In the "anthropic principle" it is proposed that the universe adapted to us, rather than the other way around. Abundant data exist which contradict such a notion.
Multiple universes have their foundation of support in string theory or the postulated existence of multiple time dimensions---but this does not support the existence of an anthropic universe.
To resolve this as it fits ID, someone (actually many scientists) will have to put forth a testable hypothesis, test it, replicate and do it again and again with the same theme but with variation as they suggest themselves by the data. That has not been done and that's why ID does not rise to the level of science.
Well, no.
The evidence for fine tuning, and the evidence for many universes, happens to be the same evidence-— the anthropic coincidences themselves.
That’s why the physicist Lawrence Krauss feels that if his many worlds version of string theory doesn’t work, the alternative is design.
Suppose someone saw a video of me taking a shot at a basketball hoop from 70 feet, and making the shot.
The hypotheses I’m lucky, that I’m good, or that I simply took many shots and the video shown happens to be of me making it, all rely on the same piece of data-— the video of me making the shot.
Of course, this is beside the point when it comes to Gonzalez, whose credentials are unassailable and has presented evidence supporting his “Rare Earth” hypothesis in Nature, among other peer-reviewed journals.
So, to paraphrase your own words, for evolution to rise to the level of science, scientists would need to repeatedly replicate random evolutionary changes from one species to another?
How exactly does one ensure pure randomness in a clinically controlled setting? There mere fact that a study is underway implies intelligent design.