WEBLOG
July 3rd 2014 (Permalink)
Strike Three…You're Out!
The final strike against the Slate article discussed in the two previous entries is its tabloid-style headline:
Do Americans Think Corporations Have the Right to Religious Freedom?
We did a survey, and the answer is no.
As explained in the previous entries, this was not shown by the survey. Putting aside the problems with how the poll's sample was selected and its size, the most it showed is that Americans think that the right to religious freedom of individuals is more important than that of corporations, not that the latter have no such right at all. The article itself, while occasionally exaggerating the poll's results, was more accurately written―see the quote in the first entry for this month, below. The language used in the body of the article is carefully and consistently comparative, since the most that the poll shows is greater support for the religious freedom of individuals than of groups.
Many people are likely to read such a headline and not bother with the article itself, thus getting a false impression about what it shows. Even those who read the entire article may be misled into misinterpreting it along the lines suggested by the headline. Perhaps the headline writer misunderstood the poll results, or the editors at Slate are willing to write false but provocative headlines in order to trick people into reading their articles.
Update (7/4/2014): I've revised the second paragraph to remove a silly, unintended implication.
July 2nd 2014 (Permalink)
Strike Two…
The same Slate article discussed in yesterday's entry also provides a good test of your understanding of the pitfalls of charts and graphs. Look at the chart below, which accompanies the article and reports the poll results. Using the terminology given in the lessons on charts and graphs from last year―see the links, below―can you identify what type of chart this is? See below the chart for the answer.
This is a gee-whiz line graph―see Lesson 1, below. The y-axis doesn't start at 1, which is the lowest number in the scale, thus visually exaggerating the differences between the results for people and businesses. Worse, the graph does not have a broken y-axis near the x-axis, which is standard practice in chartmaking when cutting off the bottom of a graph. One might excuse the truncation, but not failing to include a visual sign of it, for that is charting malpractice.
As I pointed out in the previous entry, it's not at all clear what the numbers are supposed to mean in this survey. Someone just glancing at the graph might get the false impression that, because the line for groups is close to the bottom of the chart, that must mean that most people do not believe that groups deserve "religious liberty rights". However, all of the results but one are at or above 4 on a scale of 1 to 7, which might just as well indicate differing degrees of support for such rights.
Lessons in Charts & Graphs: 1, 2, 3, 4, 5, 6, 7, 8, 9
July 1st 2014 (Permalink)
Strike One…
A current article in Slate is a good test of your understanding of the pitfalls of polling and its reporting. See if you can figure out what, if anything, is wrong with this poll before I discuss it. Here's a brief synopsis of the article, though I suggest reading the whole thing, which isn't long―see the Source linked below:
The Supreme Court ruled 5–4 on Monday that Hobby Lobby, a family-owned hardware retail chain, has a right to religious freedom. … What does the American public think about religious rights for corporations? Over the past few months, we’ve conducted a series of surveys, delivered online to more than 300 people across the country between the ages of 18 and 76. … To understand how people think about corporate rights, we asked how important they thought the religious liberty right of a company’s owner, its employees, and the company itself, on a scale of 1 to 7. … Our respondents consistently agree that religious liberty is important for both individual employees and owners. In contrast, respondents were significantly and consistently less willing to grant the same scope of protection to all the for-profit companies we presented them. …[P]eople did distinguish between for-profit and non-profit organizations, deeming the latter to be more deserving of religious liberty rights. But even when the religious liberty of a church is at stake, people viewed it as much less deserving of religious rights than its employees and officers.
There are several problems with this poll or the article that reports it:
- This was an online poll, and most such polls use self-selected samples, that is, they are not "scientific" polls that use random samples―for more on the distinction between self-selected and scientific surveys, see the Resource below. The article itself doesn't explain how the sample was taken, nor can I find a link to a longer report on the poll that would do so. In the absence of some assurance that the sample was taken randomly, I think we have to assume that it was not.
- There is no mention in the article of a margin of error (MoE) for the poll, which violates the usual American journalistic standard to report the MoE. Since MoEs can only be calculated for surveys based on random samples, this is further evidence that the sample was probably self-selected. However, even if the sample was taken randomly, 300 respondents is a small sample for a public opinion poll, which typically will have one more than three times as large. Assuming a random sample and a confidence level of 95%, the MoE would be around five and a half percentage points.
- Unfortunately, even assuming a random sample, it's impossible to use the MoE to evaluate the poll since none of the results are reported as percentages. Instead, the results are numbers on a scale from 1 to 7, which are presumably averages of the answers given by respondents. However, as far as we can tell, it may be that none of the differences reported by the poll are statistically significant. See the Resource, below, for more on MoE errors.
- Finally, the poll asked respondents to rate the religious rights of people and companies on a numerical scale, but what do the numbers mean? Presumably, a "1" would mean that respondents think that the person or company has no rights, and a "7" would mean that they do, but what does a "4" or "5" mean? Did the survey itself explain the meanings of the numbers? If so, then the article is poorly reported.
In sum, there's no reason to believe the sample used in this survey is representative of the American public, and no evidence that the differences in results are statistically significant. So, as far as we can tell from this article, the poll is worthless.
Source: Moran Cerf, Aziz Huq & Avital Mentovich, "Do Americans Think Corporations Have the Right to Religious Freedom?", Slate, 7/1/2014
Resource: How to Read a Poll
June 30th 2014 (Permalink)
A Puzzle in Wonderland Forest
There are exactly three kinds of animal in Wonderland Forest: bandersnatches, snarks, and boojums. These categories are not necessarily mutually exclusive, that is, it's possible for an animal to belong to more than one. All of the following statements are facts about the animals in the forest:
- 5% of the animals are only boojums.
- The percentage of boojums that are bandersnatches but not snarks is the same as that of animals of all three types.
- 33% of the animals are not bandersnatches.
- The percentage of snarks that are boojums but not bandersnatches is one percentage point greater than that of animals that are boojums only.
- 17% of the animals are both snarks and bandersnatches but not boojums.
- The percentage of animals that are only snarks is half that of those that are just bandersnatches.
From these facts can you determine what percentage of animals in Wonderland Forest are of all three types, that is, bandersnatches that are both snarks and boojums?
Solution to a Puzzle in Wonderland Forest
June 28th 2014 (Permalink)
Why deny the Obvious Child?
Some people say "A lie's a lie's a lie"
But I say "Why?
"Why deny the obvious child?"
―Paul Simon, "The Obvious Child"
Ads and posters for the new movie Obvious Child include the following blurb:
"THE MOST WINNING ABORTION-THEMED ROM-COM EVER MADE."
-THE DISSOLVE
Unlike the usual "Blurb Watch" item, this is not misleadingly taken out of context. Rather, it raises a different question: Just how many abortion-themed romantic comedies have ever been made? If it's the only one ever made―and I can't think of another―then of course it's the most winning one, but also the least. In that case, what's the most winning needn't be particularly winning at all.
The author of the review did like the movie, but in a later interview with its director and star he commented about the blurb:
The Obvious Child poster quotes my Sundance coverage, where I called it the “most winning abortion-themed rom-com ever made.” I feel a little strange about that―it was a little facetious.
Just "a little"? This reminds me of the following capsule review of the Australian monster movie "Razorback": "Arguably the best movie ever made about a man-eating hog." Only arguably?
Sources:
- "Razorback", Xfinity
- Nathan Rabin, "Day 2: The young and the restless", The Dissolve, 1/18/2014
- Nathan Rabin, "Gillian Robespierre and Jenny Slate on finding Obvious Child’s voice", The Dissolve, 6/3/2014
June 24th 2014 (Permalink)
Taking Gay Marriage Too Far, Again
The Associated Press reports:
BREAKING: Methodist panel overturns decision to defrock Pennsylvania pastor who married his gay son.
I laughed when they told me that gay marriage would lead to gay incestuous marriage.
Sources:
- "BREAKING: Methodist panel overturns decision to defrock Pennsylvania pastor who married his gay son.", Associated Press, 6/24/2014
- "BREAKING: Clarifies: Methodist panel overturns decision to defrock Pennsylvania pastor who presided over gay son's wedding:", Associated Press, 6/24/2014
Resource: Headline, 12/12/2013
June 18th 2014 (Permalink)
The Million "Million X March" March
Is there any possibility of a voluntary agreement to stop calling demonstrations "The Million X March", such as "The Million Man March", "The Million Mom March", "The Million Mask March", etc. Most recently, there was The Million "Pibble" March―"pibble" apparently being baby talk for "pit bull", so this was actually The Million Pit Bull March. Did a million pit bulls really march? No. How many did march? Apparently none. That's right, no pit bulls marched―or walked, trotted, waddled, or whatever pit bulls do―in The Million Pit Bull March. Zero, zilch, nada pit bulls. In fact, no dogs of any kind were allowed to march! So, did at least a million pit bull owners march? Here's a quote from an article that appeared before the event:
…Los Angeles-based comedian and pit bull advocate Rebecca Corry estimates…as many [as] 5,000 people will assemble on the West Lawn of the U.S. Capitol, asking that government resources be dedicated to eradicating dog fighting and that laws restricting the ownership of pit bulls be abolished.
So, even the organizer of the march was predicting only 1/200th of a million people would show. Of course, "The 5,000 Pit Bull Owner March" just doesn't have the same ring to it. From the article, it appears that they didn't even march but "assembled". So, like the "Holy Roman Empire", there was no million, pit bulls, or march in "The Million Pit Bull March".
What's the next Million X March? The Million Little Green Man March? The Million Bigfeet March? Why not "The Billion Bigfeet March", anyway? As far as I know, none of these marches managed to get anywhere near the million mark, so why not take propaganda inflation into account? After all, the Million Man March, which seems to have been the first "million" march, was almost twenty years ago, so we should at least be up to the ten-million mark by now. Even that march didn't live up to its hyped name, since the U.S. Park Police estimated the crowd as less than a half-million at most. So, given that the names of these marches are not constrained by the actual numbers of marchers, what's to stop some group from having a "Billion X March"?
I hereby announce The Million Anti-"Million X March" March, in which a million marchers will march on Washington to protest the continued use of the phrase "Million X March" in events in which a million Xs do not participate. How many marchers do I expect to show up? At least as many as pit bulls that showed up for The Million Pit Bull March!
Sources:
- Joel Best, Damned Lies and Statistics: Untangling Numbers from the Media, Politicians, and Activists (2001), pp. 132-137.
- Arin Greenwood, "Here Comes The Million Pit Bull March On Washington!", The Huffington Post, 4/10/2014
- Charles Seife, Proofiness: The Dark Arts of Mathematical Deception (2010), pp. 15-16.
June 16th 2014 (Permalink)
Q&A
Q: Take a scenario where a religious leader makes a statement/claim that is subsequently scientifically verifiable as not true, but in his religion there are other statements/claims that might or might not be true, but because of at least one falsehood the whole religion is then dismissed as phoney. Is rejecting the whole because of one false occurrence a fallacy, and if so, what is the name of the fallacy? It amazes me that purveyors of new religions/cults include "physical type" false claims as science invariably will refute them.―John Addicott
A: There is a mistake sometimes called the "all-or-nothing fallacy", which is a type of black-or-white fallacy―see the Fallacy, below. The black-or-white fallacy attempts to force you to choose between two alternatives when other alternatives are available; the all-or-nothing fallacy specifically tries to force you to accept or reject something as a whole when there is the additional alternative of accepting part and rejecting part.
I suspect that some religions that demand their adherents swallow doctrines whole do so because they claim to be divinely inspired. God is usually conceived as infallible, and if the scriptures or prophet of a religion are claimed to speak for God then they should also be infallible. Logically, if the scriptures of a given religion contain a false claim, that wouldn't imply that every claim they make is incorrect. However, it would entail that they are not infallible, and therefore not divinely inspired.
Now, those who recognize that a given religion contains errors, and therefore is not infallible, should not swing to the opposite extreme of rejecting every claim it makes, which would be another application of the all-or-nothing fallacy. The fact that a religion is fallible means that some of its claims are false, not that they all are.
As to why someone starting a new religion would make easily-refutable empirical claims, your guess is as good as mine.
Fallacy: The Black-or-White Fallacy
Source: "Little Hope Baptist Church", Funny Signs
Resource: A Philosopher's Fallacy: "All or Nothing", 10/1/2006. This is actually a different fallacy under the same name as the one discussed above. I point to it just to make it clear that it's not the same thing.
June 8th 2014 (Permalink)
Mystery Science Theater 2014
Every so often I post a "Blurb Watch" entry as an example of how quotes taken out of context can be misleading. Of course, not a lot is at stake with contextomies in movie ad blurbs―at worst, you might spend ten dollars and waste a couple of hours watching a bad movie―but here's a more serious one from the headline of a recent Huffington Post story:
Nasa: Strange Markings Across The Globe 'Might Have Been Made By Aliens'
In true tabloid fashion, the article itself doesn't quite live up to its own headline―in fact, it pretty much goes on to debunk it―but there were some other articles that took it seriously.
Where did they get the quote attributed to NASA that prehistoric rock carvings "might have been made by aliens"? Here's a passage from another article on the same subject:
…William Edmondson from the University of Birmingham considers the possibility that rock art on Earth is of extraterrestrial origin. ‘We can say little, if anything, about what these patterns signify, why they were cut into rocks, or who created them,’ he writes. ‘For all intents and purposes, they might have been made by aliens.’
This whole story seems to have been concocted from this one sentence taken out of context. I don't know where or how it got started, but it's difficult to see how it could be an innocent mistake. No one reading the quote in context is likely to misunderstand it:
Consider again, therefore, the desirability of establishing symbolic/linguistic communication with ETI [ExtraTerrestrial Intelligence]. It is helpful to review some parallels from human existence that pose problems for us today. One of these is “rock art,” which consists of patterns or shapes cut into rock many thousands of years ago. … We can say little, if anything, about what these patterns signify, why they were cut into rocks, or who created them. For all intents and purposes, they might have been made by aliens. Unless we find a readable exegesis of them produced at the time they were made, we will never be able to say with certainty what the patterns mean.
This is from a book put out by NASA concerning the prospects and difficulties of communication with ETIs, should there turn out to be any. The author, William Edmondson, rightly emphasizes the difficulties of deciphering even human-made artifacts, let alone those made by intelligences from another planet and a different evolutionary history. But don't take my word for it: you can read the chapter for yourself―see Source 2, below.
Sources:
- "Nasa: Strange Markings Across The Globe 'Might Have Been Made By Aliens'", The Huffington Post UK, 5/21/2014.
- William H. Edmondson, "Constraints on Message Construction for Communication with Extraterrestrial Intelligence", in Archaeology, Anthropology, and Interstellar Communication, edited by Douglas A. Vakoch, pp. 235-248 (PDF). Neither long nor overly-technical.
- Jason Mick, "Did NASA Book Say Rock Art Was the Work of Extraterrestrials? Nope.", Daily Tech, 5/21/2014.
- Jonathan O'Callaghan, "Have aliens already visited Earth? Nasa book suggests that ancient rock art could have been created by extraterrestrials", Mail Online, 5/21/2014.
- Mick West, "Debunked: Nasa: Strange Markings Across The Globe 'Might Have Been Made By Aliens'", Metabunk, 5/22/2014.
Fallacy: Quoting Out of Context
June 3rd, 2014 (Permalink)
New Book: How Not to be Wrong
In more book news, mathematician Jordan Ellenberg's long-awaited―at least, I've been waiting for it for so long I forgot about it―How Not to be Wrong: The Power of Mathematical Thinking is finally available. Judging by Amazon's "Look Inside!" feature, Ellenberg discusses some issues that we've encountered here, at least in passing, such as the voting paradox, straight-line extrapolation, regression to the mean, and Bayes' theorem.
Resource: Evelyn Lamb, "How Not to Be Wrong (Book Review)", Scientific American, 5/31/2014
Update (6/13/2014): Ellenberg has written a series of posts for Slate magazine based on his book and dealing with several subjects discussed, at least in passing, in these webpages. Check 'em out.
Source: Jordan Ellenberg, "How Not to be Wrong", Slate, 6/3-13/2014
June 1st, 2014 (Permalink)
In the Mail: Humbug! (Second Edition)
Humbug!, a book by Theo Clark and his late father Jef, is out in a second edition. I reviewed the first edition several years ago―see the Resource, below. Unfortunately for us old-fashioned folks who still think of a book as a thing made out of paper, it seems to be available only as an e-book. However, it's hard to beat for sheer entertainment value, and the price is certainly right.
Resource: Book Review: Humbug!, 3/1/2006
Online dissertation services such as MastersThesisWriting may help with your thesis or dissertation.