Of a sample of 1,555 registered nurses, 1,038 (67 percent) returned questionnaires. They responded to vignettes in which case characteristics were systematically manipulated. Three hypotheses were tested: (1) recognition and reporting would vary by characteristics of the act; (2) recognition and reporting would vary by personal and professional characteristics of nurses; and (3) recognition and reporting would vary by organizational characteristics. Analysis using ordinary least squares regression showed that case characteristics involving abuse type and level of seriousness, perpetrator status, and perpetrator psychology had the greatest effects on average recognition and reporting scores. Victim age and gender had consistently significant but small effects on both recognition and reporting, while perpetrator account had mixed effects on recognition and reporting. The addition of nurse and organizational characteristics accounted for a small amount of variance. The social construction of child abuse by nurses appeared to be more influenced by case descriptors than by the nurses' own characteristics or by their organizational work setting. Thus, more support was provided for the first hypothesis than for the second and third hypotheses. Appendixes contain supplemental information on child abuse types and perpetrators. 35 references, 3 notes, and 1 table