• Corpus ID: 16065586

Why Quantitative Probability Assessments Are Empirically Justifiable in Foreign Policy Analysis

@inproceedings{Friedman2015WhyQP,
  title={Why Quantitative Probability Assessments Are Empirically Justifiable in Foreign Policy Analysis},
  author={Jeffrey A. Friedman and Joshua D. Baker and Philip E. Tetlock and Richard J. Zeckhauser},
  year={2015}
}

• J. Friedman, J. D. Baker, +1 author R. Zeckhauser
• Published 2015
• Psychology

Aristotle counseled us to seek precision insofar as the nature of the subject permits. But how much is too much? This article provides the first systematic test of long-standing debates about how precisely foreign policy analysts can estimate probabilities. Using a data set of 888,328 forecasts drawn from a series of geopolitical forecasting tournaments, we demonstrate that qualitative probability assessments, including seven-step scales employed by U.S. intelligence analysts, systematically sacrifice accuracy. Respondents’ capacities to extract these “returns to precision” correlate more with forecasting skill, effort, and training in probabilistic reasoning than with numeracy, education, or cognitive style. Our results indicate that foreign policy analysts can parse their judgments more precisely than conventional wisdom supposes and that this ability can be cultivated. We argue that it is possible to improve the value of intelligence reports and other political commentary by supplementing qualitative descriptions of uncertainty with quantitative estimates of subjective probability.