[mattyb]

Originally Posted by moonmonkey 

No its not,
Having a gun:
1) increases the chance that someone will shoot you in a conflict because they know you are armed
2) increases the chance that you will be shot with your own gun if overpowered by an unarmed person
3) increases the chance you will shoot yourself accidentally.

Cars don't do this stuff.

Isn't LSD banned in HK? You really need to read a bit more.

Its a pity that these sorts of statements come out when history and the lives of people across the planet prove otherwise.

[Spheric Harlot]

Originally Posted by Mrjinglesusa 

No offense Spheric, but your "simple logic" is not logic at all. It's a fundamentally flawed conclusion based on a fundamentally flawed "study".

You said:

If you have a gun, you are more likely to be shot and killed by one.

You claim that this statement is supported by a "study" you cited. THE "STUDY" DOES NOT ALLOW ONE TO MAKE A BROAD GENERALIZATION LIKE YOU HAVE. Period.

Actually, that is the ONLY conclusion the study allows you to make.

The only one.

The only conclusion it DEFINITELY SUPPORTS is this broad, generalized one, because THAT IS THE ONLY FACTOR IT HAS ACCOUNTED FOR.

You really have no FECKING clue how a scientific study works, do you?

You try your DARNDEST to allow for ONE variable, and ONE variable ONLY, because then you only need one control group that doesn't match the criterion (gun-owner/non-gun-owner).

This study did just that, and found, conclusively, that gun owners are more likely to get shot and killed.

That's simple fact, and the conclusion is necessarily general BECAUSE THAT WAS THE POINT OF THE ****ING STUDY.

Now that this has been established, further studies can be set up to differentiate: white/non-white, licensed/non-licensed, sport-shooter/non-sport-shooter, safety-trained/non-safety-trained, prior convictions/no convictions, etc.

However, and this is what I've tried to tell you earlier, allowing for ONLY ONE of those variables to differentiate between the "experimental" group and the control group is really, really difficult, especially in numbers that allow you to actually draw statistically relevant conclusions.

This is why it took so long to come up with even this study.


Jeezus, man, I'm not gonna give the whole rundown on basic scientific method - WTF do they teach you people in middle school?

[Spheric Harlot]

Originally Posted by ghporter 

I'd like to point out that these Swiss rifles we're discussing are actually fully automatic rifles-machine guns. In basically unregulated, private hands. No Armageddon, no bloodbaths, etc., etc. Basically this means that, in a properly trained population, firearms seem to be non-threatening, non-dangerous items. Since the Swiss still have mandatory military service, their citizens have good training and basically enforced contact with firearms, so they're not mysterious, magical things, which probably results in less fooling around and less (if any) abuse or accidental misuse.

They're also almost *certainly* unloaded when carried outside of a firing range, and almost certainly deemed completely useless for self-defense by anyone owning one.

It's kind of silly to argue that the situation in a country where only specialists are allowed to carry handguns, but people carry unloaded automatic rifles to work, is the same as a country where any half-wit unconvicted criminal able to sign his name is allowed to purchase a gun.

[BadKosh]

Originally Posted by moonmonkey 

No its not,
Having a gun:

Originally Posted by moonmonkey 

1) increases the chance that someone will shoot you in a conflict because they know you are armed

NO it doesn't. How would they know you had a gun? So the opponent is a mind reader, or do we need to split the group into those who know you and those who don't?

Originally Posted by moonmonkey 

2) increases the chance that you will be shot with your own gun if overpowered by an unarmed person

This assumes you are carrying the gun, and it's not concealed.

Originally Posted by moonmonkey 

3) increases the chance you will shoot yourself accidentally.

Originally Posted by moonmonkey 

Cars don't do this stuff.

There would be no difference swapping car for gun in your first 'point'. Most people don't know who they get into a car accident with either. Hitting pedestrians or being in the car during the accident will change your data. The fact that innocent bystanders are involved in both kinds of events doesn't change the data either.

[Mrjinglesusa]

Originally Posted by Spheric Harlot 

Actually, that is the ONLY conclusion the study allows you to make.

The only one.

The only conclusion it DEFINITELY SUPPORTS is this broad, generalized one, because THAT IS THE ONLY FACTOR IT HAS ACCOUNTED FOR.

You really have no FECKING clue how a scientific study works, do you?

You try your DARNDEST to allow for ONE variable, and ONE variable ONLY, because then you only need one control group that doesn't match the criterion (gun-owner/non-gun-owner).

This study did just that, and found, conclusively, that gun owners are more likely to get shot and killed.

That's simple fact, and the conclusion is necessarily general BECAUSE THAT WAS THE POINT OF THE ****ING STUDY.

Now that this has been established, further studies can be set up to differentiate: white/non-white, licensed/non-licensed, sport-shooter/non-sport-shooter, safety-trained/non-safety-trained, prior convictions/no convictions, etc.

However, and this is what I've tried to tell you earlier, allowing for ONLY ONE of those variables to differentiate between the "experimental" group and the control group is really, really difficult, especially in numbers that allow you to actually draw statistically relevant conclusions.

This is why it took so long to come up with even this study.


Jeezus, man, I'm not gonna give the whole rundown on basic scientific method - WTF do they teach you people in middle school?

I have a PhD in Chemistry - I know about the scientific method. THERE IS NOTHING "SCIENTIFIC" ABOUT THIS STUDY. There was NO control group that comprises the SAME POPULATION AS THE GUN OWNER GROUP where the ONLY DIFFERENCE IS THE GUN-OWNERSHIP! How do you not understand this????

If the "gun-owner" group is comprised of 50% criminals, 30% African Americans, 40% people aged 15-21, and 75% males and the "non-gun-owner group" is comprised of 5% criminals, 10% African Americans, 20% people aged 15-21, and 50% males, YOU CANNOT SAY THAT GUN OWNERSHIP IS THE CAUSATIVE FACTOR!!!! Jesus, you tell ME I don't understand scientific method?

I'll spell this out for you: THERE WAS NOT ONLY ONE VARIABLE IN THIS STUDY.

Period.

For there to be only ONE variable as you suggest (gun owner/non-gun owner) THE DEMOGRAPHICS OF THE TWO GROUPS MUST BE IDENTICAL. This means the "study" authors have to account for race, age, sex, criminal background, activity when the killing occurred, location, etc., etc., etc. THEY DID NOT.

And how did the "study" authors verify that the "non-gun owners" did not own guns? Survey? Like I said, if someone came to my house and asked me if I owned a gun, I WOULD SAY NO and so would the majority of gun owners.

This "study" does not allow you to make ANY conclusion whatsoever about how owning a gun affects the likelihood of one being shot and killed by a gun.

BTW, they only looked at "data" from ONE CITY. How does that allow a broad generalization about gun-ownership in other cities/states increasing the likelihood of one being shot and killed by a gun? IT DOESN'T.

If you want to make a statement about the results of this "study" you can only say that owning a gun IN PHILADELPHIA, PA increases the chances you will be shot and killed by a gun IN PHILADELPHIA, PA.

I might even agree with that statement and wouldn't need a study to confirm it. LOL

[NosniboR80]

Originally Posted by Spheric Harlot 

Actually, that is the ONLY conclusion the study allows you to make.

The only one.

The only conclusion it DEFINITELY SUPPORTS is this broad, generalized one, because THAT IS THE ONLY FACTOR IT HAS ACCOUNTED FOR.

~~~~

Jeezus, man, I'm not gonna give the whole rundown on basic scientific method - WTF do they teach you people in middle school?

This is what I was trying to point out from the beginning. That "generalized", "broad" conclusion is precisely what is wrong with your statement.

I could do a study showing that being white makes you more likely to commit a white collar crime. That, however, is a useless statement, especially if you said it to a plumber or somebody like that.

Although it might be statistically accurate, it is not a fact. That plumber is not likely to commit a white collar crime, like embezzlement. He's just part of a population segmented and pinppointed by my hypothetical study. My "data" is just averaging out "likelihood" from the Ivy-league Enron/AIG/Maddoff types across a wider population.

So, instead of that plumber's "actual" 0.0001% likelihood of committing felony embezzlement, my hypothetical "study" would make it look like he's 3.4% likely to commit that crime.

Whereas, the Ivy-league stock broker who is "actually" 7.8% likely to commit felony embezzlement, would show in my hyphothetical "study" that he is (just like the plumber) 3.4% likely to do the crime.

So, according to my "study", it is statistically accurate to say that a white person is 3.4% likely to commit a white collar crime. In reality, it's a worthless statement, because the study is averaging data across many different groups.

The same result comes from your statement. The fact that the study that you referenced doesn't dig deeper means that you can't make meaningful statements from it. You don't have to do another study to dig deeper. You can factor out other variables at the same time. That's what good statisticians do.

As a trained, but inactive Marine officer, I am not more likely to be shot by a weapon than anyone else, just because I own one. I would have to be conducting more dangerous activities to become more likely to be shot than another non-gun-owning person.

As for what they taught me in middle school, I would ask you the same thing, except that I know they don't teach statistics in middle school. I had to wait until college. The same course might be enlightening to you.

I found that statistics can be extremely useful, but they do more harm than good, in my opinion.

Statistical findings do not necessarily equal facts!

[Laminar]

Originally Posted by Mrjinglesusa 

I have a PhD in Chemistry - I know about the scientific method. THERE IS NOTHING "SCIENTIFIC" ABOUT THIS STUDY. There was NO control group that comprises the SAME POPULATION AS THE GUN OWNER GROUP where the ONLY DIFFERENCE IS THE GUN-OWNERSHIP! How do you not understand this????

If the "gun-owner" group is comprised of 50% criminals, 30% African Americans, 40% people aged 15-21, and 75% males and the "non-gun-owner group" is comprised of 5% criminals, 10% African Americans, 20% people aged 15-21, and 50% males, YOU CANNOT SAY THAT GUN OWNERSHIP IS THE CAUSATIVE FACTOR!!!! Jesus, you tell ME I don't understand scientific method?

I'll spell this out for you: THERE WAS NOT ONLY ONE VARIABLE IN THIS STUDY.

Period.

For there to be only ONE variable as you suggest (gun owner/non-gun owner) THE DEMOGRAPHICS OF THE TWO GROUPS MUST BE IDENTICAL. This means the "study" authors have to account for race, age, sex, criminal background, activity when the killing occurred, location, etc., etc., etc. THEY DID NOT.

And how did the "study" authors verify that the "non-gun owners" did not own guns? Survey? Like I said, if someone came to my house and asked me if I owned a gun, I WOULD SAY NO and so would the majority of gun owners.

This "study" does not allow you to make ANY conclusion whatsoever about how owning a gun affects the likelihood of one being shot and killed by a gun.

BTW, they only looked at "data" from ONE CITY. How does that allow a broad generalization about gun-ownership in other cities/states increasing the likelihood of one being shot and killed by a gun? IT DOESN'T.

If you want to make a statement about the results of this "study" you can only say that owning a gun IN PHILADELPHIA, PA increases the chances you will be shot and killed by a gun IN PHILADELPHIA, PA.

I might even agree with that statement and wouldn't need a study to confirm it. LOL

Angry underlined bold CAPITALIZED HUGE TEXT ARGAGARRRRRRG YOU FORGOT COLOR

[The Final Dakar]

Stephen A. Smith would be proud.

[Mrjinglesusa]

Let me provide an example of why this "study" does not allow one to make the generalized statement that is being made by the authors and Spheric.

I am carrying out a chemical reaction between Compound A and Compound B to form Product P. I want to see whether changing the temperature affects the rate of the reaction. So I carry out two experiments:

1) Compound A + Compound B + Solvent S carried out at 25 C

2) Compound A + Compound B + Solvent S + a few crystals of sugar carried out at 40 C

The Solvent (sex) of the reactions is the same, they are carried out in identical reaction vessels (age), I start them at the same time (race), and use the same amounts of A and B (socioeconomic background).

I find that Reaction 2) is faster.

I CANNOT SAY THAT TEMPERATURE (Gun Ownership) IS THE CAUSATIVE FACTOR IN THE INCREASED RATE.

There is a SINGLE variable that I didn't account for (sugar = criminal background, activity at time of shooting, etc.).

That is why the "study" cited is flawed and does not allow one to make ANY conclusion about the effect of gun ownership on the likelihood of being shot and killed by a gun.

Scientific Method 101

There can be ONE and only ONE variable. By not accounting for ALL potential variables, the study is fundamentally and intrinsically flawed, thus allowing no conclusions to be drawn from it.

[Mrjinglesusa]

Originally Posted by Laminar 

Angry underlined bold CAPITALIZED HUGE TEXT ARGAGARRRRRRG YOU FORGOT COLOR

Forum posting 101: If someone doesn't comprehend your point, use bold, underlined, large text to hammer the point home.

Color confuses the issue and distracts from the message. "Oooo, pretty colors! What was he saying?"

[The Final Dakar]

Can I get a pie chart up in here?

[Spheric Harlot]

Originally Posted by Mrjinglesusa 

I have a PhD in Chemistry - I know about the scientific method. THERE IS NOTHING "SCIENTIFIC" ABOUT THIS STUDY. There was NO control group that comprises the SAME POPULATION AS THE GUN OWNER GROUP where the ONLY DIFFERENCE IS THE GUN-OWNERSHIP!

OF COURSE THERE IS.

Question is, "Is there a statistically relevant correlative relationship between gun ownership and likelihood to get shot/killed by a gun?"

Studied population is "residents of city X".

"Experimental" group (in quotes because this is a field study, not an experimental set-up) is "gun owners" within studied population.

"Control" group is "non-gun-owners" within studied population.

Result: Correlation proved.

Limitations: No insight into actual nature of causative relationship. Further studies needed.

The end.

I'll spell this out for you: THERE WAS NOT ONLY ONE VARIABLE IN THIS STUDY.

IN THIS STUDY, THERE *WAS* ONLY ONE VARIABLE because all others were eliminated.

THAT IS WHY the results are so broad.

Causation WAS NOT WITHIN THE SCOPE of this study.

For there to be only ONE variable as you suggest (gun owner/non-gun owner) THE DEMOGRAPHICS OF THE TWO GROUPS MUST BE IDENTICAL. This means the "study" authors have to account for race, age, sex, criminal background, activity when the killing occurred, location, etc., etc., etc. THEY DID NOT.

Since you're a chemist, I'll cut you some slack.

You're used to magnesium being magnesium, molar masses being identical, and oxygen having a -2 charge.

In social sciences, it's kind of a little more difficult, because, like it or not, it's not only that the demographics of two groups can never be quite identical, it's that even if you break it down to the smallest possible level for analysis, you find that - oh horror! - THE DEMOGRAPHICS OF NO TWO INDIVIDUALS CAN EVER BE ABSOLUTELY IDENTICAL.



Oddly, the implications of this shouldn't be too difficult for you, a PhD in chemistry, to grasp, since everything you're used to as predictable "fact" is, on a physical level, nothing but probability. Dumping two reactants together doesn't mean that the individual atom is guaranteed to get oxydized/reduced, any more than the fact that you own a gun guarantees that you will be shot at some point.

In fact, even electrons are often enough just portrayed as probabilities within their orbitals, IIRC.

Still, you generalize. Is that "fact" enough for you?

BTW, they only looked at "data" from ONE CITY. How does that allow a broad generalization about gun-ownership in other cities/states increasing the likelihood of one being shot and killed by a gun? IT DOESN'T.

If you want to make a statement about the results of this "study" you can only say that owning a gun IN PHILADELPHIA, PA increases the chances you will be shot and killed by a gun IN PHILADELPHIA, PA.

This is the FIRST (and only) valid criticism of the study I've seen in this thread, and it's a meta-argument.

And it took a scientist (PhD) FIVE DAYS to come up with it!

(point conceded)

You do remember the one about the train on the way to the metaphysics congress in Scotland?

A priest, a chemist, a mathematician, and a philosopher see a black sheep in a field as they pass by in the train. The priest says, "I didn't know sheep in Scotland are black!" The scientists says, "Well, actually, you can only say that one sheep in this field are black." The mathematician says, "Well, actually, you can only say that ONE SIDE of one sheep in this field is black." The philosopher says, "Well, actually, only one side of one sheep in this field is black some of the time."

[moonmonkey]

Originally Posted by mattyb 

Isn't LSD banned in HK? You really need to read a bit more.

Its a pity that these sorts of statements come out when history and the lives of people across the planet prove otherwise.

I really don't see what psychoactive drugs and my current location have to do with this.

[Mrjinglesusa]

Originally Posted by Spheric Harlot 

IN THIS STUDY, THERE *WAS* ONLY ONE VARIABLE because all others were eliminated.

THAT IS WHY the results are so broad.

Causation WAS NOT WITHIN THE SCOPE of this study.

OK, last thing I will say on this subject lest we end up in the PL.

I concede the fact that they were not looking at causation.

I still disagree with you that there was only one variable, however I understand that it is impossible to find two demographically identical subject groups.

That said, the study authors did not account for a very important, perhaps the most important variable: criminal background.

That's the fundamental problem I have with the study. That is a crucial, potentially statistically significant variable when you are looking at deaths from gun violence and the study authors do not address this variable. It's not a causative factor, it's an important demographic variable between the two groups.

I will concede that the authors cannot possibly consider ALL variables, but to not consider perhaps the most important one (criminal background) is not being intellectually honest.

We will just have to agree to disagree on this subject.

[Spheric Harlot]

Whew.

Thanks, your post was a relief: we're on the same page, now.

As I mentioned in the initial post, I was aware of the limitations of the study, and posted it merely for relevance to the side-line discussion.

I was starting to get mondo annoyed by people telling me what I they thought I was saying.

If I had the slightest interest in that kind of bullshit, I wouldn't have had myself voluntarily banned from the political lounge years ago.

Anyway,



Now: The problem with accounting for criminal background in any social study is that there is NO WAY you're going to get usable results from ANY population, ANYWHERE. It simply won't be possible, because as soon as the researcher is informed of a crime, he is legally required to report it pretty much in every first-world country, lest he be charged as an accessory... who in their right mind is gonna answer "yes" who hasn't already made a rap career out of it?

[NosniboR80]

Originally Posted by Spheric Harlot 

Whew.

Thanks, your post was a relief: we're on the same page, now.

Originally Posted by Spheric Harlot

If you have a gun, you are more likely to be shot and killed by one.

That is all.

Originally Posted by NosniboR80

If the study pinpointed Philadelphia gun-owners, then it certainly does NOT mean that Texas gun-owners are more likely to be shot by a gun.

So, are you saying that your statement should be changed to "If you have a gun in Philadelphia, you are more likely to be shot and killed by one in Philadelphia"?

That's better, more meaningful, and more accurate to the study, but it still doesn't address Mrjinglesusa and my problems with it.

Law abiding and responsible gun-owners might not be more likely to get shot by a gun at all. Other population groups are skewing the statistics.

[Spheric Harlot]

Originally Posted by NosniboR80 

So, are you saying that your statement should be changed to "If you have a gun in Philadelphia, you are more likely to be shot and killed by one in Philadelphia"?

No.

I strongly suspect that the result of this study is not limited to Philadelphia, and that other studies will bear this out.

"Same page" meant that we know what the other is saying, not that we necessarily agree. I probably should have said "same language". Sorry.

He knows what I'm saying.

You, I'm not so sure.

Originally Posted by NosniboR80 

That's better, more meaningful, and more accurate to the study, but it still doesn't address Mrjinglesusa and my problems with it.

Law abiding and responsible gun-owners might not be more likely to get shot by a gun at all. Other population groups are skewing the statistics.

That's pure speculation, and there are absolutely no facts gathered by the study referenced to back it up - or knock it down, for that matter.

It might just as well be that armed law-abiding citizens are so light-headed in their gun-carrying that they perceive a false sense of security that actually makes it MORE likely that they will provoke a situation that will get them shot, or that they will escalate an otherwise malignant but not lethal situation into something that will get them shot.

It might also be that criminals actually live SAFER with a gun, and that the gun REDUCES their risk of getting shot (though common sense and the findings about being able to defend yourself greatly increasing the risk of getting shot would appear to indicate otherwise).

What you believe, is, at this point, speculation. I'm inclined to believe the former, and not the latter, but that's because I think that "law-abiding" does not necessarily equate to "smart", and because it is my inherent bias to not like guns.

Hey, some people actually believe that yapping on a cellphone while driving doesn't deteriorate their driving skills. It's not like there's a shortage of people who overestimate themselves.


Again: WHY owning a gun makes it statistically more likely to get shot, and what factors might exempt people from the trend, or make them especially susceptible, is not the focus of this study, and it makes no assumptions about them.

[TheoCryst]

Originally Posted by The Final Dakar 

Can I get a pie chart up in here?

[moonmonkey]

Phew, i'm glad that's all sorted.

[mattyb]

You people did actually read the Objectives and Methods used for this report didn't you?

Objectives. We investigated the possible relationship between being shot in an assault and possession of a gun at the time.

Methods. We enrolled 677 case participants that had been shot in an assault and 684 population-based control participants within Philadelphia, PA, from 2003 to 2006. We adjusted odds ratios for confounding variables.

What they didn't do : they didn't ask those that hadn't been in a shooting who carried a gun what their experiences were. They didn't ask people that had been shot who didn't have a gun what their experiences were. They didn't ask the people that had been shot and were carrying a gun if they had any criminal convictions.

I just can't understand why people don't see the flaw in this study.

[Spheric Harlot]

Did you actually miss the entire discussion on this page?

[ghporter]

Originally Posted by Spheric Harlot 

Did you actually miss the entire discussion on this page?

I think the point is valid though. By constructing their study the way they did, they almost guaranteed that they'd come up with a muddied result. Instead, they could have (and SHOULD have) added an age/locale/demographic matched cohort of people who had NOT been shot to demonstrate or refute the connection they posited.

Without knowing a LOT more about their study subjects, it's impossible to determine whether they chose people with extremely dangerous jobs (couriers, armored car workers, even drug dealers) or people who lived in extremely dangerous locations in the city. And of course they didn't address WHY these individuals carried guns in the first place. Had they been shot previously and wanted some form of self defense?

Further, how much training had these people had in the use of any firearm they carried? A good training program can help make a potential shooting situation more manageable, and can also convince a potentially armed person that they just aren't cut out for the responsibility and necessary reactions involved in the process.

[NosniboR80]

Originally Posted by NosniboR80

Law abiding and responsible gun-owners might not be more likely to get shot by a gun at all. Other population groups are skewing the statistics.

Originally Posted by Spheric Harlot 

That's pure speculation, and there are absolutely no facts gathered by the study referenced to back it up - or knock it down, for that matter.

This is my entire point. The study, in my opinion, does not have much use precisely because it doesn't allow one to distinguish these kinds of factors.

If you are making a statement about the usefulness of owning a gun, it should apply to the type of person and their uses of that gun.

If my pistol is locked in my basement and only comes out when I go to the range, I am willing to bet a paycheck that my likelihood of getting shot is a lot closer to my situation if I didn't own a gun, than if I were a felon working in a dangerous neighborhood and concealed carrying my pistol all the time.

This study allows you gloss over that fact and make a statement that seems useless to me, even if it is statistically accurate.

How you store, carry, and train with your weapons are important factors in determining the likelihood of getting shot by one.

How you live your life, your job, your criminal background, and the areas that you frequent are also extremely important factors in determing the likelihood of getting shot.

THAT is all.

[mattyb]

Originally Posted by Spheric Harlot 

Did you actually miss the entire discussion on this page?

Yes, I did. Apologies.

[Spheric Harlot]

Originally Posted by ghporter 

I think the point is valid though. By constructing their study the way they did, they almost guaranteed that they'd come up with a muddied result. Instead, they could have (and SHOULD have) added an age/locale/demographic matched cohort of people who had NOT been shot to demonstrate or refute the connection they posited.

They DID!

Age: All ages.

Locale: Philadelphia.

Demographic: Philadelphia residents.


That this is too unspecific for you is a legitimate objection. But I've already pretty much described why this study looks as it does, and why it's pretty much impossible to do one that will satisfy the gun nuts, in my discussion with Mrjingleusa.

[mattyb]

The Demographic is in fact : 677 case participants that had been shot in an assault and 684 population-based control participants who live in Phil.PA

[Spheric Harlot]

That's the sample.

[Mrjinglesusa]

Originally Posted by Spheric Harlot 

Whew.

Thanks, your post was a relief: we're on the same page, now.

As I mentioned in the initial post, I was aware of the limitations of the study, and posted it merely for relevance to the side-line discussion.

I was starting to get mondo annoyed by people telling me what I they thought I was saying.

If I had the slightest interest in that kind of bullshit, I wouldn't have had myself voluntarily banned from the political lounge years ago.

Anyway,



Now: The problem with accounting for criminal background in any social study is that there is NO WAY you're going to get usable results from ANY population, ANYWHERE. It simply won't be possible, because as soon as the researcher is informed of a crime, he is legally required to report it pretty much in every first-world country, lest he be charged as an accessory... who in their right mind is gonna answer "yes" who hasn't already made a rap career out of it?

RE: the criminal background, I'm not sure how to accurately account for that but I just feel it needs to be done for a study like this to have any credibility. ghporter said it better than I did. There are too many "risk factors" that already would increase one's likelihood of being shot that were not addressed in the study, that is all.

[mattyb]

Originally Posted by Spheric Harlot 

That's the sample.

It is also the demographics used. They didn't use all Philly males as a demographic. They didn't use all Philly Police officers as a demographic, they used 677 case participants that had been shot in an assault and 684 population-based control participants who live in Phil.PA.

If I only ask criminals who have been incarcerated more than 10 years if they carried a firearm, the result would probably be quite high that they did. Would this show that carrying a firearm meant that you were more than likely to spend over 10 years in prison? Of course not.

[ghporter]

Originally Posted by Spheric Harlot 

They DID!

Age: All ages.

Locale: Philadelphia.

Demographic: Philadelphia residents.


That this is too unspecific for you is a legitimate objection. But I've already pretty much described why this study looks as it does, and why it's pretty much impossible to do one that will satisfy the gun nuts, in my discussion with Mrjingleusa.

Perhaps I was not clear enough. It was not clear to me that the non-shot control sample individuals were as thoroughly matched to the investigation sample as would be necessary to truly control the study. Did they live in the same neighborhoods? Were they in similar professions? What were they doing when the shot individuals were shot? In other words, was the control sample actually well matched, or did they take people from extremely different neighborhoods, or even completely different walks of life as their control sample?

I'm not willing to spend $30 to get the journal article so I can find this information, but it seems that it's at least possible that the control sample was not "really" a perfect match. For example, there are what might be called "expected" correlations between carrying a gun and a set of political, educational, and professional alignments, with another, rather different set correlated to NOT carrying a gun. These expectations are of course incredibly broad generalizations, but they're based on real world experience and statistical data. Just the political and professional factors are a major set of confounds that would be incredibly difficult to control for, despite the study authors' claims.

[Spheric Harlot]

Originally Posted by ghporter 

Perhaps I was not clear enough. It was not clear to me that the non-shot control sample individuals were as thoroughly matched to the investigation sample as would be necessary to truly control the study. Did they live in the same neighborhoods? Were they in similar professions? What were they doing when the shot individuals were shot? In other words, was the control sample actually well matched, or did they take people from extremely different neighborhoods, or even completely different walks of life as their control sample?

I understood what mattyb was asking, but it'd already been addressed the previous discussion with Mrjingleusa.

Specifically, concerning why these issues weren't analyzed in further detail:

Originally Posted by Spheric Harlot 

Question is, "Is there a statistically relevant correlative relationship between gun ownership and likelihood to get shot/killed by a gun?"

Studied population is "residents of city X".

"Experimental" group (in quotes because this is a field study, not an experimental set-up) is "gun owners" within studied population.

"Control" group is "non-gun-owners" within studied population.

Result: Correlation proved.

Limitations: No insight into actual nature of causative relationship. Further studies needed.

The end.



IN THIS STUDY, THERE *WAS* ONLY ONE VARIABLE because all others were eliminated.

THAT IS WHY the results are so broad.

Causation WAS NOT WITHIN THE SCOPE of this study.

As to WHY this was beyond the scope of this study - that's a real-world limitation of methodology and subject matter:

Originally Posted by Spheric Harlot 

Since you're a chemist, I'll cut you some slack.

You're used to magnesium being magnesium, molar masses being identical, and oxygen having a -2 charge.

In social sciences, it's kind of a little more difficult, because, like it or not, it's not only that the demographics of two groups can never be quite identical, it's that even if you break it down to the smallest possible level for analysis, you find that - oh horror! - THE DEMOGRAPHICS OF NO TWO INDIVIDUALS CAN EVER BE ABSOLUTELY IDENTICAL.



Oddly, the implications of this shouldn't be too difficult for you, a PhD in chemistry, to grasp, since everything you're used to as predictable "fact" is, on a physical level, nothing but probability. Dumping two reactants together doesn't mean that the individual atom is guaranteed to get oxydized/reduced, any more than the fact that you own a gun guarantees that you will be shot at some point.

In fact, even electrons are often enough just portrayed as probabilities within their orbitals, IIRC.

Still, you generalize. Is that "fact" enough for you?

and

Originally Posted by Spheric Harlot 

Now: The problem with accounting for criminal background in any social study is that there is NO WAY you're going to get usable results from ANY population, ANYWHERE. It simply won't be possible, because as soon as the researcher is informed of a crime, he is legally required to report it pretty much in every first-world country, lest he be charged as an accessory... who in their right mind is gonna answer "yes" who hasn't already made a rap career out of it?

In other words: We're probably never gonna get much more detailed studies than this one, as severely limited in its usefulness as it may be.

[ghporter]

Originally Posted by Spheric Harlot 

In other words: We're probably never gonna get much more detailed studies than this one, as severely limited in its usefulness as it may be.

Sadly, you're right. I can throw darts all day, but the authors get to stay published on this whether they were actually incredibly thorough and above reproach or the exact opposite.

The problems I cited, including using "citizens of Philadelphia" remain problems for this study. That's a big city, with a lot of different areas that are SO different as to be completely separate locales. If the authors didn't use extreme care in selecting control subjects who lived in close proximity to the study subjects, their control would indeed be worthless.

Meanwhile gangbangers continue to treat "having a gun" as if it were some magical talisman, and then using said gun so badly as to hit everything and everyone but their intended target. And adults so spooked by cultural hype around firearms will still fail to educate their kids to stay the heck away from any firearm they see, resulting in kids having horrendous accidents. If guns in the U.S. remain so mystical and relegated to the criminal-related world (bad guys or good guys), then bad things will continue to happen. If we were just less stupid about KNOWLEDGE, we might fix a lot of these problems-and more-and make real progress.

[mattyb]

And a bit more info about why banning guns doesn't work : Metropolitan Police records sharp rise in gang 'punishment shootings'

[turtle777]

Originally Posted by ghporter 

The problems I cited, including using "citizens of Philadelphia" remain problems for this study .

If that's the same folks that hang out at Paddy's Irish Pub, then yes, I can see how this MUST be twisted...

-t