Most data miners believe that there is a connection between gun ownership and gun deaths, and outside the US most believe that the connection is strong.
The previous KDnuggets poll asked
What is the connection between gun deaths and gun ownership?
Here are useful graphs and analysis of the gun ownership vs gun deaths data.
The poll results are summarized below and suggest that world-wide most KDnuggets readers believe that there is weak positive connection between gun ownership and gun deaths - more guns/capita lead to more deaths.
Outside the US most KDnuggets readers believe that the connection is strong.
There are vocal sceptics, mostly concentrated in the US, where about 8% believe in the opposite connection - more guns lead to less deaths.
|What is the connection between gun deaths and gun ownership?|
| Strong positive connection - more guns strongly increase gun deaths (113)
| Weak positive connection - more guns somewhat increase gun deaths (56)
| No connection (30)
| Not sure (11)
| Weak negative connection - more guns lead to somewhat fewer gun deaths (5)
| Strong negative connection - more guns lead to significantly fewer gun deaths (10)
If we assign a "Strong positive connection" vote a value of +2,
"Weak positive" : +1, Not sure and "No connection": 0, "Weak Negative": -1, and
"Strong Negative": -2, we get the following regional breakdown:
We note that overall KDnuggets readers believe that there is a positive connection
- US/Canada, 63% of voters, avg score: 1.0
- Europe, 23%, avg score: 1.25
- Asia, 8%, avg score: 1.56
- Other regions (Africa, Middle East, AU/NZ, Latin America), 6%, avg score: 1.64
- Overall, avg score 1.14
More guns lead to more gun deaths
and outside the US most data miners believe that the connection is strong.
The sceptics who doubt the connection or believe in negative connection:
More guns lead to less gun deaths
are most present in the US, where 8% believe there is a negative connection.
This data is also shown in the graph below.
(note: the average score was the same for US and Canada, so they were combined in this graph)
Here is the summary of my own analysis based on graphs and data for gun ownership vs gun deaths.
- Globally, there is no simple link between gun ownership and gun deaths - drug violence is the biggest factor.
- If we consider only advanced countries (GDP > US $20,000) there is a strong positive correlation (R2=0.66) between gun ownership and gun deaths
- Most gun deaths in advanced countries are due to suicides.
- If we examine only homicides in advanced countries, there is a weak correlation (R2=0.46), and US has much bigger homicide rate and gun ownership rate than its peers.
- Among advanced countries, excluding US, there is very little correlation (R2=0.04) between gun ownership and gun homicides.
Nordic dataminer, Gun ownership
I put this gun-index data and Census data by state together. The correlation of gun-index with murder rate is almost non-existent and some of the best predictive models don't even use gun-index parameter when you try to predict murder rate. People should really look this data themselves because it took me less than hour and now I KNOW what causes high murder rate.
It is unclear to me whether you really ask about correlation between gun ownership and deaths (answer: strong correlation with US, moderate correlation without US) or between gun ownership and homicides (answer: moderate correlation with US, no correlation without US). You should clarify the question. Or simply erase the graphs about homicides.
Al J, per capita vs % adult gun owners
the numbers for per capita gun ownership are
USA 89 per 100
Canada 31 per 100
What this means is clearly that an average US adult owns more guns than the average Canadian. Nearly 3 times more. (assuming only adults were surveyed in both countries ...) to me this is indeed a gun problem, more likely a gun obsession.
Ian, Yes Tom
You can clearly "SEE" a connection here, but what happens when you take a look at the numbers themselves? While it does appear that anyone selecting a negative correlation is pulling your chain, those selecting a 'weak correlation' are not in the wrong. In fact, depending on your statistical point of view, the correlation coefficient is low enough with the data provided (>20k GDP/C =~.8 w/US, ~.6 without US) that the correlation can most definitely be considered weak. Looking from the variance analysis point of view, the linear model shows some form of positive correlation at the 99.9% confidence interval, however it is quite small. Finally, correlation is not causation. There are too many other factors here to base anything on this data alone.
Causal Observer, And no, I don't have a gun.
Given how analytically irresponsibly the survey was written, the only justifiable answer is "No clear connection".
Yes, the graphs show correlations, but the survey asks us to attribute causality to the "connections", e.g., "more guns strongly increase gun deaths" or "more guns lead to significantly fewer gun deaths". What about psych drug usage (anti-depressants, Paxil, Ritalin, etc. often found in such murderers)? How does the US compare with other countries given such other possible causes? What about the Norwegian mass shooting, or inter-state comparisons? As is often the case, the design of this survey tells us more about the bias of its designer than providing useful insights. Don't you find that disappointing here, especially?
Ross Bettinger, More Guns ==> More Killings
It ought to be intuitively obvious to the most casual observer (IOTTMCO) that availability of deadly firearms is positively correlated with frequency of mass killings of civilians in non-wartime situations, e.g., most recently Newtown, CT, and other earlier incidents. There are enough mentally unbalanced people (insert psychographic profile here) who, driven by anger and unbalanced hormones, vent their aggressions on innocent victims and ultimately themselves.
The equation is simple: frequency of violent firearm use is positively related to frequency of access to weapons of mass close-range violence.
It is specious reasoning that the NRA proposes when it claims that "Guns don't kill people. People kill people." That statement must be corrected to "Guns have the potential to kill people, especially when possessed by people who who have intentions of killing people with firearms." But that won't fit into today's soundbite mentality.
Grashnakl, US/Canada Comparison
to Anonymous Analyst
Not sure where you're getting those figures. Quick googling suggests that according to the RCMP, there were 1.8 million licensed gun owners in Canada in 2010 - out of an adult population of ~28 million that means about 15% of adult Canadians owned guns.
The last figure I saw for US adult gun ownership was something like 47%.
Not saying you're definitely wrong - just saying you need to provide some numbers to back up that assertion.
Gregory PS, Editor, Percent of gun owners
Indeed, not all households in the US have guns, but statistics on households with guns are not as easily available worldwide as guns/capita. Where can you find such numbers?
In the US, about 30-34% of adults and 40-45% of households own a gun
Anonymous Analyst, Hmm.
Fact: Even though *per capita* rate of gun ownership in the U.S. is much much higher than it is in Canada, the *percentage of adult gun owners* isn't all that different. In other words, Canadians have just as much access to guns as Americans.
Fact: Gun violence in Canada is much lower than it is in the U.S.
Conclusion: What we have is not a gun problem, it's a cultural problem among a minority of gun owners.