Another Questionable Study

S

sigcurious

New member
Article

The studies conclusion is that there is a correlation between more gun laws and few firearms deaths.

Some of the cited numbers seemed pretty out of whack, so I took a closer look.
In states with sparser firearms laws, researchers reported that gun-mortality rates were higher: Louisiana (18.0 per 100,000), Alaska (17.5 per 100,000) and Arizona (13.6 per 100,000).
Click to expand...

The study covers 2007-2010 statistics. Using AK as an example, they say 17.5/100k, the FBI for 2010 says 4.4/100k murders/non-negligent manslaughter total(Source)... granted studies like this usually don't bother differentiating between suicides, negligent deaths and murders. And the FBI data is for 1 year only instead of 4, but I think it's safe to guess that even if the murder rates were trending upward or downward they would not make up for the difference between 4.4 and 17.5. I can only guess without seeing the study itself, that the remainder are negligent/accidental and suicide deaths.
 
That's my best guess. Many of the studies I've seen that are obviously slanted towards anti-gun mentalities consistently do these two things to show that less guns is "better." The first is that they use rates from "mortalities" or "gun-deaths" instead of murders/homicides, which therefore do include suicide rates and such. The second is that they only look at the "gun" deaths/mortalities, which while sometimes (but not always) higher in pro-gun states do not show that overall murder/suicide rates are generally not affected or even negatively affected by strict gun control. Suicides are generally not prevented by forcing the person to use a different means: such as jumping, sharp objects, hanging, etc. instead of a firearm.

Beyond that, one should look to what source of the actual death rates are being used. I recently did a write up on a Mother Jones article whose data on relative murder rates in pro-/anti-gun states was very, very different than the numbers given by the FBI Uniform Crime Reports.
 
So what you are saying is that their infromation, as clearly stated, may be 100 percent correctn but you don't like it because they don't focus on murders?
 
I think what he's saying is that they allow the reader to believe the deaths are the result of violent crimes by criminals against victims, whereas many may be suicides.

Statistics may be accurate, but the message is very, very different.
 
Not altogether surprising, given both the source of the article and the source of the study. Public health officials are paid to worry about the total number of deaths. I'm more concerned about the identity of the deceased, personally. Not just how many die, but who, and how. The article (to my reading) also implies causation, when correlation might (and only might) be the factor present.

Also, ALWAYS look at the definitions section of a study. There's a big difference between measuring "gun-related homicides" and "gun-related deaths."

I also note that this study "scor[ed] individual states simply by the sheer volume of gun laws they have on the books," in looking for the relation between laws and gun-related death rates. IOW, for purposes of this study, it doesn't matter whether the particular law is:
1) sentencing enhancement for using a firearm in a robbery; or
2) a prohibition on fireams with a "Hello Kitty" motif.
 
its like claiming that light beer causes fewer alcohol related crashes then people who get drunk on regular beer before driving.

its like claiming that its easier to get eloctrocuted when you take the hairdryer in the shower then if you take a toaster oven in when you take your bath.
 
I think what he's saying is that they allow the reader to believe the deaths are the result of violent crimes by criminals against victims, whereas many may be suicides.

Statistics may be accurate, but the message is very, very different.
Click to expand...

Not really...if you actually read what was written. Certainly violent crimes make up part of the data. There is no doubt about that. However, the OP did complain about the presentation of the mortality data and the data were presented accurately.

However, the information isn't new at all...
http://www.thedailybeast.com/articl...t-gun-states-from-mississippi-to-arizona.html

http://www.cbsnews.com/2300-204_162-10010004.html

http://www.thedailybeast.com/articles/2012/07/20/gun-violence.html

You know, looking at the actual JAMA study and the correlations noted, they are just that, correlations. There were exceptions to the correlations as well. If you wanted to look at firearms mortality compared to their map and US poverty, you would also find strong correlations.
http://archinte.jamanetwork.com/article.aspx?articleid=1661390
http://visualeconsite.s3.amazonaws.com/wp-content/uploads/percent_in_poverty.gif
 
Last edited:
The study covers 2007-2010 statistics. Using AK as an example, they say 17.5/100k, the FBI for 2010 says 4.4/100k murders/non-negligent manslaughter total(Source)... granted studies like this usually don't bother differentiating between suicides, negligent deaths and murders.
Click to expand...

The statement in question was taken from the abstract of the article (scientific papers usually start with a brief summary called an "abstract") and it is reasonably clear from the context that the numbers cited refer to all firearms fatalities. The analysis in the article itself separates homicides and suicides and looks at each separately.

The authors of the paper freely admit that their findings are not conclusive, and they point out in the paper some of the issues with their own study. Unfortunately, when the media get ahold of something like this they don't bother to report it objectively, but I think the authors of the study were reasonably objective.

The original paper, in its entirety, is available on line here: http://archinte.jamanetwork.com/article.aspx?articleid=1661390
 
It's co-authored by David Hemenway, who's well known for being an advocate for gun control. The data was provided by the Brady Campaign and the Law Center to Prevent Gun Violence, and the study was funded by the Joyce Foundation.

So, yes. Completely objective. :rolleyes:

As far as I can tell, they're pointing out total fatalities, not homicides or criminal uses. In fact, the only point of the paper appears to be "I found these numbers in a database, but I can't assert a conclusion."
 
I don't want to appear to be defending this study, and in fact I'm not defending it. There are too many unanswered questions and too many confounding factors to support the simple conclusion that the media (moreso than the authors) seem to be drawing from this work. Some of these issues are discussed by the authors in the paper itself, and those issues and others are expanded upon by Dr. Wintemute's commentary article that immediately follows the Fleegler et al. article in the same issue of the journal. That's what I meant by "reasonably objective."

I agree that the participation of David Hemenway as a junior author raises a red flag, but to dismiss the study simply because of that is an ad hominem argument that should be beneath us. I think most of us here bristle when our pro-gun arguments are dismissed simply because we're gun owners and/or NRA members, or whatever - it's no different when the shoe is on the other foot. The study was not funded by the Joyce Foundation, or at least that's not stated and I think they would have done so had that been the case; the authors - to their credit - do acknowledge that Hemenway (and, apparently, only Hemenway) receives funding from the Joyce Foundation.

The data provided by the Brady Campaign relate to the classification of states according to their gun laws. The data on firearm-related fatalities came from the CDC and another [presumably] government agency that I have no knowledge of. I think that one of the major flaws of the study is how one goes about ranking the states on the extent of their firearms-related legislation, but much as we all may dislike the Brady Campaign, overall their ranking is not unreasonable (except that we would redefine which is the "good" and which the "bad" end of the list, of course). For example, Brady ranks Massachusetts as their "best" state in terms of gun laws and while we may quibble with whether New York or New Jersey or some other state deserves that dubious distinction I don't think the choice of Massachusetts is far off the mark. Regardless, if you're going to use the Brady ranking (and, to reiterate, any ranking is fair game for criticism) it would seem to make some sense to have the Brady Campaign provide it.

If you actually read the study, you will see that they analyze and discuss the data on homicides and suicides separately, along with overall fatalities. They do exclude homicides that result from legal intervention and accident, both of which were a very small percentage of the total. They don't separate criminal vs. non-criminal (justifiable?) homicide, but I think a fair observer would admit that the vast majority of homicides are in fact criminal, so making that distinction would likely not change the overall results. (Still, it's a valid criticism of the work.)

There are valid criticisms that can be made regarding this study, but most of what appears in this thread doesn't fall into that category. The study should be reviewed and critiqued on its merits, not based on who did it, who funded it, or how much we may dislike the conclusions. We only weaken our own position by doing otherwise.
 
The issue I take with the study are the implications and how the data can be used.It puts data out there for other people to misinterpret and manipulate. Is it necessarily the authors responsibility to not do so? No. However, given the source and the methodology, even if the data is accurate and the stats were done properly, the pool of information was chosen in a way that implies a predetermined conclusion to be reached, rather than studying data to see where it leads.

Considering the recent issue of the "40% of sales are without background checks" number being out there and remaining out there even after the source was debunked, I think it is noteworthy when new statistics of questionable sourcing get put out there. We may take the time to look at the whole picture and the individual factors involved, but others may not.

Mass killings such as those in Columbine and Aurora in Colorado, the Wisconsin Sikh temple shooting, and most recently the Newtown, Connecticut, school massacre have renewed debate about the need for more stringent firearm legislation. Some have called for more restrictions on gun purchases.3 Others have called for arming teachers.4 It is challenging to calculate the exact number of firearm laws: a single law may have multiple parts; laws are potentially passed at the national, state, county, and city level; and there is no repository available for tallying these laws.5 The factoid that there are “20 000 laws governing firearms”5 has been erroneously quoted since 1965, but the most recent and reliable estimate, performed in 1999, counted about 300 state firearm laws.6

The real question is not about the number of firearm laws but whether the laws ultimately safeguard the citizens they are intended to protect. Although multiple studies have examined the relationship between federal and state firearm laws and homicide and suicide rates, the overall association between firearm legislation and firearm mortality is uncertain and remains controversial.7- 8
Click to expand...

I think the above from the study itself is telling of the mindset that went into the study. They start by indicating malicious acts have called into question the relationship between firearms laws and firearms deaths. Yet the parameters chosen are so broad as to include everything related to firearms both in mortality and legislatively.

I think it's disingenuous to question whether firearms laws protect people, then to lump suicide and homicides together. As suicides are their own complex issue only tangentially related to questions of firearms law and public safety.
 
I've seen the Alaska study before. What they probably fail to mention in that study, at least as far as Alaska is concerned is that Alaska has one of the, if not the highest suicide rate in the nation. If you discount suicides, Alaska has one of the lowest firearm mortality rates in the nation. And you can discount suicide, because these people would have found another way to kill themselves had the gun not been available.

Actually, a lot of these studies can usually be discounted by removing suicides. The harder it is to buy a gun, the harder it is to have a gun available. The harder it is to have a gun available, the harder it is to shoot yourself. Not to mention, many of these studies are skewed from the beginning to favor gun control, as in the methods of collecting data, or what data they use. I saw a study once that took gang violence out of the picture. Yet, gang violence is one of the greatest reasons some individuals need to ability to arm themselves.
 
As a part of one of my former occupations, I read and evaluated a lot of studies. After a while I would turn to the section which identified the sources of funding and the organization which was conducting the research.

I remember two studies. One on marjorin was funded by the corn growers and conducted by the University of Iowa. The second was on butter and funded by the Dairy Association and conducted by the university of Wisconsin. Both studies compared the benefits and negatives of the two products. You can guess what the results showed by the source of the study.

Frankly, I do not believe that any study on firearms and firearms safety can be trusted. Even the national Safety council has been caught cooking the numbers when it comes to guns.
 
I agree that the participation of David Hemenway as a junior author raises a red flag, but to dismiss the study simply because of that is an ad hominem argument that should be beneath us.
Click to expand...
Respectfully, it isn't ad hominem to note a study is conducted by someone who is a policy advocate.

This study has so many metholdolocal issues, including selected data sets, comparative problems etc that it isn't even funny.

To start with I don't see a proper notational and emphasis that the highlighted comparative states have extremely different mortality rates on all trauma.

If you experience major trauma in Mass, NJ or Ct you are going to have the best rained, best equipped, most densely allocated EMS in the US on top of you in minutes and you are going to be in a world premier trauma center minutes after that. If you experience life threatening trauma in Louisiana or Alaska that isnt going to be the case. so comparing mortality for a specific trauma between those very different states is highly problematic.

There are also severe problems with the assertion on the "amount of laws" approaching logical tautology. Conn, NJ, and Mass have more laws in general.They even have more real estate laws.

The study authors don't count extent of private security (armed and unarmed) across these states, or the relative distribution presence of secure gated communities. Indeed the guns in the hands of private security are Excluded as well. Again this creates an elitist bias seen in the example of Bloomberg, who doesn't own a gun but has around the clock protection by several guys with guns -- again guns that don't even count to the number of "households with guns" in these type of studies.

Reading the study one sees that the data was cherry picked to support these biased researchers and the interest groups they represent. the are a dozen serious methologiclal problems I haven't even mentioned

They are just as flawed and immensely biased as the research trying to show home ownership of guns puts the owner and their family more at risk. Those are obscenely bias studies which a) include active criminals who keep guns at home in the same data set as legal gun owners; and b) exclude incidents of showing a weapon and scaring off a assailant s a positive outcome.

I think these type of studies illustrate why the NRA did well to opposed federal funding. they are junk studies
 
Respectfully, it isn't ad hominem to note a study is conducted by someone who is a policy advocate.
Click to expand...

Agreed. Nothing wrong with noting the potential biases of an author of a study, and nothing wrong with scrutinizing the study for the potential effects of those biases. That was what I meant by "raises a red flag." But to dismiss the study solely because of who did it, without examining how the study was done and how the results were interpreted is most definitely an ad hominem fallacy.

The rest of your post raises some valid issues with the study, which was really the only point I was trying to make, i.e., that if we disagree with the conclusions of the study we should base our disagreement on valid criticisms rather than simply because we don't like who did it or what they found.
 
But to dismiss the study solely because of who did it, without examining how the study was done and how the results were interpreted is most definitely an ad hominem fallacy.
Click to expand...
I agree. But one must note that virtually none of the large amount of press coverage this is receiving mentions that a key author is associated with an advocacy group.

I have looked at the actual report. My view is while the researched make a cursory nod to some (not all) of the problems in their methodology and choice and application of data sets presents, that overall it really more than tilts toward advocacy derived bias.

I can go into some of the several other biases and problems with it in some notes tonight.

Food for thought, the states with the high mortality are hunting states and hunting is a sport with considerable danger of gunshot injury and mortality. Frankly I would expect Florida has drowning deaths from snorkeling then several other states
 
Food for thought, the states with the high mortality are hunting states and hunting is a sport with considerable danger of gunshot injury and mortality. Frankly I would expect Florida has drowning deaths from snorkeling then several other states.
Click to expand...

I'll look forward to reading your additional thoughts on the study. With regard to the hunting deaths issue, the article does note that they excluded "unintentional firearms fatalities" (see the section entitled "Study Population" in Methods), so presumably hunting fatalities were not included in their tally of homicides by state.
 
One of the major methodological problems, and it goes to the very root of the study's primary numbers is their means of establishment of relative gun ownership rates.

They are looking to estimate the ratio of: gun ownership rate : murder and suicide rates.

how do they establish the first number, gun ownership rate in a state? they establish it by asserting it derives for the murder and suicide rate.

Wow. the study seeks to determine x/y. how do they know what the value of y is? they know it from x/y!

They then assert this by noting five studies. three of which are from one single anti gun researcher, one of which is international. (the international one is biased and unprofessional as it includes all major industrialized countries but excludes Japan -- where the data on suicide would not "fit". )

Interestingly these referenced studies they use in their establishment of gun ownership /crime rates include felons and illegal owners.

This I really tautological at its root. count illegal owners, count felons, count active criminals. Establish that alone with legal owners as "gun owners" to establish gun ownership rates. Correlate that with gun crime rates to show "more gun owners more crime." Use those studies to establish your basis numbers for a new study with Brady campaigns assertions of what gun laws are relevant.

there isn't any doubt that a methodology like that is going to give the anti gun advocates who authored the study the results they were looking for when they started.
 
I agree with your overall criticism of how they derived firearms ownership rate by state, but I don't think you have the details correct. I don't believe they used ownership rates from any of the literature they cited. In the explanation of Table 2, they say they derived firearms ownership rates by the state-wise ratio of firearms suicides to total suicides calculated from the mean (average) rates over the years 2007 to 2010.

Percentage household firearm ownership was calculated by mean firearm suicides/total suicides (2007-2010) by state.
Click to expand...

That is, if a particular state had an average of, say, 100 firearms suicides per year and 200 total suicides per year over that period they assumed 50% firearms ownership for that state. That is an absurd way to determine firearms ownership - to note only two problems, implicit in their procedure is an assumption that (1) an individual intent on committing suicide will always use a gun if one is available, and (2) individuals who commit suicide are representative of the general population with respect to percent gun ownership. I would question both assumptions, but that's not important - what's important is that they failed to establish them as valid, which is their responsibility.

There are numerous sources of information on gun ownership rates by state, so there was no need for them to calculate it by a procedure of highly questionable validity. Because the firearms ownership rate they're using is invalid, the relationships shown in Figures 2b and 2c are similarly invalid.
 
I find it telling that so many on the other side of the debate only want to talk about gun-crime and gun-related deaths rather than crime and death in general. When did a murder committed with a knife, baseball bat, broken bottle, or any other object become preferable to one committed with a gun? Perhaps I'm behind the times, but I was always under the impression that any violent crime, regardless of the instrument used to perpetrate it, was abhorrent. Zeroing in on gun-crime seems to me like a convenient way to move the goal post.
 
You must log in or register to reply here.
Back
Top