What is a scientific model?
Here’s the central point: Science is a model of reality, not reality itself. It tries to approximate reality as closely as possible. For example, in the kinetic theory of gases, the molecules are considered as rigid spheres. However, it is not correct to say that the molecules are hard spheres. They are modeled as hard spheres and the theory works wonderfully well. Of course, it doesn’t exactly match experimental results, and we should not expect that either.
Another important point: Experiments are supreme.
If it disagrees with experiment, it’s wrong.
– Richard Feynman, late theoretical physicist at Caltech
How true. Nothing else matters. If theory predicts a rise in temperature and the experiment shows a decrease, the theory is wrong. Feynman also makes the brilliant point that we can never know that we are right. We can only know that we are wrong. Say we have a theory. We test it in a case in which it applies. If the result agrees with prediction, can we say that the theory correct? NO! Just that it isn’t proved wrong. There could be an experiment conducted in the future which may produce a result that will contradict the theory. Till then, as long as experiments keep on verifying the theory, it will be proved less and less wrong. But it cannot be proved absolutely correct.
From hypothesis to theory:
Make a guess
A theory starts off as a hypothesis a good guess. The guess needs to be checked for simple cases first and then for more intricate ones. The hypothesis needs to produce numbers which can be checked against actual experiments that can be conducted, i.e. it needs to be falsifiable.
Falsifiability: What can I do to prove it wrong?
Falsifiability is a vital criterion of a hypothesis to be even taken seriously. (The first thing to ask: How can I prove this wrong? If there is no answer, forget about the hypothesis! It’s Not Even Wrong‘ is the worst insult!)
Can it explain the known? Finding faults is not enough.
Next, we need to see how general the hypothesis is. Can it explain all of the results that are explained by the existing theory? We have to find where they clash, and find out which one prevails at that point. Does the hypothesis say something more than what the old theory says? Does it cover for the limitations of the old theory? The important point here is that it is not merely enough to show that the old theory lacks explanation for a certain phenomenon, but the new hypothesis should be able to be successful at explaining all that the old theory can explain. (Thus, creationism is NOT a theory, not even a valid hypothesis!)
Peer review, peer review, peer review.
THE most essential step for a hypothesis to become a theory is peer review. It needs to be published in a science journal. A hypothesis is rarely completely correct, but the good parts are generally noticed by readers.
It’s redundant to say that not every accepted theory is a revolution. You never know if a theory will be a revolutionary one. Don’t do science aiming for that! You’ll be disappointed.