You should distrust every analyst and trust their analysis iff the argument is sound and you reach the same conclusion by doing the same analysis process based

You should distrust every analyst and trust their analysis iff the argument is sound and you reach the same conclusion by doing the same analysis process based on data you trust. Faked data is a big problem as that pretty much requires at least some trust; but the analysis part of any decent paper should be something which should be convincing even to a "hostile" reader who doesn't want to believe the author.