Previous Month | RSS/XML | Archives | Current | Next Month

September 30th, 2009 (Permalink)

Headline

Poll: Support for Obamacare at New Low

"Only 41 percent of voters now support President Barack Obama and the Democrats' healthcare reform proposal―down from 44 percent two weeks ago and the lowest level of support yet measured by Rasmussen Reports."

I got a call this morning from someone supposedly taking a poll on health care, but I declined to participate. It's been months since I did a polling-related entry, but since we're now in the middle of a big debate about health insurance legislation, the media are getting back into election-mode poll hyping.

The above headline and first paragraph of a NewsMax article is the usual much ado about nothing: the margin of error of the poll is the standard plus-or-minus three percentage points, so the drop is within the margin. The headline and first paragraph are very similar to those of Rasmussen's own report on the poll results, so it's a case of reporting by rewriting a press release, which means that Rasmussen is partly responsible for its own poll results being exaggerated. However, at least Rasmussen was honest enough to include the following paragraph in its report:

The overall picture remains one of stability. Today’s record low support for the plan of 41% is just a point lower than the results found twice before. With the exception of a slight bounce earlier this month following the president’s nationally televised speech to Congress to promote the plan, support for it has remained in the low-to-mid 40s since early July.

Oddly, no mention of this appears in the NewsMax article.

Sources:

Resource: "How to Read a Poll: Margin of Error Errors", Fallacy Watch

September 23th, 2009 (Permalink)

Check 'Em Out

  • (9/27/2009) What should we learn from the Jaycee Dugard kidnapping case? According to Lenor Skenazy, not much:
    …[T]here is no lesson to be learned from Dugard's ordeal except that sometimes, terrible things happen to innocent people, randomly. In our blame-, lawsuit- and silly advice-obsessed country, it's a lesson we find hard to accept. … One post-Dugard advice article…earnestly suggested that from now on, we simply "never go anywhere alone." That's not asking too much, is it?

    Unfortunately, the very thing that makes the Dugard case one that we shouldn't generalize from is what makes it catnip to the news media: its uniqueness. There's an old saying that dog bites man is not news, but man bites dog is. This is because the former type of event is common, but the latter is thankfully not. The Dugard case is more "man bites shark" than "man bites dog", so it's all over the news.

    This kind of sensational coverage of an event tends to impress itself into our minds, and as a result we end up greatly overestimating its likelihood, which is what I call "the anecdotal fallacy". The chance of another case much like the Dugard kidnapping is virtually zero; of course, the probability of a child being snatched by a stranger is higher than that, but it's still very small. So small, in fact, that trying to do something to protect your children makes about as much sense as having them wear helmets all the time to fend off meteorites.

    The chance of any child being abducted and killed by a stranger is roughly one in 1.5 million…. And yet, whenever she points this out she is constantly reminded "but what if that one is yours?" "It’s as if people cannot imagine being part of the 1,499,999," she says. "They only see the one―they only see the one on the milk carton, they see one on TV…."

    What if your child is the one in umpteen million that gets clobbered by a meteorite? You need Dr. Curtis' Patented Anti-Meteorite Head Protection Device! Just four easy payments of $24.95! Do it for the children!

    Sources:

    Update (9/28/2009): Coincidentally, David Post of The Volokh Conspiracy has just written about what he calls "The ESPN Effect" in baseball, or more accurately "The Sportscenter Effect":

    I…said something like―"but it does seem like the overall level of defense is improving all over―I see so many great plays these days…" before I recognized how stupid a comment that was. Of course I was seeing more great defensive plays than I had 10 or 20 years before―because 10 or 20 years before there had been no Sportscenter (or equivalent). … I could turn on the TV and catch 20 or 30 minutes of great highlights every night, including 5 or 6 truly spectacular defensive plays…. It was just my mind playing a trick on me; I had unconsciously made a very simple mistake. The way in which I was perceiving the world of baseball had, with Sportscenter, changed fundamentally, but I hadn’t taken that into account.

    This, of course, is a sports version of the anecdotal fallacy: getting your sense of how common types of plays are from media coverage, rather than from sports statistics, or at least watching games. Sportscasters have the same interest in the unusual and sensational as other newscasters.

    Source: David Post, "The ESPN Effect", The Volokh Conspiracy, 9/28/2009

  • A short video of statistician Peter Donnelly explaining the base rate and prosecutor's fallacies, and giving a shocking real-life example of the latter:

    Via: David Smith, "Because it's Friday: Unfunny Statistics Jokes", Revolutions, 8/14/2009

    Resource: The Prosecutor's Fallacy, 10/30/2006

September 16th, 2009 (Permalink)

Documented Doublespeak

In his book Doublespeak Defined, William Lutz documented the use of the term "undocumented worker" as doublespeak for "illegal alien". Ten years after, the National Association of Hispanic Journalists' (NAHJ) advocates the use of the term "undocumented immigrant":

People who are undocumented according to federal authorities do not have the proper visas to be in the United States legally. … Terms such as illegal alien or illegal immigrant can often be used pejoratively in common parlance and can pack a powerful emotional wallop for those on the receiving end. Instead, use undocumented immigrant or undocumented worker, both of which are terms that convey the same descriptive information without carrying the psychological baggage.

The adjective "undocumented" does not convey the same descriptive information as "illegal", which is why it's a good euphemism. Moreover, the noun "worker" leaves out the important fact that these are people from other countries; for instance, Joe the unlicensed plumber would be an undocumented worker.

I have some sympathy with one of NAHJ's arguments against the use of the term "illegal alien", namely, that "the term criminalizes the person rather than the actual act of illegally entering or residing in the United States without federal documents". However, I don't know any alternative phrase that conveys the same information in a few words; "alien who is in the country illegally" is more accurate, but awkwardly long. If the choice is between "illegal alien" and "undocumented worker/immigrant", I'll take the former for its superior clarity and honesty.

"Illegal" is an unpleasant word because illegality is an unpleasant fact. Rather than attempting to obscure that unpleasant fact, shouldn't we face it squarely and try to do something about it?

Sources:

Via: Eugene Volokh, "Illegal Aliens", The Volokh Conspiracy, 9/16/2009

Reader Response (9/17/2009): Vance Ricks writes:

While I think you're right that "undocumented worker" is presented as a euphemism for "illegal alien", isn't it also true that "illegal alien" itself is a dysphemism? (Isn't that, partly, NAHJ's criticism?) In other words, calling the former a euphemism for the later suggests that "illegal alien" is the neutral starting point.

I noticed that you talked about the less-pleasant connotations of the term "illegal", but that you didn't say anything about the term "alien"! I realize that it's a legally cognizable term, but in what other context(s) do we refer to other humans as "aliens"?

William Safire claims that the phrase "illegal alien" started out as a neutral term:

In recent years, with the assertion of ethnic pride, the language has tended to purge itself of stereotyped national characteristics…. In one usage about Mexicans, the change has been extreme, from slur to neutral description and finally to euphemism: "Wetback" (from having entered the United States by swimming the Rio Grande) became "illegal aliens"; as that category came to include millions of people, it was changed to "undocumented workers." The objection to the use of this bureaucratic euphemism has led nowhere, or to what can best be called an American standoff.

And that was written almost thirty years ago! However, if "illegal alien" started out as a neutral term, it has since taken on a negative cast, or a euphemism to replace it would not be thought necessary. This is a form of the process that I call "euphemism inflation", that is, the fact that euphemisms lose value over time, so that they wear out and have to be replaced by new ones. As Steven Pinker explains:

Linguists are familiar with the phenomenon, which may be called the euphemism treadmill. People invent new words for emotionally charged referents, but soon the euphemism becomes tainted by association, and a new word must be found, which soon acquires its own connotations, and so on.

Some of the negative connotation of "illegal alien" may be due to the secondary meaning of "alien" as "extraterrestrial". However, there doesn't appear to be an alternative English word that has the same meaning as "alien" in its primary sense, so the only available replacement would be an unwieldy phrase or a neologism. Also, it's a silly equivocation for people to take offense because calling them "aliens" is accusing them of being non-human beings from another planet!

However, much of the negative charge of "illegal alien" is surely due to the literal meaning of the term, that is, a foreigner living in a country in violation of its laws. I agree with Pinker that attempting to get people to avoid using such unpleasant terms, rather than changing the unpleasant reality they refer to, is ultimately futile:

The euphemism treadmill shows that concepts, not words, are primary in people's minds. Give a concept a new name, and the name becomes colored by the concept; the concept does not become freshened by the name, at least not for long. Names for minorities will continue to change as long as people have negative attitudes toward them. We will know that we have achieved mutual respect when the names stay put.

That's why I consider "undocumented worker/immigrant" to be doublespeak, because it works largely through obfuscation. When first confronted with such a term, people don't react negatively simply because they have no idea what it means! As time passes and people begin to understand it, a euphemism becomes a dysphemism and we're back where we started.

Sources:

  • Steven Pinker, The Blank Slate: The Modern Denial of Human Nature (2002), pp. 212-213
  • William Safire, On Language (1980), p. 81
'DELIGHTFUL! AN ABUNDANT, IRRESISTIBLE FILM!' Mick LaSalle, SAN FRANCISCO CHRONICLE
September 11th, 2009 (Permalink)

Blurb Watch: Julie & Julia

Most critics seem to have divided minds about the new movie Julie & Julia: they love Julia Child as played by Meryl Streep, but aren't so fond of Julie Powell as portrayed by Amy Adams. Mick LaSalle of the San Francisco Chronicle might seem to be an exception, given the way his review is blurbed in an ad for the movie: "DELIGHTFUL! AN ABUNDANT, IRRESISTIBLE FILM!" However, that would underestimate the power of contextomy; here's the context of the blurb:

Few movies are as delightful as "Julie & Julia." … At its best, this is an abundant, irresistible film. …[I]f "Julie & Julia" focused entirely on the "Julia" part, it would be, without question, one of the best movies of the year. But there's another side to the movie, the Julie Powell side. … Alas, it takes longer to explain the one thing that's off about "Julie & Julia" than to explain the many things that are right about it. In truth, by the time the movie is over, few people will be talking about the sourness of the Julie scenes, except in passing. They'll repress them.

By dropping the "at its best" qualifier―with no ellipsis!―the blurb gives the misleading impression that the whole film is "abundant" and "irresistible", rather than just the Julia scenes. Apparently, LaSalle considers the Julie scenes not only resistible, but repressible!

Source: Mick LaSalle, "Review: 'Julie & Julia' celebrates good life", San Francisco Chronicle, 8/7/2009

September 6th, 2009 (Permalink)

New Book: Hyping Health Risks

Terence Hines reviews Geoffrey Kabat's newish book Hyping Health Risks in a recent issue of The Skeptical Inquirer. The review provides an opportunity to play "Name that Fallacy!":

Sometimes cancer does occur more often than would be expected by chance. The psychological desire to blame something obvious and identifiable for such clusters is easy to understand. It’s much more satisfying to have a known villain to blame than to put the cause down to amorphous statistical deviations from chance. The result is a search of the local environment. Inevitably, such a search yields an excess of power lines, leaky old oil tanks, microwave towers, or some such, which are promptly blamed for the cluster.

Here's a hint: "Howdy, my name is 'Tex'." Bang! Bang!

Hines mentions another fallacy:

One failing of this otherwise excellent book is that Kabat doesn’t emphasize one of the major reasons for the false belief in these various health scares: the multiple comparison fallacy. The basic idea is that if one makes enough comparisons one can find, just by chance, results that seem to show that, say, stamp collecting causes cancer.

I discussed a version of this fallacy, though not under this name, in "How to Read a Poll" (see the Resource below). As I explained there, the confidence level used in most public opinion polls is 95%, which is the same confidence level usually considered to be "scientifically significant". Given that level of confidence, we can expect the results of 5% of polls to be off by more than the usual margin of error.

The same thing can happen with scientific studies that look for statistically significant correlations. If a study checks twenty variables, we can expect one significant correlation at the 95% confidence level just by chance. For this reason, if a large number of variables are going to be checked, it is necessary to raise the confidence level in order to avoid chance correlations. I have not heard the name "multiple comparison fallacy" before, but perhaps it is well known in statistics. The fallacy is an important and common enough one that it should be named.

I would gladly review the book here if only someone would send me a copy.

Source: Terence M. Hines, "When Science Gets Distorted for Nonscientific Reasons", The Skeptical Inquirer, 7/8, 2009

Resource: How to Read a Poll: The Confidence Game, Fallacy Watch

Book: Geoffrey C. Kabat, Hyping Health Risks: Environmental Hazards in Daily Life and the Science of Epidemiology (2008)

September 3rd, 2009 (Permalink)

Reader Response

David A. Ventimiglia wrote in with the following comments about the Hitler Card fallacy. Since they are long and involve multiple points that I want to reply to, I will intersperse my comments in italics:

I've often wondered about certain variations on the "Hitler Card" that I think either are not fallacious, are less fallacious, or perhaps are not even the Hitler Card at all but only share a resemblance to it. I think sometimes I have witnessed "Godwin's Law" itself invoked fallaciously because readers mistake any argument containing the word "Hitler" to be the Hitler Card. Consider these two cases:
  1. There is the invocation of Hitler for the sake of a sort of "proof by contradiction". Someone claims a premise like (maybe this is an absurd example) "Citizens always should yield to the dictates of the person who holds the office of the head of state in their country" to which a retort might be, "We can't accept that as a premise because then it would have been okay for German citizens at the time to yield to Hitler, yet that is something we do not think was okay." The point is to demonstrate that a premise cannot be general by showing that if it were general, it would have to be applied in all cases, even to the most undesirable ones. People then try to think of the "worst person ever in the world", and that often leads to Hitler.

    I think that what you are describing is a reductio ad absurdum―perhaps that's what you mean by "proof by contradiction". In this case, the obviously false conclusion is something involving Hitler or the Nazis. I agree with you that this type of argument is not necessarily an instance of the Hitler Card.

  2. Maybe more important and more common, there is the invocation of Hitler to remind people about premises they hold, or say they hold, perhaps to catch them in a contradiction. Someone says something like "Sometimes, the press really gets out of hand, and then it's important for the government to clamp down on them" to which a retort might be, "You know, Hitler stifled the press." It sounds like the Hitler Card, which is fallacious, and it is fallacious if it is meant to say that stifling the press is bad because Hitler did it. However, I think sometimes what is meant essentially is the reverse: Hitler was bad for many reasons, among which was his stifling of the press. The point is to remind people that Hitler was bad, not because of who he was, but because of what he did. Hitler is such a villain of history that, I think, the thought of him can sometimes stir emotion without sufficient thought, such that people sometimes fail to think about why he's hated. And again, it was because of his actions. The actions are what were bad, and are what made the man bad, as they would make any person bad who commits them.

That sounds like a confusingly roundabout way of arguing. If stifling the press is bad, why not simply point that out, or explain why it's bad? Hitler brings along with him a lot of other baggage, such as the Holocaust and World War 2, which is likely to muddy the issue of censorship. Given that the thought of Hitler "can sometimes stir emotion without sufficient thought", why bring up such an emotional subject? I would still classify this argument as a Hitler Card of the Red Herring variety, that is, the introduction of Hitler is a distraction.

It's this second case that I believe I witnessed during the blog wars of the Bush years. So, President George Bush might have, say, skirted due process in dealing with "enemy combatants". A critic of the President might have said, "You know, Hitler also didn't adhere to due process", instantly violating Godwin's Law and earning public censure, at least as far as some on-line discussion forum's moderator was concerned. The moderator thought a fallacy had been committed, but I saw it differently.

"Godwin's Law" is ambiguous and does not necessarily mean playing the Hitler Card. In it's original form, it was just an ironic "law"―in the sense of a scientific law―to the effect that the Nazis would eventually be invoked in any online debate. Other forms of the "law" are intended as rules of debate etiquette, such as the one that insists that whoever mentions Hitler first automatically loses. So, some moderators of debates may punish those who first bring up the Nazis in order to maintain civility, which is not unreasonable given that people usually resent being morally compared to Hitler. Unless done very carefully, mentioning Nazism in a debate often brings it to an abrupt end, or causes it to spiral out of control as people become angry. So, while not all violations of "Godwin's Law" are fallacious, that does not necessarily mean that they shouldn't be punished by moderators.

What I inferred is that the critic wanted people to consider these things:

  1. Obviating due process is a really big deal.
  2. If you consider all the world leaders there have ever been in the last century, you'll realize that very few actually have ever had a policy of ignoring due process.
  3. Therefore, despite all the many ways in which Bush is different from Hitler, in this way at least he is more like Hitler than most other modern leaders are or were.
  4. We repudiate Hitler for many reasons, some of them much much worse than ignoring due process. But still, ignoring due process is one of his sins. Again, it's a big deal.
  5. Therefore, it'd be hypocritical to damn Hitler for (among other things) ignoring due process, then blithely permit it for George Bush.

This reminds me of a joke from the former Soviet bloc:

Question from Armenian Citizen: Is it true that last Sunday Akopian won a hundred thousand rubles in the state lottery?

Answer from Government Spokesman: Yes, it is true. Only it was not last Sunday, but Monday. And it was not Akopian, but Bagramyan. And not in the state lottery, but in checkers. And not a hundred thousand, but one hundred rubles. And not won, but lost.

If Bush had ignored due process, cancelled the elections, become a dictator, started a world war, killed six million jews, and grew a funny little mustache, then Hitler would be a good object of comparison. However, we would surely condemn Hitler much less if all he had done was ignore due process, and it isn't easy to untangle that one part of our condemnation from the rest of his sins. As a result, trying to compare Bush to Hitler on just this one point is likely to exaggerate the condemnation that Bush deserves. In other words, it's comparing apples and oranges.

So I am arguing that, ironically, sometimes it is the charge of fallacy itself (i.e., the Hitler Card fallacy) that is fallacious, that there sometimes are indeed valid reasons for invoking Hitler.

I've discussed false charges of fallacy before, under the name "fallacy abuse" (see the Resource below), and I'm sure that there are occasionally mistaken ad nazium accusations. For the reasons that I've given above, I'm not sure that invocations of "Godwin's Law" fall into this category.

Then again, because of the passions it arouses it probably is never a good idea, from a pragmatic point of view.

Agreed.

Source: Thomas Cathcart & Daniel Klein, Plato and a Platypus Walk into a Bar…: Understanding Philosophy through Jokes (2008).

Resource: Fallacy Abuse, 10/17/2008

September 1st, 2009 (Permalink)

Back to School Puzzle

Clues:

  1. Some philosophy majors study classics.
  2. All philosophy majors study logic.
  3. Some religious studies minors major in philosophy.

Which of the following propositions is a logically valid conclusion from the clues above? For extra credit, identify the syllogistic fallacies committed by the propositions that do not validly follow from the clues.

  1. All logic students major in philosophy.
  2. Not all religious studies minors major in philosophy.
  3. Some religious studies minors study classics.
  4. Some classics students study logic.
  5. All classics students are logic students.

Solution

Previous Month | RSS/XML | Archives | Current | Next Month

Solution to the Back to School Puzzle: Venn Diagram

  1. Invalid.
    Fallacy: Illicit Conversion of Clue 2.
  2. Invalid.
    Fallacy: Some Are/Some are Not, from Clue 3. "Not all P are Q" is equivalent to "some P are not Q".
  3. Invalid.
    Fallacy: Undistributed Middle Term, from Clues 1 and 3.
  4. Valid, from Clues 1 and 2. See Venn Diagram.
  5. Invalid.
    Fallacy: Illicit Minor Term, from Clues 1 and 2.

Source: Deborah J. Bennett, Logic Made Easy: How to Know When Language Deceives You (2004), pp. 17-18. The puzzle is based on a sample test question from the 1992 national teachers' examination. I have changed the class terms and rearranged the propositions, but the form of the puzzle is the same. Of course, I also added the extra credit task of identifying the fallacies committed by the wrong answers.