As a researcher who often claims things and listens to claims of peers, all I can tell you: never trust a scientist. Long gone are the days of scientific method and integrity. There are solid facts and statistics showing that only 10-30% of published academic research can be reproduced. The morale of the story: be very reserved, and don't fall for it easily.
Here's an exercise for you: whenever faced with an article making grandiose, game-changing claims (curing cancer, developing new computing technologies, etc.) - make a note of it. Few years later, check on the status or even existence of these claims.
This. Do not trust something just because it is "peer reviewed", Peer reviewed is much better than nothing. There is no doubt it is 1000x better than listening to your neighbor or pizza guy. Even so, many researchers have no idea what they are doing or how to even analyze their data. Moreso even if you want to do that you do not have time due to publish/perish. All the folk publishing half-analyzed, half-controlled, filtered results get the publications, money, and jobs. Many who try to do it right get discouraged early. It is a culture problem. Personally, I am glad I have realized this early on.
Most of what is published (at least in the biomed field) is barely worth anything. You need to assess all the nearly worthless results together and then you may have something. One paper alone is like nothing.
