2016 Volume 53 Issue 2 Pages 29-41
Motives for single and serial cases of arson in the recent 10 years in Japan were classified by using text-mining. In Study 1, the author analyzed 67 nouns in the confessions of 253 single arsonists, by using classical multi-dimensional scaling (cMDS) with weighted Euclidean distances. The coordinate data obtained from cMDS were analyzed by hierarchical clustering (Ward method, Euclidean distances). The results indicated that the 67 nouns could be classified into 7 clusters. These clusters were interpreted as single cases of arson motivated by (1) revenge, (2) suicide, (3) frustration, (4) secondary crimes, including concealment of theft or dissatisfaction of a failed crime, (5) insurance fraud, (6) vandalism, and (7) gloomy pessimism. In Study 2, 44 nouns were analyzed in the confessions of 127 serial arsonists by using the identical method in Study 1. The results showed that the motives of serial arsonists could be divided into 5 clusters: (1) frustration, related to gambling losses, among others, (2) frustration, related to working or failure in job hunting, among others, (3) fire tumult, (4) unfair resentment against the police, and (5) secondary crimes. Furthermore, this study suggested that the core motive for single arson was revenge and that for serial arson was frustration.
