One question counts above all else in natural science:
How can this idea--theory, hypothesis, claim--be tested?
Testability gains its primacy in science from our deep understanding that empirical knowledge does not rest on authority, tradition, or majority rule, but rather on the testimony of nature itself. We can picture the structure of scientific knowledge and inquiry as a three-way relation, a logical and evidential triangle, where the three vertices are (1) the scientific community, (2) scientific theories, and (3) the physical world, or nature itself.
Each of these is a necessary element of the scientific enterprise. Take away scientists, and no science will be done. Take away theories, and science loses its structure. Take away nature itself, however, and science loses not only its very subject matter, but also the independent check of evidence, against which theories are tested. Thus, if we rephrase the central question of testability as
If this scientific idea were false, how would we know it?
--it is evidence, drawn from nature, which ultimately provides the answer. The path between scientific theories and nature allows for traffic in two directions, namely from theory to evidence and back again.
In what follows, then, we shall be guided by the criterion of testability. If one trusts the judgment of the scientific community on any particular matter, one should do so only because the claims that community makes are testable against the evidence nature itself provides. "Accept this theory because all scientists do" or "everyone in a position to have an authoritative opinion agrees" or "organizations such as the National Academy of Sciences tell us" can, therefore, only be proxies for
We tested this theory and the evidence supports it.
Just as, in our shared civic life, no person is above the law, in science no theory escapes the requirement of testability. Ideas that cannot be tested are not necessarily false, but they cannot claim the standing of scientific knowledge.
All of this may strike the reader as self-evident. Of course scientific ideas must be testable. We begin here, however, because many parts of the National Center for Science Education (NCSE) critique of Explore Evolution show either an indifference to testability (when the predictions of neo-Darwinian theory fail) or, worse, a complete lack of awareness that testability is required. When evolutionary theory's perceived competitor is described as "creationism," and creationism is intrinsically unscientific, and, in public science education, unconstitutional, ordinary requirements of testability may come to be neglected.
Yet those requirements still exist.
NCSE. October 17, 2008. Critique: Explore Evolution, available online at http://ncseweb.org/creationism/analysis/explore-evolution.