4 ms·

> the overwhelming majority of papers in most disciplines will hold up very well under scrutiny I guess you didn't get the memo about the replication crisis?

by pdonis 3y ago

> the overwhelming majority of papers in most disciplines will hold up very well under scrutiny I guess you didn't get the memo about the replication crisis? But even leaving that aside, no, your statement here is not correct. Even in "hard" sciences, many papers don't end up holding up under scrutiny. That's by design. The purpose of publishing a paper is not to document a result that is already known for sure to be right. It is to push the boundaries of current knowledge. Many such pushes don't pan out. We non-scientists don't see that because we never hear about those papers. We only hear about the ones that do pan out--or the ones that somebody hyped and marketed even though they hadn't panned out yet, and then got caught. To take an example from a century ago: if you just look at Einstein's published papers on General Relativity, it looks pretty straightforward. He made a few tries all tending in the same direction and finally ended up with the right field equation, the one that we now know, a century later, to have extensive experimental confirmation. What you don't see is all the papers published by all the other physicists that were working on a relativistic theory of gravity--all of which were wrong and were dropped. But they were still published papers, because the standard of publication wasn't "this has to be right", but "this is worth considering to see if maybe it's right". > We need scientists talking directly to the public, explaining what they're doing We already have that. Scientists write books, do videos, have blogs. Unfortunately, scientists don't do a good job at communicating with the public, because they don't take care to distinguish what is actually established--by experimental confirmation and replication--from their own pet hypotheses that haven't yet been confirmed. (String theory in physics is a prime example: string theory makes no testable predictions at all, yet string theorists routinely talk about it as if it were proven fact.) > and listening to screwball ideas from the peanut gallery. No, we don't need that, because nobody from the "peanut gallery" (meaning people who are interested in science but don't have a good understanding of our best current theories) has ever, in the history of science, come up with an idea worth pursuing. Tales to the contrary are myths. For example, Einstein is often given as an example of a "peanut gallery" person who came up with a great theory; in fact he was nothing of the kind. He got a doctorate in physics right around the time he published his classic 1905 papers. He had a thorough understanding of the best current theories of the time in physics, and that was recognized by the top physicists of the time. He had close connections with top experimental physicists who kept him up to speed on the latest discoveries, and he worked with them to develop theoretical models for them (for example his work on Brownian motion and on specific heats). He was no "peanut gallery" person; he was a working scientist.