Previous Month | RSS/XML | Current


November 27th, 2020 (Permalink)

How to Fact Check Quotes, Part 1:
Four Types of Misleading Quote

Quotation, n. The act of repeating erroneously the words of another. The words erroneously repeated.1

This entry in the series on amateur fact-checking2 is the first part on how to check a particular type of factual claim, namely, quotes. In order to know how to check quotes, it's useful to know what to look for. In what ways can a quote fail to be a fact?

Below are four real-life examples of quotes. You might want to practice your quote-checking skills by trying to answer the following questions before you look at the answers, below: Who is claimed to have said the quote? Who really said it? What sort of mistake does the quote illustrate?

  1. "Where you stand depends on where you sit."
  2. "That's one small step for man, one giant leap for mankind."
  3. "This year will go down in history! For the first time, a civilized nation has full gun registration. Our streets will be safer, our police more efficient, and the world will follow our lead into the future!"
  4. "This guy looks like he's up to no good. He looks black."

There are four ways that quotes may go wrong and each of the four quotes above illustrates one of these ways:

  1. Misattributions: A misattributed quote puts the right words in the wrong mouth. The first quote is often attributed to Nelson Mandela, but it is also known as "Miles' Law" after Rufus Miles2. "Who the heck is Rufus Miles?" you may ask. You've probably never heard of him, which is one reason why the quote is so often attributed to Mandela, since Miles is obscure but Mandela is famous. As Ralph Keyes has written: "Famous quotes need famous mouths"3.
  2. Misquotes: A misquote puts the wrong words in the right mouth. The second quote is a misquote of the words spoken by Neil Armstrong as he became the first man to set foot on the Moon. What Armstrong actually said was: "That's one small step for a man, one giant leap for mankind." The version quoted in the quiz and widely published makes no sense, since "man" and "mankind" mean the same thing. Armstrong's point was that the step onto the Moon's surface was a small step for him―"a man"―but a large advance for humanity―"mankind". Armstrong was apparently misheard due to radio static4.
  3. Bogus Quotes: A bogus quote puts the wrong words in the wrong mouth. Such quotes are usually invented for propaganda purposes, and their true authors are seldom known. The third quote is often falsely attributed to Adolf Hitler5 in order to attack the policy of gun registration, but its actual author is unknown.
  4. Contextomies: A contextomy is a quote that puts the right words in the right mouth, but takes them out of context in a misleading way. The final quote was spoken by George Zimmerman on the telephone to a police dispatcher shortly before he shot and killed Trayvon Martin. The quote, which was broadcast by NBC News at the time, makes Zimmerman appear to have volunteered the information that Martin looked black. However, between the two quoted sentences the dispatcher asked Zimmerman: "Okay, and this guy, is he white, black or Hispanic?" Zimmerman answered this question: "He looks black." So, the quoted words were the right words in the right mouth, but they were misleadingly edited6.

An accurate quote puts the right words in the right mouth, but it also provides sufficient context so that the quote can be correctly understood. So, in addition to being sure that a quote puts the right words into the right mouth, the quote-checker needs to see enough of the context to be sure that it doesn't give a false impression.


  1. Ambrose Bierce, The Devil's Dictionary (1911).
  2. For earlier entries in this series, see:
  3. Ralph Keyes, "Nice Guys Finish Seventh": False Phrases, Spurious Sayings, and Familiar Misquotations (1993), p. 20.
  4. See: "A Small, Belated Step for Grammarians", Associated Press, 10/3/2006.
  5. See: Ciara O'Rourke, "No evidence Hitler made this statement about gun control", Politifact, 8/21/2019.
  6. For the full story, see: An Audio Contextomy, 4/4/2012.

November 26th, 2020 (Permalink)

Thank You!

My thanks to everyone who has read and supported this site since I last thanked you! The Fallacy Files is disengaging from Amazon, so please do not try to support the site by making purchases through any of the remaining Amazon ads. Feel free, however, to click on any Google ads. Also, if you feel generous this holiday season and wish to support the site's mission, you can donate via the PayPal button in the navigation pane to your right. Your support is appreciated!

Poll Watch
November 11th, 2020 (Permalink)

Post Mortem

In September I wrote about this election that "it ain't over 'til it's over"1, and it still ain't over. It won't be over until about a month from now at the earliest, when the electoral college meets2, and maybe not even then. In that earlier entry, I made two main claims: First, that despite the popular impression, the public opinion polls in 2016 weren't off by much; and, second, that despite being about as accurate as could be reasonably expected, they were of no value in predicting the election results.

The second of these claims needs no re-examination, since I doubt anyone would deny it. It's true that the polls seem to have erred in the right direction this time, so that if you made a bet on who would win based on them, you would have won. However, if you had done that last time you would have lost, which is no better a track record than a flipped coin.

In contrast, some people may think that the current year's polls and election results show that my first claim was incorrect. There is a widespread impression that the polls were wildly wrong last time, and even worse this time, which has led some people to pronounce the "death" of public opinion polling3.

The results of the current election do not directly affect my previous analysis, because it's possible that the polls performed reasonably well four years ago, while those of this year flopped. As I pointed out in the previous entry, Real Clear Politics' (RCP) final average of polls in 2016 was only 1.1 percentage points off from the popular vote total. As far as I know, there's no way to figure a margin of error for such averaged poll results, but surely this is a near miss.

This time around, the RCP average showed Biden winning the popular vote by 7.2 percentage points, whereas he has so far won the vote by only 2.9 points for a 4.3 point deficit4. This is a worse prediction by 3.2 points and is, thus, almost four times as bad as that in 2016. In addition, the RCP average makes a photo finish look like an easy victory, if not a landslide, whereas Biden has so far done only 8/10ths of a percentage point better in the popular vote than Hillary Clinton did. So, there's no doubt that the polls did a much worse job this year than four years ago, despite the fact that the pollsters were supposedly adjusting their methods in order to prevent a replay. What went wrong I don't know.

Once again, the computer models failed. Nate Silver's model, the one that did best in 2016―though that's not saying much―did even worse this time, giving Biden an 89% chance of winning5. That's wrong in the right direction, but it's still wrong, since it gives the false impression of a landslide.

In their defense, however, all computer models suffer from a GIGO problem: "Garbage In, Garbage Out". Such models are based on the polling results, along with other data, so it may be that most of the failure was due to the flawed polls. Still, there should be a track record of successful predictions for more than one election―I would say at least three―before even provisionally relying on such models to project election results.

If we can't rely on polls or computer models to predict the electoral results, what can we do? Crystal balls? Tarot cards? Tea leaves? If you really must guess the result of a presidential election in advance, I suggest flipping a coin; you'll be right about half the time, which is just about as good as the polls or computer models. Otherwise, just wait until the election is over, though this year even that may not work.

I doubt that political polling is dead. There will no doubt be a decline in public trust of polls, which is entirely warranted, because people were putting too much trust in them. Perhaps this election will lead to fewer polls in future elections, which would be a welcome result. I haven't even tried to count the number of presidential polls conducted this year, but it surely must be in the hundreds. Other than a lot of manufactured news stories, what did we learn from all those polls?

After the 2016 election, the American Association for Public Opinion Research (AAPOR) conducted a post-mortem evaluation6, and no doubt there will be a similar review of this year's results, so perhaps then we'll get an explanation of why the polls were so much worse this time.


  1. "It ain't over till it's over.", 9/23/2020
  2. Thomas H. Neale, "The Electoral College: A 2020 Presidential Election Timeline", Congressional Research Service, 10/22/2020
  3. See, for instance: Michael Graham, "Is political polling dead?", Orlando Sentinel, 11/5/2020
  4. "General Election: Trump Vs. Biden", Real Clear Politics, accessed: 11/10/2020
  5. "National Overview", FiveThirtyEight, accessed: 11/11/2020
  6. Ad Hoc Committee on 2016 Election Polling, "An Evaluation of 2016 Election Polls in the U.S.", American Association for Public Opinion Research, accessed: 11/10/2020

November 5th, 2020 (Permalink)

A Surprise Puzzle Prize at the Logicians' Club*

While they're still counting the votes for President, here's a voting puzzle to help you pass the time. It's a difficult one so it should keep you busy for a long while. By the time you finish, maybe they'll have solved the puzzle of who is the next President of the United States of America.

There are currently five members of the Logicians' Club, one of whom is the president. As you might expect, logicians love variables, so the five members are known only by the letters V, W, X, Y and Z, ranked by seniority. V is the most senior member of the club and, therefore, its president.

The club's treasury has swollen to $100 and, according to the bylaws of the club, the sum must be distributed to the current members whenever it hits that amount. The president, V, withdrew it from the club's bank account in the form of one-hundred silver dollars. The following procedure for distributing the money among the club members was specified in the bylaws.

V would decide upon and announce a distribution of the silver dollars among the five members. Then, a vote would be held on V's plan and, if it received at least half of the members' votes, the coins would be distributed according to the plan. However, if the plan did not receive at least half of the votes, then the plan would be rejected and V would lose eligibility. Should that happen, then W as the next most senior member of the club would come up with a new plan, and the above procedure would be repeated until a plan was accepted and the money distributed.

All members of the Logicians' Club are perfect logicians who are able to reason out all of the consequences of any plan. Moreover, the only thing they care about is maximizing the amount of money they personally receive; they do not care whether a distribution is fair, nor do they get emotional about it. They don't resent another member who receives a greater amount of money than they do, so long as they get the maximum amount they can in the situation. Finally, the vote would be taken immediately after the plan was announced so that there would be no time for any of the members to negotiate agreements with each other; besides, none of the members trusted any of the others to keep an agreement since their only concern was to maximize their own winnings.

How many dollars did X receive?

* For other meetings of the club, see:

November 3rd, 2020 (Permalink)

Zombie Voters

A terrifying headline:

Records show dead people caught voting in NYC, report says*

I thought Halloween and the Day of the Dead were over. The article beneath the headline has this sentence: "The New York City Board of Elections…received mail-in absentee ballots in the name of dead voters, including a Staten Islander who died eight years ago…*." In all fairness, dead is about as absentee as you can get.

* Annalise Knudson, "Records show dead people caught voting in NYC, report says", Staten Island Live, 11/3/2020. Shouldn't that be "Staten Island Dead"?

Previous Month | RSS/XML | Current | Top of Page