How Not to Play the Hitler Card
David Kopel has an interesting post at The Volokh Conspiracy on how to make analogies with the Nazis without playing the Hitler card―though that's not how he puts it. The problem with most such analogies is that they are farfetched because they are based on few significant points of similarity, and usually ignore significant points of dissimilarity. Check it out.
Source: David Kopel, "Enforcing Godwin's Law", The Volokh Conspiracy, 1/31/2006
Poll Watch: The Mystery Pollster on Partisan Polls
Mark "Mystery Pollster" Blumenthal has some interesting comments about polls conducted by partisans:
How reliable are partisan sponsored polls? Although most of the polls reported on in mainstream media are sponsored by the media outlets themselves, political parties, candidates and interest groups also routinely conduct their own surveys. Readers should remember that MP earns his living conducting polls for Democratic political candidates, and as such, may not be the most impartial source on this question. However, my perspective is that my "partisan" clients expect, first and foremost, that we provide accurate and reliable results, so our professional duty is to get the numbers right, without regard to their potential propaganda value. As such, data from the internal polls that drive campaign strategy rarely see the light of day.
However, as many of you know, partisan pollsters do sometimes release survey results selectively when they cast our clients in a favorable light. Academics who analyze public polls systematically…find that releases from partisan pollsters show a consistent bias. That is, publicly released pre-election polls from Democratic pollsters tend to be [a] few points more favorable to Democrats, and polls from Republican pollsters tend to be a few points better for Republicans. Obviously, pollsters debate the reasons for this pattern, but most believe "selection bias" explains a lot of it. Consider it this way: Random sampling error means that all polls have a range of variation. If partisans release only those results that fall on the positive half of the bell curve (from their perspective), the released results will show a consistent bias. Either way, data consumers should take public releases of partisan results with the appropriate grain of salt.
Source: Mark Blumenthal, "Update: MyDD Poll", Mystery Pollster, 1/29/2006
Puzzle: Your Mileage May Differ
Which is farther: a mile on land or a mile at sea?
Carl "The Numbers Guy" Bialik has posted the results from his quiz last week about innumeracy in the news. In pointing to the quiz, I mentioned that you could also use it to spot logical fallacies in a few examples. Here they are with my explanations so that you can see if you spotted them:
- Boosting the state's economy was a central tenet of Gov. Smith's campaign as a challenger in the 2002 election. His supporters note that statewide economic growth of 3.5% in 2004 was a new record under Gov. Smith.
Some of Bialik's readers noted that to evaluate the growth rate in this example one would need a number to compare it to, as the average growth rate for a state in 2004 might be the same or greater than 3.5%. However, no one seems to have noticed that Smith's supporters are trying to give him credit for the record growth simply because he was Governor when it occurred, which is an example of the cum hoc causal fallacy. To correctly credit Smith, we need some evidence that he had enacted some policies to stimulate economic growth, that they had had time to work, and that there was reason to believe that they contributed to the record growth.
- A crowd of 93,356 saw the U.S. women's soccer team beat China, 1-0, to clinch the World Cup. That was the largest crowd to witness a women's athletic event since the 1996 soccer final at the Atlanta Olympic Games, which drew a record 84,975 fans.
Bialik points out that other sources actually gave different numbers for the crowds at these two events. This confirms that one should be suspicious when reading such exact numbers: how are these crowd sizes measured, and can they be so accurate? The example does not explain how the sizes of the audiences were measured, but the existence of different counts suggests that these numbers are misleadingly precise. A numerate and logically literate reporter should have given approximations, such as: "around 95,000" and "about 85,000".
- The glaciers that span much of Greenland are melting quickly; one of them has more than doubled in speed, moving at a rate of 5.2 miles an hour, compared with 2.3 miles an hour a year earlier.
This has the same problem as the previous example: Are the measurements of slow movement really so accurate? Also, even if they are, what is the point of the additional information to the tenths of a unit? Why do we need to know that? Why doesn't the reporter simply write "approximately 5" and "approximately 2"? A. K. Dewdney calls these extra digits "dramadigits", because they are used to impress the reader rather than to inform.
- Carl Bialik, "The Results are In", The Numbers Guy, 1/27/2006
- A. K. Dewdney, 200% of Nothing (Wiley: 1993), pp. 52-53.
Resource: "Pop Quiz", 1/19/2006
Q: The whaling debate has been going on for a while and has been getting media (in Australia at least!). One argument goes along these lines (here I paraphrase a comment made by a representative of the High North Alliance, interviewed on "60 Minutes"):In Norway we eat the whale, in Australia, you eat the cow. You cannot―how can you?―tell us Norwegians not to eat the whale?
On its face, this seems reasonable. However, I suspect it is a tu quoque fallacy. Am I correct?
A: You are right, Jon. It is also a nice example of how real-world arguments can combine multiple fallacies, since it's also a weak analogy. The "you're another" accusation is based upon drawing an analogy between Australians eating cows and Norwegians eating whales, but these cases are distinct enough to make a difference to the argument. Here are some important differences:
- Cows are domesticated animals, but whales are wild.
- Cows are abundant and in no imminent danger of extinction, whereas whales are rare and some species are either endangered or threatened with extinction.
- There is evidence that whales have superior intelligence to cows.
For these reasons, the analogy between whale-eating and cow-eating is a weak one. However, even if it were a strong analogy, the argument would still commit the tu quoque fallacy that you spotted. You are hereby officially awarded the genuine Fallacy Files fallacy spotter spurious prize!
Check it Out
Jack Shafer has an article in Slate on a survey based upon a biased sample and the uncritical reporting of it.
Source: Jack Shafer, "This Is Your County on Meth", Slate, 1/19/2006
Carl "The Numbers Guy" Bialik's latest column includes a pop quiz to test your numeracy. You can also use it to test your logicality in spotting a few fallacies. Try it out.
Source: Carl Bialik, "Monitoring the Numbers in the News", The Numbers Guy, 1/20/2006
Here's an exchange between Senator Dianne Feinstein and Judge Samuel Alito from the latter's nomination hearing:
FEINSTEIN: [Y]ou went on to say that the present form, quote, "might be sustainable in its current form if Congress made findings that the purely intrastate possession of a machine guns [sic] has a substantial effect on interstate commerce or if Congress or the executive assembled empirical evidence documenting such a link. … So if I understand this, you essentially said that you wanted to follow precedent, newly established law in this area. And you left a little hedge that if Congress made findings in that law, then that might be a different situation. If Congress did make findings, would you have agreed that that statute would been constitutional?
ALITO: What I said in the opinion and what I will reiterate this afternoon is that it would have been a very different case for me. I don't think I can express an opinion on how I would have decided a hypothetical case.
FEINSTEIN: It's not hypothetical. I'm just asking you, if there were findings as you said, you might have sustained the law.
The word "innumeracy" was coined to refer to mathematical illiteracy: ignorance of basic mathematics. We need a similar word to refer to ignorance of basic logical concepts. "Illogicacy" sounds ridiculous; perhaps the word "illogicality" will serve. So, what Feinstein says is not fallacious, but is an example of illogicality, as she doesn't seem to understand what a hypothetical question is.
A hypothetical question is a question based upon a hypothesis, which is a proposition usually introduced by the word "if" or one of its synonyms. For example: "if Feinstein doesn't know what a hypothetical question is, how did she ever become a United States Senator?" One use of hypothetical questions is to ask "what if" questions about how things could have turned out differently than they did. Feinstein's question to Alito is such a "what if" question, which Alito refuses to answer on the grounds that it is hypothetical; which it is, despite Feinstein's denial.
Fallacious reasoning is one type of illogicality, but another is ignorance of simple logical vocabulary, such as "hypothetical". Knowing such basic concepts is far more important than knowing the names of fallacies, since understanding fallacies requires a knowledge of the vocabulary.
Daniel Drezner cited this as the "dumbest" thing said by a Senator in the hearing. I doubt that that's true, but it is illogical.
Source: "U.S. Senate Judiciary Committee Hearing on Judge Samuel Alito's Nomination to the Supreme Court, Part II of III", Washington Post, 1/10/2006
Via: Daniel W. Drezner, "And the Dumbest Thing Said by a Senator is…", 1/14/2006
Lost in Translation
Why can't today's computers translate as well as a bilingual human translator? Where's the universal translator predicted in old science fiction? Slate has an interesting article today by a human translator wondering when artificial translation will put him out of a job. What's the problem holding up progress?
The problem with translation software is context. When you hear or read a sentence, your brain refers not only to the spoken words but also to its accumulated experience. The words "con" and "pen," for example, have various meanings and can represent different parts of speech. But when you read "the con is in the pen," you know instantly that you are dealing with an incarcerated criminal―your life experience allows each word to contextualize the other. A computer can't do that because it has no frame of reference to help it match the contingent sense of "con" as criminal to the contingent sense of "pen" as jail. Short of being endowed with a knowledge base as vast as the human mind's, a computer simply cannot read context.
The author doesn't note that there are research projects with the goal of creating such knowledge bases. However, such an enormous amount of knowledge would have to go into a knowledge base in order to rival a human translator that the author's job is safe, though the jobs of his children or grandchildren―should they follow in his footsteps―may not be.
Why does one have to know so much to be able to translate? Shouldn't it be enough to know the grammar and vocabulary of both languages? Some philosophers and computer scientists used to think so, which is one reason why there were rosy predictions of quick progress. However, grammar and vocabulary are insufficient because of ambiguity and vagueness. Ambiguity and vagueness may seem to be things that we should want to get rid of―and many logicians of the past dreamed of artificial languages purged of fuzziness―but they help make communication efficient. If we had to speak and write with perfect precision, our messages would have to be much longer and more tedious. Moreover, for most purposes we don't need precise information. So, we learn to supply just enough information when communicating, knowing that our audience can fill in the gaps themselves from knowledge that they already share with us.
So, ambiguity and vagueness together with a fund of shared knowledge makes communication faster and more interesting. Of course, it also makes fallacies possible, but that's a price we pay to shorten and speed up our messages.
Source: Jesse Browner, "The Translator's Blues", Slate, 1/9/2006
12 found alive in coal mine
Source: Chicago Tribune, 1/4/2006, frontpage headline
Twelve miners caught in an explosion in a coal mine found alive Tuesday night
Source: Terre Haute Tribune-Star, 1/4/2006, frontpage headline
John Allen Paulos' latest Who's Counting column gives a nice example of sampling bias. I don't have anything to add to it, so read it for yourself. However, I do want to comment upon the following "Postscript on Iraqi War Dead" at the end:
Another figure in the news recently has been the number of Iraqis killed in the war. President Bush mentioned last month that in addition to the more than 2,100 American soldiers killed so far in Iraq, that there were approximately 30,000 Iraqis killed. He was likely referring to the approximate figure put out by Iraq Body Count, a group of primarily British researchers who use online Western media reports to compile an extensive list of Iraqi civilians killed. The organization checks the names and associated details of those killed. It necessarily misses all those whose names don't make it into the reports, and it makes no attempt to estimate the number it misses. The group's list contains almost 30,000 names at present.
Here is what Paulos said about the Iraq Body Count (IBC) in his previous article:
[T]he number of civilians killed…is no doubt bigger than the figure reported by Iraq Body Count. IBC is a group of British researchers who compile as extensive a list of Iraqi civilians killed as they can from published reports, hospital records, and morgue reports. They doublecheck and otherwise fully document the names and associated details of those killed, but miss those whose names don't make it into the papers or onto hospital or morgue lists. Their list contains approximately 15,000 names.
These descriptions of the IBC methodology are quite misleading and wrong about the number of names. I didn't catch it in the first article, though I was puzzled that he thought that the IBC had to be an undercount. IBC does have a list of names, but only about 3,000 of them, that is, approximately 10% of the total number of casualties they have counted. If Paulos' description of the IBC methodology were correct, then it would almost certainly be an undercount. However, what IBC does is to count the numbers of dead civilians reported in at least two news articles, not the names. When different sources disagree, the higher number is added to a maximum count and the lower to a minimum, and the IBC reports the total count as a range. So, as far as I can see, the total IBC count could just as easily be an overcount as an undercount, especially the maximum.
There are at least two ways that the IBC may overcount casualties:
- The same death may be counted more than once when it is reported more than once. Presumably, the IBC tries to avoid multiple counting by determining whether different news stories are reporting the same or different casualities, but this may not always be possible.
- Some news stories will likely report exaggerated casualty figures. Such reports may be later corrected, or competing stories may have more accurate figures. However, such reports will at least inflate the maximum number of casualties, unless IBC is careful to revise numbers downward when they are later contradicted.
Do the factors leading to under- and overcounting balance out so that the true number of casualities is within IBC's minimum and maximum? Possibly, but we can't know unless and until the new Iraq government does a count on the ground.
When I went to the local public library earlier today, I saw three different newspapers each with the main frontpage headline proclaiming that twelve miners had been rescued alive. This puzzled me as I seemed to remember having heard that all but one of the miners had died, yet here were three separate, trustworthy news sources telling me otherwise in large type. If journalists can be so wrong reporting in their own country where they speak the language, then they can be wrong reporting in a war zone in another country where they don't speak the language.
- John Allen Paulos, "Flu Deaths, Iraqi Dead Numbers Skewed", Who's Counting, 1/1/2006
- John Allen Paulos, "Misleading Numbers in the News", Who's Counting, 12/5/2004
- "Methodology", Iraq Body Count
- "Named and Identified Victims of the War on Iraq", Iraq Body Count, 9/2004
Resource: Jocelyn Noveck, "Some Say Media Erred in Mine Coverage", Associated Press, 1/4/2006
Answer to Puzzle: Your Mileage May Differ: The word "mile" is ambiguous: distance on land is measured in statute miles, while distances at sea are measured in nautical miles. A nautical mile equals 1.15 statute miles, so a mile at sea is farther than a mile on land.