[ahoy]

This SF district attorney recall thing is a really good example of the media's power to manufacture a story. Even if you accept the premise that crime stats rise or fall in response to DA policies, those stats mostly fell during Boudin's tenure.

But if you show the video of someone robbing a convenience store (in oakland! another city entirely!) enough times on loop, you can convince enough voters that Something Must Be Done.

[avalys]

Statistics are not reality. Crime stats fell because the DA was so tolerant of crime, people stopped bothering to report anything but the most serious ones.

[ajhurliman]

This seems like a huge flaw, the idea that cities and law enforcement agencies are reporting their own crime statistics. There's an obvious conflict of interest, it seems like this should be done by a 3rd party. I know it's not that simple, but at least getting some neutral indication of the level of crime in a city would go a long ways.

Crime is visibly exploding, the police are only recording the most heinous/obvious incidents, and then we're told that crime is down.

[SllX]

Do you have a methodology to go with that proposal?

When the paperwork is filled in, we have reported crime (because the police take a police report) and we have convictions (how many people were convicted is a matter of public record since courts are open). What we don’t have is a paper trail for all the crime that isn’t reported or when the paperwork isn’t filled out.

What can a third party bring to the table that the paper records cannot?

[ajhurliman]

Axon has an app for citizens to report crime and upload evidence, that seems like it would go a really long way to get numbers that are at least correlated with publicly visible crime statistics, and I think that's the main goal. The published numbers are a function of real crimes committed plus how hard the law enforcement agencies and local government are trying to suppress those numbers. The Citizen app would give us the dot product of publicly visible crime and use of the app, but I feel like that's probably more reliable than what's currently being published.

Also, Seattle has a Find-It-Fix-It app that allows people to report various things that are broken with the city, opening up that data to the public would be great. I know it's not crime-related, but getting an overall sense of the health of the city seems correlated. This would require that the app be administered by a 3rd party though.

[SllX]

I’ll give you credit, that is indeed a 3rd party addition that would bring something else to the table that I overlooked.

I can’t say I particularly like it though. I do care about suppressing crime, but not to the point of becoming a police state. That ship may have already sailed between both existing apps like the ones you listed and the prevalence of home security cameras, so I’ll say it’s at least worth considering.

> Seattle has a Find-It-Fix-It app

Is this like SF’s 311?

[ajhurliman]

Similar, yeah. Reports to it used to go completely ignored, but the city has recently been ramping up its efforts to pay attention to it.

[ahoy]

So in this land of topsy turvy logic, high crime stats is a good thing, because that means they're being reported?

[filoleg]

No. High crime stats is pretty much always a bad thing (regardless of whether there is actually more crime driving it up or due to enforcement being overzelaous). But low crime stats aren't always a good thing, like in a situation when they are low because there is no enforcement prosecution happening.

To give you an analogy that might help, high covid infection numbers are bad always, but low covid numbers can be either good (if the covid numbers are truly down) or bad (if the real numbers are just not being tracked properly/swept under the rug, and people are just not getting tested enough).

But to give a direct answer to your original question, yes, low crime numbers are bad when they are the result of prosecution and enforcement not doing their job, as opposed to being the result of crime actually going down.

[inglor_cz]

No, the real (not topsy-turvy) logic behind GPs statement is "garbage in, garbage out", or "you can't trust a statistics unless you know how reliable the inputs are".

[SllX]

You basically can’t trust statistics without statistical literacy, the source data and the time to interpret it.

Lies, damn lies and statistics is still applicable and relevant today, if conveniently forgotten.

Just to bring something else to this discussion, Chesa didn’t run on statistics. He ran on a soft-on-crime message and delivered. Since this is an elected office, he didn’t answer to the Mayor, so he didn’t benefit from the Mayor’s cover but also wasn’t in the same chain of command as SFPD, so when the voters turned, they turned directly on him and this being an elected office, he just didn’t have a good story to tell or the support he needed from the communities he needed it from. The real statistics don’t even matter at that point because statistics don’t run for office or win elections.

[kalleboo]

They never said that high crime stats is a good thing, just that the statistics are unreliable and you can't draw conclusions from them.

[mimikatz]

If an Application team is finding and fixing a lot of vulnerabilities and bugs is that a good thing or a bad thing?

The answer is it depends on the context.

[throwaway894345]

Sometimes there are two problems. In this case: (1) high crime and (2) underreporting.

[beebmam]

Is there evidence for this claim?

[kansface]

I think this is normally a good question, but honestly, in what world would I care about numbers when there is a lady screaming 'Help'/'Asshole' for hours at a time at my door, a homeless guy running a bicycle chop shop in front of my building (he uses a 3' stretch of pipe to deconstruct bikes!), people breaking into the building (taking both packages and keys!), the outside of the building is turned into a communal toilet, the local Walgreens gets looted out of business, and people get stabbed uncomfortably often just around the corner? That was my somewhat unfettered, actual pandemic experience in Hayes Valley turned Tenderloin. You could claim that's not entirely the fault of the DA, probably with merit, too, but I just don't care all the same. If the DA claims no agency, takes no responsibility for any of this mess, and the plan is even more of the same thing, whats the point?

[vineyardmike]

There is tons of AnecData suggesting this is true. The recalled DA himself admitted to not reporting his car breaking because he didn’t think it’d be prosecuted.

I can name roughly 10 people in my life who said they didn’t bother. Another 5 who went to police only to be told by them that there was nothing that would happen and that they shouldn’t waste time on finishing the paperwork. To complete the data, I know only 1 person who fully reported a crime, and it was an extreme violent crime.

[elefanten]

The pervasive testimony from local law enforcement that the DA's office won't even respond about prosecuting most offenses. Not just minor offenses -- theft, burglaries, assault and at least one manslaughter/homicide case.

The second hand testimony from residents and business owners saying the police are telling them not to expect them to be able to prosecute anyone.

But there's also the evidence of what Boudin openly said and advocated. Sometimes the message and the stated intent are worth replacing regardless of how effective they've been at their goals.

[heretogetout]

Testimony from law enforcement should be dismissed out of hand (including in court) without hard evidence backing it up. The second hand testimony could just mean the police are acting like children, we don't know.

[throwaway894345]

It's not so much "as a result of the DA's policies" in particular (although I don't doubt this contributes), but all over the country discretionary policing has fallen as a result of pressure from anti-policing activists on politicians and police departments. The result is that crime overall rises, but "lesser" crimes which are more likely to fall under the rubric of "discretionary" appear to fall because the police are effectively prohibited from looking for them. However, crimes which aren't discretionary such as homicides (in other words, police can't overlook homicides) are rising sharply. This has shown up in some of the "Ferguson effect" literature (and I suspect it will also show up in the "Minneapolis effect" literature once criminologists compile and process the data), but I don't have a link handy at the moment.

[joshuamorton]

In SF, rapes and robberies are both down in 2021/2022 compared to 2018. Murder and shootings are comparable with 2018. Car thefts are up. All of which completely goes against this narrative you've described.

The rising crime narrative is entirely an artifact of the pandemic causing historically low crim because for like 8 months no one did anything.

[throwaway894345]

> In SF

I'm not sure about SF in particular, but my account was for the US as a whole.

> Murder and shootings are comparable with 2018.

This is incorrect. SF homicides:

    Year | Homicides
    -----+----------
    2018 | 44
    2019 | 41
    2020 | 48
    2021 | 56

That's a 25% increase since 2018. Source: https://www.nbcbayarea.com/news/local/san-francisco/sf-mayor....

> The rising crime narrative is entirely an artifact of the pandemic causing historically low crim because for like 8 months no one did anything.

Again, I'm not sure about SF in particular, but nationally the "rising crime narrative" is entirely an artifact of soaring crime, especially violent crimes. For example, homicides have been falling for decades, and then the trend abruptly reversed around 2014. Criminologists have identified specific reversals in cities that experienced large BLM protests immediately following those protests, and this is all pre-pandemic data (the 2020 and 2021 data is still being compiled and analyzed as far as I know).

For example (https://www.nber.org/system/files/working_papers/w27324/w273...):

> For investigations that were not preceded by "viral" incidents of deadly force, investigations, on average, led to a statistically significant reduction in homicides and total crime. In stark contrast, all investigations that were preceded by "viral" incidents of deadly force have led to a large and statistically significant increase in homicides and total crime. We estimate that these investigations caused almost 900 excess homicides and almost 34,000 excess felonies. The leading hypothesis for why these investigations increase homicides and total crime is an abrupt change in the quantity of policing activity. In Chicago, the number of police-civilian interactions decreased by almost 90% in the month after the investigation was announced. In Riverside CA, interactions decreased 54%. In St. Louis, self-initiated police activities declined by 46%.

[joshuamorton]

> That's a 25% increase since 2018. Source: https://www.nbcbayarea.com/news/local/san-francisco/sf-mayor....

And in 2022 we're on track for somewhere between 40 and 48, precisely in line with prior years. A single outlier year does not a trend make.

> and then the trend abruptly reversed around 2014

Until 2016, when it began dropping again, and then it jumped up a ton in 2020 but we don't really have data yet to know if that was a fluke or not. (it really looks like the 2014 thing was just regression to the mean).

[upsidesinclude]

Have you been to SF? That's your evidence

[worker_person]

The evidence was stolen. Person filing the complaint was mugged. Mugger was hailed as a hero.

[mensetmanusman]

Statistics are not reality.

Cites statistics.

[quartesixte]

Crime stats as presented and as experienced can diverge quite heavily.

I think the one thing with the data is how much crime it takes to make a neighborhood feel unsafe. And my hypothesis is that the threshold is very, very low. Like double digits low. It doesn’t take much to rattle the psyche of a community.

But sure. Crime falls by 40% during a DA’s tenure (hypothetical example) so all is well right? But what is the absolute number and what does that absolute number translate to in the real world?

[derwiki]

The number of employees who quit the DA’s office because of Chesa was not manufactured, and is ultimately why I voted the way I did.

[newbie2020]

This is a great case of where statistics don’t tell the whole story. I’ve seen multiple crimes in SF that I don’t bother reporting since I know the police won’t do anything about it. So up this years’ stats by at least a few notches in the vandalism, car breakin, theft, sale of illegal goods, and bodily harm inflicted upon another categories…

[AdrianB1]

If you don't prosecute anyone, crime stats fall. Why? Because nobody will care to report crimes knowing they will not be prosecuted.

[dragonwriter]

> This SF district attorney recall thing is a really good example of the media's power to manufacture a story

Given the degree to which opposition to Boudin comes from other progressive Democrats, including politicians, in SF, the article we are discussing about a “message” to “Democrats and the nation” is an example of that.

[ardit33]

It seems that the people of SF don't agree with you. Crime didn't fell, prosecutions did. It is like the Trumpian 'if you don't test, then covid doesn't exist'.

There you go the clear graph: https://www.economist.com/united-states/2022/06/05/why-san-f...