I think you can say they are rarely correct. Scientific Papers that truly advance our understanding are rare indeed. The rest are either failed hypothesis or

I think you can say they are rarely correct. Scientific Papers that truly advance our understanding are rare indeed. The rest are either failed hypothesis or of little predictive value. Note that this is not at all a value judgement. Incorrect hypothesis are important steps in the way of learning new things, and papers of little predictive value can in volume lead us to new solutions.