Skip to main content

View Diary: Statistics, Guns, and Wishful Thinking (106 comments)

Comment Preferences

  •  Demographic data (6+ / 0-)

    correlation is not causation, but where there is no correlation you can dismiss out of hand the suggestion that there is causation.

    34% of the control group live in a household with guns.
    30% of the experimental group do.

    survivors have a 34% chance of owning guns.
    deceased have a 30% chance of owning guns.
    34%>30%.

    It means exactly what I said it means, unless you want to weasel on the difference between owning guns and having guns within the household.

    Ask your barista what her degree is in.

    by happymisanthropy on Mon Mar 26, 2012 at 02:06:12 PM PDT

    [ Parent ]

    •  Go read the original article (0+ / 0-)

      Go read the original article, particularly the section on Materials and Methods.  Then you and I can have an informed discussion.

      It does not "means exactly what I said it means"; it means exactly what the author says it means.

      "The fool doth think he is wise: the wise man knows himself to be a fool" - W. Shakespeare

      by Hugh Jim Bissell on Mon Mar 26, 2012 at 03:40:51 PM PDT

      [ Parent ]

      •  I'm the one who (5+ / 0-)

        linked to that article, remember?  You didn't.

        Ask your barista what her degree is in.

        by happymisanthropy on Mon Mar 26, 2012 at 05:50:47 PM PDT

        [ Parent ]

      •  ok, now, (3+ / 0-)
        Recommended by:
        hagagaga, gerrilea, theatre goon

        what about the demographic table am I misreading?  I only took one Stats class in college, so talk to me like I'm a kindergartener.

        Ask your barista what her degree is in.

        by happymisanthropy on Mon Mar 26, 2012 at 05:55:35 PM PDT

        [ Parent ]

        •  OK. Let's discuss (0+ / 0-)

          Please read the first sentence of the section marked "Discussion" (pg. 777) and tell me what it says.

          "The fool doth think he is wise: the wise man knows himself to be a fool" - W. Shakespeare

          by Hugh Jim Bissell on Mon Mar 26, 2012 at 06:53:23 PM PDT

          [ Parent ]

          •  asdf (7+ / 0-)
            A gun in the home is a risk factor for gun-related homicide
            and suicide among this sample of adults in the
            United States.
            So, he's 1) taken a sample of people murdered, 2) taken a control group, 3) used some kind of multivariate analysis to create a synthetic control group that demographically matches the case group, since the case group does not demographically match the United States as a whole, 4) compared the case group to the synthetic control group and found that between these two groups gun ownership is higher among the demised, and 5) he suggests that his findings can be generalized to the United States as a whole, even though they can't be generalized to the real control group that he started with.

            Am I missing anything?

            Ask your barista what her degree is in.

            by happymisanthropy on Mon Mar 26, 2012 at 07:43:48 PM PDT

            [ Parent ]

            •  Nice summary (0+ / 0-)

              You have done a good job at summarizing the study methods and findings.  A couple of important points:
              1) The case subjects (the dead ones) were not chosen due to being murdered, but chosen because they had died, and had been included in the National Mortality Followback survey, a survey of people who had died (regardless of cause) in the US.
              2) The statistical analysis used was logistic regression in order to calculare an odds ratio.  An odds ratio is simply the ratio of two odds (i.e. (the odds of having a gun and dying by gunshot) / (the odds of not having a gun and dying by gunshot)).  A value > 1 indicates that the nominator is larger than the denominator, while a value < 1 indicates that the denominator is larger than the nominator.
              3) The issue of whether a sample can be generalized to a population is tricky business, and is an issue for any study that hopes to use a small sample to make inferences about a large population.  You may fairly question whether the sample in this study allows making inferences about the population (tho' it appears that the editors of the Annals of Emergency Medicine and the peer-reviewers who reviewed this work were content that such a generalization was acceptable or believable).   Please note that almost all studies (including studies purporting to show the law-abiding nature of gun-owners) use the same methodology: draw up a sample, analyze the data derived from that sample, and then make an inference about the population.  So if that technique is not good enough for studies of deaths among gun owners, it is probably not good enough for studies of the law-abiding nature of gun owners.

              So now that we have covered those basics, do you want to discuss those demographic numbers in Table 1?

              "The fool doth think he is wise: the wise man knows himself to be a fool" - W. Shakespeare

              by Hugh Jim Bissell on Tue Mar 27, 2012 at 08:02:42 AM PDT

              [ Parent ]

              •  No. (2+ / 0-)
                Recommended by:
                theatre goon, hagagaga

                you're wrong, the two case groups represent subsets of the NMFS, not the entire NMFS data set.

                Specifically, the first case group represents people killed in homicide.  A non-random sample of the whole data set.  That's why the table is titled

                Characteristics of homicide case subjects 18 years old or
                older who died in 1993 and living control subjects.
                If they had done it the way you're saying, they only would have needed one case group to cover both homicide and suicide.

                Ask your barista what her degree is in.

                by happymisanthropy on Tue Mar 27, 2012 at 09:39:09 AM PDT

                [ Parent ]

                •  Correct (0+ / 0-)

                  You are correct.  However, this sample was collected from the NMFS data; subjects became part of the NMFS not because they were murdered, but by virtue of being dead of any cause.

                  So, are you satisfied with the study methods, findings, and conclusions, or do you have concerns that prevent you from accepting the study results and conclusions?

                  Do you still think the data in Table 1 on gun ownership negates the results and conclusions?

                  "The fool doth think he is wise: the wise man knows himself to be a fool" - W. Shakespeare

                  by Hugh Jim Bissell on Tue Mar 27, 2012 at 09:50:56 AM PDT

                  [ Parent ]

                  •  they don't negate the conclusion (2+ / 0-)
                    Recommended by:
                    hagagaga, theatre goon

                    they just show that they don't apply in the real world.

                    Ask your barista what her degree is in.

                    by happymisanthropy on Tue Mar 27, 2012 at 10:10:46 AM PDT

                    [ Parent ]

                    •  Wha? (0+ / 0-)

                      Your statement is non-sensical.

                      Either the conclusions apply in the real world, or the conclusions are erroneous.  Wiebe is not writing about imaginary life on Jupiter or Mars.  The Annals of Emergency Medicine does not publish research about fictitious worlds.  

                      Tho' I take it you no longer maintain that the data in Table 1 somehow negates the study conclusions, correct?

                      "The fool doth think he is wise: the wise man knows himself to be a fool" - W. Shakespeare

                      by Hugh Jim Bissell on Tue Mar 27, 2012 at 10:25:19 AM PDT

                      [ Parent ]

                      •  but they do (3+ / 0-)
                        Recommended by:
                        hagagaga, KVoimakas, theatre goon

                        see fit to publish articles on guns, when their specialty is supposed to be medicine.

                        Ask your barista what her degree is in.

                        by happymisanthropy on Tue Mar 27, 2012 at 11:55:14 AM PDT

                        [ Parent ]

                        •  Now you are just babbling (0+ / 0-)

                          You have yet to make one cogent argument criticizing the study on the grounds of method, findings, or conclusion.  

                          Yet you continue to impy the study is without merit.

                          Just like the conservatives do with the findings of scientists regarding the facts of climate change: you ignore or discount the facts because the facts do not argree with your opinions and beliefs.

                          Bush reminded us that "facts are stubborn things".  To me, this means the facts continue to be true, whether you acknowledge them or not.  

                          Whether you are willing to acknowledge it or not, the emprical data shows that guns are associated with human death and injury.  Those who are not blinded by their pro-gun prejudice understand this readily.  

                          Unfortunately, America continues to be held hostage by the wealth of the gun industry, ready to use its money to corrupt our law-makers, and the willingness of gun enthusiasts such as yourself to turn a blind eye to the demonstrated factual consequences of such actions.

                           

                          "The fool doth think he is wise: the wise man knows himself to be a fool" - W. Shakespeare

                          by Hugh Jim Bissell on Tue Mar 27, 2012 at 12:44:34 PM PDT

                          [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site