US STATES WITH MORE GUN OWNERS HAVE MORE MURDERS

Reuters ^ | 12/04/02 | Reuters - Charnicia E Huggins

[Pitchfork]

I'll bite on at least some of your screed:

Your harangue begins with a shot at the researcher’s use of a "presumed rate of firearms ownership, using unreliable proxies."

I read the article to which you have linked. How 'unreliable' or invalid might these proxies be? Let's look shall we?

First, it should be recognized that no "Direct measures of household firearm ownership are available". Quite charitably, the researchers did not make the point that this dearth of data (a truly sad oversight given the importance of this public policy question) was largely due to NRA resistance to documentation of weapons purchases. In the absence of this kind of data, researchers must make a choice to either abandon the project or find other indicators. Obviously to yours and the NRA’s disappointment these authors chose the latter. When creating operational measures when data is poor there are accepted standards for establishing measurement validity and reliability. These researchers appear to have followed the ‘rules’ to the letter—perhaps anticipating criticisms like yours.

Laudably, they try to employ data from the GSS (the premier national demographic and attitudinal survey with a very sophisticated sampling methodology that is first rate). Gun ownership levels in this survey are based on a self-report. Because respondents may under-report guns owned illegally or may be hesitant to report firearm ownership. The researchers compare the GSS measure with BRFFS data. This is an excellent example of using two measures to test validity and reliability.

Comparing distinct measures as means of assessing the validity of their 'proxies' is standard practice. While I would have liked to have seen some correlation coefficient for the GSS, BRFSS and their proxy figures to go along with their claim the three are "very highly correlated" (I would hope this was at least .8 or higher) I presume that this information was provided to the journal reviewers and likely edited for space this is also not an uncommon practice. I have had statistical tables edited out or reduced in article I have published in the interest of readability. I am sure the authors would provide these correlation data on request. However, it is hardly needed because in their regression analyses on a state by state basis they present models using all of them! You can eyeball them and get some sense of the high correlation between them. It seems high. If three different measures of firearms ownership, collected by three different sets of researchers using three different methods ALL point in roughly the same direction, that’s a pretty GOOD proxy! Will there be outliers? Is there measurement error? Of course! But the pooled cross-sectional time series method they use is pretty robust I’m sure they ran the typical diagnostic measures.

However, they are presented with some problems of ecological inference. Although the GSS is a national random sample it is not designed to be such on a state-by-state basis and so they cannot simply extrapolate from the surveys to other households in the state. However it appears that the GSS IS a random sample within US Census regions and so may be used as a regional indicator of firearms ownership. The study has 'lumped together states' into these regional blocks not because of some sinister purpose but because of data limitations.

Is this measure perfect? No. I would much rather have government collected data on firearms ownership levels from a national registry but I'm sure you would be opposed to such data collection. Given the circumstances the proxies they use are as good a measure as one could devise.

Interestingly of the two states you mention Wyoming and West Virginia, according to the data presented in the article you linked to WV firearms related death rate is about average while Wyoming’s (at over 3.05) is actually on the high side? So what was your point again? They weren’t looking at the murder rate in this linked article. I’ll have to read the murder rate one. Seems like there are few too many ‘hunting accidents’.

AS for SoDak's reply and the rest of the ‘data’ presented in the thread DON't make me laugh! After excoriating the Harvard researchers for unreliable measures you offer anecdotal evidence to support your claim! I'm crying I'm laughing so hard.


Some other points:

“You should also examine the FBI data on the jurisdictions where the murder rates are highest--every one of them is a major city where legal rates of firearms ownership are low. Could it be that there is something other than "easy availability" at work here?”

Obviously ‘urbanization’ is one of the multiple causes of high murder rates and should be controlled for but the fact of multiple causation does not render a relationship between availability and murder rate spurious. Perhaps urbanization is a very powerful indicator while availability is less so. Interestingly, your inclusion of the term ‘legal’ in your firearms ownership rates quote seems to indicate that you, like I have suspicions that the real ownership rates might be higher. Why not admit it! Because it would counter your argument? It appears that Cook’s index (used in the articles you supplied) was designed to estimate this and as a result probably is less accurate for rural settings. Some of this discrepancy can be seen in the FS/S indicator scores when compared to the Cook’s index. The latter seem low in rural areas compared to the former. This COULD have been controlled for through indexing or weighting but I imagine had they done this you would have called this unfair manipulation.

As for me being a ‘lefty’ well I’m an old fashioned conservative I suppose, one who is ‘data’ driven. I’ve been at odds with ‘idealist’ lefties who seem fixated on validating prescriptions of how the world “should be” and “how” we should get there for a while. I tended to associate myself with conservative ‘realists’ who seemed to respect empirical research and to instead describing “what is” and “why”. Sadly the conservative seem as bad as some lefties on the ‘should be’ side now.

I ask you this. Is there any data that would convince you that firearms were dangerous to public health or increased the murder rate? If the right study were presented would you change your mind and become a gun control advocate? I’ve read Lott’s book. It was obviously quite well done and some your critiques sound disturbingly like those leveled against Lott (on issues of aggregation for instance). The left is obviously determined not to be convinced by data. I’m becoming equally convinced of this close-mindedness on the right. You need look no further than the evolution-creation debate for evidence of this. Has gun ownership become something of a religious belief that brooks no criticism?

I would have run a few of the analyses on the data that I mentioned above (correlation coefficients between the cook, BRFFS and FS/S data) or looked at whether urbanization might be responsible for the discrepancy between Cook and FS/S but I’m certain that anything I found that you disagreed with would be summarily discounted.

Pitchfork, Ph.D.

[Pitchfork]

Two further points.

First: I visited your site. Very interesting. Sadly you don't provide information regarding who your major donors are? Any firearms manufacturers?

Second:
How do you feel about Saddam and the war on Iraq? Are you buying the adminstration's argument that saddam possessing a nuclear weapon is a threat to the US? Do you support his actions to prevent North Korean nuclear development? Did you support or oppose funds given the post-Soviet successor states to divest themselves of missles?

It seem that the US policy regarding nuclear proliferation is directly at odds with the logic of the gun lobby here at home. Internationally, we argue that the availablility of weapons increases the liklihood of use and is therefore a security threat. We've encouraged India and Pakistan to disarm and lauded countries like Argentina and Brazil who gave up thier programs. We have entered into agreements like the NPT to prevent the spread of nuclear weapons arguing that availablilty and use are causally linked! But of course, that logic couldn't apply to firearms could it?

[dirtboy]

If they wanted to skew the numbers, they would have LEFT OUT the low murder rate states like WV and WY.

Hardly. If you look at my analysis of the six high-gun ownership states, it is clear that some other factor is at play - WV and WY are FAR lower that the other four states. So by combining those two low-population states with the other four, the researchers were attempting to cover up those two inconvenient low murder rate high gun ownership rate states. Whenever I see discernable data points lumped together, nine times out of ten it means the researcher is trying to make some fudge.

One other point: I thought that gun ownership made a place safer. Isn't that the claim of the NRA etc....Even if the Harvard study doesn't prove that guns are more dangerous, it doesn't support the case that high gun ownership makes a community safer.

The Harvard study doesn't prove ANYTHING - because it clearly is not an honest study to begin with. Lott's initial study, by contrast, contradicted his initial views regarding guns - but since Lott is honest, he published the actual results rather than fudging the results to fit his preconceptions.

[Teacher317]

I don't see any BS in the article at all. Just because you don't want to believe something, don't cover your eyes. AAMOF, the last sentence could very well be the conclusion to be drawn from the article.

Did you miss the title somehow?

[meyer]

Obviously ‘urbanization’ is one of the multiple causes of high murder rates and should be controlled for but the fact of multiple causation does not render a relationship between availability and murder rate spurious.

Urbanization is a "cause"?

[challenger]

Please excuse me, but due to scheduling limitations, I can't spend a lot of time on my response to your message. Let me try to hit some of the more useful points.

[NOTE]
One of the biggest defects with a "gun death" study is the fact that it fails to address the different types of deaths as having different causes and different solutions. For instance, suicide with firearms should not really be included in a study like this, unless there is substantial evidence that the suicide happened "because of" the presence of the firearm. This is not likely to be true in many instances, since there are essentially an infinite number of means to commit suicide. Many countries with total firearms prohibitions have very high suicide rates, including "no guns since 1588" Japan. As you may know, suicide is often the largest single component of firearms deaths, even though the causation of suicide is really almost independent of the firearm. Some studies have indicated higher suicide rates in rural/small-town settings (possibly due to lack of extensive mental health facilities or isolation from loved ones), so keep those factors in mind when evaluating suicides. The more urbanized a state, the lower the suicide rate we would expect.

[Pitchfork wrote]
First, it should be recognized that no "Direct measures of household firearm ownership are available".

[My response]
Thank you for admitting the same point that I made--the measurement of firearms ownership is indirect and not necessarily very reliable. If you know many people from places like South Dakota and Iowa, you know that firearms availability is practically universal, and purchase is fast and easy for those who don't yet own a firearm. There is nothing to stand in the way of a homicide, suicide, or accident. When one must depend upon proxies that don't pass the most basic real-world tests, the study is skating on thin ice from the start.

[Pitchfork]
Quite charitably, the researchers did not make the point that this dearth of data (a truly sad oversight given the importance of this public policy question) was largely due to NRA resistance to documentation of weapons purchases.

[My response]
Despite your claims, my posts were not harangues, since I addressed the factual and methodological issues. Here, you are attacking the NRA and me on a subject that is not even important to the Harvard study. You simply wanted to denigrate the opposition. It is an incontrovertible fact that the records do not exist, so it is a real problem when the chosen proxies are inadequate, based upon real-world experience. The non-existence of the records does not give anyone a green light to invent proxies that are obviously erratic, just because they can invent or take up several proxies that are internally consistent. Precision and accuracy are not the same thing.

[Pitchfork]
When creating operational measures when data is poor there are accepted standards for establishing measurement validity and reliability. These researchers appear to have followed the ‘rules’ to the letter—perhaps anticipating criticisms like yours.

[My response]
No matter what "rules" one follows, there is no point in conducting the study using proxies that have such anomalous results. Perhaps if the researchers had a little better real world knowledge of the subject matter, they could have searched for or developed proxies that were more accurate and consistent.

[Pitchfork]
Gun ownership levels in this survey are based on a self-report.

[My response]
This is the heart of the problem. Firearms owners often refuse to report that they own firearms, because many of them are aware of the proven instances where registration has led to confiscation. I know thousands of people just on my limited mailing list who would never tell an interviewer that they own a single firearm, much less how many they own. Even people who have "on-the-books" firearms will not reveal this to telephone, mail, or government interviewers. You will never get accurate information, no matter how "sophisticated" the sampling methodology or how "highly correlated" or "standard practice" your measures are.

[Pitchfork]
You can eyeball them and get some sense of the high correlation between them. It seems high. If three different measures of firearms ownership, collected by three different sets of researchers using three different methods ALL point in roughly the same direction, that’s a pretty GOOD proxy!

[My response]
That is precision, not accuracy.

[Pitchfork]
The study has 'lumped together states' into these regional blocks not because of some sinister purpose but because of data limitations.

[My response]
Again, you engage in ad hominem attacks, indicating you don't have a substantive point. The point is not whether there was any evil intent, nor did the other serious commentators on this thread make that as their main point. We said that lumping the states together fails to tell the real story, since the lumping process fails to explain the major differences between the individual states that they listed. It implies a connection that is not there, especially if you factor out the suicide rate that should not even be included for the reasons described above.

[Pitchfork]
Interestingly of the two states you mention Wyoming and West Virginia, according to the data presented in the article you linked to WV firearms related death rate is about average while Wyoming’s (at over 3.05) is actually on the high side? So what was your point again? They weren’t looking at the murder rate in this linked article. I’ll have to read the murder rate one. Seems like there are few too many ‘hunting
accidents’.

[My response]
There is an important point here. You see, Pitchfork, this is exactly the type of deceptive statistic that can fool even someone such as yourself who has some experience reading these types of reports. Remember what I wrote above regarding the necessity of separating out the different mechanisms of death--homicide, suicide, and accidents? Wyoming appears to have a high total rate, until you break it into the separate components. Your little dig about too many "hungting accidents" simply isn't true. Examining a government source that records all types of deaths for the year 2000, Wyoming had 3 "hunting accidents" (or other accidents), 5 homicides, and 50 suicides via firearm. Notice that the vast majority (~86%) of the deaths are suicides, which would probably occur no matter how many guns are confiscated or banned. Since suicides are not really "gun deaths" (in that they are not actually caused by guns or highly preventable by gun prohibitions), the Harvard rating of 3.05 should undoubtedly be much lower. The homicide rate for Wyoming is very low and so is its accident rate, so I'm sure you can now see why I was emphasizing the murder rate rather than the total Harvard figure. The same is true of West Virginia to a lesser extent.

[Pitchfork]
AS for SoDak's reply and the rest of the ‘data’ presented in the thread DON't make me laugh! After excoriating the Harvard researchers for unreliable measures you offer anecdotal evidence to support your claim! I'm crying I'm laughing so hard.

[My response]
Don't laugh too hard at us, UNLESS you can produce more real-world anecdotes that counter what we have said. I'd be willing to bet that you can't find a single resident or ex-resident of South Dakota who would disagree with our point that firearms are almost universally available there. If you find such a person, I'd like to talk to them. Much of the data presented in this thread has been done on the spur of the moment, without any funding, unlike the Harvard study, which undoubtedly was paid. Isn't it interesting that even such quick research can find better correlations than those found by the Harvard School of Public Health? I wonder if we are getting our money's worth from the taxpayer funding? I'm not laughing.

[Pitchfork]
Obviously ‘urbanization’ is one of the multiple causes of high murder rates and should be controlled for but the fact of multiple causation does not render a relationship between availability and murder rate spurious. Perhaps urbanization is a very powerful indicator while availability is less so.

[My response]
It is interesting that urbanization correlates with high crime rates when there are restrictive, anti-self-defense gun control laws, but it often has the reverse effect in urban jurisdictions where "shall issue" CCW/Self-Defense-Permit laws are in effect. In other words, "availability" is inversely related to crime rates.

[Pitchfork]
Interestingly, your inclusion of the term ‘legal’ in your firearms ownership rates quote seems to indicate that you, like I have suspicions that the real ownership rates might be higher. Why not admit it! Because it would counter your argument?

[My response]
Actually, I only included that term to be scrupulously correct in designating the non-violent ownership of firearms. I had no qualms that a total figure for legal and illegal ownership in cities would be especially high. Let us assume, arguendo, that you are correct in thinking that real total ownership rates in restrictive cities might be substantially higher if we include those firearms owned by criminals. What is the expected order of magnitude in a city where 25% of the law-abiding population owns guns? Given a worst-case fantasy scenario, wherein 20% of the population of a major city is composed of felons, and assuming 100% of them owned guns, what is the total ownership rate? You can see the problem, can't you? It is stil only a 45% ownership rate, which only approximates the national average. It is nowhere near the ownership rates in the more wide-open states like South Dakota or even Washington, so it doesn't explain the death rates. I am sure you will agree that I am giving you the benefit of the doubt by using a figure of 20% felons. And as you can see, your point does not counter my argument at all.

[Pitchfork]
Sadly the conservative seem as bad as some lefties on the ‘should be’ side now.

[My response]
Your political paradigm seems a bit confused, but it is obvious to me that the posters on this thread have been very much evidence oriented, rather than engaging in speculation and "should be". Except of course, for those who are just being sarcastic or funny.

[Pitchfork]
I ask you this. Is there any data that would convince you that firearms were dangerous to public health or increased the murder rate? If the right study were presented would you change your mind and become a gun control advocate?

[My response]
Those are two separate questions, so I will answer them separately. Over the last two decades, I have read dozens of books, hundreds of studies, thousands of concepts and facts, and gigabytes of information--and my running scoreboard indicates that self-defense is leading by an 80-20 ratio. Therefore, it would take a massive amount of new evidence to convince me that gun control has any positive cost-to-benefit ratio. As you may know, the U.S. is approximately in the middle of the U.N. listings for murder and suicide, so we are not uniquely violent. But evidence within the U.S. shows me that jurisdictions with easy availability and even CCW/Self-Defense-Permits exhibit greater public safety. A grossly disproportionate amount of violent crime occurs in the major cities where it is impossible or difficult to defend ourselves and our loved ones. I have visited many parts of the U.S. and I have rarely seen a jurisdiction with easy availability that is wracked by crime. Having said that, I will also say I am a strong Constitutionalist (and yes, I have studied the 2nd Amendment and the law extensively), and I would need to see extremely compelling evidence that gun control would greatly reduce the violent crime rate before I would ever consider repealing the 2nd Amendment. Consider, but even then, possibly not. But I doubt that such massive new evidence will ever arrive.

[Pitchfork]
...some your critiques sound disturbingly like those leveled against Lott (on issues of aggregation for instance).

[My response]
My critique is not parallel to Nagin, Black or Donahue regarding Lott. Several people, including Lott have addressed the first two critics, and Donahue's time will come. Clearly, Lott took all 3,054 counties in the U.S. and examined them separately, which Miller, et al, did not, especially in Miller's conclusion which conflated several jurisdictions that didn't even exhibit an internal pattern of common characteristics.

[Pitchfork]
Has gun ownership become something of a religious belief that brooks no criticism?

[My response]
I, like many people, will not willingly give up my right to defend myself and my loved ones based on the incompetent and deceptive "criticism" that passes for research today. I have debated many anti-self-defense/gun-control advocates and I have always found that they are the ones who fall apart under criticism. Why? Because they have a very weak case and it is relatively easy to tear it down.

[Pitchfork]
...that anything I found that you disagreed with would be
summarily discounted.

[My response]
I don't summarily dismiss anything, if only because it is so much fun tearing it apart. No matter how clever the statistical manipulations of a study, it is worthless if it is illogical on its face or if it does not conform to the real world. Fortunately, I have a pretty extensive background and I know where to find information necessary to analyze the other side's opinions.

Yeah, I guess I didn't keep it short. Oh well, I just get wrapped up in these things. Now my day is off to a bad start.

[challenger]

Good grief! Another post I must answer before I start my day.

[Pitchfork]
First: I visited your site. Very interesting. Sadly you don't provide information regarding who your major donors are? Any firearms manufacturers?

[My response]
Wow! I'm getting the Prof. Lott celebrity treatment! Ad hominem attacks on my association with the firearms industry! First, thanks for the compliment on my interesting web site--I hope everyone who is reading this goes there and joins. Who would have thought that people would be attacking my organization for its speculative connections to the evil firearms industry? Now that's funny! The operating expenses of CSC come solely from individual memberships. CSC has yet to receive its first donation from any firearm or ammunition or accessory manufacturer--but if any of their reps read this and want to help, we will be happy to discuss this issue.

[Pitchfork]
Second: How do you feel about Saddam and the war on Iraq? Are you buying the adminstration's argument that saddam possessing a nuclear weapon is a threat to the US? Do you support his actions to prevent North Korean nuclear development? Did you support or oppose funds given the post-Soviet successor states to divest themselves of missles?

[My response]
This thread is not the proper place to discuss the issues that Pitchfork raises. If he ever sees me in a Military History forum, we'll discuss it.

[Pitchfork]
It seem that the US policy regarding nuclear proliferation is directly at odds with the logic of the gun lobby here at home. Internationally, we argue that the availablility of weapons increases the liklihood of use and is therefore a security threat. We've encouraged India and Pakistan to disarm and lauded countries like Argentina and Brazil who gave up thier programs. We have entered into agreements like the NPT to prevent the spread of nuclear weapons arguing that availablilty and use are causally linked! But of course, that logic couldn't apply to firearms could it?

[My response]
Again, not related to the topic of this thread, so I will not discuss it in depth. However, I will leave you with this political science paradigm: citizens are the sovereign power in a representative republic and they have innate, natural rights until they have individually proven that they are dangerous criminals--and even then they retain many rights. Iraq, North Korea, Pakistan, India, Argentina, and Brazil are countries that have already proven that they are unstable and in some cases, criminal. Disarming those countries has nothing in common with disarming decent Americans.

_________________________

Randall N. Herrst, J.D.
President
The Center For The Study Of Crime
http://www.studycrime.org

[gcruse]

Are you saying the headline is BS? Did you read the last line of the article? The numbers don't lie and the headline restates the numbers. The interpretation of the numbers is debatable and the last sentence acknowledges that. Just because we don't want to believe something, declaring it BS and moving on is the tactic of the blowhard.

[challenger]

Hi

You might find some interesting self-defense and public safety information in my post 159.

[meyer]

The numbers don't lie and the headline restates the numbers.

In this case, you may wish to research that assertion more deeply. Guesses and even educated guesses at the number of gun-owners within a given jurisdiction are not reliable.

[baseballfanjm]

Actually, the difference in population between those four states and six states is only about 280,000, with the six, "low gun ownership" states with more people. I'm not defending the article at all. It's just why they used a different number of states. As you said, the fact that they don't use demographics and per-captia evalutions renders it meaningless IMO. I hate gun grabbers as much as any Freeper.

[Republic of Texas]

Why don't we compare the populations of those states first. Just to be fair.

[dirtboy]

Just because we don't want to believe something, declaring it BS and moving on is the tactic of the blowhard.

Nah, we did some research and THEN declared it BS. Of the six states in the study with the highest gun ownership rate, four had high murder rates and two had very low rates. But rather than explore that discrepency in their thesis, the researchers attempted to bury it by lumping all six states together - and since WV and WY have very small populations, it did not have a significant effect on the sum of the six states. That's a classic way to fudge a study.

[Atlas Sneezed]

The actual study is now posted here:

http://www.freerepublic.com/focus/news/802586/posts?page=1

[PistolPaknMama]

So, in other words, there is definitely SOME OTHER FACTOR at play here.

Of course it is. This research proves, beyond a shadow of a doubt, that the mere presence of guns can cause death by any means. There it is in black and white. It must be true. /sarcasm

[PistolPaknMama]

I checked the US census website. I saw NO references to data about firearms ownership.

Everyone who had a firearms-related question on their census questionnaire and answered such a question, raise their hands.

(There wasn't any such question on mine and I would have answered "nunya" like I did on all the other questions except as to number in household, as provided by the Constitution.)

[AlexW]

My 20 year old daughter, just a few days ago, was telling me she had seen the Columbine movie, and was reciting all the gun and murder statistics that were cited in the movie.

I mentioned that the statistics were probably out of context, but I had no facts to refute what she said...I need help.

[FITZ]

More likely, states with a higher threat of being murdered have caused gun ownership to increase.

That's so obvious --I would have a gun but I live in an area that I believe is fairly safe and have curious children so I think for myself at this time, a gun isn't needed, I'd get one if murder and violent crime statistics in this area change.

[challenger]

Hi

It sounds like your daughter wants to believe the movie stats, since she absorbed them so quickly. If she or you would like actual data, in context, feel free to contact me. Tell me which "factoids" you have heard, and I'll try to show you where to find the real facts.

By the way, it would be fair to say that everything in the movie is either factually incorrect, emotionally manipulative, or out of context. It is not a documentary--it is a political polemic.

________________________________

Randall N. Herrst, J.D.
President
The Center For The Study Of Crime
JOIN NOW! The Premier Resource for Innovative Activists!
www.studycrime.org

[Texas Eagle]

The study findings imply "that guns, on balance, lethally imperil rather than protect Americans," lead study author Dr. Matthew Miller of Harvard School of Public Health in Boston, Massachusetts, told Reuters Health.

If the quantity of guns at hand leads to crimes committed with guns, then how come you never hear of a mass murder at a gun show or NRA convention?