I used to agree with you, but I see quite clearly that the need for funding, prestige and job satisfaction, scientists write in peer-reviewed publications both necessary extrapolations as well as overstatements of the impact of their research. If they stuck to the raw facts, it would be a dull read. You are certainly right that the whole result is relatively rarely overturned - unless it turns out that the P values didn't carry it, for example. Still, that doesn't make it 'all right' in terms of the doubt cast on science as a whole.