Confirmation bias is one of the great obstacles to making the practical case for liberty. People see in events what they want to see. I think of the joke about the person who planned to take his hyperactive dog on a train trip and asked his veterinarian for a tranquilizer. The vet mistakenly handed the person a stimulant. On the train he gave the dog the pill, prompting it to run madly up and down the aisle the entire trip. The embarrassed owner said to his seatmate, "Gosh, think what would have happened if I hadn't given the dog the tranquilizer!"

We are all subject to this bias. The best we can do is be aware of it and fight to overcome it.

When people look at statistics, confirmation bias always looms. It's just too easy to explain any statistical results in terms of what social scientists call one's "priors." But now and again we encounter an example of someone who refuses to let confirmation bias stand in the way of learning the truth. The latest example comes from Leah Libresco, a statistician who used to write at the political data-analysis site FiveThirtyEight. Libresco's example is particularly noteworthy because the issue she was considering was gun violence. If any issue is prone to confirmation bias, this is the one. It is also noteworthy that she published her article at the Washington Post.

In "I Used to Think Gun Control Was the Answer. My Research Told Me Otherwise," Libresco begins, "Before I started researching gun deaths, gun-control policy used to frustrate me. I wished the National Rifle Association would stop blocking common-sense gun-control reforms such as banning assault weapons, restricting silencers, shrinking magazine sizes and all the other measures that could make guns less deadly."