Wednesday, May 23, 2012

More on Guns vs. Cars

via Baldr with wonderful commentary

57 comments:

  1. Is he controlling for changes in passenger miles driven during the recession, and for the ongoing sales increase in guns?

    It's a simple analysis by a simple man.

    ReplyDelete
    Replies
    1. Yes, simple analysis that shows the foolishness of your argument that "cars kill more people than guns."

      The average car owner uses his car every day, the average gun owner does not. Remember we're not talking about gun fanatics now, but the average.

      If you adjusted the figures for that difference, guns would be 10 times more deadly than cars.

      Delete
    2. Blah blah blah. SUICIDES are intentional and car accidents are not. Apples and oranges Mike. Compare accidents with accident and then we will talk.

      Delete
  2. Also, the apparent drop in car deaths in the last few years is just a mild acceleration of the ongoing huge decline since the late 1960s. Omitting that trend is a way of being deceptive.

    Also, no adjustment for increased suicides (both gun and non-gun) during a particularly nasty recession? Are you kidding me?

    ReplyDelete
    Replies
    1. Yup, what we're seeing here is the difference between research and propaganda.

      Delete
  3. Perhaps drive by shooting affected the mix. :)

    ReplyDelete
  4. Mikeb, why is wonderful your favorite word? I didn't see a lot of commentary at his site, just a bunch of copy and paste from others' work and a whole lot of his wholesale swallowing of what he read without any digestion, much less analysis.

    ReplyDelete
  5. So basically what we can conclude from this graph and analysis is that government intervention to prevent car deaths started in 2006 resulting in a decline, while no intervention has been done to prevent gun deaths. Is that his point? That in 2006 the government finally started to crack down and make cars safe?

    ReplyDelete
    Replies
    1. No, what we can see is that cars are regulated heavily, seat belt laws, licensing and registration, insurance, and all the rest. The deaths from cars have already been reduced.

      Guns on the other hand are not regulated by the common sense laws that would reduce their deaths.

      If cars and guns were regulated in the same way, there'd be about 30,000 car deaths and 5,000 gun deaths.

      Delete
    2. Common sense laws? What does that mean? Criminals are already prohibited from owning guns. What other prohibition are you trying to put in place?

      Delete
    3. Scott, are you that new around here? You really don't know what I believe in?

      Delete
    4. Mikeb, you keep thinking that people will believe you. You've done nothing to gain our trust, but you keep wishing.

      Delete
    5. You want to make it so a law abiding citizen has to wade through reams of red tape and pass countless background checks just to keep an unloaded firearm locked in a safe at their home. While criminals roam the streets armed to the teeth. Why? Because you pissed away YOUR rights in your country and you are butthurt that we will not allow the same thing to happen here.

      Delete
    6. Scott, why can't you make your point without exaggerating and without throwing in a personal attack?

      "wade through reams of red tape and pass countless background checks"

      "reams?" "countless???"

      Get a grip, man. No one wants that. What we want is to ensure that you are a responsible person. Responsible people don't leave their guns lying around for thieves to take, or sell them to criminals in private sales, or, if you're an FFL dealer, turn a blind eye on the obvious straw purchase.

      Unfortunately, we know all those things go one, don't we?

      Delete
  6. Yes, you can tell from the sophisticated graph that 2006 is the year seatbelts, airbags, speed limits, drunk driving laws and crumple zones were invented.

    ReplyDelete
  7. I didn't generate the data. It's from the recent VPC report.

    As for all the "exceptions" you guys bring up which you think if not corrected for would somehow invalidate the data, such as relating the data to the recession or to gun sales, this is nonsense. And, as usual, you make the desperate plea that one type of gun-related death or another (such as suicide) shouldn't be "counted". Pitiful.

    Feel free, though, to try to make your own statistical analysis with accredited data. Could be good for a laugh.

    ReplyDelete
    Replies
    1. Yes, Baldr. It was VPC that generated a shoddy report- and it was you that parroted it.

      You can count suicides all you want. The point is it is very different than accidents with very different solutions. You talk about safety advancements in cars leading to fewer deaths and how that hasn’t happened for guns. Well, safety regulation is not relevant for suicides and homicides since the gun functioned as intended (unless you are calling a “safe gun” one that doesn’t go off when the user intends it to). So how are we supposed to see this effect when you lump accidental deaths (which happen roughly 500 times per year) in with suicides which have a magnitude 30 times greater, and homicides which are 20 times greater? Excel can handle more than two series- it is easy enough to break out suicide, homicide, and accidents to see something meaningful.

      As an aside, if a multi-variable linear regression model were to be used for this type of study, any reputable statistician would apply normalization to all the features (especially when they carry such a big magnitude difference).

      Delete
    2. Baldr,
      In the social sciences, you need to correct for confounding factors. This is something you'd learn in any introductory college stats or economics course. It is an issue that's been settled for hundreds of years, and that "Baldr Odinson" has no standing to dispute.

      Research is hard, and you've shown that you neither understand basic research issues, nor have the energy or perseverance to tackle them.

      For instance, your comment about suicide gives away your naivete. It's unconvincing to show that the number of suicides by gun are rising, and conclude that guns cause suicide. You need to use quantitative techniques, and discover natural experiments, to show the number of suicides that would have occurred if no guns existed. Some number of these people would have still committed suicide if they had no access to a gun, and the bad economy is a determinant of this. Similarly, some number of killers would have killed their wife with a claw hammer if a gun wasn't handy.

      Your source, the VPC, does not engage in research of any kind. Its job is to use $450,000 of the Joyce Foundation's money each year to repackage government statistics, adding no value whatsoever except inflammatory nonsense, and send them to gullible newspapers to get headlines.

      Finally, tour graph is not a "statistical" analysis in any sense, unless you consider a sum as a statistic.

      Delete
    3. The reason that I remove suicides from the total is that suicide is the right of every adult and a personal choice. You know the old line, our bodies, our choice? A good portion of vehicular deaths are unintended, while a good many firearms deaths are intentional acts.

      The difference here is that you're impressed by raw numbers, while I'm not. You believe that the number of gun deaths justify radical changes in our gun laws and culture, while I don't. (By the way, you don't give much of a statement about what laws you want.) I'd be happy to leave things as a respectful disagreement, but you have to insist on this self-righteous stance about your position, and you have to try to take away my rights Those I can't accept.

      Delete
    4. Greg, You are out of your mind. Suicide is not a "right." I even argue that gun ownership is not a right, in spite of the 2A, but how in the hell can you justify that whopper, suicide is a right?

      About Baldr trying to take away your rights, I don't know if that's true. Speaking for myself, I don't want to take away anything from you, not your choice to own guns, certainly. But, I would like you to have to qualify. If you do, no prob. If you don't, you have to give up your toys.

      My vision is that about half of the people who now own guns legally would continue to do so. The other half would be identified as unfit and unqualified for either mental health reasons, drug and alcohol abuse, violent or negligent behavior, etc.

      Delete
    5. What part of "my body, my choice" don't you get? That's how I call it my right and your right and any other adult's right. So yes, I'm serious about this, and I'm being rational in my thinking.

      But more importantly, whenever something is not accepted as a right, it can be regulated or banned out of existence. I can't trust anyone to be reasonable about controlling guns when said person doesn't see owning and carrying them as a right. It's that simple.

      Delete
    6. Greg, I don't remember if you ever went on the record about abortion. Are you as "pro choice" about that too?

      Delete
    7. I've told you before. Yes, I'm pro-choice, more or less. I think that in the third trimester, the only abortions should be out of medical necessity--as decided by a doctor and the woman--but in the first two, it's the woman's choice.

      Delete
  8. @TS: you don't seem to know statistical plotting, as I do. A multivariate will trace the curve about the same. What you refer to is a linear regression, which would ignore the year-by-year changes and deaden the curve (a statistical error people on your side tend to make).

    If you wish to break out the different types of gun-related deaths, as I said: feel free to do your own graph. But, unlike you gun extremists, I don't consider suicides to be extraneous gun-related deaths unworthy of counting. Remember, the VPC lumps together all causes of car-related deaths as well as all-causes of gun-related death, for this graph.

    ReplyDelete
    Replies
    1. No Baldr, I referred to multi-variable linear regression where the variables are on different scales as an example of how important normalization is. What VPC is showing is two separate single variable plots without regression analysis. Regression analysis is in fact needed to make a future prediction or predict gaps in data, or to compare the exact slope difference between the two datasets.

      Again, my point is that if you want to show these trends for guns it is best to break them out into three different categories because different solutions won’t apply equally across all three. Hell, you can even keep your “gun death total” if it makes you happy- just show the components along side it. You guys don’t want to do that because it will show a marked decrease in gun accidents which is why you want to hide it behind the suicides.

      Baldr: “If you wish to break out the different types of gun-related deaths, as I said: feel free to do your own graph.”

      Ok. TTAG already did one for accidental gun death from the 1981 to 1998 CDC data, and below are the numbers that I just mined from CDC for 1999 to 2009 (which incidentally is a steeper downward slope than cars deaths over the same period provided by VPC).

      http://www.thetruthaboutguns.com/2011/04/paul-raynolds/firearms-accidents-even-rarer-than-hens-teeth/

      2009: 554
      2008: 592
      2007: 613
      2006: 642
      2005: 789
      2004: 649
      2003: 730
      2002: 762
      2001: 802
      2000: 776
      1999: 824

      Are you still having a good laugh, Baldr?

      Delete
    2. Yes I am, TS. You continue to insist that accidents shouldn't be included in the gun-related deaths.

      Accidents aren't just due to faulty guns, a point you guys keep seeming to miss. Just as with cars, deadly gun accidents are more often due to faulty usage, poor decision making, and guns getting into the hands of those who shouldn't have them (such as children or mentally ill). Legislation, such as Child Access Protection laws, stricter background checks, and mandatory minimum training requirements can overcome a great many of these accidents.

      I honestly have no problem if you wish to do a multi-variate analysis on the same plot -- please do so -- as long as you also show the "total deaths" curve. My problem is that you guys poo-poo the whole concept of showing the comparison with deaths due to cars, as if it weren't a relevant comparison.

      Delete
    3. It is not that suicides don’t “count” if you want them to count- but rather if you lump accidents in with them it becomes impossible to analysis accidents. If you don’t want to analysis accidents, that’s fine but it becomes dishonest to compare it to another stat driven mostly by accident prevention. I just showed you the single variable plot for gun accidents vs. time, and it shows a marked decrease. How do you explain that decrease? Do you want to give credit to the industry for increasing the safety of their products? Do you want to give credit to the NRA and other organizations for safety educational campaigns? How about grass roots community efforts to stress the four rules of gun safety? How do you explain the decrease- or do you just not want to even think about it because people are still committing suicide?

      Delete
    4. Baldr: “Legislation, such as Child Access Protection laws, stricter background checks, and mandatory minimum training requirements can overcome a great many of these accidents.”

      And how do you know this if you don’t ever look at isolated accident statistics?

      Delete
    5. TS, one thing you're conveniently leaving out when you quote those accidental gun death numbers is that many more people each year are wounded, and even more than that are the number of negligent discharges which by chance hurt no one. I figure it's like this.

      600 dead
      2,000 wounded
      10,000 unreported negligent discharges

      I'm just guessin', though.

      Delete
    6. Yup, you're guessing. If a discharge is unreported, how can you know about it? And if no one is hurt, why should the gun owner tell anyone? Especially, that is, with people like you hanging around ready to tally up the incident and ask for losses of rights and more gun control?

      Delete
    7. MikeB: “…one thing you're conveniently leaving out when you quote those accidental gun death numbers is that many more people each year are wounded…”

      Whoa, where are you getting that “stat” from? The CDC reports non-fatal gun injuries too, you know, and 2010 is the lowest that they have data for. Why would you think that deaths are trending down, but injuries are trending up? And I thought saving lives is clearly the more important of the two anyway.

      Delete
    8. The stats on accidental woundings are wrong. For one thing, the oft-tried excuse they use that it was a drive by shooting or that it was some criminal that escaped sometimes works. For another, minor wounds are sometimes not reported.

      The point is, there's more damage being done than the number of deaths reflects.

      Delete
    9. Mikeb, I wouldn't give a shit about your guesses after swallowing a box of laxative. What PROOF do you have for that claim?

      Delete
    10. These are CDC stats- which Baldr called a credible source. If you are going to now claim they are wrong, I guess we can throw away this whole VPC "study".

      Delete
  9. @ the cowardly "anonymous" commenters: Again, feel free to do your own analyses. Should be good for a laugh.

    ReplyDelete
  10. @ Greg: Actually, suicide is considered illegal in most states. It's not a "right" to kill yourself in this manner.

    How very vulgar and heartless you are to suggest it, simply dismissing such a horrible act. Having known a friend who shot himself to death after suffering from depression, I'm particularly disgusted by your lack of empathy.

    ReplyDelete
    Replies
    1. 1. That's a rather useless law, no? Making suicide illegal. It's also an example of government overreach. We cut mental health services, but criminalize a choice.

      2. If someone asked me, I'd look at the reasons that the person wants to die. Some people suffer from incurable diseases. Oregon passed an assisted suicide law, no? You just want someone else to be in charge of our lives. On the other hand, if a depressed person proposed the same act, I'd suggest alternatives. Still, it's a person's choice. You don't own me. Get that through your skull.

      Delete
    2. Baldr's right, Greg. You are a heartless, tough-talking guy. I wonder if you're really that cold-hearted or if you're just trying to impress.

      Delete
    3. Baldr Odinson said..."...Actually, suicide is considered illegal in most states. It's not a "right" to kill yourself in this manner."

      I see research isn't your forte either.
      Under modern U.S. law, suicide is no longer a crime
      There has been no 'suicide law' in the US since about the 90's. There were some laws against attempted suicide, and if you attempt suicide, and someone is injured attempting to rescue you, then you could be held responsible.

      If you are so concerned about suicides, then perhaps you should focus your attention on poisons, because, you know, it is the preferred method of killing one's self. According to the CDC, Firearms are only used in about 5% of suicide attempts.

      I too, had a friend commit suicide with a firearm, and that was his choice. Of course, we were sad and had no idea that he was even considering suicide, but again, it was his choice.

      Delete
    4. Mikeb, you really don't understand how to live a principled life, do you? I believe in personal choice. Would I try to dissuade someone from committing suicide? In many cases, yes. But it is the person's choice. Why is that so hard for you to understand?

      Delete
    5. Greg that isn't hard for MikeB to understand -- it is impossible for MikeB to accept. Some people get-off on telling other people what they can and cannot do.

      Delete
    6. Greg, if you would try to dissuade someone from doing it, you're not respecting the fact that it's their right. You can't have it both ways.

      I think this is another example of where you basically agree with me but your stubborn contentious nature won't allow you to simply say it.

      Greg and Capn, are you both pro-choice on the abortion issue? That's a yes or no question.

      Delete
    7. I answered your abortion question above. It's not s simple yes/no, but I gave you my position on it.

      On the question of suicide, how does expressing my opinion to someone disrespect that person's right? If I'm asked my opinion, I'll give it. In a similar vein, if you only expressed your opinion that people shouldn't have guns, we'd have no problem. It's when you propose laws to force them into your choice that we object. The same would be true for someone considering suicide. I could say that it's a bad idea without taking away that person's right to make a decision.

      Delete
    8. Greg, You said that it's my "opinion that people shouldn't have guns."

      If you're gonna put words in my mouth, you have to be more accurate. What I say is UNFIT AND IRRESPONSIBLE people shouldn't have guns.

      Delete
    9. The problem is that you keep telling us about unfit people, but where's your talk about who is fit, in your view? In addition, you won't acknowledge that good citizens have a right to own and carry guns, so moving the goal posts of fitness is too easy for you. Until you advocate for a recognition to those two basic rights, we can't trust your motives, and we won't bargain away what we have.

      Delete
  11. Baldr,
    Multivariate does not imply nonlinearity or curves of any sort. You fail. You're ignoring the substantive critique of your meaningless chart, and deflecting with jargon you clearly don't understand.

    ReplyDelete
  12. Baldr
    You could do a linear regression on differenced time series, if you're worried about a unit root.

    Cue up the sound of crickets...

    ReplyDelete
  13. I love the cowardly anonymous commenters here who try to sound like they know something about statistics. You should know I do statistics for a living, every day.

    ReplyDelete
    Replies
    1. Not very well, I'll give you that

      Delete
    2. Baldr is butthurt because everyone's calling out the crappiness of his graphs.

      Delete
    3. Ok, but why did you wiff on the meaning of “muti-variable”? You seam to imply that it is about non-linear curve fitting- which it is not. You can apply quadratic, cubic, etc., functions to a single variable regression. Multi-variable would be if you threw in something like say… unemployment rates in an analysis of “gun deaths”- or broke out accidents, suicides, and homicides separately.

      What VPC did was show two separate single variable plots vs. time (where the time quotient was way too short) and no mathematical analysis. The analysis consisted of a guy with a beard looking at an excel plot. It is just raw numbers. And numbers that would matter for the topic of his discussion (like the effect on accidental deaths do to the introduction of transfer bars on revolvers, or Glock’s three level safety system) are going to be completely obscured by suicides- and that is IF he actually included it in the timeline.

      Delete
    4. "You should know I do statistics for a living, every day."

      That makes total sense. It explains why you are doing this CeaseFire stuff as a side gig. If I was as shitty at statistics as you appear to be, I'd be lining up other jobs as well.

      Delete
  14. Baldr here is the data behind the graph you published.

    (1) About 19,000 annual "gun deaths" are suicide.
    (2) About 9,000 annual "gun deaths" are murders where the perpetrators are hardened criminals, gang members, and/or drug dealers.
    (3) About 1,400 annual "gun deaths" are murders where the perpetrator "snapped" and committed their first felony.
    (4) About 600 annual "gun deaths" are accidents.

    Here is a news flash:
    Government laws and policies will not stop people who are determined to harm themselves or others. And people who are determined to harm themselves or others are responsible for about 28,000 of the 30,000 annual "gun deaths". Enact any gun laws you want and it will not stop those 28,000 people who annually harm themselves or someone else.

    If you could magically make all guns disappear forever, at best you might slightly reduce the lethality of some of those 28,000 annual attempts at suicide/murder. At worst the overall number of murders (as well as attempted murders) would increase because criminals would be more confident attacking unarmed citizens.

    The result? Gun control laws and policies could possibly reduce 2,000 annual "gun deaths". And they might slightly decrease the lethality of attempted suicides/murders. They could also increase the number of attempted murders annually. When you contrast this with the anywhere from 10s of thousands to millions of times annually that armed citizens successfully defend themselves with their firearms, it becomes obvious that gun control is bad for citizens.

    And this is why you should exert your effort on auto safety. Of the total annual automobile deaths, the overwhelming majority are accidents -- unintentional and unwanted. Drivers welcome actions to curb (pun intended) unwanted outcomes. Thus laws and policies to curb unwanted outcomes have a decent chance of being successful.

    ReplyDelete
    Replies
    1. Capn said, "Government laws and policies will not stop people who are determined to harm themselves or others."

      As I've mentioned before this argument is a slick attempt to put one over on us. Suicides and murders are not only committed by people who are "determined." If they were, you might have a point. But they're not and you don't.

      Delete
    2. mike - if a person was not determined to hurt themseleves they would not attempt to commit suicide. Additionally, Capn separated murders committed by felons from those committed by first time offenders that supposedly just "snapped." I would say that a convicted felon that decides to murder someone is indeed determined to hurt someone since they have to break the law again to even obtain the gun used in their crimes. He gave the benifit of the doubt to the first time offenders that they may not have been determined to commit their crimes.

      Delete
    3. MikeB, a person who isn't really determined to kill themselves -- who is crying for help -- doesn't use a firearm to "fake" an attempted suicide. They ingest medications or poisons or superficially cut their wrists and then tell someone about it before they are too ill.

      As for murderers, I separated first time offenders who may not have been totally deliberate -- like Jim said. Face it. The only possible population of "gun deaths" that any laws or public policy has any hope of affecting are those 2000 annual deaths which are accidents and crimes from first time offenders.

      Delete
  15. Mikeb, we don't know the things you said are facts. We have to see evidence before we know something like that.

    But is Scott Henrichs on to something? You keep trying to psychoanalyze me, and you hint that I'm a hidden criminal waiting to explode. Why don't you come clean about your past?

    ReplyDelete