This is critical for science to be meaningful and useful. It should be the utmost priority in any realm to ensure the input data is accurate. Tip was referring to multiple studies just yesterday citing how they were "science'd" (I believe that was the term he used) and peer reviewed/confirmed. If your input data is inaccurate, your peer-reviewed output data is also inaccurate. I'm not sure what data specifically was being used there and can't say one way or the other if that data was accurate but the point is much bigger than the accuracy of a handful of studies. The principles of science must be adhered to so it can be relied upon in all realms it's deployed. Whether it's improperly calculated instruments, human error, lax testing/recording, biased influencing, or any other inaccurate input it degrades the usefulness of the output. Perhaps, and realistically considering, science has always been this way and the innate level of ignorance borne from that has instilled too much confidence. With so much data at our fingertips now it's easy for a pro or average Joe to spot inconsistencies/inaccuracies and it certainly muddies the water of reliable consensus.