Hacker News Reader

Comment by addicted | Hacker News Reader

I keep posting this because there’s always this comment.

There’s been a pretty standard way used for decades to determine whether crime is actually rising or increasing in the US that’s immune to reporting, legislative, etc changes.

Look at homicides. Dead bodies do not get hidden in the US.

And homicides are decreasing.

View on HN

Besides that throughout the 2000s and 2010s when crime hit historic lows in the US Americans continued to say crime was increasing.

We’re talking about murders going down by an order of magnitude in many cities and people were still saying crime was increasing.

1) It's great that homicides are decreasing. But while homicides are one way of measuring crime, there are others. I think many people are saying that the divergence is increasing between homicides and other crimes. Would you agree with the GP that "crime" can be increasing even if homicides are decreasing?

2) While it's rare, there are some well documented cases where "dead bodies" get hidden in the US by unscrupulous police trying to look better in statistics. Here's Chicago from a decade ago: https://www.chicagomag.com/chicago-magazine/chicago-crime-st...

I live in Seattle. We just set a new record for homicides. I think an all-time high.

Also, I don’t think homicides is necessarily that accurate a metric. It’s immune to many things. But it isn’t immune to advances in trauma treatment. I’ve always wondered if cell phones reduced homicides by reducing response times.

I’d love to know shooting victim statistics. But that’s not as complete and comprehensive as homicide counts.

It would be reassuring if the proportion of murder to other, more common and visible crime was constant. But it's not. You may still get decrease in murders while having 10 fold increase in theft.

Homicides are decreasing exactly the same way inflation is decreasing. Down from recent highs but still quite elevated.

What about deaths classified as suicide/accidental?

Yes, that's great and all, but that doesn't change the fact that all of the car windows are smashed, and you regularly see people, sometimes the same people, smashing the windows in person.

Quoth Jeff Bezos: "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."

No they are higher than theyve been in like 30 years (since mid 90s). And they have been trending up for like 10 years, not just pandemic thing.

Not according to this chart: https://www.statista.com/chart/31062/us-homicide-rate/

Pretty clearly the inflection is at 2020. It was flat before that, so sure, you can draw a "trend" upwards since the lowest point in 2014 or so, but that's just playing games with curve fitting. I can equally tell you, with the same kind of support, that clearly the trend is downward from 1991 to today. "Just look at the numbers!"

And more importantly, the rapid rise in 2020-21 is dropping rapidly now.

That chart shows exactly what I said, we have the highest rates since the mid 90s and there is a clear upward trend since 2014 (ten years ago now). Thats not "playing games"---its a lot more true than what the OP said.

Just to emphasize, indeed "look at the numbers". The gap between 2014 and 2022 is about 2/100,000--that is over 6000 people. This isnt some statistical anomaly---it is a real increase. Crime is going up.

Why 2014 and 2022 and not 1991 and 2022? Or 2020 and 2022? Other than one confirms your prior and the other two refute it?

And regardless, the fact that the first derivative is negative clearly makes "crime is going up" a lie. Crime is going DOWN right now, as measured. You need to be more specific: "Crime over the past 9 years is going up" would be true (but again, not if you replace 9 with 2 or 30!).

Here are the actual stats:

https://www.statista.com/chart/amp/31062/us-homicide-rate/

What you say is technically true but also highly misleading. You're cherry-picking points to put together a narrative that is not the obvious shape of the data.

Peak homicides over the past 10 years are nowhere near the peak in the 90's.

And they trended downward in 2022, and mid-year 2023 estimates were downwards another 9.4% relative to 2022:

https://counciloncj.org/mid-year-2023-crime-trends/

What I said is true, and what the op said is false and the stats you posted show this. What is actually misleading is saying something outright false like the OP. The homicide rate is higher than it has been since the mid 90s. People are worked up about crime increases because they have lived through the past 4 years and feel the tension.

What about overdoses which are at all time highs? Also, there are crimes beyond murder that affect quality of life. It seems like a fallacy to argue that because murders are down that all crime is down...

What do overdoses have to do with crime?

Reply on news.ycombinator.com