll results for 1250 precincts in five partisanship groupings. The groups ranged from strong Bush to strong Kerry. The optimization model used the Excel “Solver” algorithm to determine a feasible solution: Kerry won the 2-party vote by 52.15-47.85%. The model used the final recorded 2-party vote, within precinct error (WPE) and response rates as constraints. The resulting vote shares exactly matched the 12:22am National Exit Poll “Voted 2000” shares.

 

These graphs display various components of precinct exit poll data by partisanship:

Effect of Incremental Aggregate Alpha on Precinct Category Alpha

Alpha (Kerry/Bush response) by Precinct Category

WPE by Precinct Category

Exit Poll Response by Precinct Category

Sensitivity of Kerry Vote Pct and WPE to Aggregate Alpha (K/B)

Bush Percentage of Refusers by Precinct Category Required for Recorded Vote

 

2. Location Size Exit Poll Response Optimizer

The exit polls were categorized into five groupings by size. The input constraints were location WPE and response (completion) rates provided by Edison-Mitofsky.

Kerry won the 2-party vote by 52.15-47.85%, exactly the same margin as the 1250 precinct optimization.

 

3. State Exit Poll Response Optimizer

The state exit polls were categorized into five partisanship groups using state exit poll response rates provided by Edison-Mitofsky. The state discrepancies were based on average precinct WPE.  Kerry won the 2-party vote by 52.3-47.7%.  The model confirmed the Precinct Response Optimizer: non-response rates were higher in strong Kerry states, again contradicting the non-response (rBr) hypothesis. A regression analysis  of Kerry state poll share and recorded vote vs. the state exit poll completion rate confirmed the optimizer.

 

1250 Precincts Categorized by Partisanship Groupings

 

 Given:                                       

1- Recorded 2-party vote (Bush 51.24- Kerry 48.76%)

2- Partisanship precinct exit poll response                                        

3- Partisanship "Within Precinct Error" (WPE)                                      

 

Calculate:                                          

Kerry’s true vote - aggregate and by partisanship category

Kerry 52.15-Bush 47.85%

Assume Census 125.74m total votes cast                                             

 

        2-party 

        Vote%    Vote   

Kerry   48.76%   60.70                        

Bush    51.24%   63.78                        

Total   100%    124.48               

 

Optimizer results:

(2-party)                                           

                Kerry   Bush                          

Vote            64.91   59.57m

2-party        52.15%   47.85%

Deviation      -3.39% >   3.39%

Vote Deviation -5.89     2.47

Dev /2-pty     -6.50% >   7.09%

                                             

Vote share     51.62%   47.37%

                                             

PARTISAN ALPHA                                      

Kerry strongholds: 1.052                                    

Other precincts:   1.203                                    

                                             

WEIGHTED AVERAGE (AGGREGATE)                                        

Response:    53.59%                                 

K/B (alpha): 1.165                                  

 

SHARE OF EXIT POLL REFUSERS                                         

Kerry   44.87%                               

Bush    55.13%                               

                                             

PARTISANSHIP CONSTRAINTS (1250 PRECINCTS)                                             

        HiBush Bush Even  Kerry HiKerry                      

Precincts 40   415   540    165   90      

                                                   

KERRY SHARE                                                 

Min      0%    20%   40%   60%   80%     

Max      20%   40%   60%   80%   100%    

                                                   

RESPONSE                                                    

Min      56%   55%   52%   55%   53%     

Max      56%   55%   52%   55%   53%     

                                                   

ALPHA (K/B)                                                 

Min      0.5   0.5    0.5   0.5   0.5    

Max      2.0   2.0    2.0   2.0   2.0    

                                                   

WPE                                                

Min     10.0%  6.1%  8.5%  5.9%  0.3%

Max     10.0%  6.1%  8.5%  5.9%  0.3%

 

Actual

E-M     10.0%  6.1%  8.5%  5.9%  0.3%    

                                                   

                                                    


                 OPTIMIZER OUTPUT SUMMARY                           

                                                   

         Poll     Count    Diff     Poll     Count   Diff

Kerry    52.15%   48.76%   -3.39%   64.91    60.70   -4.22

Bush     47.85%   51.24%    3.39%   59.57    63.79    4.22

Diff      4.29%   -2.48%   -6.77%   5.346    -3.09   -8.43

                                                   

Pship    HiBush Bush     Even     Kerry    HiKerry  Total

Prcts    40       415      540      165      90      1250

Votes    3.98     41.33    53.78    16.43    8.96    124.48

Pct      3.2%     33.2%    43.2%    13.2%    7.2%    100.0%

                                                   

Resp     56.0%    55.0%    52.0%    55.0%    53.0%   53.59%

                                                   

ALPHA                                              

K/B      1.50     1.17     1.19     1.08      1.0    1.17

K/ 50B   75.1     58.7     59.3     54.1     49.8    58.2

 

VOTE                                               

Kerry    0.79     14.51    24.58    11.87    8.95    60.70

Pct      19.9%    35.1%    45.7%    72.2%    99.8%   48.76%

                                                   

Bush     3.19     26.82    29.20    4.56     0.02    63.79

Pct      80.1%    64.9%    54.3%    27.8%    0.2%    51.24%

                                                   

RESPONDERS                                                  

Kerry    0.99     15.77    26.86    12.35    8.93    64.91

Pct      24.9%    38.2%    50.0%    75.2%    99.7%   52.15%

                                                   

Bush     2.99     25.55    26.91    4.08     0.03    59.57

Pct      75.1%    61.8%    50.0%    24.8%    0.3%    47.85%

                                                   

REFUSERS                                           

Kerry    13.5%    31.4%    41.1%    68.6%    100%    44.87%

Bush     86.5%    68.6%    58.9%    31.4%    0.0%    55.13%

 

VOTE DEVIATION                                              

Kerry    0.20     1.26     2.29    0.48     0.01     4.22

Pct      20.1%    8.0%     8.5%    3.9%     0.20%    6.50%

 

WPE                                                

Calc     10.0%    6.1%     8.5%     5.9%    -0.30%   6.77%

E-M      10.0%    6.1%     8.5%     5.9%    -0.30%   6.77%

                                           

 

 

_____________________________________________________________________

 

 

Location-size Exit Poll Optimization

 

Given:                                       

1- Recorded vote: Bush 50.73- Kerry 48.27%                                      

  (2 party: Bush 51.24%- Kerry 48.76%)                                             

2- Location exit poll response  rates                                     

3- Location "Within Precinct Error" (WPE)                                                                  

 

Calculate:                                          

True 2-party Vote share (aggregate and by location)

Kerry 52.15-Bush 47.85%

Assume Census 125.74m total votes cast                      

                                     

                           WPE    WPE            

Size      Weight  Votes    Mean   Median  Prct Response

Big City   13%    16.35    7.9    5.9    105   0.52

Small City 18%    22.63    8.5    7.7    236   0.54

Suburbs    45%    56.58    8.1    7.9    487   0.53

Small Town 8%     10.06    4.9    5.0    126   0.57

Rural      16%    20.12    3.6    3.6    296   0.55

                                                   

Total            125.74   7.17    6.68   1250  0.54

                                             

EXIT POLL (2-party)

                 Kerry    Bush                              

Vote              64.91   59.58                             

Vote share       52.15%   47.85%                            

Deviation        -3.39%   3.39%                             

 

Vote Dev          -5.88   2.47                              

Dev/2-pty        -6.50%   7.08%                             

                                                   

True Vote        51.62%   47.37%                            

Deviation         9.06%   -4.15%

                                     

WEIGHTED AVERAGE (AGGREGATE) RESPONSE: 53.98%                              

                                             

PERCENT OF EXIT POLL REFUSERS REQUIRED TO MATCH RECORDED VOTE                               

Kerry   44.83%                               

Bush    55.17%                              

                                             

         Rural    Town     Suburb   City     Big City

Prcts    296      126      487      236      105     

         23.7%    10.1%    39.0%    18.9%    8.4%    

Votes    20.12    10.06    56.58    22.63    16.35   

         16%      8%       45%      18%      13%     

 

Kerry NEP share

12:22am  43%      52%      50%      53%      64%     

Final    40%      48%      47%      49%      61%     

TRUE     40%      50%      49%      59%      73%     

                                                   

RANGE CONSTRAINTS                                           

 

KERRY WIN%                                                  

Min      35%      45%      45%      45%      60%     

Max      55%      55%      55%      60%      100%    

                                                   

RESPONSE                                           

Min      55%      57%      53%      54%      52%     

Max      55%      57%      53%      54%      52%     

                                                   

ALPHA (K/B)                                                 

Min      0.50     0.50     0.50     0.50     0.50    

Max      3.00     3.00     3.00     3.00     3.00    

                                                   

WPE                                                

Min      3.6%     4.9%     8.1%     8.5%     7.9%   

Max      3.6%     4.9%     8.1%     8.5%     7.9%   

E-M      3.6%     4.9%     8.1%     8.5%     7.9%   

                                                   

                                                   


                 OPTIMIZER OUTPUT SUMMARY                           

                                                   

......   Poll    Count    Diff     Poll     Count    Diff

Kerry    52.15%   48.76%   -3.39%   64.91    60.70   -4.21

Bush     47.85%   51.24%   3.39%    59.57    63.78    4.21

Diff     4.29%    -2.48%   -6.77%   5.34   p;  -3.09   -8.43

                                                   

.....    Rural    Town     Suburb   City     Big City Total

Prcts    296      126      487      236      105      1250

2-pty    29.48    12.55    48.50    23.50    10.46    124.48

Pct      23.7%    10.1%    39.0%    18.9%    8.4%     100.0%

                                                   

Response 55.0%    57.0%    53.0%    54.0%    52.0%    53.98%

                                                   

ALPHA                                              

K/B      0.792    0.800    0.840    0.835    0.800    0.82

K/ 50B   39.6     40.0     42.0     41.8     40.0     41.0

                                                   

VOTE                                               

Kerry    12.69    6.10     21.82    11.75    8.34     60.69

Pct      43.0%    48.6%    45.0%    50.0%    79.7%    48.76%

                                                   

Bush     16.79    6.45     26.67    11.75    2.12     63.78

Pct      57.0%    51.4%    55.0%    50.0%    20.3%    51.24%

                                                   

RESPONDERS                                                  

Kerry    13.22    6.41     23.79    12.75    8.75     64.91

Pct      44.8%    51.1%    49.1%    54.3%    83.7%    52.15%

                                                   

Bush     16.26    6.14     24.71    10.75    1.71     59.57

Pct      55.2%    48.9%    51.0%    45.8%    16.3%    47.85%

                                                   

REFUSERS                                           

Kerry    40.8%    45.4%    40.4%    45.0%    75.4%    44.83%

Bush     59.2%    54.6%    59.6%    55.0%    24.6%    55.17%

                                                   

VOTE DEVIATION                                              

Kerry    0.53     0.31     1.96     1.00     0.41     4.21

Pct      4.0%     4.8%     8.3%     7.8%     4.7%     6.49%

                                                   

WPE                                                

Calc      3.6%     4.9%    8.1%     8.5%     7.9%     6.77%

E-M       3.6%     4.9%    8.1%     8.5%     7.9%     6.77%

Diff      0.0%     0.0%    0.0%     0.0%     0.0%     0.0%

 

_____________________________________________________________________

 


State Exit Poll Response Optimizer

                                                                                 

Given: Recorded vote

        2-Party   Total

Kerry    48.76%   48.27%

Bush     51.24%   50.72%

 

2- State exit poll average response rate   

3- State exit poll average WPE                                                                 

                                                                                               

Calculate: TRUE VOTE

        2-Party   Total                                                         

Kerry   52.30%   51.77%

Bush    47.70%   47.21%                                                      

WPE     6.77%    7.08%

                                                                                               

WEIGHTED AVERAGE (AGGREGATE)

Response 53.32%                                                               

                                                                                               

REFUSER VOTE SHARE REQUIRED TO MATCH RECORDED VOTE

Kerry    44.86%                                                               

Bush     55.14%                                                               

                                                                                               

                                            CALCULATED TRUE VOTE       Deviation from Recorded  

State   Votes    Weight   RESP.    Kerry    Pct      Bush     Pct      Votes    Pct    WPE

Total 121,056   100%      53.32%   63,314   52.30%   57,741   47.70%   4287    6.77%   7.08%

                                                                                               

HIGH BUSH                                                                                  

UT       905      0.75%      59.6%    270      29.9%    635      70.1%    29      10.7%   6.4%

WY       238      0.20%      66.0%    76       31.8%    163      68.2%    5       6.8%    4.3%

ID       590      0.49%      63.2%    184      31.2%    406      68.8%    3       1.6%    1.0%

NE       767      0.63%      66.5%    285      37.2%    482      62.8%    31      10.9%   8.1%

OK       1,464    1.21%      53.2%    490      33.5%    974      66.5%    14      2.8%    1.9%

                                                                                               

BUSH                                                                               

ND       308      0.25%      63.0%    103      33.5%    205      66.5%    8       7.8%    5.2%

AK       302      0.25%      53.2%    126      41.6%    176      58.4%    14      11.5%   9.6%

AL       1,870    1.55%      58.3%    800      42.8%    1,071    57.2%    106     13.2%   11.3%

KS       1,171    0.97%      64.5%    445      38.0%    726      62.0%    10      2.2%    1.7%

TX       7,360    6.08%      58.3%    3,009    40.9%    4,350    59.1%    177     5.9%    4.8%

                                                                                               

SD       382      0.32%      42.7%    141      37.0%    241      63.0%    8       5.7%    4.2%

MT       440      0.36%      63.0%    170      38.6%    270      61.4%    4       2.3%    1.8%

IN       2,448    2.02%      38.6%    987      40.3%    1,461    59.7%    18      1.9%    1.5%

KY       1,782    1.47%      52.6%    712      39.9%    1,070    60.1%    1       0.1%    0.1%

MS       1,130    0.93%      49.6%    522      46.1%    609      53.9%    64      12.2%   11.3%

                                                                                               

SC       1,600    1.32%      59.4%    742      46.4%    858      53.6%    80      10.8%   10.0%

GA       3,280    2.71%      63.9%    1,402    42.7%    1,878    57.3%    36      2.6%    2.2%

LA       1,922    1.59%      47.8%    857      44.6%    1,066    55.4%    37      4.3%    3.8%

TN       2,421    2.00%      66.7%    1,043    43.1%    1,378    56.9%    6       0.6%    0.5%

WV       750      0.62%      48.7%    305      40.6%    446      59.4%    22      7.1%    5.8%

                                                                                               

NC       3,487    2.88%      52.6%    1,723    49.4%    1,764    50.6%    197     11.4%   11.3%

AZ       1,998    1.65%      57.3%    939      47.0%    1,058    53.0%    46      4.9%    4.6%

AR       1,043    0.86%      60.2%    473      45.3%    571      54.7%    3       0.6%    0.5%

VA       3,172    2.62%      56.4%    1,580    49.8%    1,592    50.2%    125     7.9%    7.9%

MO       2,715    2.24%      47.0%    1,338    49.3%    1,377    50.7%    79      5.9%    5.8%

                                                                                               

EVEN                                                                               

FL       7,548    6.24%      49.0%    3,870    51.3%    3,678    48.7%    287     7.4%    7.6%

CO       2,103    1.74%      55.5%    1,066    50.7%    1,037    49.3%    64      6.0%    6.1%

NV       816      0.67%      49.1%    438      53.7%    377      46.3%    41      9.4%    10.1%

OH       5,599    4.62%      45.0%    3,045    54.4%    2,554    45.6%    305     10.0%   10.9%

NM       748      0.62%      56.9%    400      53.5%    348      46.5%    29      7.3%    7.8%

                                                                                               

IA       1,494    1.23%      52.6%    764      51.2%    730      48.8%    22      2.9%    3.0%

WI       2,968    2.45%      55.3%    1,559    52.5%    1,408    47.5%    70      4.5%    4.7%

NH       672      0.55%      44.0%    386      57.5%    286      42.5%    46      11.8%   13.6%

PA       5,732    4.73%      46.8%    3,190    55.7%    2,542    44.3%    252     7.9%    8.8%

MI       4,793    3.96%      50.2%    2,630    54.9%    2,163    45.1%    151     5.7%    6.3%

                                                                                               

MN       2,792    2.31%      45.3%    1,575    56.4%    1,217    43.6%    130     8.2%    9.3%

OR       1,810    1.50%      53.0%    943      52.1%    867      47.9%    0       0.0%    0.0%

                                                                                               

KERRY                                                                              

NJ       3,581    2.96%      59.7%    2,085    58.2%    1,496    41.8%    174     8.3%    9.7%

WA       2,815    2.33%      53.8%    1,628    57.8%    1,187    42.2%    118     7.3%    8.4%

DE       372      0.31%      57.5%    230      61.8%    142      38.2%    30      12.9%   15.9%

HI       426      0.35%      53.4%    242      56.8%    184      43.2%    10      4.1%    4.7%

ME       727      0.60%      61.3%    411      56.5%    316      43.5%    14      3.4%    3.8%

                                                                                               

CA       12,255   10.12%     50.5%    7,413    60.5%    4,842    39.5%    668     9.0%    10.9%

IL       5,239    4.33%      51.9%    3,007    57.4%    2,231    42.6%    115     3.8%    4.4%

CT       1,551    1.28%      51.0%    979      63.1%    572      36.9%    122     12.4%   15.7%

MD       2,359    1.95%      59.4%    1,430    60.6%    929      39.4%    96      6.7%    8.1%

                                                                                               

HIGH KERRY                                                                                 

NY       7,277    6.01%      57.9%    4,729    65.0%    2,548    35.0%    415     8.8%    11.4%

VT       305      0.25%      53.1%    207      67.8%    98       32.2%    23      11.1%   15.0%

RI       429      0.35%      44.2%    270      62.9%    159      37.1%    10      3.7%    4.7%

MA       2,875    2.37%      56.5%    1,887    65.6%    988      34.4%    83      4.4%    5.8%

DC       224      0.19%      53.5%    207      92.2%    17       7.8%     4       1.8%    3.4%

                                                                                               

                                                                                               

                                                                                               

                                    CALCULATED TRUE VOTE              Deviation from Recorded    

State    Votes    Weight   RESP.    Kerry    Pct      Bush     Pct      Votes   Pct     WPE

Total    121,056   100%    53.32%   63,314   52.30%   57,741   47.70%   4287    6.77%   7.08%

 

HBUSH    3,965    3.3%     61.7%    1,306    32.7%    2,659    67.3%    54      5.4%    3.58%

BUSH     39,582   32.7%    55.2%    17,415   42.8%    22,167   57.2%    955     3.6%    3.49%

EVEN     37,073   30.6%    50.2%    19,868   53.7%    17,205   46.3%    1,397   6.8%    7.35%

KERRY    29,326   24.2%    55.4%    17,426   59.2%    11,900   40.8%    1,346   7.5%    9.07%

HKERRY   11,110   9.2%     53.0%    7,300    70.7%    3,810    29.3%    535     6.0%    8.06%

 

 

From: US Count Votes:

The exit polling firm Edison/Mitofsky (E/M), and USCV, agree that the historically unprecedented discrepancy between the exit polls and the reported vote count for the 2004 U.S. Presidential election cannot be a result of random sampling error.   This leaves either exit poll error or vote miscount as the only two possible explanations for the exit poll discrepancy.  E/M has claimed that the exit poll discrepancy is exclusively a result of “within precinct error” (WPE), and that the entire WPE observed in 2004 could be explained by a hypothetical exit poll completion rate of 56% among Kerry voters and 50% among Bush voters (herein referred to as “the E/M hypothetical”).

 

 The E/M hypothetical was widely interpreted by the media and by USCV as a claim that the 2004 exit poll discrepancy was caused by a pervasive, and on average uniform, shortfall in Bush voter exit poll response relative to Kerry voter exit poll response that was dubbed the “reluctant Bush response” (rBr) hypothesis. A recent clarification by E/M indicates that the “E/M hypothetical” should be interpreted as referring to hypothetical average (rather than constant average) partisan exit poll response rates. In this interpretation, average precinct partisan response rates may vary widely by reported precinct vote  shares, yet all of the reported WPE could be explained by partisan response rates whose average across the sample is K=0.56 and B=.50.

 

This interpretation of the “E/M hypothetical” does not depend on the “rBr hypothesis” of constant average partisan response rates which was shown by USCV to be inconsistent with the pattern of the exit poll discrepancy. However, our analysis below shows that even if the “E/M hypothetical,” is interpreted as referring to average, rather than constant average, partisan exit poll response rates, it is inconsistent with the reported WPE data. There is no configuration of partisan response rates, however varying across precinct partisanship categories, with overall averages of K=0.56 and B=0.5 that  can produce the reported values of the actual E/M exit poll data for mean WPE and median WPE.

 

Thus, neither a "reluctant Bush responder" (rBr) hypothesis interpretation nor an overall average interpretation of the “E/M hypothetical” is consistent with the WPE pattern shown by the Edison/Mitofsky exit polling data.   A larger overall average partisan exit poll response gap than that proposed by E/M could account for the observed WPE, but the large differences in mean and median partisan response bias necessary to produce the actual observed WPE levels across partisan precinct categories would need to be explained. In particular the very large mean and median WPE for precincts with over 80% reported Bush vote, and the almost zero mean and median WPE for precincts with over 80% reported Kerry vote, requires an explanation. 

 

Ten months after the election, no plausible explanation of the 2004 exit poll discrepancy, based on exit polling error, has been provided by E/M. The precinct level exit polling and official vote count data that would enable independent investigators to analyze the exit poll discrepancy has not been provided to the public.  Perhaps an exit poll explanation for the discrepancy does exist. However, a cloud of suspicion is cast on the 2004 presidential election results because the possibility that a “vote miscounts” explanation is required to generate the reported exit poll discrepancies is still open. It is a matter of the utmost national importance that detailed precinct level exit polling and election data that would allow for investigation by independent analysts, such as USCV, be publicly released.