Assuming that psychosis and violence are associated at greater than chance levels, delusions, therefore, are also likely to be widely prevalent among people who are violent and psychotic on a chance basis alone. There is evidence, however, that among people with psychosis, the violent may be more likely than a nonviolent group to have experienced delusions at some crisis in their history, such as hospital admission. The excessive association between violence and delusions has also been shown through the indirect evidence that the paranoid subtype of schizophrenia, paranoid states, or delusional disorders, all characterized by the prominence of delusional symptomatology, are generally cited as being more commonly associated with violence than other psychotic diagnoses. An instrument for the reliable and valid assessment of delusional experiences and their possible consequences was created and tested at the Institute of Psychiatry of the University of London. The spontaneous account of beliefs is an important starting point for both the original, extended Maudsley Assessment of Delusions Schedule (MADS) and the brief form for clinical use. Following a description of the evaluation of the MADS, the authors conclude that, given the typically substantial delays between the onset of a psychotic illness and serious violence, the MADS should be used to monitor critical developmental patterns of the characteristics of symptoms that may trigger violence, but as models for testing, not yet as markers for clinical practice. 2 tables and 49 references