There is a selection bias problem with the method of gathering news stories. Many defensive gun uses never make the news. Sometimes that is because the person us- ing a gun in self-defense saw no need to call the police—he or she scared off the bad guy. In some cases, the victim might not want to explain to the police that he has a gun, perhaps because he is a felon, or perhaps because he lives in a jurisdiction with very restrictive gun control laws. Sometimes the police do get called, but the officers do not find the circumstances sufficiently impor- tant to issue a press release. After all, “Man Scares away Burglar, No Shots Fired” is not particularly newsworthy, unless you live in a very small town.
In spite of the selection bias problem, there is one enormous advantage to this model of gathering data: it provides a rich set of information about motives, circum- stances, victims, and criminals. It also pro- vides a sufficiently large database (almost 5,000 incidents), randomly selected, so that some conclusions about the nature of armed self-defense in America can be drawn. Best of all, whatever the deficiencies of news reporting, the model is not completely de- pendent on the honesty or accuracy of the respondent—unlike some of the questions raised with respect to defensive gun use sur- veys.