"Statistics are no substitute for judgment" ? Henry Clay
"He uses statistics as a drunken man uses lamp-posts ? for support rather than illumination." ? Andrew Lang
"Statistical averages are one part truth and two parts lies ? all that is left should be made into hot-dogs." ? Hans Bluedorn
The proper interpretation of a statistic can be a very elusive task and it is not uncommon, in such a deceptive field, to find a fallacy poking its head from behind the protective percentages.
Below, such an instance can be seen "in action."
From the "Handgun Control Incorporated" (HCI) website:
"Does a gun in the home make you safer? No. Despite claims by the National Rifle Association (NRA) that you need a gun in your home to protect you and your family from possible home invasion, public health research demonstrates that the most person likely to shoot you or a family member with a gun already has the keys to your house. Simply put: Guns kept in the home for self-protection are more oftentimes [sic] used to kill somebody you know than to kill in self-defense; 22 times more likely, according to a 1998 study by the New England Journal of Medicine."
For the sake of brevity I will assume the research that resulted in this statistic was carried out in a scrupulous way. I will assume it is, in fact, a true representation of the facts.
(However, I have heard from various quarters that there were unscrupulous methods used to arrive at this particular "22 times more likely" number. You can, in fact, see some evidence of ambiguity in the study: How do they define "used to kill somebody you know"? Is "somebody you know" a friend? An acquaintance? Or the guy down at the Seven-Eleven? When vague terms are used, something fishy may being going on. If you wish to find out more on this you must do your own probing.)
What I am concerned with is the conclusions the people at Handgun Control draw from this number: "Does a gun in the home make you safer? No."
This conclusion, based on this number, represents what is known as the fallacy of sweeping generalization.
The fallacy of sweeping generalization is committed when a rule that is generally accepted to be correct is used incorrectly in a particular instance. The best way to explain this is with an example:
Suppose you decide to spend an evening at the opera. After sitting contentedly through most of the performance, as the opera nears its end - the part where the portly lady sings a last song and promptly dies - the man sitting next to you produces a kazoo and begins to accompany the performer. When you press him to explain his conduct he states, "Everyone has a right to free speech, don't they?"
While this man's premise - everybody has a right to free speech - is generally accepted, at least in America, this man is exercising it in an improper venue: Whatever this man's opinion of his kazoo playing, his right to express himself does NOT apply in a place where everyone has paid money to hear the lady sing. The audiences' freedom to hear the song they paid for overrides the man's right of free speech. It is important to remember that a general rule only applies where it is meant to apply. It is not necessarily appropriate under all conditions.
The fallacy of sweeping generalization is also at work when a statistical average is applied to specific people.
Example: "Divorce is rampant in America, Mary. I heard that 50% of marriages end in divorce within three years. So I've decided not to marry you because the odds are against us."
Here, a statistic is used to arrive at a conclusion, when the situation in question (between the speaker and Mary) may be quite different from the average. This couple may be much more serious about marriage than all those couples that were divorced, and consequently their chances of success may be better. They should find out what the main causes for divorce are, then determine whether these things are worth considering.
Now back to the point. In our first quote, the fallacy of sweeping generalization comes up when the HCI says: "Does a gun in the home make you safer? No."
Here, as with the divorce example, a general statistic - people in the average home are 22 times more likely use a gun for murder than for self defense - is being applied to you, the individual. Not very many homes are average, and yours may not be. You may live in the kind of home where the thought of murder is non existent, but the presence of a gun may add a level of security. But when HCI says you are "22 times more likely," they imply that having a gun around makes you that much more a murderer. In reality, this statistic (if it is a national statistic) covers a lot of groundfrom the godly Christian home to the to the inner-city kill-you-for-a-stick-of-gum neighborhood home. It lumps them all in the same average. Your household may be closer to the former than the latter.
This is like the man who, when he heard that most people die within five miles of home, said: "I'm moving."
It is applying a statistic in a way contrary to the original intention.
How do you expose a sweeping generalization?
As above, you need to find out what the rule or statistic in question is, and then show that, if the rule or statistic is properly understood, it cannot be applied to the specific case: The average family may (or may not) be "22 times more likely" to obtain adverse benefits from a gun, but that does not mean your family would fall within that average.
There may be some other fallacies in this article from Handgun Control Incorporated. I would be interested in what you think. The entire article can be found at: http://www.handguncontrol.org/facts/ib/gunhome.asp
This advertisement might already be familiar to you. It was sent to me by Amber from Decatur, Illinois.
Quote: "Show me Success! "We admit it: P*******l Homeschooling magazine isn't for everyone. It is for you if:
You would like your children to someday be in the running for full-tuition scholarships to top colleges instead of saddled with huge loans for second-rate degrees.
You care even more about their spiritual state, and want them to be ready to change the world, not to be changed by the world."
What fallacy is this? Below I have included some descriptions of fallacies; they should help you get started.
Send your answersand reasons for your answersto me and I'll print my selections next time.
1. Hasty Generalization: Generalizing about something without enough of a sample (this fallacy is similar to the fallacy of sweeping generalization):
Ex: All Fords are bad. I once owned a Ford and it was junk.
Ex. I don't think anybody lives in North Dakota. Everybody I ever talked to who was from there doesn't live there anymore.
2. Appeal to Ignorance: Saying something is true because it hasn't yet been proven false:
Ex: You can't prove you didn't steal that car, so you must be guilty.
Ex: No evidence has been found that there isn't life on other planets. Therefore, we are not alone in the universe.
3. Snob Appeal: Appealing to the inner snob. This fallacy tries to get us to do something to distinguish ourselves from everyone else.
Ex: Buy Skunk brand perfume; you will stand out from the crowd.
Ex: Marines: Do you have what it takes?
4. Generic Ad Hominem: an attempt to prove an argument false by attacking its source (its genesis), not the argument itself.
Ex: The only reason you disagree with me is because you were abused as a child.
Ex: That church cannot possibly be good. It was formed by an unregenerate heretic.
I am awaiting your responses,
I have included the following letters for your perusal and comments.
From: "Roger Van Couwenberghe" vancocon#hotmail.com
The Press Democrat 3/11/2001
Front page title: Bush Puts Strict Stamp on Office
Here is the first paragraph: "In seven short weeks of his presidency, George W. Bush has transformed how the White House and elements of the sprawling government operate in ways that contrast sharply with those of Bill Clinton and other past presidents...... But those common threads do not reveal the fundamental ways – besides ideology – that Bush differs from Clinton and other past presidents."
By blithely connecting President Clinton's presidency with "other past presidents," the reader is apt to equate President Clinton and other past presidents. The article points out the sharp contrast between Bush's style and Clinton's and other's styles. In reality, Bush's presidency differs far more sharply with just one past president: Clinton himself.
President Clinton doesn't seem so starkly errant when he's grouped together with less wacky company.
From: "Michels, Matt" mattm#ftc1.co.lsil.com
I have enjoyed this list quite a bit. Here is a bumper sticker which I have always thought had an interesting logical fallacy.
"If you can't trust me with a choice,
How can you trust me with a child?"
Trying to say that you should be "Pro-choice." This could potentially fall into a number of categories. I think it's primarily a non-sequitur but perhaps you have some others that it violates.
Just thought this might make for an interesting quiz.
From: Eugene B Sedy sedyzoo#juno.com Subject: Simple Logic?
I'm very new to the study of Logic, so I would appreciate your evaluation of the quote that follows. I've seen it many times, most recently on the Trivium Loop, and it bothers me every time I see it.
"Live simply, so others may simply live."
Practically speaking, in a free society this "simply" is not true. If I consume much it necessarily follows that others would have to work to provide me the goods I consume. Of course, in a free society, if others work for me, they will require me to compensate them. That compensation
raises the standard of living for those people who are at work to supply my consumption habits. In a controlled society, as in China, if I were not of the elite, I have no choice to live extravagantly. I would be glad to simply live, and my simple living would in no way benefit my
Copyright March 26, 2001, all rights reserved. 58726 views