PDA

View Full Version : Is Gun Control linked to Violent Crime?


Chris M
12-29-2009, 6:29 AM
This was posted in another thread, but thought I'd start a new one.

This chart shows a comparison of Gun Laws & Violent Crime Stats for each state.

The Green line indicates # of violent crimes per 2,000,000 people. (higher number = higher crime rate)
Data collected here. (http://en.wikipedia.org/wiki/Crime_in_the_United_States#States)

The Blue line indicates Brady Bunch state rating. (higher number = more gun control)
Data collected here (http://www.bradycampaign.org/stategunlaws/statescorecard/).

BTW - I gave D.C. the same Brady score as CA, for comparison purposes.

http://lh6.ggpht.com/_a-xP0_-akjE/SzlC_niuZqI/AAAAAAAAACk/YwHYq_R4_tE/chart.jpg

Coincidence? I think not.

vladbutsky
12-29-2009, 11:26 AM
As much as I would like to agree with you on the correlation, there is none there...
The problem here is in source data that has numbers from large states like CA or TX and small ones like RI put together into a single table given same level of significance. Also "average" crime rate is not a very useful number for anything. What does it mean to have "average" crime rate between Oakland or Richmond and safe rich cities like Los Altos or Palo Alto? There is no meaning for any of this.
Brady's score has nothing to do with science either so there is no point to try playing statistics to get numbers that we like. This is not a good game because it can be played any way you want or Brady wants. So if you will try to give it some sort of "scientific legitimacy", keep in mind that other side will use it too..

Chris M
12-29-2009, 11:38 AM
But, it is interesting to see that violent crime per capita is higher in those states that the Brady's deem as 'safer' due to having more restrictive gun laws.

Of course the data isn't perfect - it's coming from the Brady's and Wikipedia - not the most reliable sources - but nonetheless, it's still interesting...at least to me.

The fact the the data sets came from completely different sources, and the lines are nearly parallel tells me that there is some significance to it, and it's not just coincidence.

Hopi
12-29-2009, 11:42 AM
Without law genesis dates, it's hard to use that graph in a positive way. Couldn't it read that those areas with higher crime have more strict gun laws due to the efforts to reduce already high crime rates? Chicken and egg thing without the dates.

Rusty_Rebar
12-29-2009, 11:58 AM
I think someone could also argue that the reason there are so many gun laws in the higher crime rate states, is because of the crime rates.

So is it that the crime rate is so high in California because of the gun laws, or are there all these gun laws in California because of the crime rates?

There is really no good information on a coloration between tough gun laws and crime rates. It is apparently a difficult thing to normalize the data with so many factors involved.

Are the Police doing something different in how they investigate?

Are there less murders because the medical technology is keeping more people alive now?

Are the gun laws working?

Does the diversity of California lead to more gun crimes then say, North Dakota?

Maybe the bad guys would rather steal your stuff in sunny Los Angeles, then snowy Fargo?

Who knows? But I don't think abridging a constitutional right to run an experiment is really a great idea.

Chris M
12-29-2009, 12:08 PM
OK, I stand corrected. I wasn't thinking about the dates that the laws were put into place, and whether crime increased/decreased due to those laws. My bad.

madmike
12-29-2009, 12:18 PM
I think what the chart shows is that the Brady scores are the cause of all crime, and they should be banned by name...

-madmike.

hoffmang
12-29-2009, 12:57 PM
This may show that places with higher crime rates tend to be more open to vilifying gun ownership for whatever causes.

-Gene

LarryC
12-29-2009, 8:19 PM
Chris M - I think your analysis is good for discussion, especially among the unconverted. When I talk with my co-workers this would be an ideal topic; not because it answers every nuance of the relationship between firearms, crime and gun control, but because it casts the argument in a way that can be easily communicated and discussed over lunch (which is the length of time I have to make my arguments).

No, it will not stand up to publication standards of peer reviewed academic journals. However I would argue that half of the laws passed in this county use less evidence to support their passage than you provided.

Don't let the criticisms of other posters, no matter how appropriate, discourage you from pursuing this line of thinking. Keep building on your ideas using the observations from the responses to your post. I like your thinking.

navyinrwanda
12-31-2009, 4:52 PM
http://lh6.ggpht.com/_a-xP0_-akjE/SzlC_niuZqI/AAAAAAAAACk/YwHYq_R4_tE/chart.jpg



Would anyone know the source of the violent crime data in this chart? Or who produced this chart?

I'm trying to reproduce the results, and am having a hard time duplicating them with the data from the referenced Wikipedia (http://en.wikipedia.org/wiki/Crime_in_the_United_States#States) entry.

Thanks very much.

aileron
12-31-2009, 5:05 PM
Couldn't it read that those areas with higher crime have more strict gun laws due to the efforts to reduce already high crime rates? Chicken and egg thing without the dates.

This is along the lines of what I saw from it when looking at it....that it shows is high crime areas have lots of anti-gun laws.

Chris M
12-31-2009, 6:25 PM
Would anyone know the source of the violent crime data in this chart? Or who produced this chart?

I'm trying to reproduce the results, and am having a hard time duplicating them with the data from the referenced Wikipedia (http://en.wikipedia.org/wiki/Crime_in_the_United_States#States) entry.

Thanks very much.

I made the chart, and now that I'm double-checking it, it looks like something must have been seriously off in excel. I don't have access to the file right now, but when I do, I'll see what went wrong.

Basically, I was trying to adjust the ratio, in order make the numbers in the wiki more level with the numbers in the Brady scores.

For instance 500 violent crimes per 100,000 = 50 crimes per 10,000. The Brady's scoring system is 0-100...I wanted to adjust the ratio so that the majority of the crime data fell within that range.

RideIcon
03-04-2010, 12:30 AM
This may show that places with higher crime rates tend to be more open to vilifying gun ownership for whatever causes.

-Gene

I was thinking the same thing.

But stating that wouldn't be using their tactics against them. :p

racer98765
03-05-2010, 8:17 AM
I took the FBI's violent crime numbers for years 1999-2008 and the brady campaign's state ratings and was unable to reproduce his original graph. But with all the data at hand wanted to see what I could take from it. Going forward with the thought train that if gun legislation did work then a state with a high rating should have a low violent crime rate. So if you took the difference and the results was negative that would show that increased gun regulation has a corresponding lower violent crime rate. If it is positive then it shows that additional gun legislation does not lower the violent crime rate. You average all the differences to find the overall trend for the nation.

The algorithm is such:

State Violent Crime Rate per 100,000 = (State's Total Violent Crime Incidents x 100,000)/State's Population.

Normalize both the State Violent Crime Rates and brady campaign state ratings: Normalized Data = (Data-Min(Data))/(Max(Data)-Min(Data))
Now both sets of data range from 0 to 1 so you can do a direct comparsion, i.e. apples to apples.

Difference = Normalized State Violent Crime Rates - Normalized brady campaign state ratings

If (Difference)<0 then gun legislation does lower the violent crime rate

If (Difference)>0 then gun legislation does NOT lower the violent crime rate.

To find the overall trend for the nation you do the arithmetic mean to compute the "average": Mean Differences = Sum(Differences)/50 [Number of States]

If (Mean Differences)<0 then the overall trend for the nation is that gun legislation does lower the violent crime rate.

If (Mean Differences)>0 then the overall trend for the nation is that gun legilsation does NOT lower the violent crime rate.

I computed graphs for all the years I could get data for, 1999-2008. And the overall trend for each year was positive therefore showing that additional gun legislation does NOT lower the violent crime rate.

racer98765
03-05-2010, 8:20 AM
Graphs for years 2004-2008.

kap
03-05-2010, 8:47 AM
I made the chart, and now that I'm double-checking it, it looks like something must have been seriously off in excel. I don't have access to the file right now, but when I do, I'll see what went wrong.

Basically, I was trying to adjust the ratio, in order make the numbers in the wiki more level with the numbers in the Brady scores.

For instance 500 violent crimes per 100,000 = 50 crimes per 10,000. The Brady's scoring system is 0-100...I wanted to adjust the ratio so that the majority of the crime data fell within that range.

You can just plot one set of data on the secondary axis which would require no data transformation.

timdps
03-05-2010, 9:01 AM
Graphically, I think a double bar graph for each state would work better for the OPs table than a continuous line graph. There is no reason to link the data points from different states.

Nit has been picked...;)

Tim

inbox485
03-05-2010, 9:13 AM
There is an article similar to this from one of the Gun Rights Examiners (http://www.examiner.com/examiner/x-2879-Austin-Gun-Rights-Examiner~y2009m6d18-The-world-according-to-Brady). The correlation is pretty solid. The more violence there is in a state, the higher the Brady bunch rates the state for the gun control that gave way to the violence.

Flopper
03-05-2010, 9:24 AM
As much as I would like to agree with you on the correlation, there is none there...

Actually if the statistics are correct, there is by definition a correlation.

I think the word you were looking for was "causation."

IrishPirate
03-05-2010, 9:28 AM
I think this says it all

http://2.bp.blogspot.com/_oa9e6tcYcR8/SBKSshqnKUI/AAAAAAAADxk/bZL7kNjHILE/s400/gun_control_works2.JPG

kap
03-05-2010, 10:27 AM
Same Data as OP arranged by Brady Campaign score (High to Low). D.C. is rated a 95 in this data set based on estimates of their scoring criteria. The Brady's do not score D.C. I don't need to do a statistical test to see that there is no correlation between Brady Campaign scores and violent crime rates.

http://img242.imageshack.us/img242/5407/crimebrady.jpg

mtptwo
03-05-2010, 11:12 AM
I want to see population and median income overlaid on this graph.

BakerPD
03-05-2010, 11:15 AM
http://www.youtube.com/watch?v=ERj3QeGw9Ok&cc=1#t=22m48s

ALSystems
03-05-2010, 11:23 AM
There are a lot of good stats with graphs here:

http://www.gunfacts.info/

curtisfong
03-05-2010, 11:29 AM
I think a scatter plot of brady rating vs violent crime rate (omitting the states) would be more useful. Can somebody post a .csv of the data used in post 21?

timdps
03-05-2010, 11:37 AM
Very nice graph- thank you.

Would not mind seeing the inverse graph: order the states by violent crime rate and then add the Brady data points.

tim

Same Data as OP arranged by Brady Campaign score (High to Low). D.C. is rated a 95 in this data set based on estimates of their scoring criteria. The Brady's do not score D.C. I don't need to do a statistical test to see that there is no correlation between Brady Campaign scores and violent crime rates.

http://img242.imageshack.us/img242/5407/crimebrady.jpg

jdberger
03-05-2010, 11:40 AM
I love Calgunners..... :punk:

racer98765
03-05-2010, 11:58 AM
I want to see population and median income overlaid on this graph.

Population is already normalized by using Violent Crime Rate instead of using Violent Crime Incidences. Since Violent Crime Rate per 100,000 is computed via:

Violent Crime Rate per 100,000 = State's Violent Crime Incidents * (100,000/State's Population).

This will therefore allow the different states to be compared allowing the large differences in each state's population. I.e. a Apples to Apples comparsion.

Median Income is not included in any FBI's datasets I've seen. Also you shouldn't use data outside collected the two datasets used because your not sure how the data was collected and if you can properly compare and contrast the different datasets. You can use outside data, but that then requires additional number crunching and that usually goes beyond my comfort zone of how much massaging of the data is allowed and the numbers to actually be true.

kap
03-05-2010, 12:02 PM
I want to see population and median income overlaid on this graph.

Did that. I'll post it just for fun in a second. The real take home message is that Brady Scores have nothing to do with anything except how many gun laws on on the books.

kap
03-05-2010, 1:10 PM
I want to see population and median income overlaid on this graph.

http://img710.imageshack.us/img710/6963/bradystats.jpg

curtisfong
03-05-2010, 2:02 PM
A scatter plot of brady (x) vs all the rest of that last plot (y) (correlate by state) would be awesome!

kap
03-05-2010, 2:21 PM
A scatter plot of brady (x) vs all the rest of that last plot (y) (correlate by state) would be awesome!

Here you go. Linear trend lines tell the story.

http://img221.imageshack.us/img221/9756/scatter.jpg

curtisfong
03-05-2010, 2:28 PM
ROFL! It just doesn't get any better than that. Thanks kap!

jdberger
03-05-2010, 2:54 PM
Alrighty - explain it to someone who had to take Stats a couple times.....

kap
03-05-2010, 4:14 PM
Alrighty - explain it to someone who had to take Stats a couple times.....

Just look at the blue line. The blue line indicates the general trend of the data for Brady Score vs. Violent Crime. From this you can see that the lower the Brady Score (fewer gun laws) the lower the violent crime. Similarly, the higher the Brady Score (more gun laws) the higher the violent crime.

To be fair, the linear trendline is not a great fit for this data, but does give you a general idea of what is going on. If gun laws actually helped reduce violent crime you should see violent crime decrease as gun laws increase (i.e. the line would slant downward from left to right instead of upward). The take home message is that more gun laws (High Brady Score) do not correlate with a decrease in violent crime.

HondaMasterTech
03-05-2010, 10:39 PM
I believe the answer to the question in the thread title is; Yes, there is a link. More gun crontrol = more crime. Generally speaking.

five.five-six
03-05-2010, 10:49 PM
This may show that places with higher crime rates tend to be more open to vilifying gun ownership for whatever causes.

-Gene


true, but does it not also anecdotally mitigate the contention that restricting guns lowers crime... and to be fair, our antagonists have used far more questionable examples to buttress their position


To be fair, the linear trendline is not a great fit for this data, but does give you a general idea of what is going on.

IMO, a graph that was arranged by Brady score, rather than alphabetically would be more convincing. even better would be a 3-d grph going back several years

kap
03-05-2010, 11:47 PM
IMO, a graph that was arranged by Brady score, rather than alphabetically would be more convincing. even better would be a 3-d grph going back several years

I agree! All the graphs that I made are arranged by Brady Score.

Back data would be great, so would p-values. It's all a matter of time invested and what it would accomplish.

Anonymous Coward
03-06-2010, 7:11 AM
I stumbled across this website last week and found this interesting:
http://www.statemaster.com/correlations/cri_mur_wit_fir-death-rate-per-100-000

It shows the factors that correlate most with the firerarm deaths per capita. So between the factors they track the statistic about reading and writing levels correlates the most.

five.five-six
03-06-2010, 8:14 AM
I agree! All the graphs that I made are arranged by Brady Score.

Back data would be great, so would p-values. It's all a matter of time invested and what it would accomplish.


no, you are missing th point... the high brady scores would be on the left (or right) and be descending (or ascending)


if you are going to use questionable data correlations to change world opinion, learn from the masters

http://www.globalwarmingart.com/images/thumb/5/52/Carbon_History_and_Flux_Rev.png/401px-Carbon_History_and_Flux_Rev.png


back data, shmack data, just make up data that fits your conclusion, it's just as good

obeygiant
03-06-2010, 9:23 AM
no, you are missing th point... the high brady scores would be on the left (or right) and be descending (or ascending)


if you are going to use questionable data correlations to change world opinion, learn from the masters

http://www.globalwarmingart.com/images/thumb/5/52/Carbon_History_and_Flux_Rev.png/401px-Carbon_History_and_Flux_Rev.png


back data, shmack data, just make up data that fits your conclusion, it's just as good

:rofl:

obeygiant
03-06-2010, 9:28 AM
I stumbled across this website last week and found this interesting:
http://www.statemaster.com/correlations/cri_mur_wit_fir-death-rate-per-100-000

It shows the factors that correlate most with the firerarm deaths per capita. So between the factors they track the statistic about reading and writing levels correlates the most.

We're #1 (http://www.statemaster.com/graph/cri_mur_and_non_man-crime-murder-and-nonnegligent-manslaughter)

five.five-six
03-06-2010, 9:35 AM
We're #1 (http://www.statemaster.com/graph/cri_mur_and_non_man-crime-murder-and-nonnegligent-manslaughter)
statistics are interesting

Lynchings > Black 2 [28th of 44]
Lynchings > Total 43 [21st of 44]
Lynchings > White 41 [13th of 44]
(http://www.statemaster.com/state/CA-california/cri-crime)

kap
03-06-2010, 10:12 AM
no, you are missing th point... the high brady scores would be on the left (or right) and be descending (or ascending)

if you are going to use questionable data correlations to change world opinion, learn from the masters

back data, shmack data, just make up data that fits your conclusion, it's just as good

Aren't they? Left to right. Low to high. Ascending. Just like the glocal warming graph. This one is just not as pretty since it is real data. I have better things to do with my time than manipulate data. Showing this graph to someone who is only interested in the emotional feel-good aspects of gun control would be useless. Even if it was a firm correlation. They would just say we need to try other more creative things ... like banning all gun ownership.

Further, the correlation is not questionable, just weak. It certainly does not show any positive correlation between gun laws and a decrease in violent crime. That's a firm conclusion. Actually one of the best conclusions you can make from this graph is that gun control laws have nothing to do with violent crime rates at all. If you take out D.C. that line would be flat. That means the laws don't do squat except limit law-abiding citizens ability to own firearms.

kap
03-06-2010, 10:14 AM
Babies use guns to kill (http://www.statemaster.com/plot/cri_mur_wit_fir/hea_tee_bir_rat_per_100/flag).

curtisfong
03-06-2010, 10:41 AM
Actually one of the best conclusions you can make from this graph is that gun control laws have nothing to do with violent crime rates at all.

Exactly. Showing unbiased non-correlation is the best result, because that also happens to be the easiest to defend against gun control advocates. I will be adding this to my ammo can for future "discussions" with idiot anti-gunners.

Thank you, kap.

Mute
03-06-2010, 12:52 PM
You realize, at this point, only John Madden could explain all this.

curtisfong
03-06-2010, 12:54 PM
You realize, at this point, only John Madden could explain all this.

Unfortunately, sometimes fundamental truths cannot be properly expressed in 30 second sound bites or clever (?) talking point slogans.

dantodd
03-06-2010, 2:35 PM
Unfortunately, sometimes fundamental truths cannot be properly expressed in 30 second sound bites or clever (?) talking point slogans.

I have found that most fundamental truths can be expressed in much shorter sound bites that 30 seconds.

Zero
2+2 = 4
E=MC2
Relativity
Selective Evolution

It is highly likely that some people will not believe a fundamental truth based on a simple explanation. It is absolutely true that one can make an academic career out of expanding upon or delving into the minutiae of a fundamental truth but it does not take 30 seconds to explain one.

Centurion_D
03-06-2010, 3:33 PM
D.C with the highest crime rate? Nooo I'm shocked, I tell you shocked!!

N6ATF
03-06-2010, 4:44 PM
D.C with the highest crime rate? Nooo I'm shocked, I tell you shocked!!

It would be funny if it went from #1 to dead last after Heller. Incorporate THIS!

lioneaglegriffin
03-06-2010, 9:57 PM
It would be funny if it went from #1 to dead last after Heller. Incorporate THIS!

its crime rate has gone down...
an old article about it last year.
http://www.washingtonpost.com/wp-dyn/content/article/2009/07/19/AR2009071902154.html

and they can't figure it out. :confused:

typical. :rolleyes:
recent thing by john lott.
http://biggovernment.com/jlott/2010/03/01/more-guns-less-crime/

recent eitorial
http://www.washingtontimes.com/news/2010/mar/02/learning-from-the-dc-handgun-ban/

tada

curtisfong
03-07-2010, 1:26 PM
I have found that most fundamental truths can be expressed in much shorter sound bites that 30 seconds.


Perhaps I used the wrong word.

s/expressed/proven/g