Theory

From EvaluationWiki

A set of assertions proposed to explain an event or phenomenon or wide range of events/phenomena. These begin as conjectures-guesses-become scientifically respectable hypotheses as they gain evidential support, and acquire status as a scientific theory in the corpus of knowledge when they have proven themselves as worthy by either or both of the tests of continued confirmation and explanatory simplification. It is commonly said that scientifically respectable theories can never be proved true by observations, but can be proved false by observing a single counter-example (this is called the falsifiability criterion (Karl Popper's favorite). However, this is doubly mythological: the atomic theory, heliocentrism, and scores of others have been well-established by now; and counter-examples never disproved a theory. It takes a large and highly configured array of them to even threaten a theory, and usually a better theory to displace one already in the saddle. This criterion of merit for theories arose from accepting the core neo-positivist analysis of theories as comprising true general laws-which they took to mean precise and exceptionless generalizations-but since there are virtually none of these (all the laws of physics except perhaps the third law of thermodynamics are extremely crude approximations, almost never literally true and usually a long way from true), it's a bad model and its implications fall with it. Theories in the social sciences are not even that good, and theories about evaluation, in particular, are best seen-like the laws in physics-as more or less useful approximations or idealizations that only have even that status in some respectable proportion of cases. See nomothetic, explanation, understanding.