Previous Month | RSS/XML | Current | Next Month


July 26th, 2015 (Permalink)

New Book: Spurious Correlations

Have I mentioned recently that correlation is not the same thing as causation? Well, it isn't! Not only that, but Tyler Vigen has written a whole book to prove it. That book is titled Spurious Correlations, naturally enough. According to Scientific American's brief review, Vigen is a law school student, but he used to be an intelligence analyst for the military, which must have given him a lot of experience with spurious correlations. Just kidding!

Some famous spurious correlations are part of the folklore of statistics, such as the fact that penmanship―"penpersonship"?―is positively correlated with shoe size. While that one is genuine, some others may well be statistical urban legends, that is, spurious spurious correlations. Another famous example is the positive correlation between the number of storks and the human birthrate in Europe, which is a genuine spurious correlation according to Vigen.

In the introduction to the book―which you can read via Amazon's "Look inside!" function―Vigen discusses an interesting correlation that had serious consequences:

In 1958, William Phillips, a professor at the London School of Economics, published a paper regarding the connection between unemployment and inflation. As other economists explored Phillips's data, the correlation spread like wildfire: high inflation rates were linked to low unemployment and vice versa. The policy implications were explicit. National economies needed only to choose between inflation and unemployment, or somehow find a balance between the two. The Phillips curve, as the connection came to be called, informed macroeconomic policy decisions for years in both Europe and the United States. … It turns out William Phillips's theory on economics doesn't hold up….

This is the theory that was discredited by "stagflation"―an ugly word for an ugly phenomenon that we thankfully don't hear much about anymore. Stagflation is the occurrence of both high unemployment and high inflation at the same time, which happened in the 1970s in the U.S. A lot of pundits pointed out at the time that it was supposed to be impossible, but the negative correlation between inflation and unemployment was spurious.

As with our previous new book Spinglish, this appears not to be a serious scholarly work, but one that aims to be humorous and entertaining, though with underlying serious points. One of those is the already mentioned one about the difference between correlation and causation; another is that the emergence of "big data" has made it easier than ever to find completely coincidental correlations.

Source: Clara Moskowitz, "Book Review: Spurious Correlations", Scientific American, 7/14/2015

Resource: Cum Hoc Ergo Propter Hoc

Correction (8/2/2015): Originally, I wrote above about "the negative correlation between inflation and employment", but the negative correlation was supposedly between inflation and unemployment, that is, when inflation goes up unemployment goes down and when unemployment goes up inflation goes down. I have corrected the text above.

July 20th, 2015 (Permalink)

What's new?

I've added a new contextomy to the "Familiar Contextomies" page, this one quoting birth control pioneer Margaret Sanger out of context. Sanger seems to be a magnet for contextomies and even outright bogus quotes, as well as having statements made by others falsely attributed to her. Of course, the bogus and falsely attributed quotes are not contextomies, so I won't be including any of those. However, there may be at least one additional Sanger contextomy worth documenting.

I didn't just fall off the back of the turnip truck onto the information superhighway, so I don't think I'm naive about the misinformation you can find on the internet. Nonetheless, I'm actually rather shocked at what comes up when you search for information about Sanger. The amount of vitriol directed at a woman who's been dead for almost fifty years is astounding. While much of this mudslinging comes from the religious right, as you might suspect, a surprising amount―at least, it surprised me―comes from the far left.

Of course, Sanger said and did some things for which she is rightly criticized; I'm not talking about that. I'm referring to the contextomies, fake quotes, false attributions, and the unwarranted accusations they are used to justify. So, take this as a friendly warning: in the case of Sanger quotes, don't trust; verify.

Source: Margaret Sanger, Familiar Misquotations

July 18th, 2015 (Permalink)

Courtesy Disconnect at the IRS

What does the Internal Revenue Service call hanging up on you when you phone for help? A "courtesy disconnect". According to The Associated Press:

When too many people call at once, the IRS system hangs up on callers at the beginning of their calls, rather than have them wait on hold for an hour or more. The agency refers to these hang-ups as "courtesy disconnects"….
Source: Stephen Ohlemacher, "Hello? 8 million phone calls unanswered as IRS cut taxpayer service", The Associated Press, 4/22/2015

I guess the "courtesy" comes in from the fact that, instead of making you wait on hold for an hour, they just hang up on you. But if they know that you might have to wait that long, why not inform you of the fact, allowing you to hang up if you don't want to wait? Better yet, if they really want to be courteous, why not answer the call in less than an hour?

I wonder if the IRS has one of those messages that comes on when you're on hold informing you that "your call is important to us." If the call's so important why not answer it instead of telling us how important it is? Those messages just remind the caller that the call isn't important enough to hire enough people to answer it without putting the caller on hold.

This reminded me of Henry Beard and Christopher Cerf's complaint in Spinglish: "Are you aware that whenever companies say 'for your convenience,' they actually mean 'for our convenience'?" In this form, it's just a lie rather than "spinglish", but "courtesy disconnect" is a clever piece of doublespeak concealing that lie within a term for the practice.

To add injury to insulting our intelligence, the IRS Commissioner blames the "service"'s failures on budget cuts, thus attempting to have its poor performance rewarded with more money. If the IRS succeeds in getting its budget increased, I suppose that we can expect even longer waits on hold and fewer calls answered.


July 12th, 2015 (Permalink)

"These go to eleven."

A former philosophy student writes:

Looking though your site, I wasn't able to find any fallacy that I thought fit a relatively common scenario. Often, online reviews are driven by the assignment of a point value to the views expressed. This in itself is problematic in that people's relative perceptions of the scoring system are not always similar enough to justify drawing any kind of conclusion from the scores given, but that's not my point. Sometimes, reviewers will give something the lowest possible rating, and make a note in their review that whatever it is that they were reviewing is not worthy even of getting one star or point, because that is giving it credit that it does not deserve, and it really should merit zero stars or points.

Clearly, if something constitutes the lowest value in a system, selecting that arbitrary value is no different from selecting the lowest value if it were zero, or negative one, or even three. It seems that the designation itself, an arbitrary symbol, is being treated as if it were an objective or material measurement. Do you think that a fallacy is involved here, and if so, which one?

I'm familiar with the phenomenon you describe, having seen it frequently myself. Whether it commits a logical fallacy or not, it's at least illogical, and shows a lack of understanding of rating scales. I guess that some people assume that to give a single star is minimal praise, and that a work with no redeeming value should receive none at all. This may indicate that the rating system is unintuitive, and that instead of ranging from, say, one star to five it would be clearer if it ranged from zero stars to four.

Your question reminded me of a famous scene in the mockumentary This is Spinal Tap, where movie director Marty DeBergi is interviewing Nigel Tufnel, guitarist for the rock band Spinal Tap, who points to his amplifier:

Nigel: This is…very, very special because if you can see the numbers all go to eleven. Look, right across the board. Eleven, eleven, eleven.

Marty: And most of these amps go up to ten.

Nigel: Exactly.

Marty: Does that mean itís louder? Is it any louder?

Nigel: Well, itís one louder, isnít it? Itís not ten. You see, most blokes, you know, will be playing at ten. Youíre on ten here, all the way up, all the way up, all the way up. Youíre on ten on your guitar. Where can you go from there? Where?

Marty: I donít know.

Nigel: Nowhere. Exactly. What we do is if we need that extra push over the cliff, you know what we do?

Marty: Put it up to eleven.

Nigel: Eleven. Exactly. One louder.

Marty: Why donít you just make ten louder and make ten be the top number and make that a little louder?

Nigel: (Long pause.) These go to eleven.

July 3rd, 2015 (Permalink)

Puzzle it Out

Before you read any further, check out the interactive puzzle at the first Source listed below. There are spoilers in the explanation following the Sources, below, that may give away the solution to the puzzle. So, do the puzzle before reading further. Now, read the additional explanation at the second Source. After that, if you haven't forgotten what you were doing and why, please return here for even more commentary on the puzzle. I'll wait.


  1. David Leonhardt, "A Quick Puzzle to Test Your Problem Solving", The New York Times, 7/2/2015
  2. Steven Novella, "A Quick Logic Lesson", Neurologica, 7/3/2015

You're back! Regular readers of The Fallacy Files may not have been too puzzled by that puzzle because we've seen a version previously―see the Resource, below, for a reminder or in case you missed it. This is a good one to try on your smartalecky, know-it-all friends, though maybe not if you want to keep them.

As explained in the first Source, the puzzle illustrates confirmation bias. As Steven Novella explains in the second Source, people tend to formulate a hypothesis and then seek out evidence to confirm it. However, as the philosopher of science Karl Popper famously pointed out, it's even more important to test a hypothesis by trying to disconfirm. It's almost always possible to find some evidence to support a hypothesis, no matter how wrong that hypothesis may be. It's only when a hypothesis withstands determined attempts to disprove it that we are warranted in provisionally accepting it.

Confirmation bias is one reason why smart people often believe "weird" things, such as conspiracy theories and pseudoscience. The smarter they are, the easier it will be for them to come up with evidence that seems to support their pet theory, whereas they ignore or dismiss the mass of evidence that disproves it.

Resource: Rolling Stone's Worst-Case Scenario, 4/19/2015

Previous Month | RSS/XML | Current | Next Month