Hacker News Reader

Crime data is messy and can take decades of debate without ever coming to any conclusions. Beyond the abortion argument, I’ve heard convincing claims that the cause of a decline in crime was due to phasing out leaded gasoline.

For 2020, I blame Tiger King for glamorizing a life of crime. I’m prepared to put together some infographics about it and the theory will be as convincing as any other.

View on HN

Indeed. I am familiar with the leaded gas hypothesis. Inferring causality from observations without randomizing the intervention is hard to neigh impossible.

In Levitt's defense, he uses natural experiments, where natural variation does the work of random assignment of a subject to 'treatment group' and 'control group' and does the causal analysis. This is over and above accounting for the known hypotheses that purport to explain the phenomenon. Short of RCT it doesnt get much better than that.

Reply on news.ycombinator.com