In typical papers, what the abstract says contradicts what the graphs say, and what the conclusion says; and the graphs contradict the conclusion.

Papers where they all match are exceptional and precious. But we cannot trust that just any paper is among them, even those published in Nature.