Donald Trump Defeats the Graphs

In 2012, Nate Silver, America's favorite statistician, got some attention from outside the stat-geek community by correctly predicting the results of all 50 states in the presidential election. His model, based on the polling he had identified to be most consistently accurate, was not any kind of breakthrough in the field of statistics, but people think it is pretty cool when you get 100% on things in real life. Already a well-respected, established statistician, he parlayed his success in the election guessing game previous books and articles into an ESPN supported website: FiveThirtyEight. The website applies statistics to politics, sports, and other parts of our world, and its staff writes some pretty interesting articles--check the site out if you enjoy a good graph. They don't always predict things perfectly, but generally they do a pretty good job. This election cycle, however, that pattern did not hold.

This graph is very skeptical of Donald Trump

As late as November, Silver routinely wrote about Trump's prospects with an air of dismissiveness, explaining that those who proclaimed Trump's-really-got-a-chance! ought to calm down. Any time you ascribe an opinion to people by hyphenating what they might say and putting an explanation mark at the end, you are using language to dismiss them. The punctuation manipulates the written word to evoke some person yelling out a phrase devoid of analysis, and we typically don't respect people who do that very much. He followed by delivering a glorious array of charts, one shortly after another, reminding you that he is Nate Silver, King of the Graph, and in doing so justified his previous mocking of Trump believers. They now become people who yell things that barely mean anything (what does really really mean anyway?), while he presents so much data that disagreeing is pointless. I believed it, and having Silver, who seems more objective in his arguments than anyone else, using dismissive rhetoric made it much easier to believe. The body of the article was the most persuasive part, but that little hyphenated passage at the beginning went a long way as well, illustrating the power of language.

1 is the loneliest number...

Now, of course, the election has changed, and so has Silver's word choice. His latest article, Why Republican Voters Decided On Trump, is much more about how the man came to be the de facto nominee than a concession to Silver's incorrect predictions in the past, but there is some of both elements. The structure of the beginning of the article sets the tone: the first sentence, "Donald Trump is going to win the Republican nomination," sits atop a paragraph break, with nothing but a footnote to accompany it. The shortness and bluntness of the sentence, followed by the space that is bigger than the first section of text, gives the impression that it was forced out of Silver, then followed by a long, rueful sigh. No stats or analysis; those are relegated to the bottom of the page because they would distract from the style. He then launches into a relatively brief explanation of what he thinks got Trump on top and against all odds kept him there, and I would recommend that you read it, but it is all in the context of that first paragraph (sentence). Instead of being a masterful amassing and utilizing of data, I could not help but read this article as the thoughts of a confused, distressed statistician. Once again, the content was the valuable stuff, but the way Silver placed his words at the beginning dictated the lens with which I viewed it.

I must admit, sometimes FiveThirtyEight articles are too dry for me; statistic after statistic about something I do not know much about with all kinds of restrictions on the conclusions made from the data for this reason and that--I can't always pay attention to them. However, having gone through the process of writing this blog, these articles are not just objective summaries of evidence; like just about everything else out there, they use words not just to explain things logically but to produce emotional responses in the readers. I look forward to reading what Nate Silver has to say about the general election because he is good at predicting things, but now I also look forward to trying to look for the words that set the tone for each article.