Civilian gun possession rate for 90 most violent and 90 least violent nations on earth

[lawodevolution]

90 most violent nations: 5.78 guns per 100 civilians
90 least violent nations: 16.36 guns per 100 civilians

The top ten most violent nations average was 5.32 guns per 100 civilians
10-20 most violent nations average was 4.32 guns per 100 civilians

The top ten least violent nations average was 20.2 guns per 100 civilians
10-20 least violent nations average was 13.6 guns per 100 civilians

Violence is based on murder rate. I did not exclude very small nations unless data was lacking.

The data is listed and open to criticism, some may not like the gun possession rate data if you are used to the 2007 small arms survey.
For example, Canada's gun possession rate is underreported from gun-policy site but I took all the data from that source due to completeness.
You may check the math and report any math errors or report any errors in data, there may be errors because I copied the data more than once.

So far I have 180 countries plus about a dozen that lack gun ownership rate data. This is an ongoing project, so please feel free to comment.

----------------------------
Murder data was taken from
http://en.wikipedia.org/wiki/List_of_countries_by_inten...

and is trusted as correct unless another source indicates the government has a record of lying about murder data.
The following countries were excluded due to lack of gun possession rate data, although there may be more nations that lack data that I am not
aware of: Saint kitt and nevis, St V and Grenadines, St Lucia, Dominica, Sao Tome and Princepe, Grenada, Liechtenstein, Andorra, Monaco,
Anguilla, Bermuda, guam

Countries whose murder data was corrected with 2004 WHO data were nations that had a history of lying about murder rate that also had
either a real murder rate more than 3 times higher or less than 1/3 lower than what was found on the wiki list, although no nations
over-reported murder rate. Some of the nations were already correct on the wiki list. These nations are noted by placing after
the name.
http://www.unodc.org/documents/data-and-analysis/IHS-ra...

Gun possession rate per 100 was taken from this website
http://www.gunpolicy.org/firearms/region/pakistan

All data was taken from groups known to be anti-gun.

NULL HYPOTHESIS:
The difference in average rate of civilian firearm possession is not statistically different between the 80 most violent vs the 80
least violent nations



Raw Data organized from highest to lowest murder rate per 100,000 population:
1-5
El Salvador
Honduras
Jamaica
Guatemala
Venezuela
6-10
Cote d’lvoire
Trinidad and Tobago
Burundi
Lesotho
DR of Congo
11-15
Colombia
South Africa
Sierra Leone
Belize
Zimbabwe
16-20
Central African Republic
Rwanda
Sudan
Tanzania
Uganda 21-25
Panama
Zambia
Bahamas
Brazil
Dominican Republic
26-30
Haiti
Ethiopia
Guyana
Niger
Philippines
31-35
Kenya
Congo
Mozambique
Chad
Ecuador
36-40
Equatorial Guinea
North Korea
Cambodia
Burkina Faso
Guinea-Bissau
41-45
Malawi
Mali
Namibia
Nigeria
Barbados
46-50
Liberia
Cameroon
Eritrea
Gabon
Myanmar
51-55
Mauritania
Mexico
Papua New Guinea
Russia
Senegal
56-60
Benin
Gambia
Suriname
Swaziland
Togo
61-65
Nicaragua
Timor-Leste
Botswana
Comoros
Madagascar
66-70
Paraguay
Ghana
Bolivia
Cape Verde
Costa Rica
71-75
Kazakhstan
Nauru
Algeria
Lithuania
Indonesia
76-80
Malaysia
Antigua and Barbuda
Seychelles
Nepal
Mongolia
81-85
Kyrgyzstan
Bangladesh
Georgia
Sri Lanka
Iraq
86-90
Moldova
Estonia
Jordan
Pakistan
Kiribati
91-95
Thailand
Uruguay
Belarus
Argentina
India
96-100
Cuba
Ukraine
Laos
USA
Latvia
101-105
Albania
Bhutan
Turkmenistan
Yemen
Palestine
106-110
Mauritius
Turkey
Montenegro
Afghanistan
Djibouti
111-115
Peru
Somalia
Uzbekistan
Syria
Iran
116-120
Libya
Fiji
Maldives
Armenia
Finland
121-125
Israel
Bulgaria
Romania
South Korea
New Zealand
126-130
Azerbaijan
Serbia
Czech republic
Vietnam
Belgium
131-135
Canada
Bosnia and Herzegovina
Marshall islands
Macedonia
Slovakia
136-140
Chile
Tajikistan
Tunisia
Croatia
Solomon Islands
141-145
Malta
Luxembourg
Kuwait
Hungary
Ireland
146-150
France
Australia
Singapore
UK
Poland
151-155
China
Portugal
Cyprus
Italy
Samoa
156-160
Greece
Japan
Denmark
Iceland
Qatar
161-165
Tonga
Netherlands
UAE
Micronesia
Palau
166-170
Saudi Arabia
Spain
Vanuatu
Sweden
Germany
171-175
Bahrain
Egypt
Switzerland
Norway
Oman
176-180
Lebanon
Austria
Slovenia
Brunei
Morocco

----------------

Preliminary interpretation of data: The 180 nations will be split into two groups, the most violent half vs the least violent half.
The dividing point is at 7.3 murders / 100,000 population. Group A will have a higher murder rate than 7.3 and group B will have
a lower murder rate than 7.3

Average gun possession rate of the 90 most violent nations:
Total = 462.4
Average is 5.78


Average up the gun ownership rates of the 90 least violent nations:

total is 1309.31
Average is 16.36


I am working on the t test and will repost when I'm done. It's a lot of work.

[MyrnaLoy]

1. wrong forum

[Remmah2]

45. How can it be the wrong forum?

The forum is about politics, law and policy concerning firearms.

An absolute correct forum.

[lawodevolution]

52. When this person posted it was in the general forum

[whoneedstickets]

2. Look up the concept of spuriousness

There is likely a third variable that explains both gun ownership and lack of violence. (I'd guess development level)

Also instead of sorting the countries arbitrarily between high a low, try a multi-variate linear model with violence as the dependent variable and include gun ownership, HDI, per-capita gdp etc..

Gun ownership will likely not be a significant factor in a fully specified linear model.

[Aerows]

12. I have to question some of the data, as well

There are violent revolutions going on in some of those countries, yet their murder rate is low? I consider it pretty damn violent when people are gunning each other down in the street. And Somalia? Most dangerous country on earth, and it is listed as having a low murder rate? Okay, you might not get "murdered" outright, but "disappeared" isn't exactly an improvement over being murdered!

[lawodevolution]

14. That's probably true but I used the best data available, if you can find a reliable source

For murder data please link it. If I found the self reported data to be significantly different than WHO data I used the WHO data instead.

[kenny blankenship]

3. Yeah. Somehow I doubt the Zetas & pals in Mexico are reporting all their guns to the authorities

That is, if there are any authorities still down there to report to. Same goes for homicide capital numero-uno-with-a-bullet, Colombia.

[Pholus]

4. Correlation is not causation. Thanks for playing!

I could probably conclude FAR more strongly from the same country list that the murder rate is linked to the average income level of the population. The only countries in your bottom 90 that are even starting to leave "third world" status are:

Brazil, Mexico and Russia/Georgia.

T tests are meaningless the way you intend to apply them.

[lawodevolution]

5. I know some people will not like the results for political reasons

T test will show that the two means are significantly different or not.
This is honest data that you are welcome to check and comment on.

It is plausible that gun possession rate may effect violent crime rate but it is also possible that it has no effect.

[Pholus]

7. Political reasons? Not so much. Statistical reasons? Yes, there is MUCH to dislike.

You're attempting to link two variables without considering any other effect. One could easily ask if other factors exist which drive both variables rather than assuming some kind of link between the two.

Statistics are darned funny that way and easy to misuse. See for example (as a funny example of
abusing correlation):

http://xkcd.com/882/

[lawodevolution]

9. You may not think there is correlation but anti gun groups do think there is

So you are saying that murder rate has nothing to do with gun possession rate and the attempts to disarm civilians around the world or in other words "small arms proliferation" is futile? Most people think gun ownership rate is connected to rate of violence.

There is nothing wrong with this study. It is very simple. Order nations by rate of murder and compare the most violent half vs the least violent half based on gun possession rate. If you conclude there is no correlation that's fine.

[Pholus]

13. Your linear regression coefficient is only -0.25.

That number speaks for itself. Loudly.

[lawodevolution]

16. I think the 3 times higher average gun possession rate between the two

Groups speaks loudly. This is all the data we have available and that we have to work with.

Are you suggesting that the common opinion that more guns equals more violence is still true after looking at this data?
If not, you don't have to try to wow us with you statistics knowledge because I agree that there may be no correlation between gun possession rate and murder rate.

[Pholus]

18. I wouldn't make any kind of conclusion from this dataset on a bet.

THAT is ALL I am saying.

[lawodevolution]

20. So you would like to just delete this data because you dont like the results

Or if I'm wrong you can answer this:

Based in the data or what you think in general do more guns equal more crime, less crime or is there no correlation between murder rate and civilian gun possession rate?

[Pholus]

21. No. I am saying that for your goal, you have more work to do.

Yes, I would suppose you'd like me to render an unsupported opinion so I can be dismissed but guess what you're not getting one.

I am saying that your data do not have sufficient statistical power to make any kind of conclusion. Your methodology is incomplete in that it only contains two variables, your weighting scheme is highly suspect and even plotting what you have the scatter dominates any trends you might think you see.

Right or wrong, I'm saying you have a lot of work to do before you can cite your work product in arguments.

Edit: I ALWAYS mess up on the plural of data.... "data does" changed to "data do"

[Pholus]

6. Furthermore, your own dataset shows exceptional scatter between your variables.

The one that kind of glares at ME is:

Laos 5.2/100,000 murder rate; 1.2 guns/100 people
USA 5.0/100,000 murder rate; 88.8 guns/100 people

Modeling the murder rate as a simple poisson process implies that these two numbers are
statistically identical in murder rate but VASTLY different in gun ownership.

Furthermore, you should weight by country population instead of over country average. On the other hand, that means that China and India would really drive the numbers in a direction that you probably don't want to see your numbers go.

[lawodevolution]

8. There may be data points to the extreme on either side but we are more interested in the average

Trend and if the data is scattered like you said it will fail the t test.

[The Magistrate]

22. Another Interesting point, Sir, Is Variation In Murder rates Over Time

The countries with the highest murder rates show great fluctuation over the last decade. Many just about double their murder rates in that period, a few nearly halved their murder rates. No data is supplied indicating an fluctuation in gun supply in these places over that time, which suggests other influencesat work....

[Pholus]

27. An excellent point!

The main conclusion being it would take a far more comprehensive study than this to actually make a solid conclusion!

[joshcryer]

10. Will be moved to the Gun forum, but same analysis on "standard of living" will...

...provide just as... uninteresting results.

[lawodevolution]

11. The results are very interesting because it shows that more guns does not equal more violent crime

[Pholus]

15. The results are basically meaningless, not interesting. Either way.

I know you really want this to be this big logical statistical argument on your behalf, but the numbers just aren't there to conclude what you want to conclude.

There are lies, damn lies, and statistics.

[lawodevolution]

17. Do you think that more guns equals more violence, less violence or there is no correlation?

[Pholus]

19. The data is inconclusive. Period.

No conclusion can be drawn from this set, organized as it is.

[lawodevolution]

23. If you shake up the acceptable truth set forth by Faith based reasoning you get some interesting

reactions. Charles Darwin's theory had trouble with certain faith based groups. I'm sure any evidence he presented was also declared to "not correlate" and to be "absolutely useless to make any conclusions"

Are human beings really more enlightened than the middle ages or do we just have a different set of faith based reasoning today that is just as inflexible and unable to adapt in the presence of evidence to the contrary?

At least no one is trying to delete the thread or hide the data or send it to a less viewed forum.

[lawodevolution]

24. They hid my thread and told me I violated copyright law without explaining

how. I was told to edit it. I'm sure this had nothing to do with them not wanting this information to be seen.

[Pholus]

26. Actually, there is a LARGE difference between Darwin's studies and your own.

Please do not think for a moment that you have much of a similarity with Mr. Darwin at this point. Let me explain:

Darwin:

Performed one experiment that showed a premise (dramatic difference between the robustness of hybrids vs. self fertilized wildflowers). Didn't initially believe his own results. Repeated the experiment with several other types of plants. Realized that alternate factors (soil, light, water could cause differences) could cause the difference. Controlled for the factors by carefully planting pairs of hybrid/self-fertilized plants together so their growth experiences were identical. Still saw the effect.

Talked carefully to other scientists, namely statisticians. Realized that despite years of trials his statistical sizes were probably too small to be conclusive (Still a 5% chance of random chance governing the result from years of experiments). After decades, passed the work off to others. As the field of statistics progressed over the next half century it was possible to go back to his original work, realize that the original statisticians underestimated how important his controls were and how experimental design reduced the effects of random chance on the result. Redoing the analysis from the carefully compiled data showed that the chance of randomness was actually closer to 1 part in 10000.

You:

Apparently started with "truthiness" in your gut -- you were not trying to answer a question you were trying to support a view you already held. You threw together some numbers you got via the magic of Google and concentrated only on a single pair of variables that you felt would "prove" your premise. You did not consider for one instant that the source of the data or any variables which could bias your conclusions. What followed was an oversimplistic analysis whose only conclusion is questioned by a simple linear regression coefficient -- a technique taught in 10th grade statistics classes. When confronted with these criticisms you play the "martyr card."

Let's be PERFECTLY clear here -- at no point in the process have I criticised your hypothesis. On the other hand, your attempt to misappropriate a scientific mantle of authority through a flawed statistical study DEMANDS a strong response. Based on your inability to respond on the merits of my criticisms we both know who needs to go back to the drawing board.

That is not persecution, that is peer review.

[lawodevolution]

31. I took all the available data from all nations and found that

The most violent 80 nations have 1/3 the gun possession rate than the least violent 80 nations.

You can't check the data now that the moderators forced me to remove the data but thus is the result and you can try to deny it as much as you want but it's still the result.

[Pholus]

36. I examined your flawed data just now. You want factoids? Here's a few.

If I look at your 90 low murder rate countries and select only the highest rates of civilian gun ownership (>= 35 guns/100 citizens) then the countries can be classed as follows:

1) 29% of them have Islamic, highly repressive governments.
2) 14% of them have engaged in massive genocide against their own people in the recent past.
3) 43% of them are the U.S. or stable European countries which have not faced
significant threats to their existence in the recent past nor have they had
armed insurrections in the last 100 years which have changed their governments.
4) The balance made a peaceful transition from dictatorship to republics without
armed insurrection but are still not NICE places to live if you want to be left
alone.

So my conclusions from YOUR data are:

More than 40% of high gun ownership countries are oppressive towards the citizens owning the guns yet the guns have made no difference in resolving the situation. This flies in the face of the faith-based canard that gun ownership is some kind of magical talisman that protects the citizenry from government.

Am I cherry picking? Certainly. Would I back my own conclusion? Not a chance. But then again I understand the limits of your data.

Unlike you, apparently.

You think your number is kosher, fine. But then mine is too.

[jpak]

25. wrong forum - IBTL

yup

[Logical]

28. But the USA his a high gun rate and a high murder rate compared to modern countries. n-t

[lawodevolution]

29. "3rd world countries" are modern also

[Logical]

30. So you deny we have a high murder rate? n-t

[lawodevolution]

33. Compared to the world it is not high

[Logical]

34. Oh, it totally is high compared to the other industrialized nations....

We are number ONE in murder among industrialized nations.
We are NUMBER ONE!!!

United States 8.40
Canada 5.45
Denmark 5.17
France 4.60
Portugal 4.50
Australia 4.48
Germany 4.20
Belgium 2.80
Spain 2.28
Switzerland 2.25
Italy 2.18
Norway 1.99
United Kingdom 1.97
Austria 1.80
Greece 1.76
Sweden 1.73
Japan 1.20
Ireland 0.96
Finland 0.70

[Euromutt]

40. Last time I looked, Russia was industrialized

The Baltic republics are developed enough to have joined the European Union several years ago. But somehow, none of those countries figure into your claim.

It looks suspiciously like you're wielding a definition of "industrialized nation" that excludes any country that has a higher homicide rate than the U.S.

[lawodevolution]

41. I don't discriminate between nations, all nations are modern to some degree

We need to stop promoting the idea that some nations are better than others just because some are more wealthy or cleaner

[lawodevolution]

32. Moderators edited my edited post

I explained that they hide this thread and pu an edit request on it because of copyright ye did not explain why. I assume it was the list of rate of violence and gun possession rate data and I removed it. I posted this explanation in the OP but it was removed by them

[Tuesday Afternoon]

35. if correlation does not show causation then it works both ways and

the antis can not use it as evidence or proof as a reason to take away guns. cuts both ways.

[Pholus]

37. Unless your data set is bogus for the question you're asking in the first place.

Case in point -- this entire thread.

[Tuesday Afternoon]

43. case in point almost every thread started by anti.

cuts both ways. if you can't use data to prove it one way then you can NOT use data to prove it the other way.

[lawodevolution]

44. You're worried about this info getting out aren't you

You're on this thread like white on rice

[Pholus]

47. Yes, that's it. I am soooooo busted.

Go ahead, make your little study. It's plain you are inoculated against an honest argument.

I expected no less once this thread ended up here.

My next comment to you is prohibited by board rules. That tells me the usefulness of this is over.

[lawodevolution]

50. I guess you want me to fudge the data and cherry pick the countries included so as to make the guns

and gun owners look bad. Sorry.

[gejohnston]

38. I think everyone is reading too much in to it.

While the critiques may be far if Law intended to be a comprehensive scholarly work to develop a larger theory in criminal science. I don't think that was his goal. The impression I get is that his goal was simply to show the absurdity of the anti's claims using their own data. There was no larger question being asked.

[Pholus]

39. No, the problem was the dataset was inadequate for the question.

I won't repeat the exact problems endlessly, though apparently I need to.

It doesn't matter if you think the question is too simple to require a "comprehensive scholarly work to develop a larger theory" when in fact the question asked was not answered by the analysis chosen. What you seem to think is needless complication on a simple question, I call due diligence.

It doesn't matter if your preconceptions lead you to believe that "absurdities" lie in the data. In fact, an honest broker realizes that their preconceived notions INCREASES the burden to do a careful well thought out analysis so that your conclusions are unassailable. In case this it means that criticisms are taken as ways you need to improve the analysis, not as an impetus for crying about how you're being censored by "those people."

And another point, this discussion passed beyond the realm of honest discussion the moment the terms "pro" or "anti" are applied with the expectation of any influence on the argument whatsoever. The data are what the data are. The question and analysis either stand up to scrutiny or they don't. They don't here and THAT is the end of that.

And before your inevitable protest this is NOT an indictment of your belief system, this is merely calling shenanigans on an attempt to throw scientific weight behind what is fundamentally an unscientific endeavor.

If you start something like this, do it right or don't do it at all.

Nuff said.

[lawodevolution]

42. You can't get more simple than murder rate vs gun possession rate

[Pholus]

49. Yes. That's the point. You cannot prove that other factors are not contributing.

But you never *really* were interested in figuring that out. You want an outcome, not a study.

[Tuesday Afternoon]

51. that is bogus.

"You cannot prove that other factors are not contributing."

he doesn't have to prove that.

death by gun is what caused the death.

are you speaking to motive?

[Tuesday Afternoon]

46. sensei -- please show us the way

I wait to be enlightened......................

[Pholus]

48. I already did.

Read the thread.

This "study" is not being used for illumination, only for support. It's a waste of both our times to continue.

[lawodevolution]

53. What pholus wants is for me to cherrypick the white countries and add in suicides and other data

In such a ways to to produce a deceptive appearance that guns are a problem. Just showing gun possession rate vs murder rate for all nations doesn't help the anti gun cause

[iverglas]

54. whatever the "anti-gun cause" is ...

Just showing gun possession rate vs murder rate for all nations doesn't help the anti gun cause

... did somebody say it did?


Since you appear to be making a rather unpleasant allegation against the person to whose post you replied, you might want to alert him/her by PM, given that this thread had sunk into oblivion before you resurrected it just now and s/he won't have any notice of its rebirth.

[lawodevolution]

55. He's probably trying to edit the wiki document I linked to

;)

[Taitertots]

56. You are using flawed methodology

You are doing nothing to adjust for poverty rates, education, political stability, gun laws, and police enforcement. Your results will be illusory correlation at best.

[lawodevolution]

57. No, I'm not changing the data or cherry picking nations to make it look like guns are bad

That's what anti gun groups do. They like to correct for this or that so that they can make guns look bad. This is straight up honest, all nations compared and treated equally.