When may a scientific theory be revised?

never, once it has been published

when a single experiment’s data goes against it

when a new hypothesis provides a better explanation

after a large amount of data refutes it