Community violence as it affects child development: issues of definition.

The state of the art of definition of community violence as it relates to child development was examined in terms of the definitions used in 23 empirical studies. In all cases community violence was defined in terms of what were assumed to be measurements obtained as linear combinations of a priori numerical weighting of responses to questions--asked either of a child or of the parent of a child--about experiencing and/or witnessing and/or hearing about instances of violence. Thus, the definitions can be seen to represent the perspectives of 2 kinds of observers--the child or the child's parent--and 3 levels of closeness to violence--experiencing, witnessing, or hearing about violence. Combining these perspectives and levels, the following 8 different definitions could be seen to be used in the practice of 1 or more of the 23 empirical studies: Child Self-Report (perception) of either (1) experiencing, or (2) witnessing, or (3) experiencing and witnessing, and hearing about violence; or Parent Report (perception) of the Child (4) experiencing, or (5) witnessing, or (6) experiencing and witnessing and hearing about violence, or (7) = (1) + (4), or (8) = (3) + (6). In almost all the examples of research definitions it was assumed implicitly and without test of the assumption that different violent events were interchangeable, and usually it was assumed (again without test) that the magnitudes of different violence events were equal. Usually, an unstated theory of stress appeared to guide the measurement definition, but in one study definitions were developed and tested in terms of a clearly-stated theory of learning. It was concluded that definition of community violence is a measurement problem; that very likely it is multidimensional; that it could be more nearly solved if better attention were given to specifying it in terms of theory that can be put to test and by attending to basic assumptions and principles of measurement.