Jump to content

Archived

This topic is now archived and is closed to further replies.

TheSanDiegan

A DIY Study on Correlating Gun Violence Everyone Can Do

Recommended Posts

As many have suggested, it is critical for us as a culture to understand the root causes behind the increasing trend of mass shootings and gun violence at large in our country.

While a majority of gun-related deaths occur in large population centers (and then are skewed to poorer neighborhoods within these cities), all this tells us is that i) where you have more people you have a greater number of violent crimes, and ii) that gun-related crimes are more likely to occur within poorer neighborhoods. 

Neither conclusion is very useful. Maybe that's because correlating the number of crimes to the number of people is a "water is wet" exercise - the more people you have, the more crimes that will be committed. However, there is a seedy side to the adherence to the city-stats approach, as poorer neighborhoods of population centers are disproportionately skewed heavily towards minority populations. Thus, there is an inherent risk in inadvertently connecting the elevated crime rates in socioeconomically disadvantaged neighborhoods with the ethnicity of their residents.

By contrast, a far better metric is crime rate. And in the context of this discussion (and the many preceding it following this weekend's shootings), gun death rates in particular.

So for shits and giggles (admittedly light on the 'giggles'), I listed in a spreadsheet the ten states with the highest gun death rates, and the ten states with the lowest gun death rates. I figured by limiting this to the bottom 20th percentile and the top 80th percentile, I would see more meaningful patterns. Of course, the larger the data set the better, and IMO there is value in rolling this out to cover all 50 states. 

Anybody can repeat this, and the results will be the same.

I then made a half-assed attempt to identify correlates: based on the concern regarding city data, I thought I would look at racial demographics in these twenty states, and in particular, the respective percentages of both white and black residents. I then logged each state's ranking in median income, gun ownership rates, and gun control stringency (as ranked by the GIfford Law Center).  

Here's what the table looks like:

GDRCM.jpg

Using the spreadsheet's tools, I measured the correlation between gun death rates (dependent variable) and in turn, median income, gun ownership rates, gun control law stringency, and racial demographics. 

The highest correlation coefficient was gun ownership (.89), but this seems like another "no shit" conclusion: the more guns there are, the more gun-related crimes occur. To me, this is almost as useless as the population-based correlation mentioned in the second paragraph above.

The second-strongest correlation was the stringency of gun laws, which produced a positive correlation coefficient of .88.

The third-strongest correlation was median income, at .78.

Racial demographics on the other hand showed little to no statistical correlation (.08 and .39, respectively). This, in my opinion, was as important as anything else, as it helps provide context to the possible misuse and misinterpretation of the known correlation between gun violence and gross population and to socioeconomic factors within those population centers. This is supported by the much higher coefficient describing the correlation between median income and gun death rates. 

The nice thing is, while a formal study requires a working understanding of statistical analysis (and a lot more work than went into this little exercise), anybody with a spreadsheet can still repeat this effort and perform a cursory analysis of the data, as every modern spreadsheet app provides a bevy of analytical tools that perform all the work for you.

*****

My takeaways: 

Let me predicate this by saying I am a gun owner and I enjoy punching paper with both handguns and long arms. I have friends who are LEOs (well, one former and one current), and enjoy recreationally shooting with them. I am a staunch supporter of 2A rights. However, I am also among the vast majority of Americans who support universal background checks and using those to close gun show loopholes. I'm somewhat ambivalent about hi-cap mag laws, but am adamantly opposed to a ban of an entire class of firearms just because they look scary.

That being said, the data speaks for itself: the more stringent the gun laws in a state, the less likely you are to become a fatality victim of a gun-related crime. A correlation coefficient of .88 is very high, and speaks to a strong, positive correlation between the two variables. Admittedly, the correlation coefficient for gun ownership rates is even higher, but IMO the take-away from that is not as actionable.

In the past, I have referenced studies which demonstrated a lack of correlation between city violence and state gun control laws. However, I believe this to be a flawed approach for two reasons, as studies show that gun crime rates are much higher in large population centers regardless of their regional location. Second, as mentioned above, discrepancies in gun control laws are pretty much the domain of the states, so it only makes sense to compare state-to-state, not state-to-city or vice versa.

I am not advocating for a particular form of legislation or law other than the universal background checks and closing gun show loopholes. But the data - at least in this cursory and admittedly limited study - definitely shows that the more stringent a state's gun laws, the lower the probability of becoming the fatality of a gun-related crime.

I hope this fosters a little discussion.

St-Javelin-Sm.jpgChase.jpg 

Link to comment
Share on other sites

I think the correlation between median income and gun death rates would also account for the discrepancy between the two ethnic groups above given the measured disparity in median income between respective black and white populations in this country. 

St-Javelin-Sm.jpgChase.jpg 

Link to comment
Share on other sites

23 minutes ago, AndroidAggie said:

Needs geospatial correlation for perp and victims cross referenced with median income for zip code

Jesus ... you forgot the flex capacitor.

bsu_retro_bsu_logo_helmet.b_1.jpg

Link to comment
Share on other sites

2 hours ago, TheSanDiegan said:

As many have suggested, it is critical for us as a culture to understand the root causes behind the increasing trend of mass shootings and gun violence at large in our country.

While a majority of gun-related deaths occur in large population centers (and then are skewed to poorer neighborhoods within these cities), all this tells us is that i) where you have more people you have a greater number of violent crimes, and ii) that gun-related crimes are more likely to occur within poorer neighborhoods. 

Neither conclusion is very useful. Maybe that's because correlating the number of crimes to the number of people is a "water is wet" exercise - the more people you have, the more crimes that will be committed. However, there is a seedy side to the adherence to the city-stats approach, as poorer neighborhoods of population centers are disproportionately skewed heavily towards minority populations. Thus, there is an inherent risk in inadvertently connecting the elevated crime rates in socioeconomically disadvantaged neighborhoods with the ethnicity of their residents.

By contrast, a far better metric is crime rate. And in the context of this discussion (and the many preceding it following this weekend's shootings), gun death rates in particular.

So for shits and giggles (admittedly light on the 'giggles'), I listed in a spreadsheet the ten states with the highest gun death rates, and the ten states with the lowest gun death rates. I figured by limiting this to the bottom 20th percentile and the top 80th percentile, I would see more meaningful patterns. Of course, the larger the data set the better, and IMO there is value in rolling this out to cover all 50 states. 

Anybody can repeat this, and the results will be the same.

I then made a half-assed attempt to identify correlates: based on the concern regarding city data, I thought I would look at racial demographics in these twenty states, and in particular, the respective percentages of both white and black residents. I then logged each state's ranking in median income, gun ownership rates, and gun control stringency (as ranked by the GIfford Law Center).  

Here's what the table looks like:

GDRCM.jpg

Using the spreadsheet's tools, I measured the correlation between gun death rates (dependent variable) and in turn, median income, gun ownership rates, gun control law stringency, and racial demographics. 

The highest correlation coefficient was gun ownership (.89), but this seems like another "no shit" conclusion: the more guns there are, the more gun-related crimes occur. To me, this is almost as useless as the population-based correlation mentioned in the second paragraph above.

The second-strongest correlation was the stringency of gun laws, which produced a positive correlation coefficient of .88.

The third-strongest correlation was median income, at .78.

Racial demographics on the other hand showed little to no statistical correlation (.08 and .39, respectively). This, in my opinion, was as important as anything else, as it helps provide context to the possible misuse and misinterpretation of the known correlation between gun violence and gross population and to socioeconomic factors within those population centers. This is supported by the much higher coefficient describing the correlation between median income and gun death rates. 

The nice thing is, while a formal study requires a working understanding of statistical analysis (and a lot more work than went into this little exercise), anybody with a spreadsheet can still repeat this effort and perform a cursory analysis of the data, as every modern spreadsheet app provides a bevy of analytical tools that perform all the work for you.

*****

My takeaways: 

Let me predicate this by saying I am a gun owner and I enjoy punching paper with both handguns and long arms. I have friends who are LEOs (well, one former and one current), and enjoy recreationally shooting with them. I am a staunch supporter of 2A rights. However, I am also among the vast majority of Americans who support universal background checks and using those to close gun show loopholes. I'm somewhat ambivalent about hi-cap mag laws, but am adamantly opposed to a ban of an entire class of firearms just because they look scary.

That being said, the data speaks for itself: the more stringent the gun laws in a state, the less likely you are to become a fatality victim of a gun-related crime. A correlation coefficient of .88 is very high, and speaks to a strong, positive correlation between the two variables. Admittedly, the correlation coefficient for gun ownership rates is even higher, but IMO the take-away from that is not as actionable.

In the past, I have referenced studies which demonstrated a lack of correlation between city violence and state gun control laws. However, I believe this to be a flawed approach for two reasons, as studies show that gun crime rates are much higher in large population centers regardless of their regional location. Second, as mentioned above, discrepancies in gun control laws are pretty much the domain of the states, so it only makes sense to compare state-to-state, not state-to-city or vice versa.

I am not advocating for a particular form of legislation or law other than the universal background checks and closing gun show loopholes. But the data - at least in this cursory and admittedly limited study - definitely shows that the more stringent a state's gun laws, the lower the probability of becoming the fatality of a gun-related crime.

I hope this fosters a little discussion.

Not bad. I agree with most of what you articulated and appreciate your effort to compile some data. 

bsu_retro_bsu_logo_helmet.b_1.jpg

Link to comment
Share on other sites

4 minutes ago, BSUTOP25 said:

Jesus ... you forgot the flex capacitor.

I think where it happens is a very interesting factor

It'd probably help reveal how much is due to gang and low class violence, well to do middle class people whom I believe typify the murder suicide, and "let's go somewhere highly populated and shoot people because reasons" 

Link to comment
Share on other sites

Well done on this analysis, but obviously it's flawed because "mOrE guNs make foR a pOLiTe sOciEty" and defensive gun uses are like Jesus himself, amirite?

 

EDIT: Oh, and "shoOtiNgs OnlY haPpen iN guN-freE zoNes".  

Link to comment
Share on other sites

4 minutes ago, AndroidAggie said:

I think where it happens is a very interesting factor

It'd probably help reveal how much is due to gang and low class violence, well to do middle class people whom I believe typify the murder suicide, and "let's go somewhere highly populated and shoot people because reasons" 

Agreed. 

bsu_retro_bsu_logo_helmet.b_1.jpg

Link to comment
Share on other sites

1 minute ago, Orange said:

Well done on this analysis, but obviously it's flawed because "mOrE guNs make foR a pOLiTe sOciEty" and defensive gun uses are like Jesus himself, amirite?

Did your shift end already or is this just a smoke break?

bsu_retro_bsu_logo_helmet.b_1.jpg

Link to comment
Share on other sites

What's fascinating to me is you can throw all those correlates out the window when it comes to mass shootings (with 4 or more fatalities), as evidenced by this chart of mass shootings from 1982:

Screen-Shot-2019-08-06-at-2-00-15-PM.png

At first glance it appears tied to population dynamics - seven of the ten most populous states are among the top ten in that chart. Maybe gross numbers and ethnic and/or socioeconomic diversity? :shrug: Given it is the horrific nature of these mass shootings, I think this is the area that merits a deep-dive, inclusive of Android's suggested geospatial correlation.

St-Javelin-Sm.jpgChase.jpg 

Link to comment
Share on other sites

1 minute ago, Orange said:

Don't worry "Shoot the lice-infested homeless guy with a dog" didn't make my list.

Thanks. Always appreciate it when delusional people on the precipice of a rampage leave me off their revenge lists.  

Link to comment
Share on other sites

The white population in the U.S. has a lower murder and violent crime rate than England.  They are also the population in this country that owns the most guns.

Guns are not the problem.

Democratic big city policies that cause dependence on welfare, unemployment and poverty are the problem.

Link to comment
Share on other sites

5 minutes ago, bluerules009 said:

The white population in the U.S. has a lower murder and violent crime rate than England.  They are also the population in this country that owns the most guns.

Guns are not the problem.

Democratic big city policies that cause dependence on welfare, unemployment and poverty are the problem.

Within the ten states with the highest rates of gun-related deaths and the ten states with the lowest rate there is approximately zero correlation between the percentage of white people that live in a given state and the number of gun-related deaths that occur there. 

Poverty is certainly part of the problem though. Within the same dataset there is a strong negative correlation between median household income and gun-related death rates.

The strongest correlation was between gun death rates and stronger gun control laws; the second-strongest was between gun-death rates and gun ownership rates - the states with the lowest gun ownership rates had the lowest gun-related death rates.

I thought I'd summarize the initial post for you as you obviously didn't read it.

St-Javelin-Sm.jpgChase.jpg 

Link to comment
Share on other sites

1 hour ago, TheSanDiegan said:

What's fascinating to me is you can throw all those correlates out the window when it comes to mass shootings (with 4 or more fatalities), as evidenced by this chart of mass shootings from 1982:

Screen-Shot-2019-08-06-at-2-00-15-PM.png

At first glance it appears tied to population dynamics - seven of the ten most populous states are among the top ten in that chart. Maybe gross numbers and ethnic and/or socioeconomic diversity? :shrug: Given it is the horrific nature of these mass shootings, I think this is the area that merits a deep-dive, inclusive of Android's suggested geospatial correlation.

Would you, @AndroidAggie, or anyone else have time to separate the Gun Rate Death/100,000 column data into two: first being homicide and the second being accidental and suicide? I believe that data can be pulled from the CDC:

https://www.cdc.gov/nchs/pressroom/sosmap/firearm_mortality/firearm.htm 

I'd be interested to see how those stats play out. 

Thanks.

bsu_retro_bsu_logo_helmet.b_1.jpg

Link to comment
Share on other sites



  • Recently Browsing   0 members

    • No registered users viewing this page.



×
×
  • Create New...