GoKawiilLet the record state that I love Scott Alexander. I’ve been a reader for over a decade and count him among the sharpest and most important public intellectuals working today. I credit essays like I Can Tolerate Anything Except the Outgroup with articulating feelings I had been having for years but didn’t yet have the words to express, thereby giving me new insights and the vocabulary to discuss them. I don’t read everything he writes anymore (who has the time?), but I remain a loyal subscriber and to this day find much of value in his prolific output.
But. (With an introduction like that, you know there must be a “but” coming). I increasingly find myself in disagreement with Scott’s essays on social issues and public policy, despite broadly sharing his small-L liberal outlook. I often see him deploy arguments I consider specious in a manner I find off-putting. And I blame his devotion to The Church of Graphs.
The Church of Graphs is dedicated to the meta-belief that knowledge must be formalized and quantifiable to be worthy of consideration. It demands that its adherents reject the evidence of their own eyes in favor of official facts and figures stamped with the imprimatur of a priestly expert class. Church members typically have curious blind spots in their understanding of the world that strike non-believers as ridiculous, but many very intelligent people are believers. Scott is a Bishop.
There’s no better example of what I mean than two of his recent posts, about crime and disorder.
First we have Record Low Crime Rates Are Real, Not Just Reporting Bias Or Improved Medical Care. The title says most of what you need to know, and the article argues that falling reported crime rates are real, not a result of people being so jaded about crime and the justice system that they don’t bother reporting it anymore. The reference to medical care is addressing a popular argument that record-low homicide rates would be much higher in the absence of advances in emergency room care that turn murders into mere assaults, and Scott argues convincingly that this isn’t true. But while homicide is the single most reliably recorded crime due to the presence of a body, reported crime is down across the board. Here’s the rate of property crime, for example.
We have reason to believe that this decline isn’t solely due to reporting bias because the National Crime Victimization Survey (NCVS), which polls randomly selected residents to ask if they have been victims of a crime, also reports a large decline. These rates are per household rather than per individual, and do not include crimes committed against businesses. But this second, independent data source corroborates the narrative of falling crime rates.
Next we have Crime As Proxy For Disorder, in which Scott starts off unusually back-footed due to the amount of pushback he received on the first post. Scott’s discussion threads are quite lively and his audience enormous, so the usual concept of a “ratio” doesn’t apply to him as it would to you or I. But still: the original article attracted over 700 comments on just 360 likes at the time of this writing. Scott’s audience was not buying what he was selling. In Scott’s words:
The problem: people hate crime and think it’s going up. But actually, crime barely affects most people and is historically low. So what’s going on? In our discussion yesterday, many commenters proposed that the discussion about “crime” was really about disorder.
All the graphs say that crime is declining, but people believe it’s getting worse. This includes many members of Scott’s way-above-average-intelligence readership. Why is that? Scott proposes that maybe people’s perceptions on this topic are based on disorder, and proposes several indicators for disorder: litter, graffiti, shoplifting, tent cities. He then spends the rest of the post demonstrating with impeccably sourced and credentialed figures why none of these disorder indicators are actually getting worse, either. They’re all getting better, or there isn’t enough data to say.
Littering seems to be down
... continue reading