States With Higher Levels of Gun Ownership Have Higher Homicide Rates

[Hunter]

This came across the "wire" from Jeff Chan's Firearm Alert emails. See RKBA.ORG and http://groups.yahoo.com/group/firearms-alert/ for more details.

************************************************** ****************************
Press release, in whole. Not surprisingly the research is
sponsored by the anti-rights Joyce Foundation:

http://www.hsph.harvard.edu/press/releases/press01112007c.html

States With Higher Levels of Gun Ownership Have Higher Homicide Rates
For immediate release: Thursday, January 11, 2007

Boston, MA -- Firearms are used to kill two out of every three
homicide victims in America.. In the first nationally
representative study to examine the relationship between survey
measures of household firearm ownership and state level rates of
homicide, researchers at the Harvard Injury Control Research
Center found that homicide rates among children, and among women
and men of all ages, are higher in states where more households
have guns. The study appears in the February 2007 issue of Social
Science and Medicine.

Matthew Miller, Assistant Professor of Health Policy and Injury
Prevention at Harvard School of Public Health, and his colleagues
David Hemenway and Deborah Azrael, used survey data from the
Center for Disease Control and Prevention’s 2001 Behavioral Risk
Factor Surveillance System, the world’s largest telephone survey
with over 200,000 respondents nationwide. Respondents in all 50
states were asked whether any firearms were kept in or around
their home. The survey found that approximately one in three
American households reported firearm ownership.

Analyses that controlled for several measures of resource
deprivation, urbanization, aggravated assault, robbery,
unemployment, and alcohol consumption found that states with
higher rates of household firearm ownership had significantly
higher homicide victimization rates for children, and for women
and men. In these analyses, states within the highest quartile of
firearm prevalence had firearm homicide rates 114% higher than
states within the lowest quartile of firearm prevalence. Overall
homicide rates were 60% higher. The association between firearm
prevalence and homicide was driven by gun-related homicide rates;
non-gun-related homicide rates were not significantly associated
with rates of firearm ownership.

These results suggest that it is easier for potential homicide
perpetrators to obtain a gun in states where guns are more
prevalent. “Our findings suggest that in the United States,
household firearms may be an important source of guns used to
kill children, women and men, both on the street and in their
homes,” said Miller.

[dondo]

Not a scientific rebuttal but it seems to me states with higher homicide rates are probably those that are more densley populated. States with cities like L.A. and Washington D.C. The more densley populated the more homicides, the more poor areas in which NORMAL LAW ABIDING CITIZENS will have defensive LEGALLY OWNED guns which will usually never see the light of day. Oh the spin. A real study nees to be done on the real circumstances of gun homicides. I.e. gang violence, murders with illegally possessed guns, in comparison to murders or deaths involving legally owned guns by people who LEGALLY own them. Its just too bad that politics get in the way of maybe a real solution to gun violence. That solution is not to take arms away from citizens who have the right guaranteed to them.

[dondo]

"These results suggest that it is easier for potential homicide
perpetrators to obtain a gun in states where guns are more
prevalent. “Our findings suggest that in the United States,
household firearms may be an important source of guns used to
kill children, women and men, both on the street and in their
homes,” said Miller."


So what they are saying is that it is easier to find a gun in places where guns exist. Wow. My own hypothesis follows:
" It is easier to find cheese, toilet paper and various deli meats generally in Supermarkets or stores that sell these goods."

[BerkeleyHK]

...found that homicide rates... are higher in states where more households have guns.

If you have an anti-gun agenda then you conclude that more guns = higher homicide rate without even considering that maybe higher homicide rates cause higher gun ownership.

[Dont Tread on Me]

I don't see any proof that the gun ownership rate is the cause of the homicide rates. It is just as meaningless as discovering that people in these areas tend to drive Fords and like the color blue. You have to show a causality.

It is nice to see that one in three homes owns a gun in the US. Yea!

[mike100]

life isn't so simple that a general statement like that is meaningful. The undereducated and lower classes with a defective culture are a big part of societies problem.

[Aluisious]

Press release, in whole. Not surprisingly the research is
sponsored by the anti-rights Joyce Foundation:

http://www.hsph.harvard.edu/press/releases/press01112007c.html
This is why in countries where all firearms are illegal, homicide rates are zero. There's simply no way to kill someone without a gun. And if there was no gun around, why would you want to? It's all love and flowers as long as no evil guns are around to corrupt you.

******* IDIOTS :mad:

[kap]

Dondo, you are correct that urbanization plays a role in homicide rates. Read the discussion of the actual paper below.


This is the discussion from the actual article.

States with higher rates of household firearm ownership had significantly higher homicide victimization rates in multivariate analyses. The association between firearm prevalence and homicide victimization in our study was driven by gun-related homicide victimization rates; non-gun-related victimization rates were not significantly associated with rates of firearm ownership. This result held overall, for women and for men, and across age groups, consistent with previous ecologic work that relied on a proxy measure for household firearm ownership rather than direct survey estimates (Miller et al., 2002).

Consistent with findings from individual-level studies that found household firearm ownership was associated with lethal victimization of women (Bailey et al., 1997b; Wintemute, Parham, Beaumont, Wright, & Drake, 1999) and with studies that found a gun in the home was a risk factor for homicides in the home perpetrated by family members, intimates or acquaintances (Kellermann et al., 1993), we found that household firearm prevalence was associated with firearm homicide victimization of women in unadjusted as well as in multivariate analyses. Our finding from unadjusted analyses that women (but not men) appear to be at increased risk of homicide victimization from household firearms suggests that household guns may play a more direct role in femicides than in homicide involving male victims. Although direct information about the location of lethal shootings and the source of firearms used in homicides do not exist for our national data set, prior work on the distribution of homicide location by gender is consistent with this possibility. For example, findings from the Chicago Homicide Dataset (Block, 1987), found that more than half of all female homicide victims but fewer than a quarter of all male homicide victims were killed in a home.

Young adult males are often killed by other young adult males and are more often killed in the street than are other homicide victims (Block, 1987). Three possible mechanisms may help explain why state-level household firearm prevalence is associated with homicide victimization of young adult males in our multivariate analyses. First, states with high rates of gun ownership tend to have less stringent regulations of firearm sales (Brady Campaign to Prevent Gun Violence, 2006), perhaps making it easier to obtain guns from a variety of street sources. Second, theft in states with high levels of gun ownership may be more likely to result in a ready source of illegal firearms. And third, altercations that escalate to physical violence may be more likely to prove fatal when firearms, including household firearms, are present.

In our analyses we have the advantage of being able to use survey measures of household firearm ownership. However, because our firearm prevalence estimates come from a survey of adults we do not have direct information about firearms that minors may have in the home that are not known to their parents. In addition the BRFSS firearm question does not provide potentially important information about many characteristics of firearm availability that may be related to the rate of homicide. For example, our measure does not differentiate handguns from long guns, or provide information on the number of firearms in gun owning households, the caliber of gun(s), or the ease with which firearms can be obtained in secondary market transfers.

Our study does not establish a causal relationship between guns and homicide. It is possible that a non-causal relationship explains our findings or that the association we observe might have arisen because individuals in states with historically high homicide rates acquired more guns (than did individuals in low-homicide states), as a defensive response to actual high homicide rates in their communities (i.e. “reverse causation”). This broad notion of reverse causation, while consistent with our association between household firearms and firearm and overall homicide, does not explain why firearm ownership is not also significantly associated with rates of non-firearm homicide. Furthermore, rates of robbery and aggravated assault are not associated with household firearm prevalence, even after controlling for urbanization and resource deprivation (not shown). Since individuals who obtain firearms in an attempt to protect themselves from violence plausibly respond to non-fatal violence (which is far more common than fatal violence), the lack of association between firearm prevalence and non-lethal violent crime militates against reverse causation as an adequate explanation for our findings. In addition, although several studies have documented that individuals obtain firearms for various reasons, including self-defense, almost nothing is known about whether the specific perceptions that motivate individuals to acquire firearms for self-defense have any relationship to actual homicide rates, overall or for any group.(Azrael, Miller, & Hemenway, 2000; Hemenway, 2004; Hemenway, Solnick, & Azrael, 1995; Howard, Webster, & Vernick, 1999; Senturia, Christoffel, & Donovan, 1994; Smith, 1998; Webster, Wilson, Duggan, & Pakula, 1992).

Consistent with previous work, we found that homicide rates were higher in areas with higher rates of urbanization (Parker & Smith, 1979), robbery (Baker, O’Neill, Ginsburg, & Li, 1992; Fingerhut et al., 1992; Hsieh & Pugh, 1993; Parker & Pruitt, 2000) and resource deprivation (Land et al., 1990); like others we also find that homicide rates are higher in the South (Gastil, 1971; Huff-Corzine et al., 1986; Land et al., 1990).

A limitation of the study is that other factors not included in the analyses may affect homicide rates. In addition, the measures for the control variables that we do use are only approximations (e.g. rates of aggravated assault come from police reports which generally underreport actual incidence) and represent aggregate, not individual-level information about our dependent and independent variables.

Our aggregate measures avoid the case–control problem of recall bias (e.g. cases being more likely to recall a firearm in the home than for controls), but this advantage comes at the possible interpretative cost of assuming group-level associations reflect individual risk factors (i.e. the ecologic fallacy) (Piantadosi, 1994). Nevertheless, our results accord with finding from prior individual-level studies (Bailey et al., 1997b; Kellermann et al., 1993; Wintemute et al., 1999).

Although we do not know whether the firearm used in a given homicide came from the victim's home, our results do not strictly require this interpretation; our results are also consistent with the possibility that the ease with which firearms may be obtained by potential perpetrators of homicide is related to the local prevalence of household firearms. For some types of homicide victimization this may reflect the use of guns from the victim's home, as might be the case in intimate partner homicides. For other types of homicide victimization, as might characterize a large proportion of inner city youth homicide, this may reflect the use of guns obtained through local burglaries. Unfortunately, our group-level data do not allow examination of these speculations.

Our study has additional limitations. Firearm prevalence data for this study comes from 2001, whereas mortality data are state-level aggregates over the 3-years 2001–2003. Although our outcome data (homicide) do not precede our exposure data, we nevertheless use household firearm measures from 2001 to assess homicide not only in 2001, but in 2002 and 2003. The effect of this temporal discrepancy on our results is likely to be small since guns are highly durable. In fact, the correlation coefficient relating state level household firearm ownership from the BRFSS in 2001 compared to 2002 is 0.99. Other studies have shown that the cross sectional pattern of household firearm ownership tends to be quite constant over longer time periods as well (Azrael, Cook, & Miller, 2004). Moreover, the pattern of association between household firearm prevalence and homicide we present using aggregate mortality data (2001–2003) is very similar to the pattern when outcome data are limited to any one of these years (not shown), though subgroup analyses are limited due to small numbers in some subgroups.

Despite these limitations, our cross-sectional finding that household firearm prevalence is a risk factor for homicide victimization of Americans is consistent with many previous studies, as summarized in a recent review (Hepburn & Hemenway, 2004). Our findings that household firearm ownership rates are related to firearm and overall homicide rates, but not to non-firearm homicide rates, for women, children and men of all ages, even after controlling for several potential confounders previously identified in the literature, suggests that household firearms are a direct and an indirect source of firearms used to kill Americans both in their homes and on their streets.

[dondo]

Aren't guns illegal in Mexico? Maybe that is why there is no crime, and everyone is so happy there, and everyone rides to work on a unicorn, and at the end of that rainbow there is a pot of pesos. Why are people so bent on destroying the foundations of this country? Can you be anti gun without wanting to throw the entire Constitution down the toilet? It seems every issue is so radical these days. Got to go, my unicorn needs more glitter.......

[Dont Tread on Me]

This is why in countries where all firearms are illegal, homicide rates are zero.

I have no idea why the British need homicide detectives:)

[kap]

If you have an anti-gun agenda then you conclude that more guns = higher homicide rate without even considering that maybe higher homicide rates cause higher gun ownership.
They attempt to explain this point in the discussion - "reverse causation" in the fifth paragraph - by saying that firearm ownership does not relate to non-firearm homicide.

[dondo]

Dondo, you are correct that urbanization plays a role in homicide rates. Read the discussion of the actual paper below.


This is the discussion from the actual article.
Thanks Kap, I did not even look to the link for the whole article. I will read now.....

[kap]

Thanks Kap, I did not even look to the link for the whole article. I will read now.....
You might not be able to get it unless you have access to the journal database.

[jdberger]

These guys at the Harvard Injury Control Research Center have a history of anti-gun papers. They offer themselves up as having no stake in the gun control debate, only being interested in the "Public Health" aspects of gun ownership. Yet their scholarly works clearly show an agenda. Miller and Hemenway have authored an enormous amount of work criticising gun owners "I am the NRA" ownership and the basic theories of the Second Amendment.

Simply put, despite their protests, they are partisan hacks. Geez..they even use the Kelleman "study". I wonder if they found Bellesiles missing data, too?:p

[Hunter]

Here is Reuters article on the report:
http://news.yahoo.com/s/nm/guns_murders_dc

[hoffmang]

First, I love this line from the report:


Our study does not establish a causal relationship between guns and homicide.


And everyone realizes that the vast majority of the states that this analysis is using to say that higher firearms ownership may correlate to higher homicides are also the A and B rated Brady states, right?

-Gene

[hoffmang]

Also Alphecca does a pretty good job debunking this:
http://www.alphecca.com/mt_alphecca_archives/002976.html

-Gene

[James R.]

I stopped at, "Boston".

[DrjonesUSA]

My rights aren't dependent upon statistics or any perceived "benefit" to any other person or entity.

The moment we try to justify the "need" to own a gun, we've already lost the argument from the start.

It doesn't matter if banning all guns tomorrow would drop the murder rate to zero: it does not change the fact that I have a right to own a firearm.

[Smokeybehr]

Geez..they even use the Kelleman "study". I wonder if they found Bellesiles missing data, too?:p

I saw the cite for "Kellerman" and "Wintemute", and didn't see "Lott and Mustard" in the paper, so I knew it was biased in its view.

[jumbopanda]

My rights aren't dependent upon statistics or any perceived "benefit" to any other person or entity.

The moment we try to justify the "need" to own a gun, we've already lost the argument from the start.

It doesn't matter if banning all guns tomorrow would drop the murder rate to zero: it does not change the fact that I have a right to own a firearm.

+100

Hoorah!

[kap]

Also Alphecca does a pretty good job debunking this:
http://www.alphecca.com/mt_alphecca_archives/002976.html

-Gene
The necessity to correlate the data to urbanization and poor resources is pretty telling of the data they had.

Actually I looked at the data he presented supporting his claim that suicide deaths are likely the cause of the statistical significance of the papers dataset. Seems to be a very accurate assessment.

California is 25th in firearm death per 100,000 (homicide and suicide)
California is 41st in firearm suicides per 100,000
California is 9th in firearm homicides per 100,000 (suicides NOT included)

Ironically Texas is 20 in firearm homicides and D.C. is #1 and only has 3.8% gun ownership.

Now for a few charts to illustrate the point. So when you take out people who can still kill themselves using other methods you get a clearer picture of the story. Granted I am not going to the trouble of generating p-values, nor do I have to. P-values in scientific research are generally only required when the trends are not readily apparent. The "Instructions to Authors" for journals usually state this. Just because someone has a p-value and did statistical analysis does not mean that their analysis is better, it just means the information is not obvious. With that said it is pretty clear from my charts that states that have high gun ownership do not have higher firearm related homicide rates. In fact they are pretty even across the board. States with high gun ownership do have higher firearm suicide rates. Keep in mind I left out D.C. since they are an outlier (D.C. has the Lowest firearm ownership rate 3.8% and the Highest firearm homicide rate 57.52 and the 3rd Lowest forearm suicide rate), so including it would only make this point more effectively. So do we enact gun control to save the suicidal who can just as well kill themselves by jumping off a building? It is interesting that they found the same relationship to decreased suicide rates after Australia enacted its firearm ban. I would love to get my hands on their dataset and see the real correlation of firearm homicides (not including suicides) and firearm suicides (not including homicides) with urbanization and resource deprivation taken into account.

http://img184.imageshack.us/img184/9625/gunownershipye5.jpg (http://imageshack.us)

I know many of you already know this info, but I figured I would share my own analysis of the data since I was doing it anyways.

[Mssr. Eleganté]

Also Alphecca does a pretty good job debunking this:
http://www.alphecca.com/mt_alphecca_archives/002976.html

-Gene


Thanks for that link. I see now that the "study" included suicides as homicides. So an important lesson we can all take away from this study is that if you live in a state with a high level of gun ownership and you shoot yourself in the head, then you stand a higher chance of being killed by gun violence.

[Ubergeek]

When can we expect the 'ivory tower geniuses' at the Harvard School of Public Heath to do an analysis of homicide rates versus:

- number/density of illegal immigrants in the locale?
- voting habits in the locale? (you mean places where people vote liberals/Dhimmicrats into office might have more crime? Pshaw!)

[kap]

When can we expect the 'ivory tower geniuses' at the Harvard School of Public Heath to do an analysis of homicide rates versus:

- number/density of illegal immigrants in the locale?
- voting habits in the locale? (you mean places where people vote liberals/Dhimmicrats into office might have more crime? Pshaw!)
This information has to be available somewhere. Doesn't the US census have this type of information? Getting crime data for each congressional district would be the next part. Granted finding a "well respected" authority to publish the information would be the hardest part.

[dwtt]

Considering the source, I don't know why anyone would take this article seriously.

[hoffmang]

There are two or three reasons to take the article seriously.

1. Always understand what evidence the other side is attempting to use to discredit your side. One should always know what the strengths and weaknesses of evidence on both sides are. This is one reason I warn people to be a bit careful with the conclusions they would like people to draw from John Lott. Credibility matters.

2. Be able to intelligently debunk it in day to day conversation with the normal skeptical folks around you who only read the headlines.

3. Remember that even though we can debunk them, any proposal that would come from this evidence is really about a new amendment to the Constitution - which they'd prefer no one mention. When someone goes down that argument road with you, ask them if they also agree that Utah should be able to institute a State Religion since that must be what the First Amendment means too.

-Gene

[LECTRIKHED]

Correlation does not imply causation. This study finds a correlation, which means absolutely nothing. You could probably find a correlation that people who eat chocolate live in areas with more murders. That wouldn't mean that eating chocolate causes murder.

Most studies are bull**** or done improperly. It is very difficult to do an accurate study proving anything. It can be done, but one study means absolutely nothing in scientific research.

[Jarhead4]

This article is only out there to catch your attention. I love statistics, you make it look any way you want base on how you use your assumptions. I really like what Kap found and what the mean line indicates. It blows this article’s assumption away. I'm sure you could further reduce these numbers by seeing how many guns that were use to murder someone that were legally owned by the murderer at the time. Then look at how many murders were committed by felons with stolen guns. Where is that statistics for that?

[kap]

This article is only out there to catch your attention. I love statistics, you make it look any way you want base on how you use your assumptions. I really like what Kap found and what the mean line indicates. It blows this article’s assumption away. I'm sure you could further reduce these numbers by seeing how many guns that were use to murder someone that were legally owned by the murderer at the time. Then look at how many murders were committed by felons with stolen guns. Where is that statistics for that?

Anyone know how to get data like this? Isn't there some sort of crime statistics database that the public has access to?

[hoffmang]

The best chance you'd have to find that data kap would be to write a letter to BATFE and FBI and ask either of them if they have those data sets.

-Gene

[kap]

The best chance you'd have to find that data kap would be to write a letter to BATFE and FBI and ask either of them if they have those data sets.

-Gene
I just may try that, but I won't hold my breath. :D

[hoffmang]

Well, you may be surprised. ATF and FBI don't know or care which side of the issue you're on and if you phrase it as a FOIA request, they have to get back to you.

Might be worth calling first to see if you can be more specific and make their lives easier while saving us all taxpayer dollars.

-Gene

[AJD]

As others have pointed out, you can't assume cause and effect from a correlational study. There are too many confounding variables at play, which cannot be controlled. It shows a relationship and nothing more. That's all fine, but most ignorant folks will read this and believe they can make cause and effect determinations because they don't know anything about a correlational study or any other research for that matter. Most media organizations never explain the purpose of a correlational study when they tell you about the "latest findings" of whatever the study was on their news broadcast.

[Jarhead4]

Here is a good start to get some information from the FBI, and it's free!!

http://www.fbi.gov/ucr/00cius.htm

One of the jobs of Law Enforcement is to collect data for statistics.


Have fun Kap!!!:D

[kap]

Good find, I will see what is in this information tomorrow. Thanks Jarhead!

[kap]

This is interesting. I saw this in the intro to the FBI 2000 crime in the US (http://www.fbi.gov/filelink.html?file=/ucr/cius_00/contents.pdf) report. (link is a PDF)

Historically, the causes and origins of crime have been the subjects of investigation by varied disciplines. Some factors which are known to affect the volume and type of crime occurring from place to place are:

Population density and degree of urbanization.
Variations in composition of the population, particularly youth concentration.
Stability of population with respect to residents’ mobility, commuting patterns, and transient factors.
Modes of transportation and highway system.
Economic conditions, including median income, poverty level, and job availability.
Cultural factors and educational, recreational, and religious characteristics.
Family conditions with respect to divorce and family cohesiveness.
Climate.
Effective strength of law enforcement agencies.
Administrative and investigative emphases of law enforcement.
Policies of other components of the criminal justice system (i.e., prosecutorial,
judicial, correctional, and probational).
Citizens’ attitudes toward crime.
Crime reporting practices of the citizenry.

The Uniform Crime Reports give a nationwide view of crime based on statistics contributed by state and local law enforcement agencies. Population size is the only correlate of crime utilized in this publication. While the other factors listed above are of equal concern, no attempt is made to relate them to the data presented. The reader is, therefore, cautioned against comparing statistical data of individual reporting units from cities, counties, metropolitan areas, states, or colleges and universities solely on the basis of their population coverage or student enrollment.

Looks like firearms ownership did not make it on the FBI list of factors related to crime. This list of factor related to crime is fairly intuitive and appear to have research to back them up. I guess this paper is just trying to make the case for firearms being on the list.

After looking at this list though it occurs to me that firearms ownership may only offer a negligible benefit to this paper's statistical model in their multivariate analysis after the other factors are taken into account. Alone, firearm ownerships correlation to homicide (including suicides, which is silly and misleading) was not that great (p-value=0.05 to 0.1). When they did the multivariate analysis which included firearm ownership, urbanization and resource deprivation as factors, they had a highly significant p-value=0.001. I am willing to bet that urbanization and resource deprivation play the largest role in predicting homicide rate while firearm ownership plays little or no role. If you take away suicides from the homicide data, firearm ownership is probably not a significant factor in this model at all.

[Mssr. Eleganté]

The fact that they included suicides as homicides in this study is telling. Over half of all firearms deaths in this country are suicides. Of course areas with higher firearms ownership are going to have higher firearms suicide rates. The only way for the "scientisions" who conducted this "study" to show that firearms ownership correlates to firearms homicide was to pretend that suicides are homicides. Otherwise that data goes directly against their findings.

This appears to be a case of a study designed to arrive at a predetermined conclusion. Just like when they did a study that found that an extremely high number of people who committed suicide had purchased a gun in the month prior to the suicide. They reported that buying a gun increased your chances of committing suicide within a month. That's like saying that purchasing gasoline increases your chances of being killed in an auto accident because 90% of auto accident victims had purchased gasoline in the month prior to being killed.

These guys actually know that they are wrong, so they are designing studies to come up with results that will make good anti-gun headlines. This is no longer a case of poorly designed studies. These are deliberately designed studies created by people who have a psychological aversion to firearms. How else could they "prove" that Chicaco is safer than Billings, Montana?

If I had a neurotic fear of ambulances, I could easily conduct a study that proved that people who take ambulances to the hospital are twice as likely to die as people who go to the hospital in a private vehicle. The press might even report the study as an amazing finding, totally ignoring the fact that people who go to the hospital in an ambulance are usually in much worse shape and more likely to die than people who drive to the hospital in their car.