Distance decay reexamined

The "journey to crime," or the study of the distance between an offender's residence and offense site, has been a subject of study within criminology for many years. Implications arising from such research touches the majority of criminological theories. An overriding conclusion from this line of research is that most crimes occur in relatively close proximity to the home of the offender. Termed the distance-decay function, a plot of the number of crimes that an offender commits decreases with increasing distance from the offender's residence. In a recent paper, Van Koppen and De Keijser raise the concern that inferring individual distance decay, from aggregate-level data may be inappropriate. They assert that previous research reporting aggregated distance-decay functions conceals individual variations in the ranges of operation, which leads them to conclude that the distance-decay function is an artifact. We do not question the claim that researchers should not make inferences about individual behavior with data collected at the aggregate level. However, Van Koppen and De Keijser's analysis raises four important issues concerning (I) the interpretation of the ecological fallacy, (2) the assumption of linearity in offender movements, (3) the interpretation of geographic work on profiling and (4) the assumption of random target selection within a delimited range of operation. Using both simulated and nonsimulated data, we present evidence that reaches vastly different conclusions from those reached by Van Koppen and De Keijser. The theoretical implications of our analyses and possibilities for future research are addressed.