Donald Trump answers a question at the Republican presidential debate in Cleveland on Aug. 6. (Brian Snyder/Reuters)

Cranky Reader: Sides, you are an idiot.

Me: Cranky Reader! Welcome back. It’s been a long time since you’ve been in touch.

CR: Unfortunately, it hasn’t been a long time since you’ve been dead wrong about a presidential election.

Me: What did I do now?

CR: Remember how a few weeks ago you said to “blame the media” for Trump’s surge? And how there is this cycle of “discovery, scrutiny, and decline” for primary candidates?  You know, in your stupid book?  Well, Trump’s been scrutinized a million times since then.  And his poll numbers are just as strong!  Where’s the decline?

Me: I’ve got some thoughts on this.  Let’s start with this idea: People don’t change their minds without new information.

CR: Well, yes. In fact, that is painfully obvious.

Me: What I mean by discovery, scrutiny and decline is that, during the presidential primary, people’s opinions respond to changes in the information they’re getting. The initial surge of news coverage drives up poll numbers. Some combination of scrutiny and the discovery of the “next” candidate brings about the decline.

CR: Fine. But what about Trump?

Me: Let’s take a look at Trump’s share of news coverage and his national poll numbers as of a couple days ago, again using data from the firm Crimson Hexagon. We’ll leave aside the fact that these polls probably overstate Trump’s support.

CR: Oy, you and your graphs. So what?

Me: This graph shows one important thing: nothing has really changed.  Trump still gets a ton of news coverage, and he still leads in the polls. Or look at it a different way: Here is the average share of news coverage and average poll standing of various Republican candidates since Trump got into the race.

CR: Okay, fine. I can see there’s a correlation.

Me: It’s a huge correlation, by the way: The largest possible correlation is 1.0, and this is 0.96.

CR: But how do we know that it’s news coverage that’s driving the polls? Couldn’t the reverse be true?

Me: It’s both, but the news coverage moves first. The coverage is what drives people to give a pollster a different answer, or to tweet about a candidate, or to Google that candidate. Of course, once the newly discovered candidate is atop the polls, then the polls justify more news coverage. In some sense, the initial news coverage perpetuates its continuation.

CR: So where’s the scrutiny, though? You said yourself that there was some scrutiny.

Me: I think there has been at least some scrutiny. But what we haven’t seen is a clear trend toward the sort of negative coverage that scrutiny would bring. The firm Crimson Hexagon also codes each news story as positive, negative or neutral.  So we can figure out whether stories that mention Trump are positive, negative or neutral. The graph below subtracts the percentage of stories that are positive from the percentage that are negative, comparing Trump to the average of other GOP candidates.

CR:  It looks as though stories mentioning Trump are more likely to be negative.

Me: Yes, but what’s also interesting is that there’s no clear long-term trend — just some ups and downs. In 2011, the scrutiny phase typically entailed a durable shift in the tone of news coverage.  See the graph for Herman Cain here, for example. This amplifies what I said before: Opinions will change when information changes, and overall, there hasn’t been a sharp change in the tenor of news coverage of Trump.

CR: Wait, are you suggesting that Trump’s appeal is all about news coverage? I mean, isn’t there more to it? I thought it was his unique appeal to conservatives or the tea party. Or maybe it’s that his hostility to immigration combined with his support for popular entitlement programs appeals to big chunks of Republican voters.

Or maybe it’s just that voters are mad. I mean, Frank Luntz did a focus group with Trump supporters and found that they were “mad as hell,” and after it was over, Luntz said his “legs were shaking.”  I mean, that means something.  FRANK LUNTZ’S LEGS WERE SHAKING.

Me: I think there are three problems with these interpretations. First, Trump doesn’t uniquely appeal to conservatives.  There’s little correlation between ideology and support for Trump. He’s not a tea party favorite, either. (Update: What I mean by that last statement is that Republicans who are Tea Party supporters are not necessarily more likely to support Trump than other Republicans. His support is not a Tea Party phenomenon.  See, for example, this Quinnipiac poll.)

Second, it’s far from clear that voters really have a lot of knowledge about Trump’s positions on issues, or any candidate’s positions.  It’s possible that it’s more about personality.

Third, as voters, we can always provide “reasons” for our choices. But some research suggests that the reasons we give for liking or disliking candidates are often rationalizations of choices we’ve made for other reasons — like, perhaps, the fact that a candidate is in the news all the time.

I think the simplest counterfactual is this: Imagine if Trump’s candidacy had been covered like, I dunno, Jim Gilmore’s. Then would we be talking about Trump’s unique appeal to white nationalists or angry voters or whatever?

CR: You can use fancy words like “counterfactual,” but I still think you’re wrong. You seemed to suggest that discovery leads to scrutiny, which leads to decline. End of story. That hasn’t happened.

Me: I think the main issue is this. In our book —

CR: Your stupid book.

Me: Yes, in our stupid book, Lynn Vavreck and I found that it could take more than scrutiny to end a candidate’s surge. The media also needed to “discover” a new candidate. We wrote that “the news media had a natural incentive to move on and find a storyline that was novel and more exciting.”

But here’s the thing: Trump is good at providing a new and exciting story essentially every single day. He’s a “perpetual attention machine.” There’s no need to talk about a different candidate. And, moreover, we know Trump stories get readers, and of course news organizations want to provide content that people want to read.

CR: Why doesn’t some other candidate just do something newsworthy?

Me: The main challenge is that there are so many other Republican candidates that it’s hard to break out of the pack — in a debate, in the day’s news cycle, or what have you.

CR: You know what? I just thought of another way you could be wrong! If Trump wins the nomination, then it will kill the idea that party elites shape the nomination process. You’ve been pushing that notion forever. Now, the party insiders are nervous and cranky about Trump, but they’re powerless!

Me: Well, okay, but we debated this in 2012, remember? That’s when you thought Newt Gingrich’s victory in the South Carolina primary was also blowing this idea out of the water.  You bet that “this time was different.” How did that turn out for you?

CR: Don’t change the subject from your own idiocy.

Me: Look, all I’m saying is lots of people say “this time is different” every time an election rolls around. But the batting average of “this time is different” is pretty low. Of course, if Trump waltzes to the nomination, I’ll admit I was wrong.

CR: I can’t wait.