ABSTRACT
Any proposed law of science or hypothesis must involve testable predictions. The grouping of similar things is often done implicitly and subjectively, which inevitably creates intangible controversies in scientific inference. The notion behind prediction is causality, whereas causality is based on the similarity principle: similar conditions will likely result in the same or similar outcomes. In Bayesian statistics, data on drug effects on animals can be used as prior knowledge in deriving the posterior distribution of the drug’s effect on humans. Drug discovery and development constitute a stepwise process: a new chemical compound is tested on animals first. Subjectivity in similarity grouping reflects an individual’s belief, experience, and knowledge. In a worldwide drug development program, for example, a drug can appear to be effective globally while its effects in different countries or regions may be very different. A dynamic treatment regime is a set of rules for choosing effective treatments for individual patients.
