Via this week's C&EN, this letter to the editor:
Reproducibility of scientific findings
I found the timing and tone of the article on reproducibility particularly interesting given the current state of politics (C&EN, March 17, 2025, page 20). But that set aside, what I did not read from the author of this article was a little simple advice, like, scientists need to pick up their phones and use them for what they were initially intended for.
This advice, which I was given by professor Peter Schultz in 1992, accelerated my trajectory through the Chemistry Department of the University of California, Berkeley, and thanks to the new friend I made at Harvard University in the process, I published an article in Science and received my doctorate degree 3 years later.
What I have found since is that there is much more to the methods section than most journals allow one to write, and thus simple but important details are left out—and probably more than 90% of the time. So even though I do think there are some bad scientists and some findings cannot be reproduced, I also believe that 90% of scientists may be too intimidated to simply pick up the phone and get the details. If they do, maybe they, too, will make a new friend.
James Prudent
Madison, Wisconsin
I like the idea of picking up the phone, but...
I have to say that a phone call should not be required for normal reproduction of a paper. I think it's probably better to talk on the phone than to publish something on PubPeer first, but the goal of writing a paper should be providing sufficient information to enable independent reproduction by another scientist.
I was taught by my postdoc advisor to document EVERYTHING in the literature. (Journals are quite happy to include appendices for that sort of thing...that don't go towards word counts.) I don't understand this letter writer at all.
ReplyDeleteWhen I have had to "pick up the phone," I found a complete ass on the other end of the line who purposely obfuscated his work to fend off his perceived competition.
The majority of times I try to contact authors about critical missing information from their papers, I get zero response. Ideally all the information would be in the paper or the SI, and in the absence of that a communicative author to get the missing bits would be nice...but all too often we get neither.
ReplyDeleteThat has been my experience as well.
DeleteThe original article says, "People need to be trained in how to do experimental design, statistical analysis, data integrity". As someone with a degree in Chemistry and Applied Statistics, there already ARE people trained in experimental design and statistical analysis: they are the people who have degrees in statistics and are often already available to help you. I would encourage university scientists to contact their statistics department and see what kind of consulting is offered. Get help early in your planning. As a pure biochemist or chemist you will not have as extensive training in statistics as someone with the degree and using the wrong methods for analysis will give "results" but they will be misleading results.
ReplyDelete