The Undoing Project

Michael Lewis tells the story of Daniel Kahneman and Amos Tversky's collaboration and their work on judgment. The author's website places The Undoing Project in that biographical context. Choose it when you want a narrative about the development of ideas, rather than a step-by-step decision manual.

Distinguish the story from the evidence

A compelling account can explain how a question arose, who argued about it, and why it mattered. To evaluate the resulting claim, you still need to know what was tested. While reading, keep two notes: one about the collaboration and one about a particular study or argument. Do not let admiration for the people substitute for checking the latter.

For an original Nuxflo exercise, imagine reading a story about a teacher who notices that confident students sometimes make predictable mistakes. The observation may inspire an experiment. It does not establish how often the mistake occurs, which students make it, or whether confidence causes it.

Turn an observation into a research question

Write a testable version: “On this set of problems, how does students' stated confidence compare with the proportion of answers they get right?” Define the problems, confidence scale, and scoring rule before collecting answers. Consider whether the students already know the material and whether guessing is encouraged.

If confidence and accuracy differ, several explanations remain possible. The questions may be unusually tricky, the confidence scale may be misunderstood, or the group may have limited experience with the topic. A memorable anecdote starts the investigation; the methods determine what can be concluded.

Read collaboration without copying the personalities

Ask what a partnership contributes: independent criticism, complementary knowledge, or a willingness to revise a question. You can try one of those practices without trying to reproduce a famous friendship. For example, have a partner write an alternative explanation before either of you sees the other's answer.

For Kahneman's own broader discussion of judgment, continue with Thinking, Fast and Slow. For evaluating an empirical claim encountered in either book, use the research-claim guide.