If there is a constant fraction of any random population predisposed to commit violent crime, then something in their environment must trigger increases and decreases in the number of crimes they commit. Suppose that everyone feels most comfortable when their environment serves their personality; for example, social people favor environments with a lot of people around, and incurious people favor simple (as opposed to complex) environments. They will be most stressed when their environments vary from these “comfort zones,” and they will tend to take whatever action they can to reduce that stress: they will either find other environments or change the ones they are in.
People at extreme ends of the personality spectrum will be under more stress than others in environments that favor the average, and particularly stressed in environments that favor the other extremes. If, as psychological studies have suggested, other people are the source of the most stress in our environments, then if highly stressed people have no power to either leave or change their environments then they may resort to hurting other people: that is, committing violent crimes.
Matching arrest rates to crime rates (as fractions of a population over expected lifetimes), the data from 1975 to 2003 indicates that in a population of 1,000 people there will be 135 violent criminals and four murderers. In terms of stress, measured as variation from one’s comfort zone, my behavior model suggests that this statistically corresponds to about two standard deviations for all violent crime and three standard deviations for murder. Put another way, if the range of personalities is measured on a scale of one to five, then people more than one of these units away from their comfort zone will be stressed out enough to commit violent crime, while someone more than 1.5 units away from their comfort zone could commit murder.