"In everyday language the word 'theory' is often used to mean an idea that looks good on paper, but has little to do with reality. In science, however, a good theory is one that explains reality very well. And excellent example is the general theory of relativity, which does a superb job of describing a host of gravitational effects. It predicted, for example, that light rays can be bent by gravity; careful observation shows that this is just what happens and that the amount of bending is just what the theory predicts. Indeed, the general theory of relativity has so far agreed in detail with every experimental and observational test it has undergone."
And earlier, the previous page:
"But the approach used by scientists to explore physical reality differs from other forms of intellectual endeavor in that it is based fundamentally on observation, logic, and skepticism. This approach, called the scientific method, requires that our ideas about the world around us be consistent with what we actually observe...A scientist trying to understand some phenomenon proposes a hypothesis, which is a collection of ideas that seem to explain the phenomenon...their hypotheses must always be developed in agreement with existing observations and experiments, because a discrepancy with what is observed implies that the hypothesis is wrong (The exception is if the scientist thinks that the existing results are wrong and can give compelling evidence to show that they are wrong). Only after a hypothesis has accurately forecast the results of new experiments or observations does the scientist feel confident that the hypothesis is on firm ground..."
"A body of related hypotheses can be pieced together into a self-consistent description of nature called a theory....Without models and theories there is no understanding and no science, only a collection of facts."