Ethics and Statistics: How is Ethics Like Logistic Regression? Ethics decisions, like statistical inferences, are informative only if they're not too easy or too hard

摘要

Consider the analogy between ethical dilemmas and statistical classification problems. The mapping is not perfect- when we categorize a decision option as ethical or not (in its context), there is ultimately no "true answer," thus we can't really speak of correct and incorrect decisions, or of error rates. But there is a sense in which many ethical questions become clear in hindsight. And as with prediction, some ethical decisions are clear-cut (for most of us) while others are tough calls. Perhaps some insights from the statistics of classification can help us better understand the study of ethics.