Ed Leamer, of UCLA and author of Macroeconomic Patterns and Stories, talks with EconTalk host Russ Roberts about how we should use patterns in macroeconomic data and stories about those patterns to improve our understanding of the economy. Leamer argues that economics is not a science, but rather a way of thinking, and that economic models are neither true nor false, but either useful or not useful. He discusses various patterns in the recessions and recoveries in the United States since 1950. The conversation closes with a discussion of the reliability of econometric analysis.