"Hans Reichenbach in The Rise of Scientific Philosophy (1951, p. 89) emphasizes the necessity of a predictive quality for science:
A mere report of relations observed in the past cannot be called knowledge; if knowledge is to reveal objective relations of physical objects, it must include reliable predictions. A radical empiricism, therefore, denies the possibility of knowledge.
The concept of predictability and subsequent testability has prompted the noted scientific philosopher Karl Popper to further emphasize that if an explanation cannot be adequately tested, it is not scientific. The concept must be testable (i.e., falsifiable) to qualify. Any kind of explanation will not do; it must be amenable to a testing process. If it survives testing, it can qualify. In our magnet example, we might propose that objects of only a certain color (and not a magnet) attract iron. If a red magnet were found to work, we could further test the notion by using a wooden block of the same color as the magnet and thus disprove the color theory. Popper in his book The Logic of Scientific Discovery (1968, p. 40) is emphatic on the matter of falsification. He states:
But I shall certainly admit a system as empirical or scientific only if it is capable of being tested by experience. These considerations suggest that not the verifiability but the falsifiability of a system to be taken as a criterion of demarcation.
The idea that a genuine scientific idea must have the consistency that gives it predictive value, and the potential for falsification, has received a great deal of attention during the past few years among scientific philosophers and evolutionists. There is very little disagreement with this aspect of science as enunciated by Popper, and there is genuine concern as to how to apply this principle to the theory of evolution. The unrepeatable or untestable events postulated for evolution are not amenable to evaluation on the basis of consistency and prediction. Thus the concept of evolution as a principle of science is being questioned at a most fundamental level. Does it really qualify as a scientific principle? Some examples of deficiencies follow.
The concept of natural selection by survival of the fittest is the basic evolutionary mechanism. This concept does not qualify as a scientific principle, since fitness is equivalent to survival. Here we have a case of circular reasoning; no consistency or predictive value can be tested. According to this idea, organisms have survived through the evolutionary process because they are better fit, and the way one tells they are better fit is that they survive. A number of evolutionary scholars have labeled the principle of survival of the fittest a tautology (e.g., Waddington 1957, Eden 1967, Peters 1976). Popper (1963) attacks the unfalsifiable nature of the concept and concludes:
If, more especially, we accept that statistical definition of fitness which defines fitness by actual survival, then the survival of the fittest becomes tautological, and irrefutable.
The concept of survival of the fittest of itself does not necessarily imply any evolution. Would not the fittest survive, whether they evolved or were created? The noted evolutionist Mayr (1976, p.3) speaks of "an all-powerful natural selection." Platnick (1977) wonders if there is any difference in this kind of explanation as compared to that of an all-powerful Creator."
(Ariel A. Roth, 'DOES EVOLUTION QUALIFY AS A SCIENTIFIC PRINCIPLE?'
from: http://www.grisda.org/origins/04004.htm )