A Ni user would accept the process of evolving as true. Then they would come up with something like mutations in the genes to explain how differences come about in organisms. If their idea of mutation did not line up with the observed process of evolution (like they predicted X% mutation rate but the evidence shows otherwise), they would change it to fit how evolution works. Their concept of a mutation is now changed.
A Si user would accept mutations of genes as true. From their observation of mutations in the genes, they would come up with some theory of what affect they might have. If their theory was pointed out to be untrue (like they predicted that this would result in random harmful adaptions staying), they would modify their theory of evolution to be consistent with the fact that harmful adaptions do not stay. Their concept of evolution is now changed.
If I had actually chosen an example where the opposing evidence was the same, you would have seen that they ultimately get the same result [of their theory being correct].
Sorry if this wasn't really the best example, but I don't know any confirmed dominant Ni types in real life.