That's an excellent example of what's called a crucial test,

which is sort of the pinnacle of what scientists are all about.

If we do a test and it doesn't agree

with a particular theory that we have,

that's problematic, but that doesn't mean

that it refutes the theory,

because if you were to refute the only theory that you have,

where do you jump to?

You don't have any alternative.

If we were to do a scientific test tomorrow,

and it was inconsistent with the theory of general relativity,

then what? There is no alternative to general relativity.

In fact, when there have been experiments over the years

that seem to have been inconsistent with general relativity,

guess what? They've all turned out to be faulty.

If you had to choose between whether or not general relativity

has been refuted by your test,

or your test is flawed,

go with the fact that your test has been flawed.

In the case of Eddington's experiment,

we had two viable theories for what gravity was.

We had Newton's theory of universal gravitation on the one hand,

...