And, below that, a reply from Jonathan Simon…
The biggest thing that makes virtually any conceivable national election upset victory “believable” is saturation negative advertising, making it possible to always believe the democratic candidate was ‘just too damaged’ to win. What’s more, it is considered a truism of politics that, while the public really doesn’t like negative ads, “negative ads work”. That makes it a great cover if in fact an election will be stolen, since negative ads do not by any means indicate that a stolen election is necessarily taking place.
If an election race is a little too far apart in the spread to be comfortably stolen, expect a huge dose of negative ads. Because this is what many candidates would do even if there were no stolen election, this makes it all the better as a cover for a stolen election.
Also, as Moveon noted and complained about last fall regarding Gallup polls, the pollsters are now assuming increasingly unrealistic proportions of republicans vs. Democrats in their ordinary presidential polls, which are adjusted in case the small sample size doesn’t have “enough” women, Republicans, or what have you…. Since pollsters are in the business of predicting ELECTIONS and on selling that “intelligence” information, the polls themselves will drift in the same general direction as stolen elections because they have an economic incentive for doing so, making the stolen elections over time seem closer than they actually are.
(e.g. the polls could have a bias of 5 points because of undersampling of working and poor people who don’t answer phones as much while the election could be stolen by 8 points, leaving a 3 point “gap” we all argue about the significance of when in fact the gap is 8 points, not 3).
It’s even more insidious that you imply in your parenthesis immediately above. Because most if not all pollsters have employed the “likely voter cutoff” model in their sampling, voters with say a 50% chance of voting (based on responses to what is generally a slate of seven questions) are cut off entirely, that is excluded from the calculation of the poll’s results. Of course, by definition, half these voters will vote. And of course the votes of these “marginal” voters (less wealthy, less anchored, etc.) will be disproportionately Democratic. Steve Freeman has done an excellent analysis (rejected of course by POQ) that concludes that by excluding all but the most likely voters from their surveys, pre-election pollsters create a rightward bias of between 4%-7%, which as you point out so nicely fits the rightward shift in the “actual” tabulations deriving from disenfranchisement and computer rigging.
From the standpoint of the pollsters, their projections are right on the mark so such models will continue to be employed and will conceal election theft very nicely. It is critical to recognize that the likely voter cutoff model has no intrinsic methodological justification (it is excluding data in such a way as to distort the sample in a consistent direction) but is “validated” only because it matches bogus vote totals still assumed to be gospel. A critical aspect of the overall scam, about which consciousness must be raised.–Jonathan