[npunt]

My point was that if you baselined how good or bad you thought humans were by the crime rates in the 80s, you'd come away thinking humans were way more violent than they are. I think the same holds for 2020 - if you assess human nature by how people act when they're in the midst of a global pandemic, you'd come to the same incorrect conclusion.

Another way to put it is that much of our behavior is dictated by our environment, and if we create an environment where extreme chemical, biological, psychological, and economic stressors are kept at bay, then humans are just not prone to a great deal of violence.

[dionidium]

> I think the same holds for 2020

The homicide rate in St. Louis bottomed out in 2003 at a rate of 21.8 per 100k. By 2014 it had more than doubled to 49.9. By 2015 it had nearly tripled to 59.3 per 100k. (It’ll end this year in the 70s per 100k). The rates have been rising for a long time. It’s not just 2020.

And the trend is similar in other cities.

Source: https://en.m.wikipedia.org/wiki/Crime_in_St._Louis

[npunt]

Homicides nationwide are well off of historic highs [1].

The Vox link you previously posted shows the large spike (+~30%) being 2020 related.

The 21.8/100k you cite for St Louis in 2003 was also abnormally and suspiciously low, as 2002 and 2004 were in the 30s/100k. I wouldn't baseline anything off a single local outlier that's 50% off from other years. Perhaps a statistical / data collection anomaly? You can see in 2004 the rate was about 35/100k, while in 2013 it was 38/100k, a scant change.

You can see that data in context within a chart in this NYT article from 2015 [2] which discusses some possible factors in St Louis' persistently high crime. 2014 and onwards is likely skewed considerably by the Ferguson unrest, and it sounds like St Louis might still be dealing with that alongside other factors.

Outside the local context, none of this data at a macro scale looks like the trend we see in the 70s-80s that the lead-crime hypothesis seeks to describe though. Crime nationwide has indeed been growing slightly since 2015... to hypothesize about why we're seeing this, I'd guess it's largely due to loss of opportunity and rising income inequality leading to poverty and desperation, which our recent economic policies have only been worsening. But I'm absolutely no expert in the subject, just trying to figure things out like you are.

[1] https://www.macrotrends.net/countries/USA/united-states/murd...

[2] https://www.nytimes.com/2015/02/11/us/st-louis-puzzles-over-...

[dionidium]

It’s surprisingly hard to find good charts, but I stumbled across this thread today:

https://twitter.com/Crimealytics/status/1343955297073770496?...

These charts very clearly show an uptick in the nationwide homicide rate starting in 2014. It then levels off. And then there’s a big jump again this year.

This is entirely consistent with my claim that rates are way up nationwide and that they started going up before 2020.

[npunt]

Thanks for the followup, just seeing this now. I think perhaps we're not on the same page about the central hypothesis.

1. Is nationwide crime up 2019 -> 2020? We agree, yes it is, and quite a bit! This link, as well as previous ones show that. My hunch is covid and its economic fallout is likely the primary factor.

2. Is nationwide crime up in the last 10 years? We agree, they are somewhat.

3. Are local crime rates in St Louis up? Sure, but thats a bit besides the point I was originally making, which was related to nationwide crime and the lead-crime hypothesis, as local factors can add noise to analysis.

4. Do recent nationwide crime trends look anything like the massive bump in nationwide crime the lead-crime hypothesis seeks to explain, thus potentially invalidating the hypothesis? This is the question I was asking. I believe recent crime trends do not invalidate it.

As I stated before, "none of this data at a macro scale looks like the trend we see in the 70s-80s that the lead-crime hypothesis seeks to describe". Later [1] in the tweet thread you just posted, the author shows a chart of murder stats going back to 1960. It shows a massive bump in the 70s-80s (up to 10 murders per 100k) followed by a sharp drop through the 90s (down to fewer than 6 murders per 100k). This has not been replicated since, ergo, current crime trends are not sufficient to refute the hypothesis.

[1] https://twitter.com/Crimealytics/status/1343955297073770496

[dionidium]

> I think perhaps we're not on the same page about the central hypothesis.

Yes, I agree. I think your response here shows that we're actually closer in agreement than it first seemed. That's my fault, I think. I don't think the lead hypothesis is bunk. (My initial comment perhaps made it seem like I was saying otherwise.) It has a lot of explanatory power. I also don't think we're in exactly the same position we were in the 90s, for lots of reasons.

But I do think the lead hypothesis lulled a lot of people into repeating the claim about falling crimes rates for so long that many of them have forgotten to check to see if it's still true. And while there are lots of interesting debates to be had about why homicides have been higher recently, the fact is that they have been!