Reaction to the Latest Zogby Polls (TIA)

Reaction to the Latest Zogby Polls

TruthIsAll 12/2/07

http://www.geocities.com/electionmodel/DuKosZogby.htm

A DUer just quoted Kos regarding the latest Zogby interactive polling which show Hillary trailing all Republicans and Edwards and Obama beating them. Now many DUers laud Kos in proclaiming that the polls are bogus since they are a self-selected sample. They assert that since Hillary is winning all the other polls, Zogby must be wrong.

Kos has zero credibility in polling analytics. He has prohibited discussion of the 2004 theft and has called the work of dedicated analytic researchers (including yours truly) "crap". He ignores or fails to understand the statistical analyses which determined to an overwhelming mathematical probability that the election was stolen.

Another DUer posts daily poll stats and appears to be a Hillary supporter. But when new polls come out showing she is slipping, he blames the messenger! He denigrates Zogby's interactive polling without understanding the methodology. He claims that Zogby was wrong in projecting Kerry as the winner in 2004. In effect, he is arguing that the recorded vote was correct and therefore the exit polls (which confirmed the Zogby polls) were wrong. He must also believe there was zero fraud and that Bush won fair and square.

Now here's the main point regarding Zogby's just-released set of polls that everyone seems to miss: The only reason that HRC leads the 3-way polls with BO and JE is that the latter two are splitting the true Progressive/Liberal vote. That’s not to say that Hillary would lose the General election. I believe that she would defeat any Republican. It’s just that BO and JE are stronger candidates who would motivate the base much more than Hillary could.

Zogby polls have been proven to be very reliable. His 2004 projections closely matched the exit polls which showed Kerry to be the winner. Let's take a close look at 2004 Zogby pre-election polls in some key battleground states.

Zogby 2004 Battleground States

Interactive and Telephone polls vs. Actuals

Interactive Polling - 20 states
   Average 5.88% discrepancy between projections and actuals

  Avg        Kerry   Bush
  Poll       49.60   48.35
  Proj       51.14   48.86
  Actual     47.85   51.45
  Diff        3.29   -2.59

Telephone Polling - 10 states
   Average 4.55% discrepancy between projections and actuals

  Avg        Kerry   Bush
  Poll       48.80   47.10
  Proj       51.88   48.13
  Actual     49.20   50.00
  Diff        2.68   -1.87

Projection Assumption (2-party):   Kerry won 75% of undecided and 3rd party voters

        Actual-Recorded       Interactive Poll     Projection      Change        Telephone Poll        Projection     
                                  
        Kerry    Bush         Kerry     Bush     Kerry     Bush    Margin        Kerry     Bush     Kerry     Bush    
                                                                                                                            
Avg     47.85    51.45        49.60    48.35     51.14     48.86    5.88         48.80    47.10     51.88     48.12    

AR        45       54           47       50      49.25     50.75    7.5                                                        
AZ        44       55           46       52      47.50     52.50    6.0                                                        
CO        47       52           48       50      49.50     50.50    4.0            47       49      50.00     50.00    
FL        47       52           48       51      48.75     51.25    2.5            48       48      51.00     49.00    
IA        49       50           52       46      53.50     46.50    8.0            50       45      53.75     46.25    

MI        51       48           52       46      53.50     46.50    4.0            52       46      53.50     46.50    
MN        51       48           53       45      54.50     45.50    6.0            51       45      54.00     46.00    
MO        46       54           48       51      48.75     51.25    5.5                                              
NC        44       56           48       51      48.75     51.25    9.5                                              
NH        50       49           50       46      53.00     47.00    5.0                                              

NM        49       50           54       45      54.75     45.25   10.5            51       48      51.75     48.25    
NV        48       51           46       51      48.25     51.75   -0.5            45       50      48.75     51.25    
OH        49       51           47       52      47.75     52.25   -2.5            43       49      49.00     51.00    
OR        52       48           54       44      55.50     44.50    7.0                                              
PA        51       49           51       47      52.50     47.50    3.0            50       46      53.00     47.00    
 
TN        43       57           47       50      49.25     50.75   12.5                                            
VA        45       54           49       49      50.50     49.50   10.0                                            
WA        53       46           55       44      55.75     44.25    4.5                                              
WI        50       49           51       47      52.50     47.50    4.0            51       45      54.00     46.00    
WV        43       56           46       50      49.00     51.00   11.0                             

Zogby Final Polling vs. Exit Polls – 9 Battleground states

Kerry led by 50.2-44.8% (9 poll average).

Election Model base case assumption:   Kerry captures 75% of the undecided vote.

Result:   Kerry wins all 9 states; projected average vote share within 0.3% of the exit poll average.

Average    Kerry   Bush

Zogby       50.2   44.8   (Election Day polling)
Recorded    50.1   49.5   (actual recorded vote)
Projection  53.7   45.9   (75% undecided to Kerry)
Exit Poll   53.2   45.8   (unadjusted, based on average precinct WPE)

Difference   0.5    0.1

            Zogby Poll       Projection         Exit Poll        Recorded

St        Kerry    Bush     Kerry     Bush     Kerry   Bush     Kerry   Bush

Avg       50.2     44.8     53.7      45.9     53.2    45.8     50.1    49.5

CO        47       48       50.3      49.1     50.1    48.6     47.4    52.1   
FL        50       47       51.9      47.6     50.9    48.3     47.3    52.3   
IA        50       44       54.2      45.4     50.7    48.4     49.5    50.1   
ME        50       39       57.6      41.5     55.5    42.7     54.1    45.1    
MI        52       45       53.9      45.6     54.4    44.7     51.5    48.0   

MN        52       44       54.5      44.8     55.7    43.0     51.5    47.9   
OH        50       47       52.3      47.8     54.2    45.4     48.9    51.1   
PA        50       45       53.8      46.3     55.3    44.0     51.3    48.7   
WI        51       44       54.3      45.1     52.0    47.0     49.9    49.6   

If Hillary is winning only because Obama and Edwards split the true progressive vote (why would be another issue - the policy differences of either from hers are not that great, but I digress) then why do they all do, taken in total, about the same in head to head matchups against leading Republicans? In fact HRC is far from the worst in that area, where any splitting should show up? If Edwards or Obama could do much better in such matchups why don't polls where Hillary is not an issue show this?

TIA's methodology is affected by the fact that he uses deductive logic rather than inductive logic... He starts with a inference and then forces the evidence to fit his inference instead of collecting the evidence and then forming a inference from the evidence...Respectfully, it's sloppy social science or value laden social science and would get you laughed out of any graduate level social science research course...