What ‘Typical Insights’ Might ‘Core Statistical Analytical Techniques’ Provide to Intelligence Analysts?

This article critically examines Jack Duffield’s proposition that (all) intelligence analysts should become competent in “a core set of statistical analytical techniques” so as to address the “total information overload” they have experienced following the proliferation of ‘Big Data’. It summarises the generic (technique-agnostic) and particular (technique-specific): analytical choices and decisions that analysts need to be competent to make when using each of the 11 techniques proposed; together with the parametric and non-parametric assumptions on which each of these techniques rely; the sources of non-systematic and systematic error that analysts using these techniques need to address; and the diagnostic measures that providers and consumers of findings generated by these techniques should use to assess any flaws or contingencies associated therewith, and thereby temper any associated inferential certainty and importance. When compared to baseline statistical competencies of non-specialist intelligence analysts, these summaries demonstrate the substantial additional training that Duffield’s proposition would require. The article concludes that this may prove a big ask without an extended period of additional training, work-based experience and specialist supervision. In the absence thereof, underqualified intelligence analysts using such techniques would risk undermining intelligence analyses with multiple analytical and inferential mistakes.