What is it about?
Even if you know a lot about a person’s life, it is hard to anticipate what might happen to them in the future. You might observe a child from birth through elementary school but still have great difficulty predicting their grades in high school. Through qualitative interviews embedded within a quantitative prediction task, we develop a framework for understanding why life outcomes are hard to predict.
Featured Image
Why is it important?
Predicting life outcomes is exactly what many scientists and policymakers hope to accomplish with new machine learning tools. It would be helpful, for example, to know in advance which student is at the greatest risk of poor academic performance! We might be able to intervene. But in many settings, these types of predictions may be inaccurate. Our framework may help scientists and policymakers to reason about the level of predictability that can be expected in different tasks. We also hope to spur new research on the unpredictability of life outcomes, which we believe has inherent scientific interest.
Perspectives
Much of science focuses on what is predictable. But unpredictability may be the dominant pattern in some settings when predictions are made over a person’s life course. Our paper is a first step to better understand the unpredictability of human lives.
Kathryn Edin
Princeton University
Read the Original
This page is a summary of: The origins of unpredictability in life outcome prediction tasks, Proceedings of the National Academy of Sciences, June 2024, Proceedings of the National Academy of Sciences,
DOI: 10.1073/pnas.2322973121.
You can read the full text:
Contributors
The following have contributed to this page
