Scholion.ai

Problems in Extrapolation Illustrated with Space Shuttle O-Ring Data

Michael Lavine

Journal of the American Statistical Association · 1991 · Volume 86 · Issue 416

DOI: 10.1080/01621459.1991.10475132

Abstract

Many standard techniques for model selection, model fitting, and diagnostics can be inappropriate in extrapolation problems. We illustrate some pitfalls using a data set analyzed by Dalai, Fowlkes, and Hoadley—the data on space shuttle O-ring damage available at the time of the decision to launch the shuttle Challenger . The aim of the analysis is to estimate the probability of failure for a launch at 31 degrees, the temperature at which the Challenger was launched, and to use that estimate to decide whether the launch should have been postponed. We conclude, first, that the usual procedures of model selection, model fitting, and diagnostics do not tell the whole story about the probability of failure at 31 degrees and, second, that a decision analysis of the launch question leads to a robust conclusion that the launch should have been postponed.

Enable JavaScript for citation tools and additional paper interactions.