Implications of Biases in Sampling Techniques for Child Abuse Research and Policy (From Family Violence: Research and Public Policy Issues, P 26-41, 1990, Douglas J Besharov, ed. -- See NCJ-125593)

• NCJ Number: 125596
• Author(s): C S Widom
• Date Published: 1990
• Length: 16 pages
• Sale Source: University Press of America
• Publication Type: Report (Technical)
• Language: English
• Country: United States of America

Annotation

This analysis of the effects of sampling decisions and subsequent biases in research on child abuse emphasizes methods for minimizing biases in future research.

Abstract

The definition of child abuse that is used and the population studied can produce criterion-dependent bias. Factors involved in this type of bias include the breadth of the definition of abuse, the types of abuse covered, the types of abusers covered, the numbers of abuse episodes, the existence of neurological impairment, and the validation of the abuse by external sources. The second type of bias is method-dependent bias, which results from the way in which child abuse incidents enter records and studies. The four basic information sources are official agency records, case files from supposedly nonabused samples, treatment or parent training groups, and self-reported responses. Ways to reduce both these types of bias include the use of explicit criteria, validated information, appropriate control groups, longitudinal studies, studies of high-risk children, and multiple measures.