Assessing risk: limits to the measurement of the target behaviours

A serious problem in the assessment of the risk of violence or other dangerous or antisocial behaviour is that the target acts do not lend themselves to measurement as readily as supposed. This problem imposes limits to the validity of research in this field. The limits become particularly serious when the sole data on the behaviour are from official, national statistics on crime. Risk factors derived solely from such an association must be regarded as suspect. Criminal data alone thus pose a double jeopardy on clinical practice-calling into question on the one hand some of the so-called information based principles behind assessments and on the other offering relatively little to the direct assessment process in the individual case. Some of the reasons behind this call for caution are discussed.