Scaling Criminal Offending

This paper reviews a century of research on creating theoretically meaningful and empirically useful scales of criminal offending and illustrates their strengths and weaknesses. The history of scaling criminal offending is traced in a detailed literature review focusing on the issues of seriousness, unidimensionality, frequency, and additivity of offending. Modern practice in scaling criminal offending is measured using a survey of 130 articles published in five leading criminology journals over a two-year period that included a scale of individual offending as either an independent or dependent variable. Six scaling methods commonly used in contemporary criminological research are demonstrated and assessed using the National Longitudinal Survey of Youth 1979: dichotomous, frequency, weighted frequency, variety, summed category, and item response theory 'theta'. The discipline of criminology has seen numerous scaling techniques introduced and forgotten. While no clearly superior method dominates the field today, the most commonly used scaling techniques are dichotomous and frequency scales, both of which are fraught with methodological pitfalls including sensitivity to the least serious offenses. Variety scales are the preferred criminal offending scale because they are relatively easy to construct, possess high reliability and validity, and are not compromised by high frequency non-serious crime types.