NCJ Number
90627
Journal
Fire and Arson Investigator Volume: 33 Issue: 2 Dated: (December 1982) Pages: 3-13
Date Published
1982
Length
11 pages
Annotation
From information obtained in the Seattle Fire Department computer files on arson incidents, arson loss and building occupancy statistics by census tract were developed for 1970-79, and a prediction model for irrational arson was developed and tested.
Abstract
The information provided is valuable to fire investigators, the insurance industry, planners, and community residents, because it can identify certain trends in arson, determine if there is a need for stricter building code enforcement, identify the geographic areas with particular arson problems, and determine the types of buildings attracting commercial arson. Grass and brush fires were found to be the leading arson occupancy category for 1970-79, followed by single-family dwellings, schools, and apartment buildings. Storage buildings and restaurants/taverns were the largest commercial arson building occupancy categories during the same period. To provide more information and trends on problem fires within the various building and non-building occupancy classifications, computer statistics were generated on the ratio of serious arson fires to all arson for all major categories (the 'serious loss' ratio). The 'serious loss' ratio of Seattle arson cases has moved steadily downward over the decade after peaking in 1975, and arson losses are more serious in particular commercial structures, as numerous commercial categories have higher 'serious loss' ratios than the overall average. 'Serious loss' ratios are higher for many of the residential structures and have increased relative to the Seattle average over the decade. In testing the feasibility of a tactical prediction model of irrational arson, it was found that the level of accuracy for predicting the location of an arsonist's next fire was not sufficient for practical applications, although it was better than random predictions. Tabular data are provided. For the first part of this study, see NCJ 87279.