In this blog post, I run the unedited words of Presidential candidate Donald Trump through writing evaluation software, in order to evaluate his intellectual capacity.
The results of this test are quite revelatory, but first a little context:
I disagree with Donald Trump on many issues: I think torture is a bad thing. I believe that scientists are right that climate change is a real, man-made problem. I don’t feel that the only way to stop school shootings is to give teachers loads of guns. And I don’t think that building a gigantic wall on the border with Mexico is the best way to address the USA’s immigration problems.
In fact, I don’t just disagree. I would go further and say that I believe that Mr Trump’s pro-torture, anti-science, armed teachers, let’s-build-a-giant-wall policies are actually empirically idiotic. These strike me as dumb, foolish ideas. Moronic, simplistic reactions, rather than carefully considered solutions.
Further still, I believe that such ideas are so staggeringly and self-evidently asinine, that they are the kinds of things that only an imbecile would come up with, or believe in. These are the thoughts and beliefs that can only have emanated from someone who is clearly, dangerously stupid. Someone who is an imbecile. A half-wit. Dumb. The kind of person who tweets about sex tapes at 3am when they are running for president.
But how can I know for sure if someone is really stupid, or if I just disagree with them?
Revolutionary and counter-intuitive ideas are often considered foolish by weak minds. Could it be possible that these aren’t silly ideas, but instead are examples of maverick thinking? After all, these aren’t just positions held by Donald Trump. These pro-teachers-with-guns, anti-science, torture-isn’t-so-bad, giant wall stances actually form part of the official platform of the Republican party of the United States of America – supported by nearly 50% of the world’s only superpower. After the Brexit vote here in the UK, I spent a lot of time thinking about whether half of an entire country (52% to be precise) could be fairly deemed to be unintelligent. Certainly this was the image the UK projected to the rest of the world after Brexit: that of a majority of dimwitted rabble voting for something against their own interests, that they didn’t understand, and that they couldn’t see the clear consequences of. A childishly naive decision based on emotional, rather than rational arguments. But are these kinds of portrayals fair?
After all, I hate it when Britons reflexively declare their supposed intellectual superiority to all Americans. So who am I to have the presumption to declare half of any country’s citizens to be stupid, much less to proclaim that the potential next President of the United States is a thick-headed dullard? Maybe it’s me who’s the idiot? How could I be sure? I needed to come up with some kind of empirical test.
The Flesch–Kincaid readability test
Flesch–Kincaid is used by writers to gauge readability. It’s dead simple: you copy and paste some text into a piece of online software, and then it grades the content for its readability level. You can try it yourself here, or here. The software tells you whether what you’ve entered is written at the intellectual reading level of a comic book, the Harvard Law Review, or somewhere in between. While mostly, it looks at language and sentence structure, Flesch-Kincaid has a long history of being used as a practical test of intellectual complexity. It was first developed by the US military to assess technical manuals for everything from rifle assembly instructions to complex radar systems. They couldn’t have some idiot pressing the wrong button and starting World War III.
But luckily for us, this now can be used for just about anything – including testing the intelligence of the next leader of the free world.
The rules of the test:
In order to approach this scientifically, I determined that I would run only unedited, verbatim words spoken by Donald Trump through the Flesch–Kincaid test. This meant I could only input transcripts – either from interviews, or from off-the-cuff remarks. In this way, I could then be confident that the test would be evaluating only Mr Trump’s actual words, and not those of a speech writer. Here’s what I discovered:
Donald Trump consistently speaks at the level of a sixth grader
According to every piece of content I tested, Donald Trump’s words are firmly rooted at the complexity level of an 11-year old. In March, Mr Trump gave a long interview to the editorial board of the Washington Post, where he discussed topics ranging from law enforcement in black communities to foreign affairs with China. Here’s what Flesch-Kincaid said:
I tried re-doing the test using other Flesch-Kincaid websites and different Trump transcripts and each time I came up with the same results. 6th grade level.
But what does it really tell us that Donald Trump consistently communicates at the level of an 11-year old? People of course don’t speak in off-the-cuff conversations in the same way or at the same level as they might do if they were capturing their thoughts in writing. How valid is a readability test in determining whether or not Mr Trump has a robust intellect? I decided to test other politicians in order to be able to make comparisons.
Hilary Clinton also recently gave an interview to the Washington Post. I ran that transcript through Flesch-Kincaid as well, and came up with quite a different result. (Lower scores are better):
Not exactly Einstein-level, but at grade nine (equivalent to a 14 year old), Ms Clinton’s words achieved a score that was notably better than Mr Trump’s. Next I ran the unedited transcript from President Obama’s 2016 interview with the Washington Post. The results this time were slightly better than even Hilary Clinton’s score – but still at grade nine.
These differences aren’t vast, and the scoring system seems rather harsh, but the more tests I ran, the more I realised that I wasn’t finding anyone who scored as consistently poorly as Donald Trump. I ran politician after politician through this system, trying to find someone, anyone, at Trump’s level.
Here’s the results for Mitt Romney (2012 Republican candidate) from his conversation with the Boston Globe in 2016:
House Majority Leader Paul Ryan (from his 2016 NPR interview transcript) also was well ahead of Mr Trump.
Every Trump transcript or quote that I ran through Flesch-Kincaid was notably worse than any quote or transcript that I tested from every politician I could think of. British Prime Minister Theresa May, for example, scored more poorly than Clinton or Obama, but better than Romney (based on her 2016 interview transcript from the Spectator).
I tested dozens of interview transcripts, and I was unable to find anyone who spoke consistently at such a low level as Mr Trump. I tested George W Bush (he scored 70.9) Nigel Farage (68.7), Ted Cruz (75.4) and even Sarah Palin (72.3). All did better than Mr Trump. I was about to give up, until I finally came upon someone at Trump’s level – quite by accident.
One of the websites that I used to run these Flesch-Kincaid tests, http://www.thewriter.com/ has a section where they offer a guide to gauging their results, and I noticed that for the very lowest rating, equivalent to grade four, they state this as being the same reading level as “the lyrics to ‘Baby one more time,'” the pop song by Britney Spears.
Oh baby, baby
Oh baby, baby Oh baby, baby, how was I supposed to know
That something wasn’t right here
Oh baby, baby, I shouldn’t have let you go
And now you’re out of sight, yeah
Show me how you want it to be
Tell me baby ’cause I need to know now, oh because
On a whim, I found transcripts to a couple of interviews with Ms Spears, including one she gave to Stylist magazine about her new perfume. I ran them through Flesch-Kincaid.
I had finally found Mr Trump’s equal:
Our next President of the United States is the intellectual equivalent of Britney Spears. Oh baby, baby indeed.