Um … Yikes

Start your day with TPM.
Sign up for the Morning Memo newsletter

We have two new horserace polls out this morning (perhaps more shortly). But before that I want to note a big issue that’s been affecting the numbers in recent weeks: a major difference between the premium national phone polls and various online polls. The former have shown a relatively unchanged race, with Clinton holding a sizable lead, while the latter have shown a tight or even tied race. The problem is that the premium polls don’t come out that often. And we haven’t had many recently. Welp, this morning CNN/ORC has a poll out at it shows Trump 49%, Clinton 48%. In a four way race, Trump is up by two. There’s also a NBC/SurveyMonkey poll out showing Clinton up by 6, more in line with what we’ve seen over the last month.

So what’s up? As I’ve said before, I take the premium national phone polls more seriously than the online polls. CNN is the former, NBC/SurveyMonkey is the latter. CNN’s just one poll. But it is in line with a general tightening of the race we’ve seen over the last couple of weeks. I will be eagerly awaiting more premium phone polls to see where we are as of what is often considered the official start of the fall campaign.

The CNN poll shows a small Clinton lead with registered voters as opposed to likely voters. But we’re now in the window of time where it makes sense to start following the likely numbers rather than registereds.

Late Update: We now have a third national poll released this morning. This one from Franklin Pierce University/Boston Herald, which shows Clinton with a 3 point edge over Trump. So that leaves us with three polls: Trump+1, Clinton+6 and Clinton+3. (In this case, I’m giving the 2 way race numbers for the pollsters who provided them.)

Latest Editors' Blog
Masthead Masthead
Founder & Editor-in-Chief:
Executive Editor:
Managing Editor:
Associate Editor:
Editor at Large:
General Counsel:
Publisher:
Head of Product:
Director of Technology:
Associate Publisher:
Front End Developer:
Senior Designer: