Data-driven logic

Putting forward criteria of validity for data-driven consequence realations is the core theorical goal of ReDa.

Since what counts as a good scientific inference changes over times, differs from discipline to discipline, and sometimes from person to person, the hope to provide the unique logic adequate for all data-driven inference is probably misplaced. So our aim is to isolate conspicuous patterns of data-driven infererence which are being in the current scientific practice. “Null Hypothesis Significance Testing” (NHST), the most popular and the most controversial of all such patterns, is our natural starting point. But we also work on the lesser known “Strong Inference” introduced by Platt a 1964 Science editorial, and at Polya’s “Patterns of plausible inference“.