Trump has two speaking modes: bland versus fiery. One is scripted (his day job), the other shoots from the hip, is borderline crazy, and inspires his base. The Narrative Metrics reveal all.

Hercules shouts about health issues, the Hulk holds court over economic policies, a couple of magazine editors speak rationally, a great columnist keeps her cool, a macho Health Secretary turns out to be a narrative weenie, and all the while, the two forms of President Trump are revealed by the Narrative Index of his speeches.

Trump: Unscripted barking versus scripted restraint

So for this week, our focus is the two versions of Donald Trump when it comes to public speaking. This fundamental divide became evident during his first term as U.S. President (I wrote a series of blogposts back then as I was first developing the narrative metrics). There are two versions of him for public speaking as you can see in the graph.

You can see the divide in the two groups of speeches we analyzed just by looking at the labels for the speeches in the spreadsheet below. The UNSCRIPTED speeches were events like Coachella, CPAC, the Al Smith Dinner — all venues where everyone wants the entertainer version of Trump. The SCRIPTED speeches were all from the teleprompter — things like his State of the Union addresses, Inaugural Address, commencement speeches.

The unscripted, shooting from the hip version of Trump is loud, bellicose, braying, bragging, bullying and often a boor. That version averages over 30 for the Narrative Index — the highest average we’ve ever seen for a politician.

The other version is more “diplomatic” (though that term is a stretch for him). It is more restrained, averages below 10 (a “D” for a letter grade) and is probably controlled mostly by the writing of Stephen Miller and his other speech writers who craft the text that he reads off the teleprompter.

Robert Kennedy, Jr. may be fond of contrarian views but when we look at his narrative metrics he’s not arguing very powerfully. Why? Uncertainty is the enemy of narrative.

We now have Narrative Shouters, a Smirker, and a Wimp.

The singular narratives vs mixed messages

As we all know, Robert F. Kennedy, Jr is trying to do a makeover on American public health. But exactly which parts of it and exactly how much, tends to be unclear. Search his name with “mixed messages” and you’ll find stacks of articles with that in the title.

One of the fundamental requirements for powerful narrative structure is to know what is the story you’re seeking to tell. It’s not clear he’s always certain of that.

As a result, take a look at his Narrative Index. It’s a paltry 11.

He claims to have the singular message of MAHA (Make America Health Again) as stated in May in his first major report, but The Daily Beast and Washington Post quickly pointed out the report was riddled with inconsistencies and even fabricated material from A.I.

RFK, Jr. with the weak narrative metrics

For RFK, Jr. we opted for only articles and speeches over 1,000 words for which we could only find 10 (but that’s plenty).

What does the Narrative Index (NI) tell us? A lot. At least sometimes. We all know Maureen Dowd is a brilliant writer, but when it comes to narrative structure, she appears to be more restrained.

Smirkers versus shouters? This odd portrait of, “Tea and Tater Tots,” compliments of Chat GPT.

The deft voice of Maureen Dowd

Let’s start with the link to our KNOCKDOWN FORUM where you can share any thoughts, comments or suggestions (like “get stuffed” if you don’t like this particular post).

The numbers don’t lie: “A smirk not a shout”

Legendary New York Times columnist Maureen Dowd scores only a “B” for narrative structure with NI = 19. That’s not bad — it’s where just about all politicians score.

It’s not blazing, bellicose argumentation. But then I don’t think that is how anyone would characterize her style. She’s more understated and sly.

I asked Chat GPT to describe her style. Among several things it said was this:

Character-driven – Rather than focusing on policy or complex analysis, she zooms in on personalities. She portrays public figures almost like characters in a drama or satire, highlighting their quirks, hypocrisies, or vanity.

When I asked if she is “aggressive,” part of what it said was:

No, she’s not uniformly aggressive in the sense of being humorless or strident. Her power often lies in her underhanded jabs—a sharp line delivered with a smirk, not a shout. Her aggression is frequently couched in irony, playfulness, and theatrical flair, rather than blunt-force argumentation.

All of these replies are indeed consistent with a narrative index of 19. Particularly interesting is the line, “a smirk, not a shout.” Think about that line in relation to our narrative heavy weights (Krugman, Faust, 49 and 47, respectively).

Shouters versus smirkers. I like that. A lot.

Also, soak in the narrative beauty of her And Frequency, a near-perfect score AF = 2.48% (she really needs to work on that last 0.02%).

Could we be discerning a pattern? Might it be that we have a couple of shouters?

As stunning as Narrative Hercules (Jeremy Faust, M.D.) might be, we’ve found his match. Our new champion is not only narratively ripped, he has an Economics Nobel Prize to boot.

Narrative Hercules gets out-butted by our new Champion

Weaponizing the ABT

The ABT is now spreading into fields like business (see this list of the Top Six Models for Storytelling for CEOs), law (read this great article from lawyer Patrick Barone applying the ABT to legal argumentation), civil engineering (check out this page from folks at MIT) and lots of other disciplines. Furthermore, this fall for the 4th year in a row we’ll be teaching the ABT course at Emory University Medical School.

Now we’ve developed two simple metrics that gauge how much a given text is drawing on “the power of narrative.” You can read about them in-depth in our new book Lincoln But Trump.

The pattern: Bulging narrative metrics

So, we thought we had a complete outlier in MedPage Editor-in-Chief Jeremy Faust, author of Faust Files. He rings the NI (Narrative Index) bell at a staggering average of 47 — half again as much as the typical New York Times columnist (see graph below). We thought surely he can’t be beaten. We were wrong.

Look at his metrics. He averages 49 for the Narrative Index (versus 47 for Faust) and a super-charged 2.0 for his average AF (AND Frequency). That sort of AF score shows his narrative voice to be almost staccato, and definitely paratactic.

These results are somewhat surprising given that he’s a humble scholar (heavily informational folks don’t tend to score so high). But then knowing the power of narrative structure, it’s not surprising he has 4.6 million followers on Twitter/X.

Your interpretation: Once again, what does it mean? You tell us…

As I said last week, we’re making this stuff up as we go along. I’ve been at the metrics non-stop for a decade — since the publication of my 2015 book, Houston, We Have A Narrative, but we’re still working on making sense of it.

Back then I managed to find one person in the Democratic party, James Carville, who found it interesting. I connected with him when I hit the terrifying realization that Donald Trump’s NI average of 29 was double that of his opponent (14) who was a candidate who ran for president with no message. I talked about it in detail on Park Howell’s podcast the morning after Trump’s victory.

So, what does this new finding for Krugman mean? As always, we’re interested in your interpretation of what the data say.

Please email us any thoughts at: randyolsonproductions@gmail.com

Our Interpretation

We think these results for Krugman push back the boundaries and expectations for the highly educated crowd. The sky-high NI says that even the most erudite in our society can tap into the power of narrative structure to excite audiences with what they have to say.

It also means that Hillary Clinton, had there been people on her staff who knew these narrative metrics (there weren’t according to the bestseller “Shattered”), they could have shown her the weak scores, then pushed for revisions that could have matched her opponent (who is still dominating this country with the overwhelming power of his narrative intuition). Paul Krugman, Jeremy Faust, and I’m sure lots of others are narrative role models. BUT sadly … the vast majority of academics, scientists, and Democratic party speech givers aren’t.

And lastly, this stuff is essential for powerful communication. If you doubt it, just ask the South Park guys.

And if you want to check for yourself, here are the raw data …

Okay, who should we analyze for next week? You tell us.

We’ve established a chat thread for the discussion of THE NARRATIVE KNOCKDOWN and for you to post your suggestions of who we analyze for upcoming bouts of this weekly exercise. Let’s hear your thoughts!

“Science Publications” – Nature Climate, Bulletin of ASLO, IUCN Reports, Molecular Ecology. Error bars are Standard Error of the Mean, N = 25 for each bar, at least, letter grade ranges on the right.

As we continue to work on the interpretation of our narrative metrics, it’s clear another major editor pales in comparison to Narrative Hercules.

Narrative Hercules swings his mighty sword of BUT.

First off, the second index

You can read a great deal about both metrics in our new book Lincoln But Trump. In it, we present the famous Lincoln-Douglas Debates of 1858 as the founding data set for the NI, along with a study of the World Bank Annual Reports as the founding data set for the AF.

The unique thing about the AF is it has a very clear optimum value of 2.5%. To learn a bunch about this just ask Chat GPT (though as usual, some of what it will tell you will be wrong). For the definitive explanation of why we stick with 2.5% definitely read the book.

Also, if you’re really interested in this topic get Chat GPT to tell you about POLYSYNDETON (what Hemingway often did) and PARATAXIS (the use of short, punchy, staccato style).

The Data: Another one bites the narrative dust

First, the data. As you can see, Richard Horton, Editor-in-Chief of The Lancet, scores very similar to last week’s contender, Holden Thorp, Science EIC. They scored 18 and 20, respectively for the NI, and 2.8 and 3.0 for the AF.

BUT … as far as the numbers go, neither is in the same realm as our Narrative Hercules, Jeremy Faust, MD. His NI scores are more than double theirs. DOUBLE. Which continues to baffle us.

What does it all mean?

But what exactly does it mean?

We’re not entirely certain ourselves. But we think it’s important to look at the New York Times professional columnists.

Look at the NI average of these big-time professionals (Dowd, Kristof, Friedman, etc.). They provide the most important reference point. They are the standard against which to measure everyone. They are the ones who know how to argue most powerfully, most compellingly, most persuasively.

The editors of both Science and The Lancet are not in the same range. Which makes you think it’s a difference of worlds between the broad societal dynamics addressed in the NY Times versus the more specific disciplines of science or medicine.

BUT THEN … once again … what about the exception? Dr. Faust.

That’s where our thinking is at the moment.

THEREFORE … tune in next week for Round 3 of THE NARRATIVE KNOCKDOWN!

THE DATA. For each group we analyzed 25 articles except for Nature Climate, Holden Thorp, and Richard Horton because their articles were around 700 words instead of the usual 1,000 words we use as a cutoff. To offset this, we analyzed 50 for each.

Jeremy Faust is our Narrative Hercules. Now it’s time to bring on the contenders. Our first match is with the editor of Science magazine, Holden Thorp.

ABT is Champ!

How much narrative punch are you delivering?

It might feel a little unfair to judge the very people whose job it is to judge others as editors. Whoops, sorry.

The Narrative Index

• NARRATIVE INDEX (NI) = (BUTs / ANDs) x 100

Here is the letter grade curve:

Last week we stumbled across the mind-blowing scores of Jeremy Faust, MD. He averaged a 47 for the 25 editorials we analyzed. We’ve looked at several thousand authors over the past decade but have never seen anyone average much over 30 for the NI. It spun our heads around.

But … he’s an editor. Maybe editors just have a brain for narrative structure when they do their own writing. To test this, we calculated the NI for 25 editorials from Holden Thorp, the current editor of Science magazine.

Here’s what we found:

Let’s not bicker over who out-argued whom

Dr. Thorp scored a solid “B” for the Narrative Index. Is that a bad score? Not for a scientist. Is it a good score? Is he hoping to reach a very broad audience with his editorials?

To put the scores into the bigger picture, here they are with the values we presented last week.

The Narrative Index for various publications. (Error bars = standard error of the mean, N=25 for each venue.)

Clearly, Dr. Thorp is not arguing with the strength of a NY Times columnist. BUT … NY Times columnists, in turn, don’t come close to arguing with the strength of Dr. Faust.

Next week on The Narrative Knockdown

Pretty simple. Just look at the BUT/AND ratio for the four main types of editorials in the NY Times. Committees present weak arguments, professionals kick ass, the others are in between.

Two obvious patterns: The committee effort (Editorial Board) presents the weakest arguments. The professional columnists present the strongest arguments. Don’t believe it? Do the analysis yourself, it’s dead simple.

Four categories of editorials

You’ll find four main types of editorials in the New York Times. Three of them are labeled “Opinion” pieces. They are written by: 1) members of the NY Times Editorial Board, 2) Guest individuals, and 3) Professional columnists of the NY Times.

The fourth type of editorial is actually called an “editorial.” It is written by the dozen or so members of the NY Times Editorial Board. This means it is more of a committee effort.

In my new book, Lincoln But Trump, I present two metrics for QUANTITATIVELY assessing the narrative strength of any document. The first is the AND Frequency (AF) which is just the abundance of the word AND in a text.

The second is the Narrative Index (NI) which is this:

An article this week in The Atlantic claims Barack Obama gave a “showstopper” speech at the 2024 DNC. Wrong.

They are of course great and wonderful people (my favorite president of all time), but that’s no reason to view them completely uncritically.

Poor structure: ending with the starting

Context is everything.

That’s what you learn when you spend 5 intense years teaching the ABT Narrative Framework. That’s what Matthew David and I have done since April, 2020 when we ran the first of 42 rounds of our ABT Narrative Framework course.

If you’re not familiar with the ABT and the two narrative metrics you can read about them in the new book, Lincoln But Trump: Narrative Metrics and Similarities Between Heroes and Villains (now available in paperback). You can also see it in the graphic (below) of the Top 6 Storytelling Models for CEOs.

One of the key, recurring mistakes we came to observe is the tendency to end an ABT with the very material that ought to open it. Over and over again we see the recurring pattern of having the last four or five words of the THEREFORE material at the end of an ABT actually be the very over-arching topic that ought to open the ABT.

And that is first sad thing to say about Barack Obama’s largely wandering 2024 DNC speech. I don’t care that Mark Leibovich this week in The Atlantic labeled Obama’s speech a “showstopper.” As history now knows, that show (the 2024 DNC) was already stopped before it ever started, given what it failed to accomplish (a victory for the Democrats).

So without dissecting the entire speech, just consider this one tragedy — Barack Obama ends his speech by saying this in his final major paragraph:

A restoration of what Lincoln called, on the eve of civil war, “our bonds of affection.” An America that taps what he called “the better angels of our nature.” That’s what this election is about.

If that’s what this election is about, why didn’t you tell us at the start?

Context is (again) everything.

How you set up a good speech — yes, how you FRAME it — is no different than telling a good joke. If you don’t set it up properly, you don’t have any impact.

The Lincoln quote was potentially powerful, had it been used properly. It should have been his first paragraph. Martin Luther King, Jr. opened his “I Have A Dream” speech by (correctly) pointing to Lincoln in his first paragraph.

Obama should have opened by simply saying, “You want to know what this election is about?” And then the Lincoln quote, which he then could have messaged around for the entire speech.

But he didn’t. He gave a largely dull speech then ended with his theme as only a final postscript. His wife, who spoke before him, was at least a half level better in terms of narrative strength, looking at the narrative metrics.

Most disappointing of all is to have it called a “showstopper” based on what? Cheering and applause?

The wife wins

In the end, neither speech is all that amazing — a conclusion supported by the mediocre scores for their narrative metrics.

MICHELLE: AF – 3.1%, NI – 15

BARACK: AF – 3.7%, NI – 10

Both AND Frequencies are over 3.0%. That’s not good. In fact, a score of 3.7% is downright bad. “Good” is close to the optimum of 2.5%. That’s what you get with tight editing, as you would find in superior publication such as … The Atlantic.

Neither come close to a score of 20 for the Narrative Index which is the lower bound for most good speeches (Barbara Jordan’s legendary 1976 DNC speech scored 39). Yes, they are beloved personas and rightfully so, but it doesn’t help to view their communications completely uncritically. What they delivered simply weren’t great speeches. This is important to realize at a time when the Democrats desperately do need some great speeches.