ABSTRACT
When asked to produce a chapter for David Canter’s Feschrift, I had to think long and hard about what part of his eclectic and wide ranging contribution to psychology I would focus on. In the end, I chose to concentrate on his forensic work and in particular, his pioneering role in developing replicable structures of offending behaviour through application of multidimensional scaling. In this chapter I attempt to address the Achilles heel of such work, the choice of an appropriate index of object association that minimises artefact distortion in the coordinate space. In doing so I have developed a new coefficient of dichotomous association that, I argue, utilises the best features of the Jaccard index as well as minimising distortion due to skew in the data. This makes the coefficient particularly useful when exploring the radex structure of behavioural or symptom data. I name the coefficient the Jaccard-Canter Index (JC), and I hope David approves.
