I don’t really understand how we can make conclusions about the entire universe when literally all of the data is collecting from one point in space. Couldn’t there be local effects that obfuscate global behavior? Is there an implicit caveat asterisk on all such statements that is like “as far as we can possibly tell from the data we have” and the reality is we really, really don’t know for sure?
Science at its core runs on a repeated process of collect some data, explain it, collect some more data, explain it as well, and keep repeating forever. The scientific model you are likely more familiar with refines this a bit where you use the current model and data to direct where you spend your time/funding collecting more data, but that is mostly a heuristic. Sometimes a person going off on their own will find something amazing to bring back, but we have to remember that many others spend their entire lives and end up finding nothing new or noteworthy.
One problem with science literacy is that too many people treat the existing models as true, even when they aren't. In some cases we know they aren't correct because there are discrepancies in experiments that shouldn't exist, but no model better fits the data and resolves the discrepancies.
At no point can science say "this model is how reality must work". But that is somewhat scary. It is scary to think the truth is unknowable and at best we will have an ever better approximation, so people find it simpler to treat that approximation as truth.