QnAs with Adrian E. Raftery

Abstract

Adrian Raftery has a penchant for probability. Over a three-decade career spanning several disciplines, Raftery, a professor of statistics and sociology at the University of Washington, Seattle, and a recently elected member of the National Academy of Sciences, has used statistical tools to unravel uncertainty in all manner of projections. Named the world’s most-cited researcher in mathematics between 1995 and 2005 by Thomson ISI, Raftery has solved a raft of real-world problems related to weather forecasting, population estimation, and personalized medicine. Modeling studies help researchers make reasonable predictions of future scenarios, but Raftery says knowledge of the level of uncertainty surrounding the underlying numbers lends such models the mantle of meaning. Raftery discusses his diverse projects with PNAS. Adrian E. Raftery. > PNAS:In your 2012 PNAS Inaugural Article you reported a method for projecting countries’ future population sizes. Can you explain how this method differs from existing methods? > Raftery:This is a collaborative effort with the United Nations (UN) Population Division, which routinely makes population projections—numbers of people, fertility, and mortality rates—for all countries. The projections are widely used for research, policymaking, and monitoring international development goals, among others. For such projections, the UN had long been using demographic models, which take a deterministic approach, rely on assumptions, and result in absolute values without any rigorous assessment of uncertainty. We …