NCJ Number
83661
Journal
Evaluation Review Volume: 6 Issue: 3 Dated: (June 1982) Pages: 393-402
Date Published
1982
Length
10 pages
Annotation
A novel solution is proposed for the statistical problems involved in evaluating the significance of variations in the number of domestic child homicides, so that the search for causes in a community's child protection system is facilitated.
Abstract
To examine whether the child homicide (filicide) rate in the District of Columbia was random or related to relevant legislative reform events over a period of time, a monthly time series of homicides suspected of having been perpetrated by a parent or parental surrogate on a victim under age 18 was assembled from the medical examiner's log and reconciled with administrative data furnished by the police homicide division. This monthly time series of frequencies was then revised into a binary series of 'active' months to increase the reliability of the analysis. Since the clusters of 'activity' resemble those of a stationary Markovian sequence, it was possible to apply the algorithm ZERON to yield the distribution of the number of busy months in a period of consecutive months. The exact probabilities associated with various numbers of active months in which one or more filicides were reported in the course of 6-12 months were calculated and accumulated into annual and semiannual control charts for the period prior to the Prevention of Child Abuse and Neglect Act of 1977. The charts were used to evaluate the binary filicide series for 60 months prior to the 1977 legislative intervention as well as the series over the subsequent 48 months. The question of randomness was thus examined within the context of two different legislative frameworks. The findings determined that the legislative reform did have some impact on the rate of domestic child homicides. Graphic data and 12 references are provided.