One of the common responses to my piece from last week on the widespread — albeit preliminary — reported decline in murder and crime could be summarized in the below comment from Twitter: “How accurate can recent burglary/larceny crime stats be when we know they're not being recorded in many locations where such theft won't be prosecuted? I imagine the same goes for some other stats here.”
Thank You very much for this! Your take on what can be gleaned from the NCVS was more or less where my thinking had taken me, but I have more confidence in you than myself on these things. :)
In your last paragraph, you wrote: "Crime data is inherently flawed...", and that made me think of the quote attributed to George Box - "All models are wrong, but some are useful". Crime data, like any data can be useful, even though it can never be perfect.
7 percent of identity thefts are reported to law enforcement per BJS. Do we create policy based on the 7 percent? Crimes reported to law enforcement are filled with endless problems which is why we have the National Crime Victimization Survey (which is routinely ignored).
Per the FBI's data from 2022 and their slight decrease in violent crime, we all recognize that violence "may" have gone up considerably for a multitude of reasons, underreporting may be the tip of the iceberg. I just did an article on family members and non-strangers being responsible for most violence. How many of these events were reported? I'm guessing that the number is quite low. It's the bulk of violent crime.
Yet crimes reported to law enforcement is the hand dealt, warts and all. It's all we have beyond the ignored National Crime Victimization Survey. There is a point where if crime statistics do not show a clear pattern of increases or decreases, does it really count?
As always, thanks for your analysis. You do great work. But I believe that we need to rethink what we report and how we report it while understanding that the caveats would be endless.
An example: Vermont (via the Associated Press-today) is reporting a huge increase in violence in one of the lowest crime states in the country per FBI numbers. Small numbers allow for large percentage increases or decreases. Yet that wasn't mentioned.
As you point out, staff shortages is likely to be part of the issue, but I also believe that the lack of police legitimacy--trust in law enforcement--is also to blame, especially in minority and immigrant neighborhoods. Residents in these neighborhoods tend to have negative interactions with police, so they don't often report crime.
You are absolutely right. However, I would caution against saying: "Both perpetrators and victims likely know that clearance rates for property crimes are usually pretty low throughout the country." A huge portion of perpetrators, those who are young (16-25ish) and live in poverty, haven't any idea about clearance rates (or, for that matter, the penalty for the crime they're committing).
Noooo this cant be true because all my liberal friends say that crime is down across all major cities. Doesnt match what I see with my eye but thats okay because....um.....well.............you know trump was racist right???
New here - appreciating the context and consideration you bring.
You describe the reality of how the choices of the people and prosecutors impact crime stats, but I offer the discretion of the responding officer weighs heavily on the resulting stats. Sure, people decide whether or not to call the police, but the crime stats don’t come from the calls for service. Crimes ‘known to police’ come from a cop making the choice to pull a case number (how varies by department and system) and then to attribute an offense to that case number’s report. Seeing Seattle’s drop in the percent of property crime calls resulting in a report is certainly influenced by slower response times - if dude is still there, it’s much more likely a report will be taken - but regardless of when a response occurs, the choice to turn a call for service into a crime report should hold steady. It’s the patrol cop or community service officer who makes that choice - and they stand squarely at the point where the other perceptions and decisions of the people and prosecutors come together. But they make that choice under the pressure of a backlog of other calls for service holding, a field supervisor pushing for the next call to get handled, and a salty detective who will ask why a report was taken if it’s got low solvability factors. Cops hand out call for service numbers (rather than case number) for expediency and so people feel better served. But it only becomes a crime for stats purposes if a case number and associated report are filed in the first place.
Stilll if we judge by the kinds of crimes that can’t be easily covered up or ignored, serious bodily injury, death, stolen cars, the property crime rate is usually something individuals are responsible for preventing...or used to be. “Property crimes are up because I left my $3000 electric bike in the front yard overnight and now it’s gone.” (Actually saw that complaint on Nextdoor recently). Frankly, I’d have been too embarrassed to report my failing. Pretty soon, I guess, we’ll become like Victorian England, expecting children to be hung for stealing bread. As long as I know the “bigs” are coming down, I’ll take care of my own property.
Really good analysis Jeff. But there are independent studies that show crime reporting is lower than even the NCVS indicates. Also, ORC is just one type of retail theft and big retailers just one type of victim. How many ground floor retailers in Seattle or SF or SJ have closed? Thousands between all three and most are not Walgreens or CVS or Dicks.
Yet we still measure commercial theft by population which is wrong. Commercial theft rates should be measured by the number of businesses. The SF Chronicle recently got this very wrong when they claimed shoplifting had dropped from 2022 to 2023. They failed to note a larger increase in retail closures. So yes, there were fewer shopliftings, but there were also fewer victim businesses to victimize. Meaning shoplifting hadn't declined at all.
It’s not just reporting times or prosecutions that measure law enforcement effectiveness and perhaps it’s effect on reporting. How about looking at confidence in policing by looking at trusted profession research? It’s declining along with reporting
And I still believe murder is down - to some degree - because emergency medicine is improving. But we’ll see.
I think you have been in research long enough to make credible inferences in your analysis based on things you know outside of the official statistics you so accurately report as well as ones your have cleverly devised and track on your own. I predict your multi city murder tracking figures will make you far more relevant than the FBI. Timeliness is also a statistical point of reference- is it not?
Thank You very much for this! Your take on what can be gleaned from the NCVS was more or less where my thinking had taken me, but I have more confidence in you than myself on these things. :)
In your last paragraph, you wrote: "Crime data is inherently flawed...", and that made me think of the quote attributed to George Box - "All models are wrong, but some are useful". Crime data, like any data can be useful, even though it can never be perfect.
7 percent of identity thefts are reported to law enforcement per BJS. Do we create policy based on the 7 percent? Crimes reported to law enforcement are filled with endless problems which is why we have the National Crime Victimization Survey (which is routinely ignored).
Per the FBI's data from 2022 and their slight decrease in violent crime, we all recognize that violence "may" have gone up considerably for a multitude of reasons, underreporting may be the tip of the iceberg. I just did an article on family members and non-strangers being responsible for most violence. How many of these events were reported? I'm guessing that the number is quite low. It's the bulk of violent crime.
Yet crimes reported to law enforcement is the hand dealt, warts and all. It's all we have beyond the ignored National Crime Victimization Survey. There is a point where if crime statistics do not show a clear pattern of increases or decreases, does it really count?
As always, thanks for your analysis. You do great work. But I believe that we need to rethink what we report and how we report it while understanding that the caveats would be endless.
An example: Vermont (via the Associated Press-today) is reporting a huge increase in violence in one of the lowest crime states in the country per FBI numbers. Small numbers allow for large percentage increases or decreases. Yet that wasn't mentioned.
As you point out, staff shortages is likely to be part of the issue, but I also believe that the lack of police legitimacy--trust in law enforcement--is also to blame, especially in minority and immigrant neighborhoods. Residents in these neighborhoods tend to have negative interactions with police, so they don't often report crime.
You are absolutely right. However, I would caution against saying: "Both perpetrators and victims likely know that clearance rates for property crimes are usually pretty low throughout the country." A huge portion of perpetrators, those who are young (16-25ish) and live in poverty, haven't any idea about clearance rates (or, for that matter, the penalty for the crime they're committing).
Noooo this cant be true because all my liberal friends say that crime is down across all major cities. Doesnt match what I see with my eye but thats okay because....um.....well.............you know trump was racist right???
New here - appreciating the context and consideration you bring.
You describe the reality of how the choices of the people and prosecutors impact crime stats, but I offer the discretion of the responding officer weighs heavily on the resulting stats. Sure, people decide whether or not to call the police, but the crime stats don’t come from the calls for service. Crimes ‘known to police’ come from a cop making the choice to pull a case number (how varies by department and system) and then to attribute an offense to that case number’s report. Seeing Seattle’s drop in the percent of property crime calls resulting in a report is certainly influenced by slower response times - if dude is still there, it’s much more likely a report will be taken - but regardless of when a response occurs, the choice to turn a call for service into a crime report should hold steady. It’s the patrol cop or community service officer who makes that choice - and they stand squarely at the point where the other perceptions and decisions of the people and prosecutors come together. But they make that choice under the pressure of a backlog of other calls for service holding, a field supervisor pushing for the next call to get handled, and a salty detective who will ask why a report was taken if it’s got low solvability factors. Cops hand out call for service numbers (rather than case number) for expediency and so people feel better served. But it only becomes a crime for stats purposes if a case number and associated report are filed in the first place.
Stilll if we judge by the kinds of crimes that can’t be easily covered up or ignored, serious bodily injury, death, stolen cars, the property crime rate is usually something individuals are responsible for preventing...or used to be. “Property crimes are up because I left my $3000 electric bike in the front yard overnight and now it’s gone.” (Actually saw that complaint on Nextdoor recently). Frankly, I’d have been too embarrassed to report my failing. Pretty soon, I guess, we’ll become like Victorian England, expecting children to be hung for stealing bread. As long as I know the “bigs” are coming down, I’ll take care of my own property.
Really good analysis Jeff. But there are independent studies that show crime reporting is lower than even the NCVS indicates. Also, ORC is just one type of retail theft and big retailers just one type of victim. How many ground floor retailers in Seattle or SF or SJ have closed? Thousands between all three and most are not Walgreens or CVS or Dicks.
Yet we still measure commercial theft by population which is wrong. Commercial theft rates should be measured by the number of businesses. The SF Chronicle recently got this very wrong when they claimed shoplifting had dropped from 2022 to 2023. They failed to note a larger increase in retail closures. So yes, there were fewer shopliftings, but there were also fewer victim businesses to victimize. Meaning shoplifting hadn't declined at all.
It’s not just reporting times or prosecutions that measure law enforcement effectiveness and perhaps it’s effect on reporting. How about looking at confidence in policing by looking at trusted profession research? It’s declining along with reporting
And I still believe murder is down - to some degree - because emergency medicine is improving. But we’ll see.
I think you have been in research long enough to make credible inferences in your analysis based on things you know outside of the official statistics you so accurately report as well as ones your have cleverly devised and track on your own. I predict your multi city murder tracking figures will make you far more relevant than the FBI. Timeliness is also a statistical point of reference- is it not?
Have a Merry Christmas Jeff