No one has directly observed either electrons or fairies. Both of them are theoretical constructs, useful to explain observations that might be difficult to explain otherwise. The "theory of fairies" can actually explain more things than the "theory of electrons". So why do we believe in electrons, but not in fairies?
Is the issue a political one, where the "electron" fans got the upper hand in the nineteenth century, so by the twentieth century the "fairy" fans were a scorned and persecuted minority? Or, have we proved for sure that fairies don't exist?
No, to both. The real difference is that for electrons, we have accumulated a set of quite narrow and specific rules about how electrons will behave under various circumstances. Those rules let us make very specific predictions about electron behavior, and about the observations that will result. If those predictions don't come true, we know that either we didn't set up the circumstances correctly, or there is something wrong with the rules. But over many decades, we have repeatedly fixed problems with the rules, so we can now make really good predictions about electrons, especially in certain highly contrived circumstances (i.e., circuits).
Fairies are much more free. A fairy does what it decides to do. We haven't been able to find any useful rules for predicting how a fairy will behave under particular circumstances, or even for telling when a fairy has been involved in a particular observation. (At least I don't know of any such rules. I stand ready for correction on this.) Over many, many decades, it has not been possible for people to try out pretty-good sets of fairy-prediction rules, find out where they make mistakes, and replace them with better sets of rules.
It's always possible that there really are fairies. But the theory of electrons has been far more successful because it makes testable predictions. Because it doesn't make testable predictions, the theory of fairies hasn't enjoyed the same process of incremental improvement. So we have lightbulbs and microprocessors and the Internet, all based on electrons, and no fairy processors.
The scientific method is an amazing procedure for incrementally improving certain kinds of theories: those that make testable predictions. A theory that doesn't make testable predictions could still be true, but it doesn't participate in the scientific method. (There are people who believe that the only truths are scientific truths, but this is essentially a religious faith on their part.)
The theory of evolution is a scientific theory, because it implies a large number of specific testable claims. The specific rules leading to testable predictions have been tested, modified, and refined over many decades (roughly as many as the theory of electrons). Simple versions of the rules for evolution have been tested and refuted long ago, and replaced by better, more specific ones, just as they have for electrons. We are about as confident in the theory of evolution as we are in the theory of electrons.
The theory of intelligent design could be true. The biological world is a marvelous place, with truly amazing complexity. The theory of evolution assumes certain random processes for generating mutations. If an Intelligent Designer could influence those random processes, then perhaps both theories could be true simultaneously. But the theory of intelligent design does not make testable predictions, just like the theory of fairies. The Designer does what He does because He decides to, not because He is governed by rules. (See Matthew 4:5-7.)
The scientific method is an enormous intellectual asset to the human race. All citizens should understand what it can and cannot do, and all children should be taught to appreciate and apply it. It is important for them to understand why the theory of electrons is a scientific theory, while the theory of fairies is not. Likewise, of course, for evolution and intelligent design.
In the debate between evolution and intelligent design, I believe that we scientists are missing an important opportunity to educate people about the difference between "truth" and "scientific truth". There is a perfectly reasonable role in society for faith in truths that are not scientifically testable. But we and our children need to understand and respect the difference.
While one driver in this controversy comes from people with a fundamentalist religious agenda, there are scientists on the other side who pursue an essentially religious belief that "There is no Designer." Occam's Razor is a useful piece of practical advice about preferring simpler theories, but it has no more empirical content than the Apostle's Creed.
An advocate for Intelligent Design provided the following quotes from leading evolutionary biologists:
These are not scientific conclusions. These are statements of sincere personal belief by these authors, who doubtless feel strongly that their beliefs are consistent with their experiences as scientists. But they are essentially statements of faith, and they are out of place in a textbook. (An interesting essay by Michael Ruse, Is evolution a secular religion?, sheds historical light on evolution and this kind of "religious atheism".)
On the other hand, scientists in general are very careful to avoid making religious (including atheist) statements in a scientific context. Simpson's and Wilson's books are opinion pieces, not textbooks, so expressions of personal faith are appropriate. Futuyma's book is a leading textbook, but the above quote seems not to appear in the current (third) edition.
There is a widespread claim that evolutionary biology textbooks argue against Christianity, or against religion in general. You can read (here) a careful examination of a number of accused textbooks, refuting that claim.
In our pluralistic society, some people believe in God or in the absence of any God (both faith-based positions), while some people don't know and some don't care. In our pluralistic society, it is important to teach science without imposing on others one's own religious beliefs, including the belief in the absence of God.
Science education is about teaching the scientific method, and about teaching some of the knowledge that the human race has acquired by applying the scientific method. It would be a useful part of a science class to teach the distinction between theories that are scientific because they make testable predictions, and other theories that could be true, but are not scientific because they don't. Comparing fairies with electrons, or comparing evolution with intelligent design, should be an opportunity to teach better science.