Does Strengthening Self-Defense Law Deter Crime or Escalate Violence?


PDA






bayesian
January 2, 2013, 07:15 PM
The title is a paper by Mark Hoekstra at Texas A&M. Below I am linking to the paper but I can give you a basic run down of their analyses and results (this is related to my day job). What they take advantage of is something akin to what is sometimes referred to as a 'natural experiment'.

They use state level FBI crime data for 2000-2010. They look at the effect of 'stand your ground' laws on murder and non-negligent homicide. Specifically, they look at homicide rates prior to and after the passage of the law, within states. They compare pre/post law, and to ensure that there's not some other factor that might be creating the increase apart from the law change, they also look at homicide rates in adjacent states that did not change their law at the same time. They also look at other crimes that one would assume unaffected by stand your ground, such as motor vehicle thefts (and find no effect, as expected).

They find an 8% increase in the murder and non-negligent homicide rate, pre versus post SYG passage. They also find no decrease in robbery or aggrevated assault as might be expected if these laws had a deterrent effect.

Now, homicides include those that are self defense, so they look at how the numbers pre/post SYG differ for murders versus homicide.

The evidence they present show the increase in homicides are almost entirely falling into the murder category. What their data suggest (there's quite a bit more in the linked paper), is that 'Stand Your Ground' appears to be associated with an escalation of confrontations that would otherwise occur, but in SYG states, these are more likely to involve deadly force.

Here is the linked paper.
http://econweb.tamu.edu/mhoekstra/castle_doctrine.pdf

I'm a scientist my nature and profession, so I think that data like these are important to be aware of and engage with. So, there you have it.

If you enjoyed reading about "Does Strengthening Self-Defense Law Deter Crime or Escalate Violence?" here in TheHighRoad.org archive, you'll LOVE our community. Come join TheHighRoad.org today for the full version!
Texan Scott
January 2, 2013, 07:58 PM
Considering the FBI stats show a steady overall decrease in violent crime, somehow I question their math, methodology, and motive. I would point out that to isolate the effect of SYG laws on shootings, they'd a) have to count ONLY self-defense or claimed syg shootings; drug-related, domestics, suicidws, police shootings, etc would have to be discounted, not included; and b) an increase in syg shootings would likely mean more shootings on lawful self-defense, not an increase in criminal homicides.

Sounds like another case of fashionably anti-gun academia numbers fraud to me.

bayesian
January 2, 2013, 08:02 PM
Sounds like a load of manipulated statistics in which they include all the legal shootings by police and citizens as murders !

See, I think the gun community is better than just dismissing quite well done analyses of crime data. And I will note, that if someone is killed, justified or not, it *is* a homicide. It may not be murder, but it is, and is reported as a homicide.

I also believe that we don't have any adjacent states that don't have similar laws

Right, that is actually the idea. So, if you think that there is something about SYG, then for a single state, you can always worry that in addition to passing SYG, there was something else that might have happened (e.g., bad economy) that happened around the same time, and that is the real reason for the increase in homicide. So, you see if adjacent states that did not pass SYG at the same time, and you see if they saw an increase also.

If they did, then there's maybe something other than the law at work. If they didn't, then that makes you more confident that there's something state specific like SYG.

Ky Larry
January 2, 2013, 08:05 PM
As Mark Twain said:" There are lies, damn lies, and statistics."

IIRC, in the 1930s the city of New York had an increase in the number of sexual assaults in the cities public parks. Some bean counter also noticed there was an increase in ice cream sales at the same time. Some one propossed that ice cream sales be banned in the cities parks as a means of curbing sexual assaults. They finally realised that the numbers went up in warm weather and fell during cool weather. Rape and ice cream have nothing to do with each other except that people use the parks more in summer so there were more potential victims out in the parks and that people like ice cream in the summer.

bayesian
January 2, 2013, 08:11 PM
As Mark Twain said:" There are lies, damn lies, and statistics."

I know that people love to quote this, but the way research works is that we don't just publish any old thing because that is the result we want. If it was, I wouldn't read anyone else's research, and heck, I wouldn't do any because I could just figure out a way for my preconceived ideas and biases to determine the results. The actual way that research works is that if you think analyses are flawed, then you poke holes in the methods, what's left out and what's not considered.

So, here's my hobby horse: I think that there's pretty good data suggesting that all pro gun legislative initiatives are not all good. And I think this is the strongest data that I've seen that telling people 'You can stand your ground' leads a portion of people involved in violent altercations to *not* do what they might have otherwise, and figure out a way to de-escalate the situation.

DeMilled
January 2, 2013, 08:17 PM
Here is a little Q&A with the lead author of the paper.

http://www.dallasnews.com/opinion/sunday-commentary/20120622-point-person-our-qa-with-mark-hoekstra-on-castle-doctrine-laws.ece

ETXhiker
January 2, 2013, 08:19 PM
Okay, based on this guy's posts and his new membership, I'm calling TROLL. :barf:

PowerG
January 2, 2013, 08:35 PM
You're a scientist by profession? I just did a cursory run-through of the data for Florida, then compared it to the per capita homicide rate going back to 1960 for that state. The authors of the study use an increase that does indeed coincide with the passage of the SYG law and use it to support the statistical extrapolation for that state, while not noting that the 2010 and 2011 rates (posted while the law is in effect) are very near historic lows (5.2/100k both years, compared to 5.0 in 2005). Also not mentioned is that for the last three years the rate is less than half of the homicide rate as recently as 1990. The authors also freely concede that certain assumptions they make in the study could possibly have huge effects on their numbers; i.e. the number of justifiable homicides that were mistakenly included in criminal homicide totals. Color me "not convinced".

bayesian
January 2, 2013, 08:37 PM
Okay, based on this guy's posts and his new membership, I'm calling TROLL.

That's fair (ish). So, why did I just join? Well, I've lived in non-CCW friendly states for quite a while but now I live in Georgia. Grew up in MI where we hunted and grew up with firearms. I am a Walther fan, starting with the Walther PPK that my dad brought home from the 'old country' while visiting during the 40s.

I'm an occasional range visitor but after considering and acquiring my CCW, I decided to become more proficient and in the process of looking for online resources, I found THR and while lurking for a while, I recently did decide to register.

I would say I'm pro-gun but with a nuanced view that comes from working in the social sciences as a statistician, and being a parent.

Ky Larry
January 2, 2013, 08:37 PM
bayesian, I stand corrected. No number cruncher ever had an ax to grind. They are all as honest as saints. Mark Twain was obviously an idiot and a subversive. From now on I'll believe any numbers I read and ignore liars like Mark Twain, Will Rogers, and H. L. Menchken. Thank you for pointing out the error of my beliefs.:rolleyes:

locnload
January 2, 2013, 08:43 PM
"Figures don't lie, but liars figure". The liberal drones have got to have a continous supply of "statistics" to feed the uninformed public so when they do a poll they get the results they want. Take the murder rates from corupt Chicago and blame it on the legally armed folks in the "free states" and viola, you've got proof that guns cause crime. Then do a poll and ask, "do you think we need to consider common sense gun laws to prevent children from being slaughtered"? Joe and Jane Citizen nod their heads and say "ah-ha". Put it together in a nice pakage and send it off to CNN, ABC, NBC CBS, and now even Fox News, and you have an evening newscast that the 1960s staff at Pravda would be proud of. :fire:

DeMilled
January 2, 2013, 08:45 PM
See, I think the gun community is better than just dismissing quite well done analyses of crime data. And I will note, that if someone is killed, justified or not, it *is* a homicide. It may not be murder, but it is, and is reported as a homicide.



I believe you are simply wrong when you say "quite well done analyses of crime data". The man looked at numbers and then drew a conclusion, however there is a bit of speculation in the authors work.

From this article
http://www.opb.org/news/article/npr-stand-your-ground-linked-to-increase-in-homicides/
Hoekstra checked to see whether police were listing more cases as "justifiable homicides" in states that passed stand your ground laws. If there were more self-defense killings, this number should have gone up. He also examined whether more criminals were showing up armed.

In both cases, he found nothing. There were small increases in both numbers, but it was hard to tell whether there was really any difference.

So if the numbers on justifiable homicide and criminals using lethal force don't explain the rise in homicide, what's causing the increase?

"One possibility for the increase in homicide is that perhaps [in cases where] there would have been a fistfight ... now, because of stand your ground laws, it's possible that those escalate into something much more violent and lethal," says Hoekstra.

Tell me, bayesian, as a scientist by nature and profession, how large of a role does speculation play in your work?

Here is another guy, from the linked article who likes to speculate and draw conclusions based on "Well, here is what I think could happen."

Stanford law professor John Donohue, on the other hand, praised the study done by Texas A&M's Hoekstra. Donohue has been studying crime and violence for more than two decades and is working on his own independent analysis of stand your ground laws. So far, he says, he's getting the same results Hoekstra did.

"The imperfect but growing evidence seems to suggest that the consequences of adopting stand your ground laws are pernicious, in that they may lead to a greater number of homicides thus going against the notion that they are serving some sort of protective function for society," he says.

He goes on to deliver this little gem.
""I've been hearing from defense lawyers around the country that if they happen to have a criminal defendant in a stand your ground jurisdiction, pretty much no matter what happens, you can say, 'Well, I shot the guy, but I felt threatened and had a reasonable basis for fearing injury to myself,' " he said. "


I have quoted three well spoken, educated men whom all have the same slant to their views on self defense. They have couched their message in the language of academia but the message is still there. "Stand you ground laws/CCW laws are bad because a gun, even when carried by a law abiding citizen, might maybe be used to kill someone and we all know that if someone is killed then it is a bad thing."

All these arguments ignore the simple fact that there are many instances where using a gun to defend yourself, even if it kills the criminal attacking you, is perfectly acceptable. If the criminal didn't want to be shot, and maybe killed, he should have refrained from the act that led to the gun owner using the gun to defend them self.

ETXhiker
January 2, 2013, 08:56 PM
Bayesian, if you are sincere (I doubt it), read John Lott's "More Guns, Less Crime." It should satisfy your statistician side and answer all of your concealed carry, 2nd Amendment questions. After that, if you want to stay and talk guns as an enthusiast, then welcome.

Art Eatman
January 2, 2013, 09:34 PM
One problem I have with causality in this is that most SYG laws have been passed since the beginning of this economic malaise. I am reading of an increasing number of break-in entries into homes.

Granted I don't know how effective the SYG laws are insofar as repelling such break-ins; they've never been necessary from a legal standpoint.

"The times they are a-changin'," so I keep the salt-shaker handy when folks go to interpreting the results of statistical surveys. Too many unincluded data points, possibly.

Maybe a better question would be how many SYG shootings were not justified? And if you're a homeowner who made a successful defense, why would you care about what is done elsewhere, away from your home?

Some questions just are not worth asking. Survey results don't necessarily have any utility in the real world.

Q: Do CHL and SYG laws enhance my probability of surviving some sort of robbery? I think that they do, so why would I care about others surveys? My opinion comes from my personal survey through seventy-eight years of life. :D (Thirty years of it quite near the Mexican border.)

bayesian
January 2, 2013, 09:43 PM
Q: Do CHL and SYG laws enhance my probability of surviving some sort of robbery? I think that they do, so why would I care about others surveys? My opinion comes from my personal survey through seventy-eight years of life.

Q: Does drinking raven's blood at midnight cure me of cancer? I think it does so why would I care about other's evidence? My opinion comes from my personal survey through 78 years of life.

Ok, maybe I'm getting banned now, but I couldn't resist.

Look folks, we need to engage with people that deal with data. Justing saying something is 'just statistics' isn't the answer. I am honestly not trolling. But I tell my kids that it is the evidence in favor of their beliefs that they need to look at most carefully because it is that evidence that they are most likely to misread and misunderstand.

Is this board about preaching to the choir or is it about dealing effectively both within the pro-gun community and with the anti-gun community?

If it is all about preaching to the choir, then I guess sing it!

Alaska444
January 2, 2013, 09:43 PM
Duplicate post, sorry.

Alaska444
January 2, 2013, 09:48 PM
The title is a paper by Mark Hoekstra at Texas A&M. Below I am linking to the paper but I can give you a basic run down of their analyses and results (this is related to my day job). What they take advantage of is something akin to what is sometimes referred to as a 'natural experiment'.

They use state level FBI crime data for 2000-2010. They look at the effect of 'stand your ground' laws on murder and non-negligent homicide. Specifically, they look at homicide rates prior to and after the passage of the law, within states. They compare pre/post law, and to ensure that there's not some other factor that might be creating the increase apart from the law change, they also look at homicide rates in adjacent states that did not change their law at the same time. They also look at other crimes that one would assume unaffected by stand your ground, such as motor vehicle thefts (and find no effect, as expected).

They find an 8% increase in the murder and non-negligent homicide rate, pre versus post SYG passage. They also find no decrease in robbery or aggrevated assault as might be expected if these laws had a deterrent effect.

Now, homicides include those that are self defense, so they look at how the numbers pre/post SYG differ for murders versus homicide.

The evidence they present show the increase in homicides are almost entirely falling into the murder category. What their data suggest (there's quite a bit more in the linked paper), is that 'Stand Your Ground' appears to be associated with an escalation of confrontations that would otherwise occur, but in SYG states, these are more likely to involve deadly force.

Here is the linked paper.
http://econweb.tamu.edu/mhoekstra/castle_doctrine.pdf

I'm a scientist my nature and profession, so I think that data like these are important to be aware of and engage with. So, there you have it.
First, they cite the Trayvon Martin as an example of these laws gone wrong. First of all, the accused is not a reliable witness as evidenced by the perjury charge against his wife and perhaps himself at some time on his financial recourses. But for purpose of argument, let's take him at his word of that is shown to be accurate. When you are flat on your back with a stronger, quicker and taller person is beating your face in allegedly stating I am going to kill you, what part of China do you retreat to? He had no opportunity to retreat at that point if his testimony reflects the reality of the situation.

1 The most publicized case is that of Trayvon Martin, an unarmed teenager who was shot and killed by a neighborhood
watch volunteer (Alvarez, 2012).

Secondly, this is a retrospective observational study based on unproven ASSUMPTIONS when in fact, there could have been many other socioeconomic factors that prompted the data sets. In addition, as a "scientist my[sic] nature and profession," you should be intellectually honest and explain the limitations of retrospective, observational studies which are only HYPOTHESIS GENERATING but cannot prove cause and effect.

Thus, the crucial identifying assumption is that in the absence of the castle doctrine laws, adopting states would have experienced changes in crime similar to non-adopting states in the same region of the country. . . Our data allow us to test and relax this assumption in several ways.

You cannot TEST in a retrospective observational study, that is the hallmark of prospective, randomized and controlled trials thus making an internally incorrect statement related to a very weak study design.

In addition, they use estimates of homicides that they further report could have been misrepresented in the annual estimates and monthly estimates. Since the justice system works at a snail's pace, many of these cases could later have been dropped or changed. They did not take that into effect nor did they confirm individual cases leading to possible reporting bias.

The study also picked types of crime not shown to decrease with CCW laws. For instance, in Alaska, the only violent crime that was reduced by constitutional carry was rape, that was NOT one of the crimes measured in this biased report. In addition, the FBI does not keep stats on "home invasions" and does not differentiate that from "robberies" leading to confounding and possible misleading data samples.

http://www.homeinvasionnews.com/home-invasion-statistics-and-definitions-from-the-fbi/

They further use ASSUMPTIONS to ESTIMATE the degree of under reporting of justifiable homicides lending to further researcher predisposed bias.

We then use those estimates, along with assumptions about the degree of underreporting, to determine if the entire increase in criminal homicides can be explained as (misreported) legally justified homicides.


They in addition use some variables from the US census, but they did not adjust for population density changes between these states in the study period.

In addition, we have two variables from the American Community Survey of the U.S. Census Bureau that measure local legal opportunities, including median family income and the poverty rate. We also have data on the share of white and black men in the 15-24 and 25-44 age groups for each state over time (American Community Survey, 2000-2010). Finally, we measure the generosity of public assistance in each state by measuring per capita spending on assistance and subsidies and per capita spending on public welfare (US Census, 2000 –2010).


They further ESTIMATE data models that have not been independently verified for accuracy.

Formally, we estimate fixed effects ordinary least squares (OLS) panel data
models, where we follow convention and use the log of the outcome per 100,000
population as the dependent variable.17

A further ASSUMPTION is the ASSUMPTION that crime rates in adopting states are similar to non-adopting states.

Since we primarily rely on specifications that include state fixed effects and region-by-year fixed effects, the identifying assumption is that in the absence of the castle doctrine laws, adopting states would have experienced changes in crime similar to non-adopting states in the same region of the country.

I could continue point for point, but the bottom line is this is NOT a definitive prospective, randomized and controlled study but it is a very potentially biased report based on ASSUMPTIONS, SPECULATION and ESTIMATION.

Now what pray tell do you wish to "educate" THR with this less than inspiring biased report.

The conclusions of this "study" are not justified based on the limitations of a retrospective, observational study which is ONLY HYPOTHESIS GENERATING, it cannot prove causal relationships, but that doesn't stop these "scientists." That is intellectual dishonesty.

We are unable to think of any confounding factor that would fit this
description, and thus we interpret the increase in homicides as the causal effect of castle doctrine.

Once again, retrospective, observational studies CANNOT determine causal effects BECAUSE they do not control for confounding factors. This statement is pure garbage as far as a "scientific conclusion" giving further evidence of the bias of these "investigators." This is nothing but a hotshot, pseudo-scientific report.

(I am a retired internal medicine specialist and understand the conclusions you can draw based on study design. This report is intellectual dishonest and does not reflect accepted scientific conclusions.)

ETXhiker
January 2, 2013, 09:53 PM
Enough, with the troll.

DeMilled
January 2, 2013, 09:54 PM
Look folks, we need to engage with people that deal with data.

I agree.
Go find a paper that deals in facts, not assumptions and opinion, and please share it with us.

bayesian
January 2, 2013, 10:00 PM
Ok, I'll bow out of the thread.

I didn't mean this to be troll.

Texan Scott
January 2, 2013, 10:00 PM
Bayesian, part of the resistance you're encountering (and part of the reason people are calling you a "troll" and treating you as an outsider) is that the nature of your posts suggests that perhaps you do not share some common beliefs many of us hold; among these, belief that society's needs (as perceived by some of its individual members) are never an acceptable basis for violating the basic civil and human rights of any of society's members.

See, if you're trying to persuade me that having a codified legal right to protect myself and my family from criminal harm is bad for society, you've already lost. Even if I were absolutely persuaded of the accuracy and validity of this study (and clearly I'm not), if made to choose between the rights and safety of my family (or even just my own) and the greater good of society on a statistical level, I'm gonna throw society under the bus every time.

The right to self-defense is an inherent human right. I refuse to give up mine to make someone else more comfortable, "sound public policy" be darned. That's the real obstacle your argument is facing.

browningguy
January 2, 2013, 10:51 PM
Figures never lie, but liars can figure any way they want.

bayesian
January 2, 2013, 11:43 PM
I agree.
Go find a paper that deals in facts, not assumptions and opinion, and please share it with us.

Ok, I can't quite resist noting this: There's no such analysis of any kind of data that does not start with some assumptions. It is entirely legitimate to question those assumptions but impossible to analyze data without making some assumptions.

Correlation/Observation/Causality: It is correct to note that, say it all together: "Correlation is not causation". But... causation implies correlation (see Hume...). So, the task of a statistician is to try to eliminate alternative causal factors as the reason for the relationship. As one does this, you get greater confidence that the remaining factor is causal. Astronomy is entirely observational, no experiments are possible but there is a fairly good grasp of a wide range of good causal relationships. So, alternative reasons for the increase in the homicide rate are things like local factors not specific to the state (dealt with by looking for similar increases in adjacent states). Also, looking at crime statistics that one would expect to not be associated with SYG (e.g., motor vehicle thefts), another way to try see if there are other factors at work.

Note that there is no way to even address these kinds of questions in a purely experimental, non-observational way, so we have to deal with the data that we've got. And I'm of the view that these guys have done a pretty decent job of it. That was my original intent.

I'd be more than happy to see the opposite result. Indeed, I'd be delighted. But can I ask: Is it really hard to imagine among the total people out there getting CCWs, that a subset might end up with a chip on their shoulder because they have a weapon?

Massad Ayoob, writing an introduction to one of his books quotes the old Heinlein quote that an 'Armed society is a polite society' and he notes that one implication of this quote is the idea that individuals carrying have a unique obligation to de-escalate situations. My read on this research is that there may be a subset of the population that does not really take this obligation to heart.

Is that really hard to imagine?

Alaska444
January 2, 2013, 11:48 PM
Dear Bayesian, I thought you were talking about scientific proof, not conjecture, assumptions and IMAGINATION. Good grief, I can imagine a world that doesn't have false so called science as well, but shucks, there is a whole bunch of that out there in this world.

In any case, retrospective, observational studies are useful in generating a hypothesis but they have no ability to answer that question because of potential confounding bias that cannot be controlled for in a retrospective study.

The accepted method of eliminating unmeasured or confounding bias is to use randomization in a prospective controlled trial. That is the only scientific method established to reveal causality. I know you understand this simple elemental truth on types of scientific studies and the hierarchy of scientific evidence. To state causality in a retrospective observational study is simply being intellectually dishonest. Even if there is no way to perform a randomized and controlled trial, you are still limited in the conclusions you can derive from this type of study. Causality is NOT one of those conclusions you can derive. Practicing medicine on shoddy studies is very poor medicine my friend.

The best form of study for this type of issue is a prospective analysis of the events including those shown in other studies to be effected in other studies such as rape for instance using the actual statistics for each state before and after. That is the best you can get in studies that do not lend themselves to a randomized and controlled trial.

Once again, if you wish to dispute these simple and basic principles of of scientific studies and evidence, I would be more than happy to debate those facts utilizing respected public sources.

Thank you again, but this is a poorly designed study that is unable to support the conclusions that the authors put forth.

Thank you and by the way, I was a very well respected internist in my time. Do you really believe you can draw a conclusion that you are glad I was not your doctor because I speak the truth about scientific evidence? I appreciate your PM, but please feel free to speak openly to everyone instead of taking cheap shots at me in a PM. I very much understand scientific evidence and I am not easily swayed by poorly performed and designed studies.

PM: Really? It is that simple?

(I am a retired internal medicine specialist and understand the conclusions you can draw based on study design. This report is intellectual dishonest and does not reflect accepted scientific conclusions.)

I am glad that you aren't/weren't my internist. I would hope that my dr. is swayed by evidence.

Texan Scott
January 3, 2013, 09:19 AM
Is it really hard to imagine among the total people out there getting CCWs, that a subset might end up with a chip on their shoulder because they have a weapon?

Bayesian,

It's not difficult to imagine; it might even seem like a reasonable thing to imagine. What perhaps many of us react to badly is the idea of our rights being restricted in reality (with real consequences innocent people may take to the grave) on the basis of what you can imagine someone doing.

The basic principle in contention (as it seems to me) is whether society at large ( a simple majority thereof) has the right to infringe the liberties of citizens who have done no wrong or harm. In some cases, perhaps... but not when the right in question is a basic human right and civil liberty guaranteed by the highest law of the land (superseding any law the simple majority may pass).

I write this, not as a professor or economist or medical professional, but merely as a middle-aged man who has once been forced to defend against a home invasion robbery - with a gun.

Steel Horse Rider
January 3, 2013, 09:41 AM
Crime statistics that do not separate justified shooting from criminal use of a firearm is a useless basis for analysis. Many leaders in the 2nd Amendment surrender movement like to throw around the statistics of how many "children" are killed or injured by firearms without separating out the gang and drug related crime involving those under 21 years of age.

Carl N. Brown
January 3, 2013, 10:44 AM
I downloaded the Hoekstra paper 13 June 2012 (at least the abstract and the NBER working paper). Even with publication in a review, I remain unconvinced.

Following passage of Stand Your Ground or Castle laws, there will be an increase in manslaughters "adjudicated" in police report as justifiable homicide and in manslaughters claimed to be justifiable under SYG or castle "adjudicated" in police report as manslaughter. That is because most manslaughters would not be ruled justifiable homicide until much higher in the criminal justice system: coroner, medical examiner, district attorney, grand jury, trial judge, trial jury and appellate court usually handle adjudication of manslaughter as criminal murder or justifiable homicide.

What Stand Your Ground or Castle laws actually increase are the number of cases adjudicated at the police investigation level, cases that in most jurisdictions still would be decided by the district attorney declining to press charges or by the grand jury returning no true bill.

Homicide rates of States with No Duty to Retreat, Presumption of Reasonable
Fear, and No Civil Liablity in Self-Defense Laws ("Stand Your Ground")
bold ital marks years with SYG
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Arizona 7.0 7.5 7.1 7.9 7.2 7.5 8.5 8.7 7.0 5.8 6.4
Florida 5.6 5.3 5.5 5.4 5.4 5.0 6.2 6.6 6.3 5.5 5.2
Kentucky 4.8 4.4 4.7 4.4 5.7 4.6 4.2 5.1 4.7 4.3 4.3
Louisiana 12.5 11.2 13.2 13.0 12.7 10.0 13.1 14.7 12.2 11.8 11.2
Mississippi 9.0 9.9 9.2 9.3 7.8 7.4 8.0 7.0 8.0 6.6 7.0
Tennessee 7.2 7.4 7.3 6.8 6.1 7.2 7.0 6.6 6.6 7.4 5.6
Texas 5.9 6.2 6.0 6.4 6.1 6.1 5.9 5.9 5.6 5.4 5.0

Madcap_Magician
January 3, 2013, 11:10 AM
I am interested in the proposed mechanism by which SYG laws increase murder rates.

Eleanor416Rigby
January 3, 2013, 11:12 AM
I'm also pretty new here. I think it's good for the choir to hear a different preacher once in a while. However, I do not agree with the findings of the article suggesting a causal relationship between stand your ground and homicide (or anything) based on a non-scientific study.

Even if it is true that some people go out and get CCW with chips on their shoulders and end up more likely to get into fights and shoot somebody, it is not the fault of or caused by society/law recognizing the right to self defense. It is the fault of those individuals having a bad attitude. (Their parents, teachers, community, etc. indirectly, too.)

The rest of us are innocent until proven guilty, so please (politicians and anti-gun people) don't try to take away the rights of the innocent based on the failings of those individuals.

Coltdriver
January 3, 2013, 11:19 AM
Anecdotaly speaking I believe the Colorado law that allows a person to protect themselves at home has resulted in fewer home invasions, with the exception of those that occur in the drug trade.

Skribs
January 3, 2013, 11:39 AM
Wait a minute...you're telling me that when people can legally stand their ground, they legally stand their ground instead of running away from bullies? And then when said bully doesn't stop, the situation escalates? Amazing! I'm so glad we have statistics to tell us this. What it tells me is that SYG laws allow people to actually defend themselves.

Like others, I think it's possible for someone to get a CCW and then provoke a fight so they can shoot someone else. However, by and large the legal system will figure out what happened. Maybe not every case, but it can be very clear. Usually people who have a sort of warped sense of reality (not psychotic, just a bit delusional) will say something that will turn heads.

Art Eatman
January 3, 2013, 12:36 PM
"I think it's possible for someone to get a CCW and then provoke a fight so they can shoot someone else."

Sure. But as near as I can tell from press releases from such as the Florida Department of Law Enforcement (FDLE) and the Texas Department of Public Safety (DPS), this does not happen. Or, so far, it has not happened.

From decades of reading about homicide events and all the data, the only thing which seems to affect the crime rates where firearms are used is demographics: Fewer young males in the 16 to 25 age group means fewer events. Plot the curves, these last thirty or so years. The rise and fall is overlapping for both data sets.

Skribs
January 3, 2013, 12:51 PM
Art, I've read a few articles of it happening, but by and large it's the same "he attacked me, it was self defense when I stabbed him 40 times after he passed out."

All SYG laws do is prevent some justified self defense homicides from turning into legally unjustified homicides because of "duty retreat" laws.

PedalBiker
January 3, 2013, 12:56 PM
I am interested in the proposed mechanism by which SYG laws increase murder rates.

Criminals, knowing they could be faced with an armed homeowner, form larger groups with more effective and therefore lethal tactics.

There is no free lunch. Having a gun at home is not magic.

It's not like having a .38 revolver in Mexico is going to make you safer from the cartels.

Guns in the crime debate are a red herring, there are much larger social issues that dominate.

A single criminal can only do so much. A group of criminals can do more. A cartel of criminals is nearly unstoppable and a government run by criminals can spark world wide warfare.

The issue of self defense is a moral one. It is immoral to prosecute people for exercising self defense. Don't fall for these "ends justifies the means" studies. They are a road to ruin. Sometimes you just have to do what's right and deal with the consequences.

bayesian
January 3, 2013, 01:47 PM
So, reading the comments is helpful in understanding where people are coming from. There's some suggestion that I am attempting to fool folks, or somehow skew the arguments in an anti-gun way.

I guess I mentioned this to a couple of folks in PM, mainly because I didn't want to clutter this thread with a bunch of my responses but I do at least want to mention a few things.

First, studies that are 'retrospective' and observational are not necessarily 'unscientific'. Archeology is necessarily retrospective and astronomy is exclusively observational. The issue is whether you approach things in a systematic way, take steps to minimize the influence of bias (meant in several ways, not just political), and adhere to some basic steps for analyzing and reporting your results. Many areas of social science, economics, public health, etc can be quite scientific.

Second, while correlation does not prove causation, the presence of correlation often implies some kind of causal factor. Research is aimed at identifying those causal factors. When your field does not allow true experiments, then your fall back strategy is to assemble a set of possible causal factors, and start analyzing your data in several ways to see which of those factors is most consistent with the data. If all goes well, one factor might come out as the most consistent with many of the relationships in the data, and that is the one that you tentatively identify as the causal factor. Even in experiments, this is essentially what we do, so all conclusions are necessarily inferential , probabilistic, and provisional. That does not mean that the results are 'speculation' (although that has an important role in thinking about all of the different possible causal factors at work). It is incorrect to say that randomized experiments are the sole way that we have to identify Cause. It is the quickest and most efficient in some cases, but not the only way.

Only mathematics deals in 'proof', science always deals in uncertainty, and I understand that there may be less certainty in the analyses than one might like, but I think it is unwise to require airtight evidence for arguments against one's favored beliefs, while requiring a much lower standard of evidence for that supporting one's beliefs. I will note that the methods that they use are actually very similar to those used in epidemiology to identify effective methods for encouraging vaccinations, reducing e. coli infections, etc. If these types of analyses work in those contexts, then it is not clear that they yield entirely invalid results when applied to this issue.

If I got a wee bit hot in the back and forth, it is probably because off the cuff rejection of evidence (e.g., Libertard...) is a good way to wind up poorly informed and poorly prepared to have an influence on future legislation and policy decisions.

Double Naught Spy
January 3, 2013, 02:32 PM
Second, while correlation does not prove causation, the presence of correlation often implies some kind of causal factor.

Playing both sides of the fence with a bias toward your purpose. First you pay lip service by noting that correlation does not prove causation, then note that correlation OFTEN implies some kind of causal factor. You are using "often" as a value statement to indicate that your statement is more commonly the norm than the exception when you have no basis for this claim.

An equally valid statement would be

Second, while correlation does not prove causation, the presence of correlation often DOES NOT imply some kind of causal factor.

My rewording of your claim is equally valid but completely changes the apparent validity of your claim.

You see, the blanket claims about correlation's relationship with causation are far too generalized to have actual meaning. I could say "A lot of things correlate without any relationship to causation" and the statement is so ambiguious to be meaningless. What is "a lot" and is that number even relevant?

Carl N. Brown
January 3, 2013, 03:25 PM
Cheng Cheng and Mark Hoekstra, "Does Strengthening Self-Defense Law Deter Crime or Escalate Violence? Evidence from Castle Doctrine".

The first version I downloaded (35 pages) was created 5-29-12; the linked version (43 pages) was created 12-17-2012. It has been submitted for comments and suggestions to seminars (2012 Annual Meeting of the Southern Economic Association and the 2012 Stata Texas Empirical Micro Conference). Changes include some rewording, extending the data to include 2010. Added notes include "7 These laws also typically state that the protections do not apply to those who are committing a crime at the time, or who instigated the conflict." The initial paper refers to Stand Your Ground as starting with Florida in 2005. The revision notes that some states had extended Castle Doctrine before 2000, essentially SYG not called SYG, and revised their table. They also note: "Our findings contrast with those of Lott (2010) in More Guns, Less Crime, who reports that castle doctrine laws adopted from 1977 through 2005 reduced murder rates and violent crime."

They seem to overlook the fact that for every justifiable homicide reported by police, four or more homicides are found justifiable as they are processed through the courts, and that in any given year fifteen to twenty of the fifty states do not report a UCR justifiable homicide stat. The revision notes that according to researchers the FBI UCR justifiable homicide stat is underreported. It removes a garbled description of FBI handbook standards for reporting justifiable homicide.

My main problem with this article is that it depends on what many consider the least reliable statistic in the FBI Uniform Crime Reporting system: the justifiable homicide table. The second problem is the fuzzy definition of what states were SYG and non-SYG for comparison purposes. It would be interesting to read this article after it has been shopped for comemnts at an American Society of Criminology meeting. Or two.

bayesian
January 3, 2013, 03:34 PM
First you pay lip service by noting that correlation does not prove causation, then note that correlation OFTEN implies some kind of causal factor.

Yeah, I think I stand by that statement.

You are using "often" as a value statement to indicate that your statement is more commonly the norm than the exception when you have no basis for this claim.

Consider a mechanic that is quite good at diagnosing problems. I think it is safe to say that, within that domain, a small number of correlations are considered, any one of which might be causal. Further testing hopefully lets him narrow in on the best one. Someone knowing little about engines may consider a large number of correlations, perhaps only a few of them are causal.

Saying that I'm treating 'often' as a dodge is like saying that I refuse to say how long string is. Ok, you got me. It is 17.36 inches. Exactly.

Slipknot_Slim
January 3, 2013, 03:47 PM
Strengthening self defense law strengthens your defense if you are involved in a shooting. It transfers more of the burden of proof from the defense to the prosecution. As a gun owner, that's my primary concern. If I use my gun in self defense, will I be able to return home when it's all said and done.

Alaska444
January 3, 2013, 03:48 PM
Yeah, I think I stand by that statement.



Consider a mechanic that is quite good at diagnosing problems. I think it is safe to say that, within that domain, a small number of correlations are considered, any one of which might be causal. Further testing hopefully lets him narrow in on the best one. Someone knowing little about engines may consider a large number of correlations, perhaps only a few of them are causal.

Saying that I'm treating 'often' as a dodge is like saying that I refuse to say how long string is. Ok, you got me. It is 17.36 inches. Exactly.
Dear Bayesian, that is intellectual dishonesty as I have stated earlier. I never stated that retrospective, observational studies were not scientific, I simply stated the true fact that retrospective, observational studies are limited in the scope of conclusions to that of a HYPOTHESIS GENERATING STUDY. They cannot answer questions of causality.

In addition, this particular study is even lower on scientific validity because of the large amount of ASSUMPTIONS and speculations. I guess you wish to throw in IMAGINATION as well. That is not science my friend.

Lastly, as others have pointed out the FBI data base is a poor choice for discerning justifiable homicides.

Garbage in my friend, garbage out.

Alaska444
January 3, 2013, 03:52 PM
I guess I mentioned this to a couple of folks in PM, mainly because I didn't want to clutter this thread with a bunch of my responses but I do at least want to mention a few things.

Looks like you can't even tell the truth in what you are stating about your PM's. Remember, you sent me a very nice and informative PM. Let me refresh your memory:

Really? It is that simple?

(I am a retired internal medicine specialist and understand the conclusions you can draw based on study design. This report is intellectual dishonest and does not reflect accepted scientific conclusions.)
I am glad that you aren't/weren't my internist. I would hope that my dr. is swayed by evidence.

You are only partially correct in your last sentence of you PM. I am NOT swayed by poorly designed and executed studies of which this one is one of the worst I have ever read. It is only pseudo-science.

In any case, enjoy your short association here on THR.

Take care

Alaska444
January 3, 2013, 04:03 PM
Second, while correlation does not prove causation, the presence of correlation often implies some kind of causal factor. Research is aimed at identifying those causal factors. When your field does not allow true experiments, then your fall back strategy is to assemble a set of possible causal factors, and start analyzing your data in several ways to see which of those factors is most consistent with the data. If all goes well, one factor might come out as the most consistent with many of the relationships in the data, and that is the one that you tentatively identify as the causal factor. Even in experiments, this is essentially what we do, so all conclusions are necessarily inferential , probabilistic, and provisional. That does not mean that the results are 'speculation' (although that has an important role in thinking about all of the different possible causal factors at work). It is incorrect to say that randomized experiments are the sole way that we have to identify Cause. It is the quickest and most efficient in some cases, but not the only way.

Your INFERENCE that retrospective, observational trials are capable of producing causality is once again intellectually dishonest, unless of course science allows you to use your imagination. Your treatise above is simply describing how retrospective, observational trials are utilized to produce HYPOTHESIS of the issue at hand.

Only mathematics deals in 'proof', science always deals in uncertainty, and I understand that there may be less certainty in the analyses than one might like, but I think it is unwise to require airtight evidence for arguments against one's favored beliefs, while requiring a much lower standard of evidence for that supporting one's beliefs. I will note that the methods that they use are actually very similar to those used in epidemiology to identify effective methods for encouraging vaccinations, reducing e. coli infections, etc. If these types of analyses work in those contexts, then it is not clear that they yield entirely invalid results when applied to this issue.

Some retrospective, observational studies are better than others. The real issue with the study in question is the LARGE amount of speculation and assumptions utilized in the analysis which increases the risk of confounding bias. In addition, there are elements of this study that are truly troubling such as not taking into account population changes during the study period. Secondary analysis of "population markers" such as welfare enrollment adds further confounding variables to the equation.

No, this is a poorly designed and implemented study. Your defense of this study is once again intellectually dishonest in many of the claims you are conjecturing.

If I got a wee bit hot in the back and forth, it is probably because off the cuff rejection of evidence (e.g., Libertard...) is a good way to wind up poorly informed and poorly prepared to have an influence on future legislation and policy decisions.

No you are getting hot because you have been exposed as an intellectually dishonest commentator. There are many very well informed folks on THR who have only a modicum of formal education. Please keep your subtle ad hominem fallacies to yourself.

Skribs
January 3, 2013, 04:16 PM
Alaska, science does allow us to use our imagination. Oh I'm sorry, that was science fiction.

Alaska444
January 3, 2013, 04:24 PM
Alaska, science does allow us to use our imagination. Oh I'm sorry, that was science fiction.
+1, so true.

bayesian
January 3, 2013, 04:54 PM
I am puzzled by the 'intellectually dishonest'. Really, I'm genuinely puzzled. I've not said anything in any of the comments that I don't actually believe.

I suppose interested parties can read the thread and they will draw their own conclusions. I don't want to argue that any single paper is either 1) the greatest thing since sliced bread, or 2) the final word on any topic.

Some retrospective, observational studies are better than others. The real issue with the study in question is the LARGE amount of speculation and assumptions utilized in the analysis which increases the risk of confounding bias.

Granted. Indeed, the possibility of confounds is the reason for modeling the effects relative to changes in adjacent states that did not see the legislative change.


In addition, there are elements of this study that are truly troubling such as not taking into account population changes during the study period.

Ah, but so that's the point of also look for changes in things such as motor vehicle thefts. If there are changes, such as a downturn in the economy, or other changes to the population, then if one observed changes to the motor vehicle thefts, then the argument that the homicide changes were related to SYG would be severely undercut. Similarly, if adjacent states show changes at the same time, again, under cuts any association between SYG and homicide rate. This is hypothesis testing.

I don't think the paper is the be all end all, but I thought this was a fairly interesting way of looking at the possible effects.

bayesian
January 3, 2013, 05:01 PM
Carl N. Brown wrote:
They seem to overlook the fact that for every justifiable homicide reported by police, four or more homicides are found justifiable as they are processed through the courts, and that in any given year fifteen to twenty of the fifty states do not report a UCR justifiable homicide stat. The revision notes that according to researchers the FBI UCR justifiable homicide stat is underreported. It removes a garbled description of FBI handbook standards for reporting justifiable homicide.

Thanks for the comment :)

So, my impression was that they were using, for the main body of their analysis, the total for murder and non-negligent homicide. My understanding is that if someone is killed in justifiable self defense, regardless of whether this is simply not pursued by the prosecutor, or one is acquitted, that the event would still fall under the non-negligent homicide category. But I'm not a lawyer, so I could be wrong.

Alaska444
January 3, 2013, 05:02 PM
I am puzzled by the 'intellectually dishonest'. Really, I'm genuinely puzzled. I've not said anything in any of the comments that I don't actually believe.

I suppose interested parties can read the thread and they will draw their own conclusions. I don't want to argue that any single paper is either 1) the greatest thing since sliced bread, or 2) the final word on any topic.



Granted. Indeed, the possibility of confounds is the reason for modeling the effects relative to changes in adjacent states that did not see the legislative change.




Ah, but so that's the point of also look for changes in things such as motor vehicle thefts. If there are changes, such as a downturn in the economy, or other changes to the population, then if one observed changes to the motor vehicle thefts, then the argument that the homicide changes were related to SYG would be severely undercut. Similarly, if adjacent states show changes at the same time, again, under cuts any association between SYG and homicide rate. This is hypothesis testing.

I don't think the paper is the be all end all, but I thought this was a fairly interesting way of looking at the possible effects.
Your language betrays your real thoughts on this study. Your last quote states it all:

I don't think the paper is the be all end all, but I thought this was a fairly interesting way of looking at the possible effects.

Let me rephrase that, I don't think the paper is the be all end all, but I thought this was a fairly interesting way of looking at the HYPOTHESIS GENERATED in this study.

Once again, looking at what you are stating, you are over and over again confirming the HYPOTHESIS generating value of retrospective, observational studies. To believe it goes beyond that stage is intellectually dishonest whether you are aware or not.

PowerG
January 3, 2013, 06:59 PM
A visual representation...

Chart from the TAMU paper:

http://img.photobucket.com/albums/v461/PowerG/syg_zps4ff57c54.jpg



Actual homicide rates 1960-2011 for FL:

http://img.photobucket.com/albums/v461/PowerG/rate_zps9e40482b.jpg

http://www.politifact.com/florida/statements/2012/mar/23/dennis-baxley/crime-rates-florida-have-dropped-stand-your-ground/

Alaska444
January 3, 2013, 07:16 PM
Hey, you are showing real data of before and after. But that is not what the study did. They compared the after to a theoretical compilation of expected values based on a number of assumptions. You really shouldn't use real data, it might confuse the issue.:neener:

Zoogster
January 3, 2013, 07:27 PM
I have said it in more detailed ways before and show some statistics to demonstrate it and various rates.

The vast majority of the homicide rate centers around gang members killing gang members, and drug dealer disputes or murders.
Such a large majority that it drowns out other factors that could otherwise be considered.

The self defense laws and concealed carry have virtually no impact on either one, assuming that not many lawful concealed carriers are violent gang members or involved in drug dealing.


As a result concealed carry rates or stronger self defense laws compared the homicide rates mean almsot nothing unless you can cleave all the gang and drug related homicides off the statistic.


Many places with higher homicide rates simply have higher numbers of gang members. So all other factors amount to very little.

The homicide rate has been trending down since the 90s, when the gang killings reached thier peak.
It has been trending down for whites since 1980, but went way up in the mid 1990s overall all, because it was mainly inner cityblack and hispanic gangs killing eachother.
Really nothing has changed for the majority since 1980, with or without concealed carry, or stronger self defense laws when it comes to homicide rate.

So arguments for or against are just plain misleading when using the base homicide rate.

bayesian
January 3, 2013, 07:27 PM
Ok, so first note that the first figure is in log units on the y-axis, so e^1.7 ~ 5.4, the scale on the second is compressed but that is in the ball park.

So what am I missing, these look like they agree?

PowerG
January 3, 2013, 08:09 PM
They certainly do agree, but...the upward trend that occurred just after the passage of the SYG law is, in the consideration of longer term trends, just as easily attributed to a statistical anomoly as it is to any change in law. There are much larger upticks in the early and late '70's, which I would suspect have more to do with economic factors than any change in self defense laws; in case you forget there were economic perturbations in the years focused on in this study also. If the study documented a valid increase in homicides due to the SYG law then it's likely the increase would continue into the present homicide rate...but the fact is that the 2011 rate is comparable to the all time lows, in fact the difference is statistically insignificant. Was the SYG law repealed? Has an attempt been made to explain the return to normal homicide rates? Why have both the homicide and violent crime rates in FL been trending down generally (see the link I posted) since the uptick? Despite the authors' claims, I don't see this study as "empirical" in any way. YMMV.

If you enjoyed reading about "Does Strengthening Self-Defense Law Deter Crime or Escalate Violence?" here in TheHighRoad.org archive, you'll LOVE our community. Come join TheHighRoad.org today for the full version!