To help clarify whether firearms were used in crimes or just possessed it might be helpful to take a look at emergency department visits for gunshot injuries. The rough breakdown is 1/3 fatal and 2/3 non fatal. Just over half are suicides. The rest are criminal or accidental. It’s an increasing number. Further take a look at the increased number of patients dying while in transit to EDs. University of Washington has those numbers and they’re increasing.
As criminal justice policy is really a state issue I’d be curious to know which states are declining or increasing and then try to see what’s happening in those states that may be contributing to the changes.
Hi Jeff! Great post. Quick question regarding SRS reports. Do you know if BJS used those SRS-style reports into account when they produced any state-level estimates? Or was the SRS info only used to produce National estimates?
I'm curious how confident you are that the recently flat trend in squishy, often unavoidably subjective violence statistics is real even as homicides have gone up. The minimal crimes closest to not counting at all are typically the most numerous. Not once have I found it useful to count "one punch plus one murder equals two crimes." Politics, including national politics, influences crime counts in many ways, as you know, having discussed local drivers of spurious crime stat changes in the context of New Orleans.
Even victimization surveys are very subjective -- less subject to local politics, but more subject to subjective personal judgments that may also conform to trends. Take https://bjs.ojp.gov/violent-victimization-race-or-hispanic-origin-2008-2021 . The 20/1000 violent crime rate is about five times higher than police report stats (I think this is an apples to apples comparison -- correct me if wrong), a point Krugman should have addressed while claiming violent crime is probably under 5/1000. But also look at racial disparities. Theres a 7:1 black:non-black rate disparity in homicide rates, and a 3:1 violence disparity per the FBI, but blacks and whites are victimized by overall violence at essentially the same rate per crime victimization surveys. 3:1 versus 1:1 is not remotely close, and I don't think it's going off on a limb to say that 1:1 is too low, presumably due to subjectivity. Again, the most frequent crimes and the most frequent victims have the greatest influence. I don't know how the NCVS statistics account for ongoing abuse victims who claim double or triple digit assaults per year, but the infrequent scenario can meaningfully affect totals.
The squishy spurious influences on violence statistics are different today than they were a decade ago, which to me seems like an adequate explanation why the statistics have become decoupled from homicides. But you may have reasons I don't know of to place more trust in that flat statistical trend.
To help clarify whether firearms were used in crimes or just possessed it might be helpful to take a look at emergency department visits for gunshot injuries. The rough breakdown is 1/3 fatal and 2/3 non fatal. Just over half are suicides. The rest are criminal or accidental. It’s an increasing number. Further take a look at the increased number of patients dying while in transit to EDs. University of Washington has those numbers and they’re increasing.
https://jeccm.amegroups.org/article/view/7345/html
As criminal justice policy is really a state issue I’d be curious to know which states are declining or increasing and then try to see what’s happening in those states that may be contributing to the changes.
Hi Jeff! Great post. Quick question regarding SRS reports. Do you know if BJS used those SRS-style reports into account when they produced any state-level estimates? Or was the SRS info only used to produce National estimates?
I'm curious how confident you are that the recently flat trend in squishy, often unavoidably subjective violence statistics is real even as homicides have gone up. The minimal crimes closest to not counting at all are typically the most numerous. Not once have I found it useful to count "one punch plus one murder equals two crimes." Politics, including national politics, influences crime counts in many ways, as you know, having discussed local drivers of spurious crime stat changes in the context of New Orleans.
Even victimization surveys are very subjective -- less subject to local politics, but more subject to subjective personal judgments that may also conform to trends. Take https://bjs.ojp.gov/violent-victimization-race-or-hispanic-origin-2008-2021 . The 20/1000 violent crime rate is about five times higher than police report stats (I think this is an apples to apples comparison -- correct me if wrong), a point Krugman should have addressed while claiming violent crime is probably under 5/1000. But also look at racial disparities. Theres a 7:1 black:non-black rate disparity in homicide rates, and a 3:1 violence disparity per the FBI, but blacks and whites are victimized by overall violence at essentially the same rate per crime victimization surveys. 3:1 versus 1:1 is not remotely close, and I don't think it's going off on a limb to say that 1:1 is too low, presumably due to subjectivity. Again, the most frequent crimes and the most frequent victims have the greatest influence. I don't know how the NCVS statistics account for ongoing abuse victims who claim double or triple digit assaults per year, but the infrequent scenario can meaningfully affect totals.
The squishy spurious influences on violence statistics are different today than they were a decade ago, which to me seems like an adequate explanation why the statistics have become decoupled from homicides. But you may have reasons I don't know of to place more trust in that flat statistical trend.