Kelly Roskam of the Johns Hopkins Center for Gun Violence Solutions discusses a Supreme Court case that will decide if a federal law prohibiting possession of firearms by people subject to domestic violence protection orders is constitutional

    • Jeremy [Iowa]@midwest.social
      link
      fedilink
      English
      arrow-up
      8
      arrow-down
      6
      ·
      edit-2
      1 year ago

      It’s interesting that you support a Johns Hopkins piece with… a John Hopkins piece. That’s a bit like doubling down on Everytown.

      That particular claim is built upon some incredibly sketch analysis. This is the most common backing source.

      The methodology:

      An 11-city case–control design was used; femicide victims were cases (n = 220), and randomly identified abused women residing in the same metropolitan area were control women (n = 343). Co-investigators at each site collaborated with domestic violence advocacy, law enforcement, and medical examiner offices in implementing the study. Sampling quotas for cases and control women in each city were proportionately calculated so that the cities with the highest annual femicide rates included the largest number of cases and control women.

      There’s already a flaw here - bias in selection. By prioritizing 11 of the ~20k cities, towns, and villages in the US which has the highest counts of domestic violence murder of the female, they’re skewing away from instances where there’s… less murder. Of course your homicide rates are going to report higher, no matter what the risk factor.

      It gets better, though - they skew numbers further by eliminating those with a history of abuse and those just too old to care about:

      Two exclusion criteria, age (18–50 years) and no previous abuse by the femicide perpetrator, resulted in the elimination of 87 additional cases (28.3% of 307 cases), with 59 (19.2% of 307 cases) eliminated solely as a result of the latter criterion.

      It’s interesting they don’t actually note what those cities are - it would be good to know if there are other notable stats e.g. crime rate, poverty, safety nets, so on. Heck, they recognize such:

      Another limitation was that we excluded women who did not reside in large urban areas (other than Wichita, Kan) and control group women who did not have telephones. We also failed to keep records of exactly which proxy interviews (estimated to be less than 10% of the total) were conducted in person rather than by telephone, and thus we cannot evaluate the effects of this source of bias. Finally, we have no way to compare the control women who participated with those who did not, and women living in the most dangerous situations may have been less likely to participate as control women. If so, true exposure to the risk factors of interest among women involved in abusive intimate relationships may be greater than our control data suggest, thus inflating our estimates of increased risks associated with these exposures.

      I’m going to go out on a limb and suggest Wichita isn’t a model of prosperity and social safety nets.

      That brings us to another flaw - this study isn’t interested in identifying the spread and impact of all risk factors but instead is only interested in confirming presence of an already-suspected risk factor - another problem they recognize:

      The interview included previously tested instruments, such as the Danger Assessment,16,17 and gathered information on demographic and relationship characteristics, including type, frequency, and severity of violence, psychological abuse, and harassment; alcohol and drug use; and weapon availability. … Perhaps the most important limitation of the study is its necessary reliance on proxy respondents for data regarding hypothesized risk factors for intimate partner femicide cases.

      This flaw entirely precludes consideration for the whether or not the presence of the firearm was material in the person’s decision to murder e.g. impulsivity, whether or not they’d have just used another implement, etc.

      That brings us to the most egregious flaw - simple, classic misleading through emotional appeal. Setting aside the selection bias and risk of over-representation, what is the actual rate and actual factor? You’ll note none of the studies seem to actually address this. Going with Violence Policy Center’s analysis of 2019 data, they at least provide numbers:

      In 2019, there were 1,795 females murdered by males in single victim/single offender incidents that were submitted to the FBI for its Supplementary Homicide Report.

      So, in 2019, a given woman was subject to odds of five ten thousandths of a percent (1,795/~330 million) likely to be murdered in domestic violence. If we extrapolate up to an expected life span of, say, 80 years, a given woman has been exposed to an ~0.04% total likelihood of being murdered in domestic violence. Oh, but that would hypothetically only be ~0.009% without those firearms; clearly they’re the problem.

      This source also provide a breakdown of implements:

      Nationwide, for homicides in which the weapon could be determined (1,566), more female homicides were committed with firearms (58 percent) than with all other weapons combined. Knives and other cutting instruments accounted for 19 percent of all female murders, bodily force 10 percent, and murder by blunt object five percent. Of the homicides committed with firearms, 65 percent were committed with handguns.

      Despite the arguments made regarding how firearms are the devil for making murder so easy, fists and knives gave an incredible showing of ~1/3 the murders. Notably, John Hopkins provides no hyperbole about knives. Weird, that. Notably absent is any implication of the presence of any of those items increasing risk.

      This data also highlights clear skew toward some states regarding domestic violence homicide rates. Want to place a bet on where significant portions of the John Hopkins data came from?

      For that year, Alaska ranked first as the state with the highest homicide rate among female victims killed by male offenders in single victim/single offender incidents. Its rate of 5.14 per 100,000 was more than four times the national rate. Alaska was followed by New Mexico (2.64 per 100,000) and Nevada (2.28 per 100,000). The remaining states with the 10 highest rates, all of which had female homicide victimization rates higher than the national rate, can be found in the chart below.

      Ultimately, we’re left with not a lot of support for Johns Hopkins’ stance - which makes sense, as they can’t really seem to support it either.

      I’m all for addressing domestic violence, but let’s not lie to ourselves and pretend it’s all sunshine and rainbows without firearms, and let’s not thoughtlessly share the conclusions of biased sources as if they’re meaningful - we’ve had enough erosion of sense over the last decade.

      • Doomsider@lemmy.world
        link
        fedilink
        arrow-up
        7
        arrow-down
        3
        ·
        edit-2
        1 year ago

        This is a lot of splitting of hairs on your part. Are you a social scientist and a statistician? If not, I will defer to the experts on this. The amount of unreported domestic abuse dwarfs the amount that is reported. Also, solely focusing on deaths is a misnomer. Being threatened by an abuser with a gun is rather common and also detrimental to the mental health of the victim.

        • Jeremy [Iowa]@midwest.social
          link
          fedilink
          English
          arrow-up
          3
          arrow-down
          4
          ·
          1 year ago

          This is a lot of splitting of hairs on your part.

          I’m not sure I’d consider criticism of Johns Hopkins tendency to make assertions not supported by underlying sources and tendency to use sources with glaring methodological flaws and myriad biases to be merely splitting hairs - the distinctions highlighted are both substantial and serious.

          Are you a social scientist and a statistician? If not, I will defer to the experts on this.

          I am a software engineer. Analysis is my bread and butter.

          You’ll note my criticism isn’t of their ability to compute statistics, but rather the methodology used for identifying data points for consideration having flaws skewing outputs and for their survey being an exercise in confirmation bias.

          Feel free to defer to others - however, please understand you’re also waiving your right to reference or discuss this study when you decide you aren’t going to bother to understand it and what it’s actually stating. I’m not comfortable opting to skip the critical thinking phase, but you do you.

          The amount of unreported domestic abuse dwarfs the amount that is reported.

          Nifty. I’m not sure how the homicides would be under reported, though - given that’s the subject.

          Also, solely focusing on deaths is a misnomer. Being threatened by an abuser with a gun is rather common and also detrimental to the mental health of the victim.

          You may have meant methodological flaw.

          Either way, given the subject was deaths as raised by Johns Hopkins, feel free to provide them such feedback.

          I’m sure they’ll get right on it.