1. There’s a stat Michelle Alexander cites in her recent New Yorker interview that isn’t correct, and it is important to see why.
She claims that 5% of all arrests are for violence, and drugs is the biggest engine of arrests.
Here’s why that’s not right, and why it matters:
She claims that 5% of all arrests are for violence, and drugs is the biggest engine of arrests.
Here’s why that’s not right, and why it matters:
2. She’s right that arrests for SERIOUS violence are about 5% of all arrests, at least in 2018: 520K out of 10M.
Drugs, meanwhile, come to 1.6M out of 10M.
BUT.
Look right below “violence.” Other assaults? 1.06M.
So violence = 1.58M, drugs = 1.65M.
Drugs, meanwhile, come to 1.6M out of 10M.
BUT.
Look right below “violence.” Other assaults? 1.06M.
So violence = 1.58M, drugs = 1.65M.
3. Moreover, if you view gun cases as essentially violent crimes, then violence comes to 1.75M.
If you think of DUIs as a form of potential but unrealized violence (after all, if the driver injured or killed someone, it would be “violent”), it comes to nearly 3M.
If you think of DUIs as a form of potential but unrealized violence (after all, if the driver injured or killed someone, it would be “violent”), it comes to nearly 3M.
4. But return to the basic definition (serious violence + other assaults). Given the reporting uncertainty in the UCR—uncertainty the FBI provides no measure of—it’s clear that drugs and violence are roughly *equal* entrants into the criminal justice system.
5. Alexander’s claim that drugs led the way in 2010 and do so today is even more problematic.
In 2010, there were 1.8M arrests for violence, 1.6M for drugs.
In 2002, there were 1.9M for violence and 1.5M for drugs.
In 2010, there were 1.8M arrests for violence, 1.6M for drugs.
In 2002, there were 1.9M for violence and 1.5M for drugs.
6. Now, to be fair, Alexander’s claim shifts between arrests and convictions, and outcomes differ here.
In 2006, the last year with good data, ~20% of all convictions were for violence, ~33% were for drugs.
BUT.
That’s a MUCH tougher stat to interpret.
In 2006, the last year with good data, ~20% of all convictions were for violence, ~33% were for drugs.
BUT.
That’s a MUCH tougher stat to interpret.
7. Some fraction of those drug convictions aren’t really abt drugs—they’re plea deals around violence.
No BJS dataset provides good info on how charges shift from arrest to arraignment to plea. But some “drug” cases are really deals involving dropping violence charges.
No BJS dataset provides good info on how charges shift from arrest to arraignment to plea. But some “drug” cases are really deals involving dropping violence charges.
8. So again, with convictions as with arrests, we find ourselves in the same place:
Drugs are not some unique entrant into the criminal justice system. They are roughly on par with violence, often even SMALLER THAN violence.
Drugs are not some unique entrant into the criminal justice system. They are roughly on par with violence, often even SMALLER THAN violence.
9. Now, I know the obv pushback, and it’s (fairly) valid: even if violence makes up 15% of arrests, that means 85% of arrests are for non-violence. Isn’t that still bad?
It’s a great question to ask.
But the violence-vs-non-violence q is a MUCH diff issue than “center drugs.”
It’s a great question to ask.
But the violence-vs-non-violence q is a MUCH diff issue than “center drugs.”
10. If we want to talk about violence vs. ALL non-violence, there is a lot to unpack there. It’s a tricky issue.
But anything we say about DRUGS specifically we could also frame using VIOLENCE specifically, since the impacts appear roughly the same.
But anything we say about DRUGS specifically we could also frame using VIOLENCE specifically, since the impacts appear roughly the same.
11. Yes, drugs are a non-trivial driver of criminal histories and records. But they aren’t some unique outlier. They matter as much as violence, and as much as property crimes.
A drug-centered discussion is still going to mislead, at EVERY level of the process.
11/11
A drug-centered discussion is still going to mislead, at EVERY level of the process.
11/11
12/11. Important point about arrests vs. ppl arrested. The UCR measures arrests. If Joe get arrested 5 times in a year, that's 5 arrests, but 1 guy.
Prob a LOT more multi-arrests under drugs than violence. So int terms of PPL, viol matters more still:
Prob a LOT more multi-arrests under drugs than violence. So int terms of PPL, viol matters more still:
13/11. Ok, one more pt on why all this matters. By centering drugs, we mislead ppl into thinking we solve this by focusing just on the (easier) case of drugs.
These polling results are linked. Bc ppl overstate the role of drugs, even libs refuse to rethink how we address viol.
These polling results are linked. Bc ppl overstate the role of drugs, even libs refuse to rethink how we address viol.
