Thank you to everyone who has supported The Fallacy Files, whether by clicking on Google ads, buying books or other items from Amazon links, or posters and art prints through Allposters. Your support is appreciated!
Update (5/11/2017): The Fallacy Files is no longer an Allposters affiliate, but you can still support the site through Google ads, Amazon purchases, or PayPal donations―see the "Donate" button in the right-hand navigation panel. Thanks for your support!
National Treasure: Book of Secrets
|"HEART-POUNDING ACTION! TOTAL ENTERTAINMENT FROM START TO FINISH!"
PETE HAMMOND, MAXIM
|Even though it's all wildly unbelievable and rather hokey at times, National Treasure: Book of Secrets manages to surpass the first one for pure movie fun and some decent heart-pounding action, a totally safe family entertainment from opening credits to closing logo.|
What I want to call attention to about this blurb is not the usual capitalization and exclamation points which were not in the review, but the difference between "total entertainment" and "totally safe family entertainment". The adwriter has turned an adverb that modifies "safe" into an adjective that modifies "entertainment", with no indication of the change or the omission of the words in between. "Totally safe family entertainment" is simply a characterization of the content of the movie as appropriate for children and adults both, whereas "total entertainment from start to finish" seems to mean that the movie is totally entertaining, that is, entertaining from beginning to end with no boring parts. It's a five-star blurb for a three-star movie.
- Ad for National Treasure: Book of Secrets, Dallas Morning News, 12/28/2007
- Pete Hammond, Review of National Treasure: Book of Secrets, Maxim
Check it Out, Too
Law professor Orin Kerr asks the question "Should the LSAT have a 'logic games' section?", and answers: "no". As I'm not a lawyer, I take no position on the issue of whether this type of puzzle is a good test of the reasoning abilities needed by lawyers. I've posted a few "logic games" similar to those on the LSAT on this weblog because of requests from some prospective law students for help with this kind of problem. As long as the test continues to use these kinds of puzzles, those who wish to attend law school will need to know how to solve them. Those who have no intention of becoming lawyers can work such puzzles for fun and exercise of their reasoning ability. So, I'll post another puzzle in the near future.
Source: Orin Kerr, "Should the LSAT have a 'Logic Games' Section?", The Volokh Conspiracy, 12/27/2007
The Hot Hand Fallacy is! This fallacy is only half new, since it's what used to be called "the Reverse Gambler's Fallacy" and shared a page with the Gambler's Fallacy. I had heard the phrase "hot hand" before, but was not aware until recently that "hot hand fallacy" was used as a name for this fallacy by some psychologists. This is a much better name for the fallacy as it's more memorable and, if you know what it means to be "hot" in gambling, the name reminds you of the nature of the mistake. I have also moved the Gambler's Fallacy, alongside the new fallacy, under Probabilistic Fallacy in the taxonomy. When I created the entry for Probabilistic Fallacy I forgot to put the Gambler's Fallacy as a subfallacy of it.
Check it Out
A recent "Numbers Guy" weblog entry concerns a pet peeve of mine, namely, the way that movie studios hype their box office returns. He discusses the fact that the studios never adjust their figures for inflation, so that almost every year they're able to trumpet record-setting receipts for some movie. If inflation were taken into account, records would be set much less often, and the studios wouldn't be able to make a big deal out of them.
However, even if a movie sets an inflation-adjusted box office record, there's something logically wrong with selling it based upon how much money it makes, or how many tickets it sells. This is a type of bandwagon appeal, and the studios are hoping that the more popular a movie appears to be, the more people will jump on the bandwagon.
If you tend to like movies that are popular, then the fact that a movie is successful may be a good reason to see it. Or, it's possible that you might want to see a popular movie for social reasons; for instance, to use as a conversation opener. Absent such circumstances, however, there's no reason to see a movie just because a lot of other people are seeing it.
Source: Carl Bialik, "Box-Office Records Are the Stuff of ‘Legend’", The Numbers Guy, 12/17/2007
A Logic Puzzle Present
Five children took turns opening their presents on Christmas morning. Each child received exactly one present, no two of which were the same, and each present was wrapped differently―one was wrapped in green paper. From the following clues determine the order in which the children opened their presents, what present each child received, and how the presents were wrapped.
- The present with a red bow on top―which was not the video game―was opened after the puzzle was unwrapped and before Alan opened one.
- Betty was the third child to unwrap a present.
- The gold box was opened after the train set was unwrapped, and before Chuck got a present.
- The present wrapped in silver paper was opened immediately before the book was unwrapped, and Dotty got her present immediately before the video game was opened.
- Eddie opened the gold box if and only if it was the third present unwrapped.
- If Chuck was the last child to open a present, then Betty was the first.
- The gold box was opened after the present tied with a red ribbon, which was not the doll.
Bad Science Reporting
Honey Beats Meds at Soothing Kids' Cough
With many children's cough syrups being pulled from the market because they don't work, an old folk remedy―honey―may work just as well or better, researchers report. In a study of kids having trouble sleeping because of cough, a research team at Penn State College of Medicine compared the effectiveness of a little bit of buckwheat honey before bedtime versus either no treatment or dextromethorphan (DM), the cough suppressant found in many over-the-counter cold medicines. "Honey provided the greatest relief of symptoms compared with the other treatments," concluded lead researcher Dr. Ian Paul, Penn State's director of pediatric clinical research. … Paul's group found that honey was more effective in reducing the severity and frequency of nighttime cough compared with DM or no treatment. Honey also allowed the children to sleep. Moreover, DM was not much better at reducing cough than no treatment, the researchers found. … Charlotte Jordan, a project manager of research at the National Honey Board, believes the finding confirms what your grandmother told you. "This is a really exciting finding," she said. "For a long time it's been folklore medicine to use honey when you have a cough or a cold, but it's exciting to have a scientific study to back that up."
Unfortunately, the study in question doesn't support the claim in the headline. It's true that the parents of the children who received honey reported greater improvement than those who received DM, but the difference was not statistically significant―a fact that was not mentioned in the body of the article. Moreover, the honey group also reported greater negative side effects than DM, a difference that was not statistically significant, either, but if you're going to report insignificant differences then you should at least be even-handed. What was statistically significant was the difference between honey and no treatment at all.
Also not mentioned in the article is that the study was funded by the National Honey Board. Now, I have no reason to think that the source of funding affected the quality of the research; the fact that the results of the study don't make a compelling case for using honey instead of a cough syrup containing DM suggests that it didn't. The study does make a good case for further research, since a larger study might be able to show a statistically significant difference between honey and DM. However, the source of funding may explain why the study is being used to hype honey as an alternative to cough syrup, though it doesn't explain why the media buy into it.
- "Hitting the Headlines: Honey 'better than cough medicine'", National Library for Health, 12/5/2007
- Ian M. Paul, et. al., "Effect of Honey, Dextromethorphan, and No Treatment on Nocturnal Cough and Sleep Quality for Coughing Children and Their Parents", Archives of Pediatrics & Adolescent Medicine, Vol. 161, No. 12, 12/2007
- Steven Reinberg, "Honey Beats Meds at Soothing Kids' Cough", U. S. News & World Report, 12/4/2007
White House Mum? Who's that? Barbara or Laura?
Poll Watch: Compare & Contrast
Here are reports of the same poll from two different sources:
|Huckabee new GOP leader in Iowa Poll||Iowa Contests Tight in Both Parties|
|"Mike Huckabee has leaped ahead of Republican presidential rival Mitt Romney in Iowa, seizing first place in a new Des Moines Register poll of likely Republican caucus participants. … "||"The races for both the Republican and Democratic nominations here are toss ups as voting approaches, a double-dose of fluidity unseen in decades. … A poll released Sunday by The Des Moines Register shows both races in dead heats."|
|Obama pulls ahead for Democrats in Iowa Poll|
|"Barack Obama has pulled ahead in the race for Iowa's Democratic presidential caucuses, while the party's national frontrunner Hillary Clinton has slipped to second in the leadoff nominating state…."|
So, who's right? The Des Moines Register, reporting on its own poll, in the left hand column; or the Associated Press, in the right hand column?
Though, as usual, a thousand people were polled, half were Republicans and the other half Democrats, so the results have a margin of error of ±4.4%. Since the margin of error applies to the result for each candidate, the difference between two candidates must be greater than 8.8% in order to be significant. So, differences of less than 9 percentage points may simply be the result of sampling error.
On the Republican side, Mike Huckabee (29%) and Mitt Romney (24%) are tied for the lead. All other Republicans are significantly behind, with only Rudy Giuliani (13%) in double digits. On the Democrat side, Barack Obama (28%), Hillary Clinton (25%), and John Edwards (23%) are tied, with all other Democrats significantly behind in single digits.
Thanks and praise to the AP for reporting a poll accurately and without hype!
- "About the November 2007 poll/methodology", Des Moines Register, 12/1/2007
- Thomas Beaumont, "Huckabee new GOP leader in Iowa Poll", Des Moines Register, 12/2/2007
- Jonathon Roos, "Obama pulls ahead for Democrats in Iowa Poll", Des Moines Register, 12/2/2007
- Liz Sidoti, "Iowa Contests Tight in Both Parties", Associated Press, 12/2/2007
Resource: How to Read a Poll, Fallacy Watch
Update (12/3/2007): John Congdon writes:
There was an interesting bit on the radio this morning about the Iowa poll you cite in yesterday's weblog entry. A political analyst said something like "The results are within the poll's margin of error, but the fact that Obama polled higher than Clinton may indicate a shift in voter support." At least they mentioned margins of error, but this sounds to me like the statistician's equivalent of having your cake and eating it too.
Indeed. This is typical of the reporting―with the exception of the AP―of this particular poll, and of polls in general―not excepting the AP. All reputable news media seem to have a policy of mentioning the margin of error (MoE)―though some relegate it to fine print at the bottom of a story―but the reporters treat it as if it were technical data for statisticians, rather than important information that the reporter and reader need in order to evaluate the significance of the results. So, the reporter will write the story as if the poll numbers were perfectly precise. This time the AP did not do this, which we can only hope will begin a trend towards better poll reporting.
- The Numbers Guy's latest weblog entry concerns whether a difference in poll results that is less than twice the MoE should be called a "statistical tie" or "statistical dead heat". Some pollsters—including the Mystery Pollster—object to these terms on the grounds that people may misunderstand them to mean that the poll shows a tie, when it really just fails to show who is ahead.
I didn't use the qualifier "statistical" in the above entry, but I did call the results of the Register poll a "tie", which should be understood to mean that the poll does not indicate who is ahead. Of any two candidates, it's virtually certain that one is ahead of the other, but the poll isn't precise enough to say who leads.
Source: Carl Bialik, "What’s a Statistical Tie, Anyway?", The Numbers Guy, 12/6/2007
- Howard Kurtz had the following question and answer session with a reader:
Gainesville, Va.: Howard, I think it is time to send political reporters and commentators to a remedial course in statistics. You wrote that Obama could as easily be tied with Hillary because his lead of 28 percent to 25 percent is within the margin of error. What the confidence interval means is that if the confidence interval is smaller than the difference between two means we can be 95 percent (not 100 percent) sure that the opinions held in the universe will reflect the higher response to be greater than the lower. If the difference of the two poll results are within the confidence interval, it may mean we can only be sure that one is greater than the other with 80 percent or 75 percent confidence. It doesn't mean they are tied. It really would be helpful if you and your colleagues boned up on this question before you made statements that aren't true.
Howard Kurtz: I stand by what I said. Obama's lead is within the margin of error. In fact, the poll's margin of error was over 4 percent, which makes the 3-point lead even more of what I would regard as a statistical tie.
And I stand by what Kurtz said, too. See the Resource below for a full discussion of this issue. The only thing that I would add to that discussion is that there are so many polls done now that the 95% confidence level may well be too low. At that level, one in twenty polls can be expected to be wrong by more than the MoE. If pollsters were to adopt the 75% confidence level, then one in four polls could be off by more than the MoE. In other words, to gain precision one would have to give up accuracy, and that would be a high price to pay. It would be better if pollsters would increase the precision of their polls by increasing the size of their samples, instead. This, of course, is expensive, which is why they're reluctant to do it, but perhaps fewer polls with larger samples and a 99% confidence level would cost no more than current polling.
Source: Howard Kurtz, "Media Backtalk", The Washington Post, 12/3/2007
Resource: The Big MOE, 9/3/2004
- Eddie received a train set that was tied with a red ribbon.
- Dotty opened a puzzle that was inside a gold box.
- Betty received a video game that was wrapped in silver paper.
- Chuck unwrapped a book that was topped with a red bow.
- Alan opened a doll that was wrapped in green paper.