The polls were not biased before the 2012 election, but they are now
It is Donald J. Trump versus the world, and this makes the current round of polling different from previous elections.
Trump has certainly attracted the anger of Democrats, but many in the GOP establishment dislike him as much as – if not more than – their Democratic opponents. And for this reason, we must be far more cautious in how the polling data is interpreted.
The GOP establishment (GOPe) – and its media outreach team at places such as National Review, the Wall Street Journal, etc. – is angling for influence with Trump, and its primary means of running this influence operation is by attempting to show that Trump's message and approach are failing with the public, and thus, he needs the GOPe in order to avoid an election apocalypse.
Consequently, we have a GOP presidential candidate who is undoubtedly the most feared nominee ever by both Democrats and the GOPe. The GOPe had no great love for Ronald Reagan, and Democrats were strongly opposed to him, but that was nothing like the vitriol at present towards Trump. Trump threatens the globalist empire of which the GOPe are charter members, and for that they will not go down without a fight.
Thus, both Democratic- and Republican-leaning polling operations are working overtime (along with their useful idiot RINO/Mondale Democrat proxies at the bastion of corrupt journalism) in a classic pincer attack, which gives the illusion of polling data obtained from both sides of partisan divide appearing to converge on a result of Trump being behind in the polls by somewhere between 5% and 12%.
When both sides hate you, that is reason enough to be very skeptical of polling data, which – as recent columns have shown – appears to be massively biased against Trump.
Some respond that these are many of the same polling operations that provided the 2012 election data, and those projections were fairly accurate.
Indeed they were, but the news is that the 2012 pre-election polls were not – repeat: not – significantly biased. They accurately reflected the public's composition, and the same methods used to show current polling data is biased indicate that the 2012 pre-election polls were largely unbiased.
A few examples are in order.
In the two months prior to the election, polling data unequivocally and consistently showed a 4% to 8% Democrat advantage – averaging 6% – in party affiliation over the Republicans.
Thus, when the Washington Post/ABC News November 4, 2012 election poll was made up of 35% Democrats and 29% Republicans, this was bang on the mark for being representative. The poll predicted a 3% Obama win, 50% to 47%, which was essentially a perfect prediction.
Next we have a Monmouth University poll released November 5, 2012 comprising 35% Democrats and 31% Republicans that predicted a tie between Obama and Romney at 48% each. Since this poll actually appeared to be biased towards Republicans by 2% – having a 4% Democrat advantage, rather than what should have been 6% – it under-predicted the Obama vote. Add 2% to Obama's side, and we effectively arrive at the election result.
The last NBC News/Wall Street Journal poll before the election predicted a 1% Obama win. The poll had a built-in 5% edge to Democrats, less than the 6% it should have (i.e., liberals were slightly under-represented). Once again, add 1% to the Obama side to account for the slight Republican bias, and the correct electoral result is achieved.
Fast-forward to the current cycle, and what we see is that the past two months of party affiliation average just a 1% Democrat advantage (including trading the lead back-and-forth between parties, something not seen in late 2012), not the absurd pro-liberal range up to 10% or more that all polls are using in an attempt to give Clinton the edge.
Consequently, polls were largely unbiased in the last stages of the 2012 election, and they yielded accurate results, particularly when any slight biases are corrected for. Polls today are overwhelmingly biased against Trump, and when that bias is corrected, Trump is tied with Clinton, or may hold a modest lead.
"Conservative" commentators who repeatedly regurgitate this obviously flawed data without critical thought are on the countdown to zero credibility, if they aren't already there. Or perhaps the #NeverTrump movement can draft Walter Mondale as their preferred candidate in Cleveland and relive their mid-1980s electoral dreams. After all, he's only 20 years older than Clinton.