President Trump says Rasmussen has been the most accurate pollster - He's Right
Bottom Line: Polling is probably the most widely misused, confused and abused information that's actually useful. Part of it is politics, part of it is that even some pollsters don't know how to properly apply their own information. Nevertheless, the arguments persist. One recurring argument is President Trump's approval rating. The President has routinely touted his Rasmussen approval rating, as its most commonly been his highest – as high as 51% recently. He's also taken the time to point out that Rasmussen's the most accurate.
Meanwhile turn on any of the alphabet networks and you're bound to see abysmal ratings in the high 30's and have a completely different sense of reality. So, what's real and what isn't? Who was the most accurate in 2016 and generally? As the headline suggested and as is often the case (once you get beyond a bit of bombast) - Trump's right.
Hillary Clinton won the national popular vote by 2.1%. Of the 11 accredited national pollsters only one was exactly right... Rasmussen which had Hillary winning by 2%. The average national pollster was off by 2.5% and the two worst pollsters were IBD/TIPP and Monmouth missing by 4%. In fact, Rasmussen was also the most accurate in 2004 as well. Aside from evidently sampling a bit better than most Rasmussen does something that no other pollster does at all times. They only sample likely voters. As I've indicated for months now, the more informed and engaged a voter is – the more likely they are to approve of the President. The inverse is also true. That's why Rasmussen's approval ratings are consistently the highest and it's why some pollsters have Trump in the high 30's. Those are all, and have been, adult only samples.