Wikipedia:Reference desk/Archives/Science/2023 December 26

= December 26 =

Why so many unpublished studies?
I was reading this article from Scientific American, How Two Pharmacists Figured Out That Decongestants Don’t Work, and something caught my attention:


 * "the oral decongestant monograph panel reviewed a few published studies and multiple unpublished studies for phenylephrine".

Why would there be so many unpublished studies? I'm not so concerned about the particular drug in question but science in general. What's the point of conducting a study if it doesn't get published? Do such studies get rejected because of some sort of flaw? Did the authors not find a publication willing to publish their study? Do scientists conduct studies with no intention of publication?

Sorry, I'm asking too many questions and speculating on the possible answers I guess what I am really asking is Thanks! Pealarther (talk) 11:04, 26 December 2023 (UTC)
 * 1) How common is it to have unpublished studies?
 * 2) What are the most common reasons for a study not to be published?


 * The vast majority of science done by industrial firms is never published. Suppose you want to find a new decongestant and test thousands of potential drug candidates. Most won't work and these negative results will sit in company databases but are unlikely to be suitable for inclusion in reputable journals. Some aspects of such work may be reported in internal documents but that's not "publication": secrecy is the norm so as not to alert competitors to what has been tried and failed. Mike Turnbull (talk) 11:47, 26 December 2023 (UTC)
 * It is not difficult to get results published because there are competing vanity publishers who have no interest in whether a book is suitable for publication or good enough to sell. In professions that are obsessed more with quantity rather than with quality, the Publish or perish phenomenon arises and may contribute to a Replication crisis where reported results of scientific studies are difficult or impossible to reproduce. Some reasons for rejection of results from Wikipedia include unverifiable claims, lack of peer review, outdated material or conflicts of interest that may lead to covert advertising. Scientists who find the peer-review and publication process to be too slow often choose to release results in Preprints. There are research results in mathematics, physics, astronomy, electrical engineering, computer science, quantitative biology, statistics, mathematical finance and economics that have been published only on the online-accessible preprint server arXiv that receives about 16,000 articles per month. Philvoids (talk) 12:06, 26 December 2023 (UTC)


 * It's one of the known issues with clinical trials. Historically, an organizations (public or private, academic or commercial) could conduct a clinical trial, and be left to their own devices and schedule on when or whether to publish their results, and how to conduct their analyses of the data.
 * This led to a number of problems, in various flavors of publication bias and selective reporting. A drug company might choose not to publish (or to slow-walk publication of) a result that failed to show efficacy for their novel drug candidate. An academic lab might lose interest in a study that failed to produce a novel result, and shelve their data in order to preserve and pursue funding for more 'interesting' studies. A clinical trial group might choose to only report a subset of their findings to amplify or suppress certain results. Worse, they might opt to re-analyze their data to identify 'effects' that they hadn't planned for or predicted when setting up the trial: HARKing and other sorts of post-hoc analysis.
 * It's only fairly recently that clinical trial pre-registration has become the norm (at least in North America and Europe). TenOfAllTrades(talk) 14:45, 26 December 2023 (UTC)


 * Trials that fail to show a useful result are particularly liable to never be published. This is a very real problem and probably affected a number of those studies. And some of those test may have been done because they hadn't heard about other failed trials so it was a wast of effort too. Documenting negative results is very important but it is only recently that people have started to recognize this and there are now journals that have been set up specifically to publish them. The bias in favor of only positive results also contributes to the replication crisis mentioned above. Hopefully the trial pre-registration mentioned in th previous response will also contribute towards rectifying this problem. NadVolum (talk) 11:01, 27 December 2023 (UTC)

max use temperature in Ar/N2 > max use temperature in vacuum
Hi. This is site lists boron nitride crucibles' maximum use temperature as 900 °C in air and 1800 °C in vacuum. "maximum use temperature in vacuum is higher than maximum use temperature in air" makes perfect sense.

However the max use temperature in Ar/N2 is actually higher than the vacuum temperature. This fact seems counter-intuitive (at least to me).

Q1: What is the reason behind this? (max use temperature in Ar/N2 > max use temperature in vacuum)

Q2: Is there any high temperature crucible materiel for which this isn't true? That is to say, for this hypothetical material: max use temperature in vacuum > max use temperature in Ar/N > max use temperature in air.

Liberté2 (talk) 19:59, 26 December 2023 (UTC)


 * 2: no. Inert gases valence being complete, the vicinity with their molecules act like an isolator, a bit as a "punching ball" effect. --Askedonty (talk) 22:15, 26 December 2023 (UTC)