The phenomena of interest to victimology are events in which persons, as individuals or groups, have been affected by acts of other persons or groups which are defined as wrongful by institutionalized criteria. To qualify, the effect of the offense on the victim must be harm, attempt or threat of harm, or being placed in special danger of harm. While it is possible to define purely objective indicators for the normative components of this definition, such as victim, offense, and harm, these concepts remain nonetheless normative; hence, there are extrascientific evaluative components intrinsic to any data for victimology. The source of these evaluative judgments, however, need not be the data recorder; the recorder's data may reflect evaluative judgments of others. Recorders, compilers, and users of records of victimology data are members of organizations who apply organizational procedures and norms to observations. Many of the properties and problems of data sources for victimology can be highlighted by considering the importance of the time dimension for the data and its users. The time dimension is expressed in the recorder's closeness to the events being recorded, whether experiments can be contrived to provide data, the presence of an event's durable physical traces, the existence of a continuous victimizing state, and the continuing consequences of victimization. With relatively few exceptions, criminology has until recently been largely dependent upon intervening agencies for data on victimization. More recently, the victimization survey, which determines the extent and character of victimization from surveys of citizen population samples, has been used to develop information on victims, offenders, and relationships between them of far greater scope and detail and in more direct usable form than official records. A bibliography contains 36 listings.