Crunching the Numbers on Sounding Presidential: A Guest Post

One of the really fun innovations in this election cycle is the extent to which the speech of the candidates has now become grist for statistical analysis. For instance, the Times’ “Caucus” blog reports that Reagan’s name was invoked 53 times last night, and by this measure Romney beat McCain 19 to 12. The Times has now set up a “Transcript Analyzer” where you can play the same game on any word of your choosing. (I already checked: No-one said “sex”, even once; this was the Republican debate, after all.) Earlier, the Times presented an analysis of who was mentioning whom, so you could figure out the rivalries among the candidates.

One of my favorite economists, Bruce Sacerdote, and his student Owen Zidar have kicked it up another notch. In a new study, they crunch a database containing all the words used in the speeches made by the major candidates (through to Jan 5). They have some pretty interesting findings:

Obama and Clinton rarely mention other candidates by name — less than twice per 10,000 words.

When they check the context in which candidates typically name their opponents, it is typically negative. This can be interpreted as a measure of negativity, and by this measure, the most negative campaigners were Mitt Romney and John Edwards.

They construct a Reagan-ometer, measuring the similarity of a candidate’s words with the former President, relative to the usage of those same words by Martin Luther King, Jr. Interestingly, the two most Reagan-esque speakers are the Republican frontrunners, John McCain and Mitt Romney. Barack Obama and former preacher Mike Huckabee more closely resemble the oratory of Martin Luther King, Jr.

They propose a more standard left-right ranking assessing the closeness of one’s speech to Bill Clinton versus George W. Bush. Many of these finds are as expected, as Republicans were typically more like Bush than Democrats. The one surprise was that Giuliani was not only more Clinton-esque than his fellow Republicans, but also more so than John Edwards.

Focusing on alternative brands of “Democrat,” they also ask whether one’s speech more closely resembles (Bill) Clinton, or former Democratic President LBJ. It turns out that Hillary’s speech is much more similar to her husband. This Clinton-ometer ranks Hillary as a true Clinton-ite. But perhaps this is a bit unfair, as I’m sure that living with someone tends to make your speech particularly similar.

And the most popular words on the campaign trail so far? Care, war, health, Iraq, security, energy, economy, jobs, life, and tax.

Read the cleverly titled paper, “Campaigning in Poetry…”


Regarding word-frequency in this campaign, there's a problem with the sort-order in the final table in the source paper, which led to an erroneous top-ten list in Mr. Wolfer's article.
The actual top-ten ranking is:
america (1571)
country (1360)
american (1333)
world (1312)
care (1073)
work (1063)
government (1059)
war (1033)
today (947)
americans (886)

It appears there's still no change (631), nor any hope, which was evidently used less frequently than "trade" (257), the least-frequent word cited.

Johnny E

People who want to evoke the Reagan Era should be careful, especially since they're looking at the myth, not reality. His administration had something like 200 officials in trouble for ethics or criminal violations or profiting from their position in government. His Justice Dept. had its own problems. And then there was Iran/Conta.


In Comtempt Of the Public, Reagan's team is eclipsing its predecessors when it comes to unethical conduct, corruption. A Baedeker of Scandal -

Newsday - Long Island, N.Y.
Author: By Mark Green and Peter H. Stone. Mark Green,
Date: Jun 28, 1987

Abstract (Document Summary)

THE REAGAN administration is proving - both in volume and variety - to be the most unlawful and unethical in memory. It has given us, among others, the Iran-contra and Wedtech scandals. Six "independent counsels" are investigating current and former administration officials for possible criminal conduct. Some 200 Reagan
administration officials have been charged with unethical or corrupt conduct, according to a compilation recently released by Rep. Pat
Schroeder (D-Colo.).......



In regards to Johnny E and his opinions concerning Ronold Reagan.

I remember very well sitting on a ship with other troops awaiting the word from newly elected president Reagan to go in nd get our people.Unfortunately president Carter never formulated a coherent plan to get our people.The same day Ronald Reagan became our commander and Chief Iran let our people go.What a coincidence.

Of course President Reagan and John Mc Cain are two differnt people but I believe they both share a love of their country and a desire to serve it.


I agree with DMart, how come change didn't ring up there high. I know it would probably be an incredible amount of work, but it would be interesting to see which words were used most relative to the size and diversity of the audience.


Since I preach, I understand the art of the message. People have certain words to connote safety in the religious word. This seems to be the same for the politic word also.

Johnny E

So what's the big deal, the parties hire the same speech writers from campaign to campaign.

I'm always amazed at how long the current administration has been telling us the "oceans don't protect us anymore", starting the day after 9/11. Any historian, businessman, or naval expert knows that the oceans are highways, not barriers. The oceans didn't protect the Native Americans from the Conquistadors, the US Navy in Pearl Harbor, or merchant shipping from U-boats within site of the East Coast.

It's frustrating when articles like this leave out candidates -- Ron Paul, for example. It's discrimination and poor journalism, in my opinion.


What is Hiliary's attack dog up to now? Bill is gainfully employed and is being investigated for creating telephone lists of elderly folks to fleece their savings. He has also been active in finding positions for his friends in the United Nations. Is he financing Hiliary campaign fund with ill gotten gains?

Justin Wolfers

The top-ten words I listed are the top ten issue-related words. The authors dropped prepositions and pronouns from their analysis. The top ten non-issue related words were: America, country, American, world, work, government, today, Americans, believe, and right.

kate niederhoffer

In light of today being Super Tuesday: not only the content of candidates' speeches are grist... James Pennebaker's blog (UT Austin) explores things like pronoun usage as a marker of candidates' ability to connect with the audience and negations as manifestations of struggles with power. The findings above become even more interesting when coupled with his more stylistic findings showing candidates' personalities, inner motives, etc. based on HOW they're saying what they're saying.


I'm surprised that the words, "change" or "hope" weren't used more ofen...curious.

Denis Jonnes

I don't get what is "freakonomical" about the Super Bowl? or why it warrants attention of any freakonomist? To be honest, I don't find anything on this blog in the least 'freaky'--at least in the sense that I have always used that word: wild, weird, far out, mind blowing. It's just more chewing of the not-really-much-to-say cud.
And is it "freaky" that Hillary Clinton's sounds more like her husband's than LBJ's. We need the NYTimes to tell us this? Jeesh, give us a break!

Just another cretonomist

Alix Martin

the study could be made better with a Markov Chain model, where information conveyed by words would be analysed in the context of the previous n words, and not derived from frequency only

Johnny E

In regards to #10:

Are you saying that Reagan made a secret deal with Iran so he could get elected? If he didn't they might have come home earlier.


A great resource - many thanks!