An important fact is that Americans also think that crime, and specifically violent crime, are on the rise. This is contrary to actual data. So the question are "does news distort our views?" and "does news make us feel more unsafe than we actually are?". Certainly the answer to both is "yes"
https://www.pewresearch.org/short-reads/2024/08/29/the-link-...
https://news.gallup.com/poll/652763/smaller-majorities-say-c...
Just in my household alone, we've been victim of 4 assault+ attacks in the past 5 years. Two resulted in arrests, zero resulted in charges.
Is that normalized in data from 20 years ago?
Finally the same reasons that cause underreporting today existed in the past.
1. Not reporting low level crimes feeling as it has no point.
2. Criminals not reporting when they are the victims of a crime out of fear.
3. When the victim takes the matter of into their own hands
I'll even suggest that due to racism and lack of accountability in the past it's possible more criminal complaints were ignored compared to today.
They use FBI data for reported crimes as well as BJS data which interviews people and ask them if they have been the victim of a crime.
It's probably worth taking a minute or two to read their short explanation as I think it is not what people are assuming.
If you don't know how to evaluate that how can you claim there's a massive under reporting of violent crime?
The last crime I was the victim of was having my car broken into around 2003. I didn't report it but I was angry. Something bad happened to me through no fault of my own and nothing I could have done reasonably would have prevented it.
There are emotions involved and I'm sure for a while I felt as though crime was everywhere and I would be the victim again.
> Something bad happened to me through no fault of my own and nothing I could have done reasonably would have prevented it.
It shouldn't happen period. Effectively, at least in the USA, we've given up on this kind of crime and just expect it. Most people take the POV it is their fault (they left something visible in the car). That's BS. You are not responsible for the thief's actions. The thief is. Period!
There are places in the world where this mostly doesn't happen at all. (Japan, Singapore). You should be able to leave stuff in your car and not have it be stolen.
We, as a society, have given up on even trying to enforce this in any way shape or form AFAICT. My belief is, most of this type of crime is by just a few people repeating the same crime. Honeypots would catch those and lower the numbers by 90-95% IMO.
I asked chatgpt about why low level crime is low in Japan and it gave three interesting reasons.
1. Collective culture - basically focusing the harmony of society over the self. Not doing things that would harm others.
2. Fear of shame by others if caught
3. Communities are against anti social behavior
Assume these are correct for low levels crimes. Does this sound like America? We are a country where the freedom to be an asshole has high value. Individualism and personal freedom are over community.
Think about loud cars, it provides personal pleasure at the expense of others.
I also think about how much HOAs are hated (though some support them obviously because they exist*).
There's also a movement against empathy, another reason you might not want to cause harm to others.
Of course, it applies both ways. We should not trust the reports of violent crime going up in the late pandemic era. The politics of "Defund the Police" suggests that this was a deliberate tactic.
> Just in my household alone, we've been victim of 4 assault+ attacks in the past 5 years. Two resulted in arrests, zero resulted in charges.
Second, you have reported those, so they have been recorded and accounted for in this data.
If you click a few of the links on the Pew site you'll land here[0]
| The FBI publishes annual data[1] on crimes that have been reported to law enforcement, but not crimes that haven’t been reported.
| BJS, for its part, tracks crime by fielding a large annual survey of Americans ages 12 and older[2] and asking them whether they were the victim of certain types of crime in the past six months. One advantage of this approach is that it captures both reported and unreported crimes.
[0] https://www.pewresearch.org/short-reads/2024/04/24/what-the-...
[1] https://cde.ucr.cjis.gov/LATEST/webapp/#/pages/explorer/crim...
[2] https://bjs.ojp.gov/data-collection/ncvs
[Side note]: In [0] you will also notice that they mention that the FBI changed the definition of rape in 2013. The former definition was limited to women, specifically vaginal penetration, and "forcible". In 2013 this expanded to include vaginal and anal, remove sex, and removes "forcible". Also to note that the MeToo movement started in 2006 and gained full attention in 2017. There's a commonly held belief that the rate of reporting substantially increased due to the change of definition and the greater public attention given to the subject. Believe that or not, but this is context needed to evaluate that data.
My anecdote would suggest there’s no crime, but we know that’s not true and why we have statistics.
Not sure you can do much with that in a discussion/argument when there is no proof of that provided other than a personal anecdote.
Edit: Actually maybe I misread it, they are saying its always been flawed. But still the general point of the thread still holds I think.
One proxy for crime statistics is homicide. Any other stat can be juked, but homicide cannot because bodies stink. To someone who was insisting to me that the UK crime rate has increased over time, I pointed out that it was possible, but unlikely. Homicide is at a multi decade low in absolute numbers, and much lower in per capita numbers. That’s likely to be correlated with violent crime.
So in the case of the UK both homicide and violent crime have followed a gentle downwards trend line, and we can trust the first line completely. Therefore, the second is likely to be true.
Did it convince the guy I explained this to? No it did not. “You don’t live in a rough area like I do”. Ok then.
Think about how afraid people were of kidnappings in the 90s, or drugs in halloween candy, etc. All overplayed, and in the case of halloween candy, a complete fabrication.
It's gotten much worse with social media. Whereas you'd normally only get your hit of "the world is on fire" once or twice a day at your favorite news station now you can get it 24/7/365.
Mitigation matters.
This misreasoning or misrepresentation is at the heart of a lot of nonsense to do with interventions and effects. See COVID and what not.
It is very hard to measure crime rate given exposure.
And finally rates do not matter much. What matters is interventions and policy in reaction to each observation. People feel that each murder not resulting in back propagation through policy intervention space to update the weights to prevent it from happening. Instead only narrative space weights are being updated.
The idea that rare events below some threshold on a per capita per year basis must be ignored is pushed around too and people can see through it.
How much has spend on crime cameras gone up over the last few decades and how do you factory that into crime data? How have people moved? We don't have contact rates for people and criminals so it's all very hard to estimate.
Just because many basic people simply can not study rare events without getting upset doesn't mean we should not do so and try to drive the rate of bad things happening to zero.
It's particularly bad in the UK where rates are actually up yet the media seems to always pick some dimension where arrests or crimes are down and push that. People no longer trust the data process let alone the reasoning behind most media reports. They might not be able to describe it but they can feel it.
I get irritated about this as it's the same contact process and susceptibility problem in epidemiology and yet some how politicized the other way.
> What is important is crime rate given exposure
There's tons of sites that show this to varying degrees, here's a few[0,1,2]. Crime is actually fairly localized. This is either entirely surprising or unsurprising to people. That the difference can be just a few blocks. There's quite a bit of research on the topic, and crime is even the main topic of Steve Levitt, that guy from Freakanomics.
But I'm not sure this is really all that related to the topics. Considering that crime is localized and that victims follow a power distribution (a few people are victims to many crimes while a lot of people are victims to few/no crimes) then that only ends up highlighting the distortion even more.
> What matters is interventions and policy in reaction to each observation
[0] https://crimegrade.org/crime-map/
Exposure to what?
> What matters is interventions and policy in reaction to each observation
Responding to individual incidents usually produces terrible legislation. It's hard to get legislators and the public to back off and stop demanding harmful and expensive reprisals, to instead do something that actually works.
e.g. the very successful https://en.wikipedia.org/wiki/Violence_Reduction_Unit
Sorry, just a pet peeve.
Is it contrary to actual data, or is crime that doesn't get to the point of homicide (so violent attacks, burglaries, shoplifting, petty crime, etc) conveniently under reported, and people have seen nothing come out of reporting it (if police ever comes to begin with) and has given up even trying to get anything done about it?
"When the data and the anecdotes disagree, the anecdotes are usually right."
-Jeff Bezos https://lexfridman.com/jeff-bezos-transcript/#chapter6_amazo...
What is the context or background for this statement? It's completely insane to think something like this
> It's pretty wild for a random HN user to label one of the top 10 most successful technology leaders of our time as "insane"
Here, let me go. I think it is insane that Elon Musk took so much ketamine that he kept peeing his pants and then kept telling people about it. I think it is insane he's been promising that FSD is <2 years away for the past decade.
Given two sets of data: The statistics and your anecdotal evidence.
The probability of anecdotal evidence being correct must be lower than than the anecdotal evidence being incorrect when it conflicts with the statistics since the statistics come from anecdotal evidence.
Therefore, all other things being equal, if you are presented with data that conflicts with your anecdotal evidence the data has the higher probability of being correct.
After all nothing is more trustworthy than statistics. They should never be conflated with lies or damned lies.
You shouldn't trust your perception, and shouldn't trust statistics. You should question both.
My personal perception can be right as well as statistics the issue is when I use my own limited experience to generalize.
Why would these be in conflict unless you take your personal experience and draw a conclusion beyond that?
>After all nothing is more trustworthy than statistics. They should never be conflated with lies or damned lies.
Why can't I trust statistics if they are sometimes lies?
It's not a dig against math or statistics. It's a dig against the people using statistics. What do you do when, statistically, the majority of statistics you're given are lies, wrong, irrelevant, misleading, or out of context? You ground yourself in direct experience.
Even simple events are difficult to evaluate. One podcaster I've listened to for a decade had an issue on a flight. He bumped into a passenger in front of him and words were exchanged. Eventually he was deplaned by the pilot. In his telling of the story, it was completely outrageous and unwarranted. But obviously to others involved, he was completely in the wrong, enough to be deplaned.
So no, "direct experiences" are definitely filtered.
For another example, my daughters feel unsafe walking alone at night in my city's downtown area. Whereas I, walking at the same time in the same area, don't feel any danger at all. Every human experience is mediated by our past history.
Also by others who may have an agenda.
What's more likely. That a person is afraid of walking alone in your city at night because of a personal experience (like being the victim of a crime) or what they heard from other people?
We can redefine this with more statistical language. The claim would then be "When the likelihood of a sample is low, the sample is likely correct." Which is nonsense.
It is worth noting when you continuously are getting low likelihood samples, but the usual conclusion is that you have biased sampling. Maybe the model is bad, but in some sense that's not so different.
Let's go back to normal language. Anecdotally, a HN user might think a $100k/yr salary is not very much. But the data suggests that it is. Is the anecdote right?! No, we are just biased because we're comparing incomes in tech and often around the Bay area.
I think it is no surprise that one of the richest men in the world is out of touch. He is, by definition, a statistical anomaly. ¯\_(ツ)_/¯
Let's redefine it with this language:
"When statistics are used as PR, data gathering is a joke, state and police competence is cratering, and direct experience and observation tells you things are getting worse, things are getting worse".
The problem is the claim doesn't distinguish biased sampling from a biased model. The former is far more likely.
Just ask yourself which is more likely: "my experience is abnormal" or "everyone is experiencing the same thing I am"? In your friend group the latter might be more likely because you're similar locally and culturally. But across the state? Across the country? Across the globe? No, you'd be crazy to think experiences are typical. There's way too many variables at play and even if we were clones we should expect differences.
So no, it doesn't make sense