| Summary/Abstract |
In a true experiment conforming to the criteria of a randomized controlled trial (RCT), we found that intelligence analysts assign significantly more credibility to secret intelligence than to identical open-source intelligence. However, this was true only when the intelligence estimate constituted a ‘complex’ problem characterized by a high degree of uncertainty and not when the estimate constituted a ‘simple’ problem characterized by a low degree of uncertainty. Moreover, we found that intelligence analysts are significantly more confident in their own assessments when they process secret intelligence and more uncertain when they process identical open-source intelligence.
|