Scientific knowledge becomes more trustworthy through processes that expose ideas to checking. Researchers describe methods and evidence so others can critique them. Peer review is expert evaluation before or around publication; it can identify weaknesses, but it cannot guarantee a study is correct. Replication asks whether independent work can produce compatible evidence.
Science is refined when stronger evidence changes the balance of support. Disagreement is not automatically failure: it can reveal methodological differences, uncertainty or a need for new tests. Confidence should match the amount, quality and consistency of evidence.