Well now isn’t this interesting? From Reuters:
Donald Trump’s support has surged and he is now running nearly even with Democrat Hillary Clinton among likely U.S. voters, a dramatic turnaround since he became the Republican party’s presumptive presidential nominee, according to a Reuters/Ipsos poll released on Wednesday.
The results could signal a close fight between the two likely White House rivals as Americans make up their minds ahead of the Nov. 8 election to succeed Democratic President Barack Obama. As recently as last week, Clinton led Trump by around 13 points in the poll.
Meanwhile, this from the Los Angeles Times today:
Ray Fair of Yale University says that if his election forecasting model is correct, the Republican nominee is likely to win the presidency by a convincing margin. John Sides of George Washington University agrees; the “fundamentals,” he says, give the Republican about a 60% chance of winning. Alan Abramowitz of Emory University gives Republicans a solid shot at the White House too; his model gives the GOP the edge — but in “a very close election.”
These guys generally have the best track record of model-based election predictions, although I’m a skeptic of many of these models. (Among other things, the sample size, especially for presidential elections, is still rather small.)
My model says there’s still a 5 percent chance Trump is not the GOP nominee, and a 30 percent chance that the Democratic nominee is going to be Joe Biden. How does my model work? Sorry—it’s a secret proprietary multivariate digression analysis.