This paper presents empirical tests of Quetelet's biological explanation for the age crime curve, originally proposed in 1831. It was presented at the American Society of Criminology annual meeting in 2013 and represents a significant advance over the preliminary theoretical framework developed in Arnold (2012), Some Preliminary Thoughts on a Biological Explanation for the Age Crime Curve. Quetelet (1831/1984) proposed that the age distribution of crime was caused by developmental changes in strength, passion, and reason over the life course. This paper simplifies Quetelet's model to two biological trajectories — physical strength and brain capacity — and develops a mathematical framework for testing whether these trajectories can explain the age crime curve. The model proposes that criminal propensity should be directly related to strength and inversely related to brain capacity, and that the resulting Z-score trajectory should produce the age crime curve as a cumulative probability distribution of criminal propensity in the population. Using arrest data from the National Incident Based Reporting Service (NIBRS) for the United States from 1996 to 2006, a mathematical model was developed using linear Z-score trajectories for ages 0-18 and 46-98, with a curvilinear trajectory for ages 19-45. The model achieved an explained variance of 99.9% for both male and female age crime curves — a remarkably high level of fit for a two variable biological model. The projected age crime curve was superior to alternative curves generated using standard statistical software, explaining 92.4% to 98.3% of variation across the three age sections. These findings directly address the conclusion of Sweeten, Piquero, and Steinberg (2013), who argued that it was highly unlikely that one or two variables operating alone could create the age crime curve. The results presented here demonstrate that two biological variables — strength and brain capacity — operating through a cumulative probability distribution framework, are capable of reproducing the age crime curve with a very high degree of accuracy. Three conclusions are drawn from these analyses. First, the age crime curve appears to be a cumulative probability distribution. Second, the Z-score trajectory for the age crime curve follows a linear trajectory from ages 0-18, a curvilinear trajectory from ages 19-45, and a linear trajectory from ages 46-98. Third, the results are consistent with Quetelet's biological explanation for the age crime curve. While correlation does not prove causation, the data suggest that developmental changes in strength and brain capacity could explain the age crime curve. This paper forms a critical step in the intellectual development leading to Arnold (2016), The Criminological Puzzle, which provides a more complete theoretical and mathematical treatment of the age crime curve solution. It connects to the theoretical groundwork in Arnold (2008), Agency in the Life Course of Criminal Offenders, and the broader framework of The Physics of Living Systems.
Thomas K. Arnold (Sun,) studied this question.