The process of radicalisation has received wide attention over the past decade. As the number of violent extremist offenders grows, the potential diffusion of radical ideologies inside prisons is gaining attention. Offender attribute data, both pre-custody and in-custody, routinely collected by Correctional Service Canada, were explored to determine whether violent extremist and mainstream offenders differed (that is, could be clustered); if so, what attributes have values that were systematically different for the two groups, and did those attributes lend themselves to predicting other offenders at risk for radicalisation. Results from the pre-custody attributes show minute differences between the two groups. The in-custody attributes show visible, although still weak, differences. Combining the two data sets provides further evidence for differences, with some interactions between the two sets of attributes. Definitive answers about radicalisation were hampered by the small number of radicalised offenders (less than 1 percent) and several major differences in the offender population as a whole that obscure smaller distinctions. Nevertheless, the analysis suggests some attributes that may differentiate violent extremist and mainstream offenders. Although unanticipated, it also demonstrates that the entire offender population separates well into three clusters, and allows the qualitative pattern of attribute values that differentiates them to be determined.