Its validity decreases as the complexity of your subject increases.
In physics, you can take simple measurements that validate inductive thinking; for example, how long it takes an apple to fall from a tree. In evolutionary theory, you cannot; for example, the genetic sequencing of retroviruses that existed a billion years ago.
In science, you pursue the hypothesis that is potentially provable; that is, there is the possibility of proving it. Proof consists of something that is experimentally repeatable. The so-called proof of evolution is all of biological history.
Well, now, we have two problems:
No one has ever witnessed all of biological history; therefore, we don’t have even a single experiment to begin to build a proof.
If we had a (past-only) time machine and could live billions of years, we could witness all of biological history but we would still only have a single experiment and could not “perform” additional experiments; therefore, we could not establish experimental repeatability and, therefore, could not build a proof.
Now, what evolutionists do in the face of this insurmountable obstacle is, like cheaters often do, to tell a lie. They say:
Well, you see, we are using inductive logic (many instances constitute proof).
The truth is:
Evolutionists use inductive interpretation (many instances inspire an opinion).