Actual 2012 election result: Obama 51.1%/Romney 47.2%, 3.9% spread.
At 3.2%, this is a wider poll points spread than 2012.
IBD was #1 in 2012 and Rand -- now the USC Dornsife/Los Angeles Times "Daybreak" poll -- was #4. Both show a Trump lead.
Can the majority of the polls be wrong?
Of course, if they are done badly! I don't see that it makes any sense to weight the worst performing polls in 2012 as much as the best ones, but the news stories weight them higher!
Gallup -- the "gold standard" for presidential election polling between 1936 and 2012 -- gave up on presidential election polling after their dismal 2012 performance. They (apparently) think it's impossible to do this accurately at a reasonable cost today. Most, if not all, of the major alphabet polls showing a Clinton lead are done much the same way as Gallup did its 1936 poll -- telephones and live, probably majority young female, interviewers -- but with a much smaller sample size.
(I have a red lizard mailbox flag. I didn't know it was political.)
Here is the 2012 graph for comparison to the 2016 final linked in the opening thread.
https://alpdata.rand.org/?page=election2012
Although Obama didn’t have a dramatic trend downward at the end, Romney was clearly surging. It wasn’t enough and the differential held. Pray that the 2016 poll is as accurate as this one.