Does this refer to Donald Trump’s status as a nationalist leader, a parallel to pessimistic polling, or both? Given the status on the RCP aggregation of national polling after the convention, I’d guess that the primary purpose of Trump’s tweet is the latter — although with his new campaign leadership, it might be a toss-up:

https://twitter.com/realDonaldTrump/status/766246213079498752

This appears to be a response to an embarrassing exchange yesterday between campaign spokesman Michael Cohen and CNN’s Brianna Keiler over the current state of polling. After the change of leadership at the top of the campaign yesterday, Keiler asked the obvious question as to whether the “shake-up” would help turn around the polling. Cohen denied that there was a “shake-up,” even though the campaign itself made clear that Stephen Bannon and Kellyanne Conway had been brought in to take it over. Keiler noted that the campaign was down in the polls, to which Cohen objected as well:

Michael Cohen and @brikeilarcnn talk polls on CNN https://t.co/6OZtrfIwim https://t.co/QARiwckSY8

— CNN Politics (@CNNPolitics) August 17, 2016

Keilar: “Well let me ask you about this — so you say it’s not a shake up, but you guys are down. And it makes sense that there would … ”

Cohen interjects: “Says who?”

Keilar: “Polls. Most of them. All of them.”

Cohen: “Says who?”

Keilar: “Polls. I just told you — I answered your question.”

Cohen: “Which polls?”

Keilar: “All of them.”

Says who? As Keilar notes, since the end of the conventions, it’s pretty much every pollster in the field. Here’s the RCP average tracker since Hillary Clinton’s acceptance speech:

In fact, there hasn’t been a national poll in RCP’s aggregation showing Trump in the lead since July 26th, the second day of the Democratic convention. Even the Trump-friendly LA Times poll has shown Hillary with a narrow lead in its last two iterations, an eight-point shift away from Trump.

Later, the argument shifted from “says who?” to “who cares?” Eric Bolling tried to argue that polling means nothing in the face of rally attendance, an argument that left both Dana Perino and Greg Gutfeld aghast on Fox’s The Five (via Rightscoop):

https://www.youtube.com/watch?v=XoiL6gEr0Do&feature=youtu.be&t=5m6s

Perino’s correct; the Romney campaign used the exact same argument in 2012, and that’s when the polls had it a closer race than it turned out to be. Bolling also is correct that polls are reflective of a slice of time, but rallies mean even less than that. It’s a much smaller slice of time, limited to one very small piece of turf. It’s an event, an anecdote, not data.

As Perino notes, the argument ended badly for Republicans four years ago. Romney pulled tens of thousands into a Hillsborough County, Florida rally in the last week of the campaign, and lost the county by six full points. He lost Colorado despite drawing so many people into a Red Rock Canyon event that it backed up the interstate highway feeding into it. Polls have their problems, and they can change on the turn of a dime, but they are at least data rather than anecdotes — and the aggregation right now looks extremely poor for Team Trump.

That brings us back to Trump’s point on Brexit. Just how badly did the polling miss on the UK’s referendum? It wasn’t terrible, but it wasn’t good either:

Trouble with Trump's "I'm Brexit" argument on polls (better than denialism) is that final avg there was w/in MOE: pic.twitter.com/zgiYNUiHGt

— Guy Benson (@guypbenson) August 18, 2016

The margins were narrower than we’re seeing in the Trump-Clinton aggregation now, but Brexit won by almost four points, 51.9/48.1. British pollsters did largely underestimate the support for a nationalist cause, and even toward the end they missed on it by six points in the aggregate — outside the MoE for most polling. Here in the US, we saw examples of underestimation of Republican support in the 2014 midterms as several Senate seats went from toss-ups to runaways on Election Night.

It’s not inconceivable that Trump might have more support than current polling shows. However, there’s no evidence of that either, which is why the campaign has shifted leadership and is retooling now for the rest of the general-election campaign. To simply assume that polls mean nothing is equally silly as claiming that the election is over in August based on polling now.