Perhaps I can get some clarifications on the article.
For example, the article states the sample used was:
Why is this not a biased sample? Shouldn’t they sample 7,000 people, not 7000 people that claimed they’d been arrested?
The other thing I didn’t understand:
Given that more than a third of black, hispanics, and whites were arrested by 23, wouldn’t the arrests of asians and native americans have to be practically zero for there to only be a third of the population that had been arrested by 23?
Finally, assuming the above issues are just misunderstandings on my part, an arrest is not a conviction. For example, if the entire Delta Delta Delta fraternity is arrested the night of a particularly rowdy party for public drunkenness and disturbing the peace, and all charges are dropped the next day as they’re let out of the drunk tank, that’s not going to stop them from getting a job or living a successful life, but it will inflate the statistics. So the followup question would be - How many of these arrests resulted in a plea bargain or trial? (i.e. were not just summarily dropped)