Skip to main content

View Diary: A closer look at DGU numbers (117 comments)

Comment Preferences

  •  One of the major problems we face in law is that (2+ / 0-)
    Recommended by:
    bigjacbigjacbigjac, bontemps2012

    we can't be masters of all realms of knowledge.  I mean, not to defend that particular quote, that number smells pretty screwy to me.  But in general, it's a huge issue that we're called, every day, to argue and, even scarier, make judgments about things with which we do not truly understand.

    I don't envy trial judges.  They get presented with complex science and all sorts of other stuff and, in general, they really can't look shit up - our system places the burden entirely on the parties.  If the state forgets to describe how a certain kind of DNA test works, the judge may be able to ask for clarification during a bench trial, but once the evidentiary phase is closed or if there's a jury involved, tough luck.

    So if a case is poorly argued, that judge may have absolutely no fucking idea what they're doing.  I have a friend doing a judicial clerkship in another state doing primarily work in disciplines where bad lawyering is alarmingly common and the kind of stuff they have to deal with is just astonishing both for how bad it is and also how bad of a position the judges are in.

    "The first drawback of anger is that it destroys your inner peace; the second is that it distorts your view of reality. If you come to understand that anger is really unhelpful, you can begin to distance yourself from anger." - The Dalai Lama

    by auron renouille on Wed Jan 30, 2013 at 10:45:59 PM PST

    [ Parent ]

    •  That is true (1+ / 0-)
      Recommended by:
      bigjacbigjacbigjac

      The opposing council and/or other interested parties should have presented information or filed briefs properly critiquing Kleck's data and conclusions. Of course, it is also possible that a critique of Kleck's data was presented and subsequently ignored.

      I don't know enough about the case to determine what information the judges had access to, and only saw Kleck's number mentioned in the majority opinion. If it was truly presented to the Court uncontested, then it was a major failure on the side of the gun control advocates.

      •  You have no idea what range of data was actually (2+ / 0-)
        Recommended by:
        ER Doc, fuzzyguy

        presented to the Court and yet you firmly conclude that the process and data followed by the majority was flawed.

        Does your argument strike you now at all ironic?

        •  The number ended up (4+ / 0-)

          in the opinion ... and the number is deeply flawed.

          So it is reasonable to conclude that something went wrong there.

          I hope that the quality of debate will improve,
          but I fear we will remain Democrats.

          Who is twigg?

          by twigg on Thu Jan 31, 2013 at 06:22:46 AM PST

          [ Parent ]

          •  So you say ... (1+ / 0-)
            Recommended by:
            fuzzyguy

            but brilliant researchers in the field disagree. For example:

            Journal of Criminal Law and Criminology Fall, 1995 Guns and Violence Symposium, Page 188

            Who am I to believe?  Hmmmm?

            •  Sounds like he should have waited few years (0+ / 0-)

              and responded after Kleck was disputed by other researchers. Nothing in your link does anything to refute the critique of Kleck's numbers.

              Your whole argument boils down to: "This guy said Kleck did a good job, and he can't be wrong because he doesn't like guns."

              Also, and I'm giving you the benefit of the doubt here, all you have to do is a little more digging to find a follow-up:

              http://scienceblogs.com/...

              The usual criticisms of survey research, such as that done by Kleck and Gertz, also apply to their research. The problems of small numbers and  extrapolating from relatively small samples to the universe are common criticisms of all survey research, including theirs. I did not mention this specifically in my printed comments because I thought that this was obvious; within the specific limitations of their research is what I meant by a lack of criticism methodologically.
              (J of Criminal Law and Criminology 86:2 p617-8)
              Does that change your mind at all?
              •  Yea ... shame on (1+ / 0-)
                Recommended by:
                fuzzyguy

                him for dying of leukemia and not waiting for more research.

                Fuck.  For realz?

                •  you missed his reported followup (0+ / 0-)

                  deliberately?

                  Are you really discussing in good faith here? Ignoring the whole second half of the response that shows your link to be essentially useless regarding the validity of Kleck's results?

                  Feeling a bit polemical are you?

                  •  Oh please ... now you're in (0+ / 0-)

                    the Land of Oz.  

                    I'm not discussing in good faith?

                    That is well-deserving of a fuck you too.

                    •  What else am I supposed to think (1+ / 0-)
                      Recommended by:
                      madhaus

                      when you ignore the meat of a response when you reply? Especially when you've been so aggressive yourself about calling out perceived bias, and so-called polemics?

                      We got it, this one guy, respected in his field, said some nice words about the Kleck study. Yes, he later clarified that it suffers the obvious extrapolation issue which is one of the two critiques under discussion but sure, Kleck did a good job with his survey methodology.

                      Do you have anything else to add to the discussion?

                      •  Yea ... (0+ / 0-)

                        keep shooting at phantoms ... I get your agenda. Never have I suggested Dr. Kleck's results are without error ... but you keep arguing that strawman to the point of nauseating annoyance.

                        And as far as your dedicated scruple to statistical integrity ...
                        explain why the fuck have you been silent as a timid church mouse to this serial polluter of your diary and cause:

                        http://www.dailykos.com/...

                        But, it's your diary and I give you now the last word.

              •  My whole argument (0+ / 0-)

                boils down to nothing that you suggest.

                This diary and ensuing comments have no interest in making a balanced assessment of Dr. Kleck's work, preferring instead to assault him with all sorts of ignorant innuendo and polemical diatribes.

                To counter that -- and not his data or conclusions per se -- I provided the contemporaneous commentary of the most brilliant statistical criminologist at that time ... and someone who was a self-avowed anti-gun nutter (now sadly deceased).

                The work obviously continues, but to suggest that there is no merit or validity or scruple in the methodology followed by Dr. Kleck is to deny the exact opposite conclusion of someone who knew infinitely more about the subject than most of us every can or will.

                •  once again (0+ / 0-)

                  ignoring that your contemporaneous commentator refutes your own claims by noting that the standard issue with extrapolating from small numbers to large populations is so obvious he thought it not worth mentioning.

                  I'm not saying he methodology was entirely flawed, I'm saying that his lack of consideration of false positives for no other reason than he didn't measure them combined with the extremely low positive responses makes his extrapolation worthless. Essentially, the error bars make any number of his meaningless.

                  Additionally, external validation against other crime states and gun death rates also dramatically refute a 2.5 million number. But again, jump into one of those threads with alternate data if you have any.

                  Why do you insist only giving weight to the first part of Wolfgang's words, and not his followup? Is it because only the first part suits your agenda? Seriously, you've ignored it like 5 or 6 times now.

                  •  Jesus H Fucking Christ (0+ / 0-)

                    Do not accuse me of pursuing an agenda when we are discussing data and analysis in a diary and ensuing comments that has sought to savage Dr. Kleck's work when an eminent and contemporaneous opponent of his conclusions would not do so.  

                    Under-sampling is a methodological issue -- admittedly.  So now please share with us your same scruple of academic criticism of Dr. Hemenway's methodological deficiencies.

                    And in any case, current credible researchers (Kleck and Hemenway included) put the range of annual DGU between 100,000 + and some 2.5 million, with the more broadly accepted range of some 300,000 to 800,000.  

                    These are statistically significant numbers -- compared to any criminal violence metric you prefer to use. Your apparent preference to quibble completely misses the forest for the trees -- both as a matter of science and policy.

                    •  it's not quibbling (0+ / 0-)

                      it's using external validation to provide a better estimate.

                      ~300-400 justifiable homicides annually is a well established, non-survey based statistic.

                      20% mortality rate from gunshot wounds is also a fairly well-established statistic based on hospital records and national crime statistics.

                      This implies 1500-2000 justifiable shootings.

                      According to Keck's own statistics, DGU use involved wounding/killing the attacker ~8% of the time.

                      This implies 19k-25k total DGU (1500-2000 / 0.08).

                      The only questionable stat here is really Keck's, because it is based on the same problem with small numbers. However, instead of using the small numbers to extrapolate, we are using the percentages of self-reported wounding in conjunction with statistics that are vetted much more thoroughly.

                      If you want to throw out Keck entirely, you can. Otherwise, you have to argue against the real data that we do have.

                      This is how you do data-driven critique.

                      Your turn.

              •  There are oodles of follow-up (0+ / 0-)

                ... and that is not remotely my point.

                The rejoinder that you provide is equally off-point.

                The assertion is that Dr. Wolfgang found Dr. Kleck's methodology perfect and wholeheartedly endorsed his conclusion.  This is what is known as a strawman.

                In response, Dr. Wolfgang suggested that all statistical analysis of large quantities of data have inherent methodological difficulties, specifically including under-sampling, that are simply part and parcel of the "usual criticisms" and this proposition is so self-evident as not worth mentioning.

                Dr. Wolfgang, however, found Dr. Kleck's methodologies to be sound and careful (not perfect), with resulting conclusions that were reasoned and meritious (not absolute and immutable).

                As with Dr. Wolfgang, if you are experienced in science and statistical analysis, why are we having to argue these points.  They should be, as he says, self-evident (at least to those who might be trained, however modestly).

                •  Because apparently (0+ / 0-)

                  I need to explain the problem to people like you.

                  Small positive sample size + false positive rate + extrapolation to large numbers ==  worthless result

                  I'm not sure why you think Wolfgang's comments are the be all and end all, especially since he's not saying what you think he's saying.

                  The criticisms by Cook and Hemenway are valid and relevant to this particular statistical issue, and I'm not sure why you think you get to dismiss them just because you have someone else who you think says something differently (even though he actually isn't).

                  •  Ok ... you keep (0+ / 0-)

                    putting up absurd and wholly unsupported assertions -- Wolfgang's comments are not the be all and end-all.

                    But, keep studying the concept of strawman arguments.

                    Most peer criticisms of all researchers in the field of statistical analysis have validity -- Dr. Wolfgang's precise point.

                    The criticisms of methodology and extrapolation against Cook and Hemenway are equally valid -- unless you wish cherry pick with your academic scruple.

                    •  still waiting (0+ / 0-)

                      for you to contribute the data.

                      If you want to critique the critiques, feel free.

                      Find a thread regarding false positives and extrapolation or if you wish, tackle the external validation critique.

                      Or, conversely, just repeat yourself over and over. Your choice.

        •  I suggest you reread what I posted (2+ / 0-)
          Recommended by:
          gramofsam1, Sandino

          The only number I saw quoted in the majority opinion was the 2.5 million from Kleck.

          The post you responded to said quite clearly that if no other data was presented, then the opposing council dropped the ball.

          But if the opposing data was presented, but not at all mentioned in the majority opinion, then clearly the judges didn't do a very good job of analysis.

          Trying critiquing what I write instead of what you think I wrote.

          •  Yes, I believe I understand your (0+ / 0-)

            argument and argument structure to which I responded:

            If this, this, this and possibly this, then the Justices sucked.

            And my other principal take-away from your comment was

            I don't know enough about the case ... and only saw Kleck's number mentioned in the majority opinion.
            So I'll stand by my irony assessment about your own methodology... thanks though.

            Dr. Kleck's methodologies, in contrast, have been remarked upon with glowing esteem, by the way. Even from a leading and now deceased criminologist who even effectively called himself an anti-gun nutter.  http://www.saf.org/...

            •  see above (0+ / 0-)

              not so glowing after all.

              And you continue to misstate my argument regarding the Supreme Court opinion.

              I'll try to be a bit more clear:

              Court heard Kleck's study and either:

              1) opposing council did not refute Kleck's study with Cook and/or Hemenway, 2.5 million ends up in majority opinion.

              Conclusion for 1) opposing council sucked

              2) opposing council did refute Kleck's study with Cook, Hemenway, or just a plan ol' logic check using external validation similar to what's been done in this thread, 2.5 million ends up in majority opinion anyways.

              Conclusion for 2) majority judges have problems with their analytic skills

              If you wish to dispute the validity of 2, have at it. Find a thread where the external validity is under discussion and jump in with any data you have.

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site