This thread is based on an assumption that Donald Trump is behind. That is merely an assumption. Several current polls place Donald Trump in the lead. The LA Times tracker has given him a small but persistent lead. It's methodology is different than most polls. If it's methodology is correct, Trump may very well be in the lead. If it is flawed, then, maybe Donald Trump is currently behind. In 2012, the Times tracker was claimed to be an outlier. Turns out it projected Obama's final margin. Salt to taste.
The problem about LA times tracker is that they ask for people's self-reported votes in 2012 and use 26% of Obama voters and 23% of Romney voters sample (close to the final share of 52-47%) and then ask who they would vote for while assuming their sample is still representative of the population.
Since humans have a bias toward claiming "I voted for the winner", and some suffer from bad memory, that would overweigh Republicans in their samples. Essentially, some people that claim they voted for Obama probably were Romney voters.
Most scientific online polls with traditional weighting have Hillary ahead.
LA Times last poll in 2012 was in August. I'm not sure how they can claim they got their margin in 2012 exactly correct. They had Obama +2 and the final margin was +3.9.
I doubt they used the same method back then. The August 2012 poll says they teamed up with Greenberg Rosner, a Democratic pollster. Greenberg does not use tracking method.
Poll
