And on the issue of credibility of specific scientists-- data replicated by multiple independent researchers has historically had the most reliability. So that is one way you can assess. Note that I am not saying multiple people making the same conclusions about one set of data, consensus. Consensus is not a real level of evidence. Independent replication, though, especially with different instruments, increases the chances of reliable observations.
When a person is not sure, perhaps because the subject in question is not in their field of expertise, reverting to reason instead of evidence is not likely to lead that person in an accurate direction. Reason is full of pitfalls, mainly hidden in faulty premises-- hidden until new observations help us see more of what is going on.
To use an extreme analogy, suppose someone doesn't trust explanations and evidence about the earth being roundish and thus says "I don't know which scientists to trust, so I'm going with the earth being flat. It looks flat to me." That person is going to make navigation errors. And of course, there are people out there like that!
In my medical career, I have seen reason lead to significant errors in standard practice, only undone finally by evidence. And I've seen how hard it is to get people to change when reasoning has gotten ahold of them. This has made me very stubborn on the issue of careful observations and wariness with chains of reason, if any part of the chain is not tested with observations.