Hacker News Reader

I haven’t read the paper and by your admission you haven’t either. You skimmed it. Do you find it reasonable to believe that you could come up with potential objections to the paper that the authors didn’t also think about?

You made a claim in #1 about the likelihood of people calling police when they think police won’t do anything about it. That claim sounds plausible but is it correct? Do know of studies on this?

If you, a non expert, can think immediately of three thoughts then it seems highly likely the authors - who spent effort and time on the topic - also thought about these same thoughts. Maybe these things aren’t addressed because it’s obvious to them that these aren’t valid objections and it’s obvious to their target audience.

I get my car broken into biyearly. After the first time, I don't report because it's a waste of time. Same when bicycles are stolen from my friends. I don't need a study to say the sky is blue.

But, I do have studies showing data is often made up or cherry-picked. So when a study disagrees with what most of us are experiencing, when cops have publicly announced they won't investigate crimes because the prosecutor won't prosecute, I'm going to need a lot more evidence that crime is actually decreasing.

https://en.m.wikipedia.org/wiki/Replication_crisis

https://crosscut.com/2014/12/d-word-seattle-confronts-dilemm...

Here’s an interesting Twitter thread that claims cops commit orders of magnitude more physical assaults than arrests for serious crimes. There is a possibility that having fewer police leads to less violence in the community.

https://twitter.com/equalityAlec/status/1430159902551445514

> You made a claim in #1 about the likelihood of people calling police when they think police won’t do anything about it. That claim sounds plausible but is it correct? Do know of studies on this?

There's a number of studies, some of them contradictory (perhaps because they're looking at different populations of people). Hagan et al[1] found that in urban African-American communities, cynicism towards police and the legal system was a major factor in whether or not someone called the police after being a victim of a crime (in a way that is actually a bit counterintuitive to me).

Anyways, I owe the authors an apology. I took another look at the paper and noticed that they (sort of) addressed the issue of under-reporting, as was mentioned in the article.

Tautologically, they point out that underreporting of the crime rate doesn't impact the accuracy of statistics showing that the number of reported crimes went down. However they do admit that this makes it harder to make a causal link.

> Maybe these things aren’t addressed because it’s obvious to them that these aren’t valid objections and it’s obvious to their target audience.

That's an interesting point, and I suppose it's hard to say without being an expert. OTOH, from talking to people I know who have left academia[2], the 'publish or perish' mentality makes it easier to put blinders on when faced with something that's inconvenient.

[1]: https://doi.org/10.1073/pnas.1722210115

[2]: Admittedly this is going to be a biased sample

Pertaining to your last paragraph. I very much agree with what you wrote. I think a better way to state my point is this. I try to read academic papers with skepticism and when I think of an obvious potential objection I think it highly likely the authors are aware of this and didn’t address it because their knowledge of topic is such that they know it’s not really an objection. I then try to find out whether or not my potential objection really is valid or invalid. I assume the latter when I don’t have more information or the time to delve deeper into the topic.

Thanks for the link to the study.

"The 'publish or perish' mentality makes it easier to put blinders on when faced with something that's inconvenient."

I'd like to provide a competing take on this.

In my experience, it's the job of the peer reviewers to force the authors to test alternative hypotheses, especially if it's something obvious or inconvenient. I once published a methods paper, where the reviewers got aggrieved I didn't mention a complementary, but mostly unrelated, method in my discussion. At first I was annoyed, because it was OBVIOUS to me that their objection was flippant.

However, once I thought about it, I realized that the reviewer was probably representative of the audience that's going to read my paper. Their criticism while tiresome was entirely valid. So I sat down and spent a week coding a series of experiments that showed how the two methods (mine, and theirs) were different and how they could complement each other. That translated into a paragraph in the discussion section and a couple of supplementary figures, and overall a better paper.

So to go back to the original point of the GP ("maybe things got left out because they are obvious"):

Obvious things, unless it's something really trivial, are only obvious to the authors of the paper. If you expect an educated reader to raise an obvious/natural objection, you should really address it in your text head on. And if you don't, then hopefully the peer review will catch it.

PS

This is not meant to be a commentary on the paper linked in this thread. I have not read it, and have no opinion on it. I just wanted to opine on the general topic of peer review and whether "obvious" things should go into an academic paper.

Correlation is not causation, but there certainly is a lot of evidence that causally links poverty with crime and drug-use, and poverty from accumulating petty fines and late-fee penalties. It seems to me the burden is on the broken-window excuse of a hypothesis.

I've spent weeks in a middle-class community never seeing a cop. Surprise surprise, the crime-rate is low...Yet I knew that plenty of recreational drugs were flowing through this white-middle class community...it was barely hidden. Meanwhile, in poorer communities like where I grew up, police are everywhere patrolling, waiting to harass anyone they don't like the look of... hell I was interrogated for sitting in my front yard once.

I've been blessed to run in many different social circles in life. Poor black kids, incredibly wealthy white kids, I've seen it all. My experience is completely in line with yours.

The wealthy white kids I was around were loaded with hard drugs. Coke use was rampant. I could open random drawers at my friends place and see bags of the stuff. Heck, some were outright dealing the stuff.

Meanwhile my poorer black friends pretty much only smoked weed, while occasionally dabbling in some other minor drugs (mushrooms, MDMA, etc...).

Of course the only ones with stores of police harassment were my poorer friends. My rich friends got away with anything they wanted

Statistical speaking, areas with higher rates of blacks have higher violent crime rates so it is possible cops are patrolling the black areas at a higher rate. This would lead to increased police which would lead to increased harassment just because there are more cops.

I don't think this works as a causal explanation. A decrease in petty fines shouldn't be significant enough at changing poverty to be reflected in crime rates in just a couple months. You'd also expect a delay as the fines that would otherwise be demanded in that time wouldn't be collected for some months, I assume.

As an alternative, maybe people who are mistreated by police become immediately more likely to commit crimes? Or maybe police who would otherwise be doing proactive policing are doing something more beneficial instead? Those are just speculations, though.

Many people cannot pay their fines, or it requires significant tradeoffs. And it is so easy for things to roll out of control.

Once the due date is past, the fine is doubled or triple. A $200 fine can become 400 or $600. Now its harder to pay, and so it doesn't because your car broke down and the kids are hungry. A warrant is issued for your arrest. They arrest you while at work. Even though you are out again the next morning, your work fired you, and won't give you a good reference. You get a knock on your door. Rent is due. Then two months. You feel fucked. A neighbor offers you relief, a fun night with some meth. You consider, but decide not too. You get kicked out for not paying rent. Now you are back living at your mom's. You feel like a loser, but at least its a roof. But she's an ass. She forgets to tell you of a phone call inviting your to an interview.

But you have to make it to court to plead guilty/mercy for not paying your fines, but your sister calls you right before. She just got beat up down the street, and needs you to pick her up. Your are 5 minutes late, and the judge doesn't care. On the way home, you get pulled over for expired tags. New ticket. I guess you won't pay the car payment this month. Your car get repossessed. It costs 200 +%40/day + $400 in late payments. You borrow some money from the neighbor. He's creeepy but get your car back.

Insurance is due. Tags are not paid. You are afraid to drive. You don't have a job. Meanwhile your creepy neighbor will forgive your debt if you get on your knees. You decide to not pay him, but he grabs you anyway. Rapes you. Police don't believe you. You are a poor slut. Your friend says come over and says let's party. You do. You get high. You keep getting high. You sleep with someone because they have a hit. He says, damn you hot, you could make some good money, and leaves a little for the road. You start jonesing, then take him up on the offer. Next thing you know you are a hooked junky prostitute. You sell your car for next to nothing and a blowjob. Your mom kicks you out. Now you live by the river with Frank. Frank protects you but he's an alcoholic.

I made this up, but this is how it works. THis is how it worked for people I know.

I'm not disputing that there is a link between fines and poverty or between poverty and crime. But I can't fathom how 5% of assaults and burglaries are a direct result of getting a fine in the last two months. Even in your perfect storm story, it's three or four months before you commit a crime. If you got fined before the slowdown, the fine still applies and still doubles. If you would have been fined during the slowdown but aren't, you would have had a month to pay anyway. You can also set up a payment plan to pay your fine over a period of a couple years.

I don't doubt that some small portion of the effect is exactly this. But 5% reduction in a month seems like far too much. I'm not even sure if there was a big reduction in fines- how many fines come out of proactive policing?

I don't know, it's not totally impossible. It just seems less likely than the null hypothesis of random variation in the data, while a more direct relationship could compete.

This is a good point that I missed in the article. I was thinking that the reduction in crime occurred over a relatively long period of time, say a couple of years. I see now it was little over a month. Frankly that is not enough time for what I laid out, or enough time with ideas of under-reporting because everyone knows the police won't do anything because of their slow down.

So I am not sure what it is.

Most of those same criticisms can be levied against the "broken windows" theory on which the studied policy was based, though.