are notorious for offering risky predictions (and are thus extraordinary examples of scientific theory). In terms of relativity, one would expect that subatomic particles would increase in mass as they approach the speed of light. This is a risky prediction, because it is clear how to test it and it represents an unexpected observation per (then) conventional theory. We have observed in accelerators that indeed particle masses do increase. This corroborates the theory. So in this case we can imagine a test wherein a given observational result would have refuted the theory.

Similarly, we have the eclipse experiment of 1919. If Einstein's theory of gravitation were bogus, star positions as recorded during the eclipse would be spatially equivalent to their positions recorded prior to the eclipse. But according to Einstein's theory, the mass of our sun would be sufficient to "bend" the light of distant stars as it passed the sun. This theory by its nature gives rise to such a risky prediction. It is therefore attributed as a scientific theory. The severity of the test one can devise for a theory defines its scientific nature.

Science education needs to do a better job explaining some things. As another poster points out, there is no such thing as a fact in science, that is, nothing you can be sure of. You can think of anything in science as a proposition and a probability that it's true or false.

For example, is George Bush the president of the US? (Notwithstanding stolen elections, etc). The probability of this being true is very close to 100%, and the probability of it being false is close to 0%. But, it is never zero. There is a non-zero probability that any arbitrary crazy-ass thing you can think of is true.

The problem, though, is that the above is no way to run a thinking mind in the real world from day to day. You have to make real binary decisions about whether to do something or not do something, and you can't be constantly confused by doubting reality, so most of these probabilities from an epistemological standpoint can just be rounded off to a 0 or a 1 so you can get some actual work done.

Evolution is both a fact and a theory. From a fact perspective, evolution as an explanation for life has a probability so close to 100% that the remainder can just be discarded. From a theory perspective, evolution has tremendous predictive power.

In any case, I think the epistemological underpinnings of science are something that most people don't understand, and the idea that nothing is certain in science can be frightening and confusing (and misleading if used for devious purposes as the IDers do).