"The thing I have noticed is when the anecdotes and the data disagree, the anecdotes are usually right. There's something wrong with the way you are measuring it," - Jeff Bezos
That’s a pretty weird take to me. Crime stats and polling of people’s perception of crime show this as clearly the wrong approach for instance.
Most studies show that no matter what direction crime is going in, a substantial majority of people think their neighborhoods are safer and that everywhere else is basically a war zone that is getting worse. There’s a total disconnect locally/nationally in perception that is also detached from crime stats.
All of this is to say that the anecdotes are basically all but worthless in the case of understanding how bad crime is on any appreciable scale beyond a few blocks of one’s neighborhood.
>Crime stats and polling of people’s perception of crime show this as clearly the wrong approach for instance.
The only crime stat you can trust is murder and that's because bodies can't be hidden (easily).
Everything else gets swept under the rug.
When I wanted to report my car broken into I was hung up on three times because of a poor quality line, which was fine before I told them what I was calling for. When I went there in person I had to wait 40 minutes for someone to take my report and give me a reference number for my insurance.
One need only download a community based self reporting app like "citizen" to see how much crime really exist.. It's more than you'd like know. Like the previous poster said, not everything is reported.
I'm sure apps like this are targeted towards an already paranoid and crime-obsessed demographic, and they're going to be full of false positives. Not sure you can take that sample as representative.
But this hasn't changed over decades. Policing never was 100% effective and crime was always underreported. Yet if you ask people it's crime getting worse YoY.
But as long as the stats are not reliable, that's unknown territory. Maybe the police is getting more lazier, maybe less, maybe there are less people reporting because it is seen as useless, maybe there are more because they get tired of it, one can simply not get a picture independently of where the trend is going.
Policing was effective enough that I never had to find someone to unlock a case for me at the store unless I was purchasing something particularly valuable. It was effective enough that you didn't see videos of people looting stores or driving around breaking into cars with impunity.
I don't know, these things were cliche plot devices in 1970s-1980s films and it appears everyone was convinced America headed into Escape From New York future. Naturally, few people had cine/video cameras on them at any time.
And yet, if more people you know report being victims of crime, and the official statistics point to a decrease, the people upthread are happy to declare a decrease.
> The only crime stat you can trust is murder and that's because bodies can't be hidden (easily).
Seems to be a significant decline since the 90’s, followed by an increase since 2016, and a Covid dip followed by a resumption of the increase. It’s still a lot lower than it was in the 90’s.
This is actually a perfect example of Bezos being right on the money. You are measuring it wrong. The first and last articles focus only on violent crime, which is not what most people who are complaining about crime mean here. 538 is a little better, but their charts only go through 2019 before it became a major issue again. Only Vox seems to get it closer to correct (though still hung up on the violence thing):
> One theory that came up again and again is that city residents and visitors are, to some extent, conflating actual violent crime with broader indications of urban disorder.
If you are a leader like Bezos or a city politician you need to meet your customers / constituents where they are and fix the problems they want fixed whether or not it they are saying precisely what they mean. The anecdotes are right and the statistics are wrong.
It’s just a refutation of naive bias towards statistics, which is rampant in big organizations (see the McNamara fallacy). This is codified in the idea of being “data driven”, which is the right thing if your data is a true proxy for the thing you care about; in practice it often isn’t and you have to incorporate some more flimsy or subjective signal to better understand a problem.
Bingo. The obsession with quantification is a crutch that allows uncreative people to delegate their decision making to a mechanical analysis of raw data, rather than a first-principles understanding of their problem space.
It's probably related to how much or little we read about crime more than any true crime level. If we see 10 news articles everyday about crime, then we think there are a lot of crime around. If we read zero articles about crime, then it barely exists in our perception. What happens in reality does not affect our perception as much, as we probably seldomly see it for ourselves and when we see it, it would be difficult to objectively and statistically judge the crime level's direction with such few data points and biased experiences.
It's similar to Hans Rosling's comments about poverty in the 3rd world. It often sounds like poverty is increasing as time goes by, but if looking at statistics, overall poverty is decreasing and have been doing so for decades.
Exactly, in the case of crime there are so many more vicarious anecdotes. If people were only allowed to discuss crimes that they were personally victim to, we would not be under the impression that crime is worse than ever.
When I have to spend 10 minutes finding a store employee to unlock a case for me to buy underwear and socks at Target when I didn't have to five years ago, I conclude that crime has gotten worse.
When I see videos on the internet all the time of criminals just walking into stores and grabbing whatever they want while the security guard looks on and does nothing because the police will side with the criminal if he touches them, it is perfectly valid to assume that crime is worse.
When the police do nothing to enforce the law I do not trust the statistics because they are based on reports to the police.
Maybe the statistics are right, over the whole city, but where I live, crime has gotten worse.
Where do you live? It sounds quite bad. Has SF got that bad? I haven't been there since I worked there in 2015, but I read the authorites gave up on some crime?
I noticed neither any crime or homeless at that time. But people now seem swear it has gotten to be a big problem.
I was last there a couple years ago when the "SF is a lawless wasteland" nonsense was gaining a lot of traction. I saw some needles, some foil, a broken window or two, but otherwise it was a very calm and inviting city. Any city due to greater density will have more observable crime, but people are crapping their pants over exaggerations.
> When I see videos on the internet all the time of criminals just walking into stores and grabbing whatever they want while the security guard looks on and does nothing because the police will side with the criminal if he touches them, it is perfectly valid to assume that crime is worse.
How many videos have you seen on the Internet of stores just calmly going about business with no shoplifting going on? The number of videos on the Internet is not an indication of any overall trend. The stuff you're seeing makes it onto YouTube because they are outliers.
Why have I only started seeing these outliers in the last few years? Also, the worst part of the videos isn't that they happen, but that they are allowed to happen. If the criminals weren't calmly going about their business without even wearing a mask and instead running from the security guard it wouldn't be so frustrating. If the news stories had a mugshot of the police arresting them quickly since their face is all over the internet it wouldn't be so infuriating. If the police didn't allow the criminals to run open air fencing operations right outside BART stations, I wouldn't be this mad.
I dunno, it seems in line with most stereotypes being more true than false [1]. “Common sense” is often derided in online spaces like this, but when there hasn’t been a massive media / social effort to convince the population otherwise [2], it’s pretty reliable.
Taking a quick glance at the articles you linked shows the same behavior as those reporting on the economy - defining disingenuous targets so they can claim their headline is true. To tie this back to anecdotes, I think it comes down to trust. When my neighbor says they’re afraid to lose their job due to housing, food, childcare being a lot more expensive I dont see any motivated reasoning behind that statement. On the other end, economists (and all the articles you linked) have many incentives to distort the truth. On average anecdotes are going to come from a more truthful place - both because you trust the source and know their biases.
I definitely understand where you are coming from and would never deny these things are happening. But my point is that argues for too much trust in anecdotes and people’s perceptions.
We all know memory is incredibly faulty, for instance. Yet people have a very high perception of their own memory’s accuracy. It’s kind of in the same vein. It’s not that people can’t remember things accurately, it’s that we need to start from a place of skepticism when depending on it. Same thing goes for people’s perceptions of crime, the economy, etc. Their anecdotes and lived experience, insofar as they can even accurately explain their lived experience, needs to be put into context re: its value for determining “reality.”
That being said I would never undermine the value of how people feel. If people don’t feel safe, that is a bad thing too. And we can cite all the stats in the world we want but ultimately feeling unsafe is not a good thing and that perception needs to be addressed.
You're making the exact same mistake as the poster you're answering to warns about - you're mixing the objective reality ("absolute crime numbers") to percieved reality of population.
If the population percieves themselves unsafe and unhappy, your numbers don't really mean much to them because to restore happy society you need to look at *perception* and fix the reasoning behind it. Making the crime stats number go down won't do that by itself.
> you need to look at perception and fix the reasoning behind it.
Many times that perception is shaped by the media we consume, which has no obligation to have any connection to the reality on the ground. At that point whatever is done to improve the reality doesn't have to have any impact on people's perception.
Which means that the fix is to change the media not the metrics. That's the gist of the argument - if your improvements of "reality" aren't making people happier, you're changing the wrong metric. In this case you need to change the media, not crime stats.
Changing the underlying reality without improving the perception leads to a political disconnect. Improving the perception without improving the underlying reality just makes people happier, but not better off. You have to do both if your objective is to improve people's lifes.
I agree. So, which one do you prioritize? Currently, it seems to me as if people's (mis)perceptions are leading them to vote against policies that will make them better off, so we need to correct perceptions first.
I don't really watch any broadcast/cable news these days. Was at a restaurant that had it going on their TV and my goodness all it was saying is "FEAR FEAR FEAR FEAR". Absolute mind rotting garbage.
I was at an airport the other day and I was struck by how I saw four different new stations all covering one tornado in a town half way across the country from me. This was top line national news for like 20min with reporters standing by downed trees and going on and on about the “utter devastation.” It seemed pretty small potatoes tbh but maybe I was missing context?
You would think there is some sort of tornado epidemic nationwide based on the way they were acting. Tornadoes happen all of the time, and they are very tragic for those involved, but y’all aren’t hearing about every house that burns down in my city lol
In a different response I actually remarked on how I would never talk down to people who feel unsafe because we can cite all the stats we want in the world, but their feeling secure is very important as well.
That's fair that perception is often wrong - but that's a different issue.
Perception is also in large part the very thing that matters when it comes to crime. That is, do I get to live in peace or in constant worry? Do I get my property priced "fairly" when I sell or dramatically underpriced because of this perception that the area is unsafe? To summarize, in what you describe "crime stats" are ignoring half or two thirds of the problem.
In San Francisco, "crime stats" are further muddied because of massive underreporting and cherry picking the definition. So called quality of life crime might be considered irrelevant because it rarely causes massive loss of property or injury. But it does make life extremely stressful for the locals (depending on the neighborhoods where it might be "tolerated" i.e. left rampant, or might not be tolerated.) In this case, "crime stats" deliberately not measuring anything very relevant.
See also recent discussion of the squatting issue in Spain.
Most studies show that no matter what direction crime is going in, a substantial majority of people think their neighborhoods are safer and that everywhere else is basically a war zone that is getting worse. There’s a total disconnect locally/nationally in perception that is also detached from crime stats.
All of this is to say that the anecdotes are basically all but worthless in the case of understanding how bad crime is on any appreciable scale beyond a few blocks of one’s neighborhood.
* https://www.pewresearch.org/short-reads/2016/11/16/voters-pe...
* https://fivethirtyeight.com/features/many-americans-are-conv...
* https://www.vox.com/future-perfect/23663437/crime-violence-m...
* https://www.pbs.org/newshour/amp/show/as-concerns-grow-aroun...