An Empirical Study of the Impact of Reliability Values on Threat Assessment

This article presents two empirical studies investigating the extent to which intelligence analysts are able to integrate new information about the reliability of intelligence sources into their judgment. The first study used a limited number of pieces of intelligence, and binomial reliability values. The second study included a larger number of sources, and reliability values distributed on a 5-point scale. A comparative analysis of the results suggests that analysts are less likely to significantly modify their judgment if they are faced with a large amount of information to analyse, possibly showing the vulnerability of analysts' to heuristic biases in this case.