I was looking into this based on the reports that reduced UCR reporting explains the apparent 2023 crime reduction, not actual decreases in crime. I see the logic, but assumed that raw reported incidents would be used to compute rates based on covered population. It's the same principle that enables the DOJ survey (touted as more accurate) to state crime rated despite only surveying 250k people from 150k households.
I shared this with my parents, and my mom caught that in the tablet here are bot city and county agencies... how does the UCR account for double counting? For example, I see Houston TX (city) and Harris TX (county), and Houston is in Harris county. Have you looked into definitive population covered to ensure it means what we think (a population of 311M *individuals* is covered by reporting agencies, none counted by multiple agencies)?
There is a long term trend of crime falling that was reversed in 2020 with the George Floyd riots. Then in 2021 you had the reporting problem. By 2022 things were reverting to the trend but perceptions always lag reality. And there are a lot more guns out there. Last I checked it was 58 straight months of a million plus legal sales and an unknown number of legal private sales. Presumably, this creates a deterrent effect since bad guy don't/cant participate in the legal market.
Love it Jeff. You make the complex (relatively) simple. Love to refute myths with facts!
"It’s a tiny difference and I’m not sure why it exists, but these things happen with crime data."
What happens to crime data is a crime. lol
Thanks for the analysis, you do make it much more understandable.
I was looking into this based on the reports that reduced UCR reporting explains the apparent 2023 crime reduction, not actual decreases in crime. I see the logic, but assumed that raw reported incidents would be used to compute rates based on covered population. It's the same principle that enables the DOJ survey (touted as more accurate) to state crime rated despite only surveying 250k people from 150k households.
I shared this with my parents, and my mom caught that in the tablet here are bot city and county agencies... how does the UCR account for double counting? For example, I see Houston TX (city) and Harris TX (county), and Houston is in Harris county. Have you looked into definitive population covered to ensure it means what we think (a population of 311M *individuals* is covered by reporting agencies, none counted by multiple agencies)?
There is a long term trend of crime falling that was reversed in 2020 with the George Floyd riots. Then in 2021 you had the reporting problem. By 2022 things were reverting to the trend but perceptions always lag reality. And there are a lot more guns out there. Last I checked it was 58 straight months of a million plus legal sales and an unknown number of legal private sales. Presumably, this creates a deterrent effect since bad guy don't/cant participate in the legal market.