PDA

View Full Version : gun control quote



nAz
03-22-2003, 06:13 PM
one of the best punch lines I've seen in a long time...

Marine Corps General Reinwald was interviewed on the radio the other day by a female interviewer concerning guns and children. Regardless of how you feel about gun control this is one of the best comeback lines of all time. It is a portion of a National Public Radio (NPR) interview between the female broadcaster and US Marine Corps General Reinwald as he was preparing to sponsor a Boy Scout Troop visiting his military installation.


FEMALE INTERVIEWER: So, General Reinwald, what things are you going to teach these young boys when they visit your base?

GENERAL REINWALD: We're going to teach them climbing, canoeing, archery, and shooting.

FEMALE INTERVIEWER: Shooting! That's a bit irresponsible, isn't it?

GENERAL REINWALD: I don't see why, they'll be properly supervised on the rifle range.

FEMALE INTERVIEWER: Don't you admit that this is a terribly dangerous activity to be teaching children?

GENERAL REINWALD: I don't see how. We will be teaching them proper rifle discipline before they even touch a firearm.

FEMALE INTERVIEWER: But you're equipping them to become violent killers.

GENERAL REINWALD: Well, you're equipped to be a prostitute, but you're not one, are you?

The radio went silent and the interview ended.

You gotta love the Marines! /ccboard/images/graemlins/smile.gif

eg8r
03-22-2003, 07:27 PM
HA HA HA. That is the funniest thing I have ever heard.

This is also another subject that I love to debate. The reporter sounded so ignorant when she said they were equipping the children to be violent killers. What a joke. I am sure she was sent in to try and prove a point maybe get in the last word, but that never happened. This is hilarious.

Following her same logic, maybe we should not let children aspire to be commercial airline pilots. That would be unreasonable, those children might become terrorists and fly them into a building.

eg8r

Hopster
03-22-2003, 07:47 PM
That ones been around for a while. Ive heard it before somewhere.
I wonder why they never mention that 40-50 years ago they used to teach marksmanship in schools ? Kids were taught responsible firearms training and you didnt have half the nonsense you have today.
Ahhh, let me not get started, i hate the anti gun people with a passion anyway.

nAz
03-22-2003, 08:25 PM
Rosie O'Donnell blames CRIME on guns....Thats like blaming FORKS for her being FAT


/ccboard/images/graemlins/smile.gif

Hopster
03-22-2003, 08:57 PM
Yeah, she sort of shut her hole when it came out she had armed guards for her and her kids.
Think Rosie can say the word : Hypocrite ??? /ccboard/images/graemlins/laugh.gif

eg8r
03-22-2003, 11:00 PM
[ QUOTE ]
Rosie O'Donnell blames CRIME on guns <hr /></blockquote> It is funny to know that her bodygaurds carry guns. What a hypocrite.

eg8r

cycopath
03-25-2003, 07:28 PM
<blockquote><font class="small">Quote Hopster:</font><hr>Think Rosie can say the word : Hypocrite ??? /ccboard/images/graemlins/laugh.gif <hr /></blockquote>Think Rosie can say the phrase: "Fatass ignorant dyke."?

I'm sorry, but moron gets me so wind up.

TomBrooklyn
03-26-2003, 03:20 AM
She lost a sweet job as K-Mart's spokesperson for whom she was doing a series of TV commercials that were getting a lot of network airtime. She probably didn't realize that most, though not all K-Marts sold firearms in it's stores. LoL.

Urban dwellers tend to have a much different view of firearms than more rural dwelling folk. In NYC, most people's primary thought-association with guns is criminal use. In other places it is hunting. In NYC, people think owning a gun is a big deal. In rural areas, having several firearms in the household is thought of as a routine aspect of life.

TomBrooklyn
03-26-2003, 04:55 AM
There is evidence that The Gun Debate pendulum may finally be feeling the effects of having swung too far towards the control side.

Congressmen Wilson of South Carolina has introduced a bill called the CITIZENS SELF DEFENCE ACT OF 2003, part of which I reprinted below. The whole bill is a little longer and can be seen here. (http://www.theorator.com/bills108/hr648.html)

========
108th CONGRESS, 1st Session H. R. 648
-----------------
A BILL
To protect the right to obtain firearms for security, and to use firearms in defense of self, family, or home, and to provide for the enforcement of such right.

SECTION 1. SHORT TITLE. This Act may be cited as the `Citizens' Self-Defense Act of 2003'.

SEC. 2. FINDINGS. The Congress finds the following:

(1) Police cannot protect, and are not legally liable for failing to protect, individual citizens, as evidenced by the following:

(A) The courts have consistently ruled that the police do not have an obligation to protect individuals, only the public in general. For example, in Warren v. District of Columbia Metropolitan Police Department, 444 A.2d 1 (D.C. App. 1981), the court stated: `[C]ourts have without exception concluded that when a municipality or other governmental entity undertakes to furnish police services, it assumes a duty only to the public at large and not to individual members of the community.'.

(B) Former Florida Attorney General Jim Smith told Florida legislators that police responded to only 200,000 of 700,000 calls for help to Dade County authorities.

(C) The United States Department of Justice found that, in 1989, there were 168,881 crimes of violence for which police had not responded within 1 hour.

(2) Citizens frequently must use firearms to defend themselves, as evidenced by the following:

(A) Every year, more than 2,400,000 people in the United States use a gun to defend themselves against criminals--or more than 6,500 people a day. This means that, each year, firearms are used 60 times more often to protect the lives of honest citizens than to take lives.

(B) Of the 2,400,000 self-defense cases, more than 192,000 are by women defending themselves against sexual abuse.

(C) Of the 2,400,000 times citizens use their guns to defend themselves every year, 92 percent merely brandish their gun or fire a warning shot to scare off their attackers. Less than 8 percent of the time, does a citizen kill or wound his or her attacker............

Wally_in_Cincy
03-26-2003, 07:25 AM
funny as hell, but unfortunately not true.

http://www.snopes.com/military/reinwald.htm

Wally~~likes guns, doesn't like Rosie O'Fat

Ross
03-26-2003, 11:16 AM
<blockquote><font class="small">Quote TomBrooklyn:</font><hr> (2) Citizens frequently must use firearms to defend themselves, as evidenced by the following:

(A) Every year, more than 2,400,000 people in the United States use a gun to defend themselves against criminals--or more than 6,500 people a day. This means that, each year, firearms are used 60 times more often to protect the lives of honest citizens than to take lives.

(B) Of the 2,400,000 self-defense cases, more than 192,000 are by women defending themselves against sexual abuse.

(C) Of the 2,400,000 times citizens use their guns to defend themselves every year, 92 percent merely brandish their gun or fire a warning shot to scare off their attackers. Less than 8 percent of the time, does a citizen kill or wound his or her attacker............ <hr /></blockquote>

Tom, these estimates have been shown to be grossly inaccurate but unfortunately are still widely quoted. I will explain why in detail below, in case you are interested. This is a statistical explanation of why we know the estimate is certainly grossly inflated. (Most of this explanation is borrowed from an article in "Chance Magazine," a journal for statistics teaching). There are other arguments that show that the quoted number leads to almost laughable contradictions of known facts that I won't go into here.

The 2.4 mil figure came from an extrapolation from a national random-digit-dial telephone survey of 5000 dwelling units. In this survey slightly over 1% of the individuals reported that they themselves had used a gun in self-defense in the past year. So the extrapolation was that 1.2% of 200 million adults is 2.4 million.

Of course random samples are used all of the time to estimate population trends and can be very accurate at doing so. For example, the Gallup polls etc are usually accurate at predicting presidential voting within 2 or 3 percent even though they are based on a sample of only approximately 1000 likely voters.

A problem arises however when random samples are used to estimate rare events. All surveys have problems with accuracy: people consistently misreport such things as whether they use their seat belt, whether they voted, etc. These inaccurate self-reports become a major problem however when estimating rare events because misclassifications are assymetrical.

Assume for example that a trait exists in .5% of the population. Out of 1000 people, 5 will have the trait, 995 will not. So there are 995 people in the survey who, through reporting inaccuracies, can be misclassified as false positives, while there are only 5 who can be misclassified as false negatives. Now lets assume that reports on the survey are 99% accurate (and this would be a minor miracle). Of the 995 without the trait, there would be about 10 false positives. There will likely be 0 false negatives since there are only 5 people who could do so and the reporting is 99% accurate. So add the 5 true positives to the 10 false positives and you get an estimate of 1.5% of the population overall having the trait. In other words, even if the accuracy of reports were 99%, the survey results would overestimate the true incidence by 3 times. In real world surveys, we don't get near 99% accurate self report so the overestimates are much larger.

There is another issue in estimating rare events. If something occurs 40% of the time, overesimating it by 2 or 3 percentage points is not a big problem. But say something occurs .1 % of the time. Then overestimating it by only 1 percentage point (getting 1.1% instead of 0.1%) results in a 10-fold overestimate! Again in real life, the overestimate is likely to be much larger that this.

In May, 1994 ABC news and the Washington Post did a random-digit-dial survey of 1500 adults. They were asked "Have you yourself ever seen anything that you believe was a spacecraft from another planet?" Ten percent said yes. These 150 individuals were then asked "Have you personally ever been in contact with aliens from another planet or not?" and 6% of those responded yes. Extrapolating to the US population as a whole should we conclude that 20 million Americans have seen alien spacecraft and 1.2 million Americans have been in actual contact with being from other planets?

The exact same flawed methodology was used in estimating the annual rate of gun use in self defense and the results are equally invalid.

So whichever side you take on the gun debate, the 2.4 million self-defense/year estimate should not be given any credibility.

eg8r
03-26-2003, 12:58 PM
[ QUOTE ]
A problem arises however when random samples are used to estimate rare events. <hr /></blockquote> Which event are you considering rare? The event happening that an innocent person protects himself with a gun, or a criminal has used a gun. These are different. Another issue is your definition of rare.
Statistics is a game that can be played to get the information you want to see. There are many variables that are often overlooked. In a lot of the surveys that are anti-gun biased the surveyor will only quote registered gun owners as the population for people who have defended themselves with a gun. This grossly underestimates the amount of people that have illegal possesion of a gun.

[ QUOTE ]
Assume for example that a trait exists in .5% of the population. Out of 1000 people, 5 will have the trait, 995 will not. So there are 995 people in the survey who, through reporting inaccuracies, can be misclassified as false positives, while there are only 5 who can be misclassified as false negatives. Now lets assume that reports on the survey are 99% accurate (and this would be a minor miracle). Of the 995 without the trait, there would be about 10 false positives. There will likely be 0 false negatives since there are only 5 people who could do so and the reporting is 99% accurate. So add the 5 true positives to the 10 false positives and you get an estimate of 1.5% of the population overall having the trait. In other words, even if the accuracy of reports were 99%, the survey results would overestimate the true incidence by 3 times. In real world surveys, we don't get near 99% accurate self report so the overestimates are much larger. <hr /></blockquote> This is all well and good, but a real true survey will offer some sort of sampling error. Did this magazine when referring to the example you quoted ever talk about sampling errors, especially measurement error. Meausrement error is one blinding variable that has been omitted from the study you quoted. There is one portion of measurment error quoted (Interviewer error) but that might be a small part of the whole.
There are never going to be accurate statistics on this issue because you will have hard time getting people to be honest when their actions were illegal (ex. defense using a gun in which they are in possesion of illegaly).

Another reason why these statistics will be off, is because a certain percentage of people will not report any problems if they came out on top and were not hurt. There are instances when someone is trying to break into a car and the owner of the car shows a gun and the burgular runs away. The owner might not report this happening.

[ QUOTE ]
So whichever side you take on the gun debate, the 2.4 million self-defense/year estimate should not be given any credibility. <hr /></blockquote> And neither should the person that says there is no credibility in the estimate.

eg8r

Hopster
03-26-2003, 02:42 PM
Well guys as an avid target shooter i try to keep up with most of the control issues but sometimes it gets ridiculous with all the figures that go back and forth.
But this much i have read numerous times: The areas that have allowed concealed carry have seen significant drops in street crimes such as muggings , rapes, violent street stuff,etc.
Another good figure is that most of the anti,s have said that the people who are issued concealed,s will on the whole act irresponsibly. The exact opposite has been true. The number of arrests of people with concealds is astnoishingly low.
I havent gottern around to getting mine yet but the old adage seems to be holding true. An armed society is a polite society. lol

Ross
03-29-2003, 01:47 AM
<blockquote><font class="small">Quote eg8r:</font><hr> Quote: Which event are you considering rare?
<hr /></blockquote>
Americans defending themselves with a gun in the past year.

[ QUOTE ]
The event happening that an innocent person protects himself with a gun, or a criminal has used a gun. These are different. <hr /></blockquote>

Yeah, I would say so...

[ QUOTE ]
Another issue is your definition of rare. <hr /></blockquote>

I used the term "rare" in reference to a low prevalence rate. I.e., the 1.2% prevalence rate under discusion.


[ QUOTE ]
Statistics is a game that can be played to get the information you want to see. <hr /></blockquote>
Exactly! I couldn't have said it better. That is why you should not accept statistics at face value without thinking them through. And that is the reason behind my post explaining that when random sampling is used to estimate events in the population that have a low prevalence rate, the estimates are almost always greatly inflated.

[ QUOTE ]
There are many variables that are often overlooked. In a lot of the surveys that are anti-gun biased the surveyor will only quote registered gun owners as the population for people who have defended themselves with a gun. This grossly underestimates the amount of people that have illegal possesion of a gun. <hr /></blockquote>

Sure there are other flawed surveys. How is that relevant to my post? I was speaking very specifically about the survey that is the source of the posted claim "Every year, more than 2,400,000 people in the United States use a gun to defend themselves against criminals--or more than 6,500 people a day." That conclusion is clearly an invalid interpretation of the survey results.

[ QUOTE ]
This is all well and good, but a real true survey will offer some sort of sampling error. Did this magazine when referring to the example you quoted ever talk about sampling errors, especially measurement error. Meausrement error is one blinding variable that has been omitted from the study you quoted. <hr /></blockquote>

I'm not sure what a "real, true" survey is, but if you are critisizing the survey then you are making my point.

And just to clarify, measurement error is NOT a type of sampling error. They are completely distinct issues. Measurement error is the result of incorrectly assigning a value to an individual. For example, you measure a person's height as 5' 8", when the person is actually 5'9". Or a person reports that he/she "Has used a gun to defend his/herself in the past 12 months" when the person actually has not. Or vice versa.

On the other hand, sampling error is the error due to randomly selecting subjects for your sample. You may, just by chance, get a higher or lower proportion of subjects with the trait in question than is true for the population at large. Sampling error is a random quantity with a probability distribution determined by the sample size.

Second, MY WHOLE DISCUSSION was about the problem of measurement error. In this case, measurement error consists of misclassifications that result in false positives or false negatives.

[ QUOTE ]
There is one portion of measurment error quoted (Interviewer error) but that might be a small part of the whole. There are never going to be accurate statistics on this issue because you will have hard time getting people to be honest when their actions were illegal (ex. defense using a gun in which they are in possesion of illegaly).
Another reason why these statistics will be off, is because a certain percentage of people will not report any problems if they came out on top and were not hurt. There are instances when someone is trying to break into a car and the owner of the car shows a gun and the burgular runs away. The owner might not report this happening. <hr /></blockquote>

Your examples show you completely missed the statistical point. Let me try again with a simple example. Assume your survey had 100 respondents. And assume for the moment, that indeed about 1.2% of all adult Americans use a gun to defend him or herself each year. That means you would expect 100 x .012 or about 1 person in your sample who meets this criteria. In this is the case, then, the GREATEST POSSIBLE NUMBER of false negatives will be 1.

Well that leaves an expected 99 Americans in the sample that did not use a gun for this purpose. Each of these 99 is a potential false positive. Do you see the asymmetry here? The only way the survey would fail to exaggerate the true incidence here would be if 100% of people who met the criteria lied (in this case the one person who it happened to lied) and no more that 1% of the 99 who didn't use a gun lied. That would yield 1 false negative and 1 false positive and therefore an accurate prevalence estimate.
Let's take more realistic figures. Assume that 95% of the people in both groups answer the question accurately. That would yield an expected false negative rate of 0 (.05 x 1, rounded) and a false + rate of 5 (.05 x 99, rounded). So if the population rate were indeed about 1%, you would expect the survey to yield an estimate of 6% (1 true positive + 5 false positives/100). So the estimate would overestimate the true incidence by a factor of 6!
<blockquote><font class="small">Quote RossNC:</font><hr>Quote: So whichever side you take on the gun debate, the 2.4 million self-defense/year estimate should not be given any credibility. <hr /></blockquote> And neither should the person that says there is no credibility in the estimate.

eg8r <hr /></blockquote>

I'm not sure I'm the one with the credibility problem. I was just pointing out that the prevalence estimate being cited was almost certainly a gross overestimate due to some well-established statistical issues with estimating rare events. I teach statistics at Duke University so I do have some expertise in this.

In fact this exact issue is at the heart of the debate about mammograms for women and prostate test for men. The vast majority of women who get mammograms don't have cancer. But even if only a small percentage of these are incorrectly diagnosed, the number of false positives may end up very large compared to the number of true positives. So you can end up with a lot of healthy women getting invasive follow up tests.

I could give you many more examples of this phenomenon. I just part with one coincidentally related to the discussion at hand: The NRA reports a 3 million dues paying members. That is about 1.5% of American adults. However national random phone surveys have consistently found that 4 - 10 % of Americans claim to be current active members. Those are sample overestimates ranging from 166% to 566%!

Finally, as to the issue of credibility, I have to say eg8r, you sometimes seem to be the one that cherry picks statistics that fit with your beliefs and seemingly take little notice of those that don't. For example, in a previous CCB post I read a surprising interpretation of crime statistics regarding England vs. the US and the rest of the world. Curious, I went to the original source to see if the interpretation was sound. I found, and posted, that for the one crime number that tends to be reported consistently enough to be compared across countries --murder -- England had a prevalence that was only 1/3 of the US. Your responsed to my post, but completely ignored that fairly striking contrast. I can't read your mind, but I would be willing to wager that if the stat had instead said the exact reverse -- that England had a murder rate 3 times that of the US -- you just might have paid a bit more attention to the stat, and maybe even quoted it as further evidence that gun control doesn't work.

I believe if we really want to be intellectually honest in our understanding of the world around us--- and you've posted statements which imply that you are in this camp ---then we have to be open to whatever we see, whether it jibes with what we believe to be true or not. Otherwise, there is no learning and no growth -- just self-righteous certainty that we are right!

Just my .02 worth!

Ken
03-29-2003, 07:16 AM
Your explanation about the possible errors that can enter a statistic are interesting. There is real relevance when considering something like the interpretation of an x-ray. I do feel, however, that when asked whether I used a gun to protect myself last year that I could get the answer right and so could 100 percent of the population. It's not that hard a question.

Looking at a current situation, one might go to Iraq and ask a sample of the soldiers whether they are male or female. The female is rare but I think everyone questioned could get that answer right and it would take a very inept questioner to make a mistake in recording the answers. I would predict that the sample could be done with 100% accuracy. You can give me a theoretical dissertation on how inaccuracies could occur but I would still consider the statistic that resulted to be pretty good.

I know it's typical for college professors to sit at their desk and figure everything out without feeling the need to leave their office. That is where inaccuracies occur and they don't even realize it. There's more to this than just theory.
KenCT

Ross
03-29-2003, 10:02 AM
<blockquote><font class="small">Quote Ken:</font><hr> Your explanation about the possible errors that can enter a statistic are interesting. There is real relevance when considering something like the interpretation of an x-ray. I do feel, however, that when asked whether I used a gun to protect myself last year that I could get the answer right and so could 100 percent of the population. It's not that hard a question.<hr /></blockquote>

Ken, you are correct - it isn't that hard a question. But neither is whether or not you are an active NRA member, or whether you have communicated with aliens. Both of these are significantly over-reported in telephone interviews. In fact it has been found that accuracy rates range from 85% to 98% for even simpler questions such as "Do you have a drivers licence?", "Do you own a car?", and "Do you own a home?"

Why the inaccuracy in self-report? There are a number of reasons. It has been repeatedly found that some people will give an answer that they think makes them look good in the eyes of the interviewer. Or the person may not intend to lie but he shades the truth in a way that makes him look better.

For example the respondent may perceive they he used a gun in self-defense while the other parties perceived it as an unprovoked escalation of a simple altercation. (How many times have you seen jerks in the pool hall that always see themselves as the victim!) Or the respondent may be pro-gun and therefore deliberately want to inflate figures that would support his stance. You might think the reverse is true, an anti-gun respondent could lie and distort the survey. But remember that for an anti-gun respondent to falsely deny the event he/she would have to be in the 1% for which the event actually occurred. Also, the anti-gun person would have to own a gun and to have used it as well!

Another issue is called telescoping. When you ask respondents if something has happened to them over a particular time period (as was done in the survey in question) there is a tendency to over-reporting by including events that occurred outside that time window.

Finally, the percentage of American adults with either Alzheimers, schizophrenia, or antisocial personality disorders is approximately 2 to 3%. Many of these people live at home and pick up the phone and respond to questions just like everyone else.

[ QUOTE ]
Looking at a current situation, one might go to Iraq and ask a sample of the soldiers whether they are male or female. The female is rare but I think everyone questioned could get that answer right and it would take a very inept questioner to make a mistake in recording the answers. I would predict that the sample could be done with 100% accuracy. You can give me a theoretical dissertation on how inaccuracies could occur but I would still consider the statistic that resulted to be pretty good. <hr /></blockquote>

You are probably correct there. But if they were doing a random phone interview (not asking the questions in person) of American's in general, who knows, some respondents might lie just to mess with the interviewer. /ccboard/images/graemlins/smirk.gif

[ QUOTE ]
I know it's typical for college professors to sit at their desk and figure everything out without feeling the need to leave their office. That is where inaccuracies occur and they don't even realize it. There's more to this than just theory.
KenCT <hr /></blockquote>

Whoa, Ken! Where did that come from? You don't know anything about me but you are stereotyping me based on the fact that I teach statistics at a university? /ccboard/images/graemlins/confused.gif Excuse me for thinking that having studied statistical issues makes me an appropriate person to comment on these issues.

At the beginning of your post, you said that the statistical points I made were interesting and you could see how they would apply to issues like x-rays. That being the case, the only way to defend the numbers in the gun-use case was to argue that you think responses to this question in random telephone surveys is probably 100% accurate. Don't you see how that could be seen as a stretch to defend a figure that happens to agree with your view on guns? Honest question, not being sarcastic.

eg8r
03-29-2003, 10:31 AM
[ QUOTE ]
I used the term "rare" in reference to a low prevalence rate. I.e., the 1.2% prevalence rate under discusion. <hr /></blockquote> I guess 1.2% is rare if you count all 100% of scenarios. What is that percentage across all countries. You might find that 1.2% is quite high compared to other countries. Like I said before Stats is a game you play to get a desired result. It is the same as in engineering, keep testing (Qual testing) until you get the right answer.
[ QUOTE ]
Sure there are other flawed surveys. How is that relevant to my post? <hr /></blockquote> It is relevant in the fact that it is a gun survey. There are a ton of gun surveys in which an anti-gun surveyor will gather just the information to prove his/her point. In my quote you will see I am referring to the attributes of the survey. This survey included plenty of the same attributes. One major attribute not included would the amount of times this sort of event happens that is not reported. This is very similar to rape cases, many go unreported.
[ QUOTE ]
And just to clarify, measurement error is NOT a type of sampling error. <hr /></blockquote> http://www.dssresearch.com/library/general/sampling.asp These people seem to think so. I also have SQE Statistical books that disagree with you also.
[ QUOTE ]
Your examples show you completely missed the statistical point. <hr /></blockquote> No it does not miss the statistical point. I have not stated whether I believe the survey or not. My example shows an attribute the survey did not take into consideration. You seem to be caught up with sample size. Did you go back and check to see if that is a statistically good sample to go with given the level/% of correct the interviewer was looking for. Just an example, when we DD250d missiles to the Government, we would have to run destructive tests on 4 out of 200 missiles sold. This is a statistically sound number of missile tested to give us a 99% accuracy level for the lot. I don't know the right answer to what I asked you, and for all it is worth, they might have chosen just a random amount of people to survey also.
[ QUOTE ]
I'm not sure what a "real, true" survey is <hr /></blockquote> I would think a real survey is an unbiased survey that looks at all the angles and allows the numbers to make the decision. Once this is formulated, then the sampling error should be listed to let the reader see what the chance of error is going to be for any point. A biased surveyor will only choose the attributes to prove their point. Once again, I am not agreeing with the survey or disagreeing with it.
[ QUOTE ]
I'm not sure I'm the one with the credibility problem. <hr /></blockquote> I am also sure that you would not be the one to state it if you did.
[ QUOTE ]
Finally, as to the issue of credibility, I have to say eg8r, you sometimes seem to be the one that cherry picks statistics that fit with your beliefs and seemingly take little notice of those that don't. For example, in a previous CCB post I read a surprising interpretation of crime statistics regarding England vs. the US and the rest of the world. Curious, I went to the original source to see if the interpretation was sound. I found, and posted, that for the one crime number that tends to be reported consistently enough to be compared across countries --murder -- England had a prevalence that was only 1/3 of the US. <hr /></blockquote> I may have missed that point, however I did address the fact that, first off, that site did not go far enough back in time to a point prior to the illegalization of private handguns. I also point out the fact that there are once again, criminal activities that happen and are never reported. I don't remember your reply to those.
[ QUOTE ]
but I would be willing to wager that if the stat had instead said the exact reverse -- that England had a murder rate 3 times that of the US -- you just might have paid a bit more attention to the stat, and maybe even quoted it as further evidence that gun control doesn't work. <hr /></blockquote> Well, I am not sure if this is correct or not. I still read these things with an open mind and try to see all the data that was used to develop the end result. My argument was not to prove that england has a higher or lower amount of gun crime than the US. My argument is that the gun crime rate when up quickly when the private ownership of guns was made illegal. You gave us your source for you argument and I asked a few questions about it. I don't remember if you bothered to respond. Why do they only go back to 1998 (that is as far back as I could see on that site). When I am comparing a change that has happened, I need data to represent what happened prior to the change as long as the data taken to represent what has happened after the change. I have done this sort of work for only about 4 years, at LM during missile component analysis, but I think that amount of time has helped me understand this idea fairly well.

I will not state I agree/disagree with all the graphs here, what I do like is that there is a broader time range given in the analysis. http://www.gunsandcrime.org/crvsgraf.html

eg8r

Ken
03-29-2003, 10:36 AM
Ross, Now I have to say that I agree with you. There are many reasons why I would not rely on those gun-use statistics. I mistrust the means of sampling since I don't think people will tell the truth when they cannot be anonymous especially about an act that might be illegal. Anyone who calls me can find out who I am.

I also believe that asking if someone used a gun for self-defense invites liars to try and inflate the numbers since it will be perceived that the person asking is in favor of concealed-carry laws and is trying to amass statistics to back that up. It's too obvious by the question. I would suggest that the survey must alternate questions with something like "did you ever pull a gun out to intimidate someone?". Now there is a negative connotation where the pro-gun people who actually used a gun in self-defense might be inclined to lie and say they didn't. The answer to both questions should be the same in most cases and follow up questions can determine if the threat of violence actually existed.

I would perefer a study that is totally anonymous and contains follow up questions eliciting details from the responder. Tell them to read all the questions first and they might be inclined to tell the truth if it seems too difficult to make up answers for all the next questions, based on a lie in the initial question.

I just threw that professor part in because I've seen so many professors who think all they need to do is think about something and they feel they can find the correct answer without actually getting any data. Of course, statisticians are not in that group since they need data to work with and will have their graduate students go out and get them.
KenCT

Wally_in_Cincy
03-29-2003, 10:45 AM
<blockquote><font class="small">Quote TomBrooklyn:</font><hr> (2) Citizens frequently must use firearms to defend themselves, as evidenced by the following:

(A) Every year, more than 2,400,000 people in the United States use a gun to defend themselves against criminals--or more than 6,500 people a day. This means that, each year, firearms are used 60 times more often to protect the lives of honest citizens than to take lives.

(B) Of the 2,400,000 self-defense cases, more than 192,000 are by women defending themselves against sexual abuse.

(C) Of the 2,400,000 times citizens use their guns to defend themselves every year, 92 percent merely brandish their gun or fire a warning shot to scare off their attackers. Less than 8 percent of the time, does a citizen kill or wound his or her attacker............ <hr /></blockquote>

Ross,

Even though I am a proud NRA member and an advocate of concealed carry, I have read these statistics many times and always thought they sounded way out of whack.

Without going into statistics, margin of error etc. , which I have never studied, I look at it this way. Do you personally know anyone who, in the last year or 5 years for that matter, who has defended themselves with a weapon? I don't and I know well more than a hundred people.

If I go back 10 or 20 years I can name a couple of people who I know personally who have had this happen. But from extrapolating my personal experience the number doesn't come close to 2,400,000.

I say if even 50,000 or 5000 people per year can defend themselves with a weapon, then the CCW laws are a good thing. But Dr. Lott should re-evaluate his statistics, for the sake of credibility.

heater451
03-29-2003, 12:29 PM
quote=Ross]. . .And just to clarify, measurement error is NOT a type of sampling error. <hr /></blockquote> <blockquote><font class="small">Quote eg8r:</font><hr>. . .[http://www.dssresearch.com/library/general/sampling.asp These people seem to think so. I also have SQE Statistical books that disagree with you also.<hr /></blockquote>That site seems to clearly define "Measurement Error" and "Sample Design Error" as separate errors--however, I think that what Ross refers to as "sampling error", the site calls "Random Error".


===============

Ross
03-29-2003, 01:18 PM
Ken, I give you a lot of credit for being open in this discussion.

You are correct that it is difficult to get accurate responses to questions about illegal activities or about issues that are politically or emotionally loaded. This is not as big a problem if the trait/activity is more common (like, say, 30%). So what if we miss by a few percentage points - we still get a pretty good picture of what is going on. However, when talking about things that happen 1% of the time, even small response biases or sampling error quickly overwhelms the small number we are trying to estimate.

Your suggestions for follow-up questions, consistency checks, wording questions to minimize response bias are all excellent ones. Many of these have been incorporated into the National Crime Victimization Survey. This survey is done by face-to-face interviewers going to the homes of the respondents. As much as possible the race of the interviewer is matched to the interviewee to help promote trust. Respondents are encourage to talk about their victimization. Questions are asked about events during the past 6 months instead of a year to reduce the problem of telescoping I mentioned before. The data collected include type of crime, month, time, and location of the crime, relationship between victim and offender, characteristics of the offender, self-protective actions taken by the victim during the incident and results of those actions, consequences of the victimization, type of property lost, whether the crime was reported to the police and reasons for reporting or not reporting, and offender use of weapons, drugs, and alcohol. Basic demographic information, such as age, race, gender, and income, is also collected. Proxy interviews are done for children if the parent prefers and for disabled household members and for those household members that are unavailable.

Of course this survey is not 100% accurate either. Under-reporting seems to be more of an issue than over-reporting in the face-to-face interviews about victimization, especially about sexual victimization. NCVS has continuously refined their methodology to reduce this underreporting though. And in recent years the improvements have yielded higher report rates, especially for sexual crimes.

And, yeah, some professors sit in their ivory towers and spin theories that have not met the test of the real world. I am the opposite - I doubt every theory and continually put them to the test of real world data. That is why it raised my hackles a bit when you put me in that camp. That and the fact that I grew up with guns on a farm in a small town in Texas, have hunted deer, geese, quail, etc. and own a gun currently. I didn't really want to hear how I had no real world experience! Thanks for retracting that Ken. /ccboard/images/graemlins/grin.gif

Here is an example from pool of my tendency to question pet theories. Almost every pro will tell you the break is a great advantage. And they can give very convincing arguments and anecdotes to show this is so. However, the few times I've seen where people have calculated the win percentage of pros when breaking vs. not breaking, they have consistently found a slightly higher win percentage when not breaking! But since it doesn't "feel right" no one is really willing to take this empirical finding seriously. Go figure.

eg8r
03-29-2003, 10:59 PM
<blockquote><font class="small">Quote heater451:</font><hr> That site seems to clearly define "Measurement Error" and "Sample Design Error" as separate errors--however, I think that what Ross refers to as "sampling error", the site calls "Random Error". <hr /></blockquote> I know this means nothing, but go back to the site. You will see the main heading on the top right. Then the subheadings fill the page. Sample design error and measurement error are all types of sampling error.

<blockquote><font class="small">Quote wally:</font><hr> If I go back 10 or 20 years I can name a couple of people who I know personally who have had this happen. But from extrapolating my personal experience the number doesn't come close to 2,400,000.
<hr /></blockquote> This is exactly why I say statistics is a game people play to prove their agenda. The numbers are off by a large amount, but there are others things that are missing also, ex. Instances when a victim defends themselves with a gun but never reports the instance to the police. Another good example is if the instance is reported and shown as a burglary, but the burglar gets away, and no mention of the gun.

eg8r

eg8r
03-29-2003, 11:09 PM
[ QUOTE ]
I just threw that professor part in because I've seen so many professors who think all they need to do is think about something and they feel they can find the correct answer without actually getting any data. Of course, statisticians are not in that group since they need data to work with and will have their graduate students go out and get them. <hr /></blockquote> Present company not included, this does happen often. Staticians are not exempt by no means, as they can manipulate the data to prove their point. There are professors out there working their butts off to extend the research and really care about what the results are, and there are also professors out there with an agenda and are only looking for the information to prove or solidify their agenda. This is true for all professions.

When I was doing statistical analysis/SPC for LM in Orlando, I was approached many times by engineers who were very sure of themselves. You know this when the sentences starts off with, "I need you to prove our processes are good, or I need you to prove the supplier is not meeting qual standards, etc." Most times the engineer was trying to cover his butt, other times they just wanted to know the truth so if need be, they can schedule a change board and fix the problem. As the QE, I requested the data that I needed and let the numbers speak for themselves. For ex, we had an issue with a paint process. The paint was bubbling during bake, and the paint guys were blaming this on silicone in the air and poor filtering of the air. Well, when we were done the analysis, it turns out the level of the silicone in the air was not the problem, and well within the tolerances of the application, so something else was the problem. Sure did not make the mfg group happy, they really wanted to spend the money to get the new state of art ventilation system.

eg8r

Ross
03-30-2003, 09:42 AM
<blockquote><font class="small">Quote Wally_in_Cincy:</font><hr> <blockquote><font class="small">Quote TomBrooklyn:</font><hr> (2) Citizens frequently must use firearms to defend themselves, as evidenced by the following:

... Of the 2,400,000 times citizens use their guns to defend themselves every year, 92 percent merely brandish their gun or fire a warning shot to scare off their attackers. Less than 8 percent of the time, does a citizen kill or wound his or her attacker............ <hr /></blockquote>

Ross,

Even though I am a proud NRA member and an advocate of concealed carry, I have read these statistics many times and always thought they sounded way out of whack.

Without going into statistics, margin of error etc. , which I have never studied, I look at it this way. Do you personally know anyone who, in the last year or 5 years for that matter, who has defended themselves with a weapon? I don't and I know well more than a hundred people.

If I go back 10 or 20 years I can name a couple of people who I know personally who have had this happen. But from extrapolating my personal experience the number doesn't come close to 2,400,000.

I say if even 50,000 or 5000 people per year can defend themselves with a weapon, then the CCW laws are a good thing. But Dr. Lott should re-evaluate his statistics, for the sake of credibility.

<hr /></blockquote>

Like you Wally, I also can't think of more than a couple of incidents during my lifetime where I have heard of a friend or an acquaintance or a friend of a friend using a gun in self defense. I go to a pool hall daily and there are a lot of stories bandied aboutinvolving violence and mayhem, but that is not a story I hear about.

Contrast this to the situation with fatal auto accidents. The 2,400,000 number being quoted for annual gun self-defense is almost 50 times higher than fatal automobile accidents (50,000/year). Yet most of us have personally known a handful of people who were killed in auto accidents. And I hear stories about such accidents on a fairly regular basis. (In fact, unfortunately, only last week a one-pocket player who came through my local ph a few times was killed in such an accident.) Are we to believe that for every story of a fatal car crash we hear, there are 50 gun self defense incidents that no one bothers to mention?

Based on the same survey these authors also claim that 8% of the time the person shoots and injures or kills the attacker. Well 8% of 2,400,000 is 192,000. But mortality, hospital, and emergency room data show that the TOTAL number of non-fatal gunshot injuries treated each year is about 100,000 and 30-35,000 people are killed each year with guns. And almost all of these are assualts, suicides, and accidental shootings (just read your daily newspaper for details). So where are the 192,000 shot offenders being treated and buried?! I guess all of the self-defenders are bad shots and just giving their attackers self-treatable flesh wounds? /ccboard/images/graemlins/smirk.gif

Wally, I don't know why the number continues to be quoted when it is so patently obvious a gross exaggeration. It does reduce the credibility of those that keep quoting it. Not so much for the lay-person since number checking is not his/her job, although we should all think a bit more about the numbers we read. But media and professional organizations do (or should) have the expertise and they certainly have the responsibility.

03-30-2003, 10:22 AM
or hitting your intended target /ccboard/images/graemlins/wink.gif