Slashdot Mirror
How Many Scientists Does It Take To Write a Paper? Apparently, Thousands
An anonymous reader writes: The Wall Street Journal takes a look at the current spike in number of contributors cedited in scientific journals. The problem is highlighted by a recent physics paper which credits 5,154 researchers. The journal reports: "In fact, there has been a notable spike since 2009 in the number of technical reports whose author counts exceeded 1,000 people, according to the Thomson Reuters Web of Science, which analyzed citation data. In the ever-expanding universe of credit where credit is apparently due, the practice has become so widespread that some scientists now joke that they measure their collaborators in bulk—by the 'kilo-author.'"
Yes, there's a trend going upwards but there are only 1,400 papers with 50 or more authors. In 2009 about 1 million biomedical papers were published. So if we make the unlikely assumption that all the high author number papers are biomedical, that means that a whopping ~0.15% of the papers published each year have more than 50 authors. Not exactly a big deal.
soylentnews.org
I find a greater red flag to be a large number of reference sources, without specific denotations of which data is pulled from which source, making it impossible for the reader to establish context in how the source data was established and therefore if it is properly employed. But that is a stray from the topic at hand.
References;
1. Me
When everyone has been credited on dozens or hundreds of scientific papers it dilutes the perceived value of the average researchers involvement in each paper on which they claim credit. In the competitive worlds of grant application and academic positions, this means you have to be more worried about getting credit on as many studies as possible than about actually doing meaningful research on a single issue in order to make it through the initial review/HR screening process.
Is it just my observation, or are there way too many stupid people in the world?
But the question is actually if it makes sense to have many authors on a paper. If you have 10 or more it should already raise a warning flag.
One of the examples they highlight is CERN, where thousands of researchers do indeed contribute. Since the current way of assigning credit in academic circles is to put people on a paper, it's hard to see what else can be done. Yes, it's a bit weird but it happens very rarely. I don't see what the "warning flag" is being raised in aid of. There's no reason to think the science is worse in a large author count paper. If I was interviewing someone with only authorship in high author count papers then I'd ask the the appropriate questions. Then again, I'd likely ask those questions anyway. I've interviewed first authors (of a paper with under 5 authors) who couldn't explain the analyses described in the article. There are warning flags, but the number of authors likely isn't one of them.
soylentnews.org
The Human Genome Project, assembling work from thousands of researchers, developers, and technicians worldwide had hundreds of authors.
http://www.nature.com/nature/j...
It can make sense in such a large project to list as many of the contributors as possible.
Scientists have played pranks with co-author names. The famous Alpha-Beta-Gamma paper comes to mind. Low temp physicist Hetherington had included his dog as a co-author so that he could use "we" in the paper.
The spoof paper authored by S Candlestickmaker done by a student of S Chandrashekar was very famous. The student later lamented that spoof paper was his best known contribution to science than his PhD or his entire research career. It is telling I remember S Candlestickmaker but not the student's name.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
But are all of those authors, and not data contributors?
But are all of those authors, and not data contributors?
Collecting the data is the actual work. Any idiot with a computer can make the analysis. And draw the wrong conclusions from that.
Of course news about a fake are Fake News.
You want to increase the level of bureaucracy and publish-or-perish mindset in research? Private research certainly isn't doing anything that doesn't increase next fiscal quarter's numbers, let's continue to hobble public research, sure...
It typically works the other way. The grad student writes the paper and his adviser wants his name included because he "advised".
Research is the purpose of universities. Teaching is a secondary activity which is fully satisfied if it provides the next generation of researchers. Training for jobs should never have been done by universities but since it has it sure as he'll must not become the core focus.
There is no research without value. Knowledge is the most valuable asset there is and it is all practical. But sometimes you need to wait 200 years to be able to use the value.
Tenure is the single most important aspect of universities. Academic freedom cannot exist unless it is absolute. That includes freedom from market forces.
People like you are the result of the dumbing down of humanity but don't try to now dumb down the last bastions of intellectualism as well.
Unicode killed the ASCII-art *
It is common practice for grad students to include more renowned professors in the lab etc on papers out of respect of their funding. Since your worth as a Primary Investigator (grant holder) is based on the number of papers on your resume, It is the only way of having a career. I've seen papers with 20+ PH.Ds on it where most of them, I know for a fact, have done practically nothing for the paper.
I needed to fix that for you. In general a well funded lab is so busy the PIs do almost nothing buy advise and review to make sure the grant work stays on progress. In reality Grads and Post Docs do 90% of the work in hopes of one day moving to a career in management aka a PI position. I've never meet a PI that has time to work the day to day minutia of science discovery in fact they often lament if they wanted to keep doing science they would have never started a lab.
Momento Mori
With academic papers the author list is usually a flat list. Sometimes the first authors are the key ones but many papers simply have the authors in alphabetical order. Combine that with a writing style that ephasisies what was done over who did it and it's pretty difficult to figure out who the main drivers of the project were and who were just workers.
Contrast that to movie or videogame credits where there is a structure that tell you who played (or in the case of videogames voiced), who the people driving the project were, which category each of the workers fell into and so-on.
note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register
Absolutely agree that it's much ado about nothing. AND bad statistics ! CERN as an example is a lot of nonsense... it's a HUGE project with a HUGE population of PhD's, grad students, undergrads, managers, technicians, and everybody else. All working towards a common goal. And the science developed by those thousands of authors is an enormous collaboration, enabled by ... yeah, you guessed it, the World Wide Web. Which was INVENTED at CERN to enable... Collaboration.
WSJ, you look like a bunch of idiots. Stick to talking about stocks and rich people stuff. You suck at science.
Science today is judged by two metrics: papers published and students graduated.
It's important to actually understand that statement if you want to understand some of the quirks and problems with scientific culture.
You do not get credit for projects, advancements, talks, transition to industry, programs, results, etc. The government granting agencies only track papers published and students graduated when comparing different granting offices. Put another way, the government internally sets funding targets for each sub-field based on papers published and students graduated. Thus only papers published and students graduated are meaningful to science (again, not results, but papers).
Papers and # of PhDs became the currency of science, and are used to judge everything from the readiness of a student to graduate to the differential societal contribution of different scientific fields.
This has led to a situation where if you want to graduate students in fields like particle physics, you need to include them on the very rare papers that come out. Failing to graduate students would lead to a decrease in funding. For a student to get "credit" for working at CERN or NASA, that student needs to be on a paper. It's as simple as that.
It is common practice for more renowned professors to include grad students, PhD's in the lab etc on papers just to get their names out there. Since your science worth is often based on the number of papers on your resume, it's a great way of starting a career by having your professor include your name in grad school and beyond. I've seen papers with 20+ names on it where most of them, I know for a fact, have done practically nothing for the paper.
Common practice or not, those people don't belong in the list. It demeans the value of those that actually did work on the paper.
The cesspool just got a check and balance.
There have been efforts underway to standardize acknowledgement of data that came from other scientists. Most of us in the field have been calling the concept 'data citation' for a while (but it also refers to the act of linking, plus the string of text in the paper ... so it's a bit of a polysemous term at this point).
The basic idea is that for each grouping of data (I won't get into trying to define what a 'data set' is) that's being released, the group that's doing the release puts out a web page of information describing the data.
It would have the DataCite fields to specify how the data should be listed in the reference section of your paper, plus the w3c DCATterms to explain how to obtain the data. The DataCite schema allows you to acknowledge many different roles for people, allowing you to more clearly describe what different people's contributions were ... instead of just a long list of people, you'd have something more like movie credits.
This would solve much of the super collider issues, as you'd separately acknowledge the people who obtained & processed the data, who might have had no hand in the specific research that the paper describes. In my opinion, the authors should be people who agree with the findings that are being presented -- the folks who made the data should be acknowledged, but if they've had no chance to review the research that's being published (or have no ability to understand the researcher), they should not be listed as an author.
If you're interested in the topic, here are a few links that might be of interest:
If you're interested in participating in these efforts, either find a group in your research area, or for wider efforts, the Research Data Alliance's Data Citation Work Group.
I should also mention that there are similar efforts going on with scientific software. I've participating in some workshops (eg, RENCI's on data & software), but I'm not as active in that field. Some RDA have discussed starting up a group on software issues, but I think they'll be focusing more on Software Carpentry issues; for software citation I'd suggest contacting the Software Sustainability Institute.
ps. I've been included as a 'co-author' on papers where I've never had a chance to review the paper first. I think that all journals should check with all listed authors if they approve of the paper. (I've also peer reviewed a paper that had so many grammar errors in it that I suspect that none of the co-authors (most were native english speakers) had reviewed it) ... and it didn't reference the co-authors' earlier related articles). PeerJ does this and it also requi
Build it, and they will come^Hplain.
This has been my experience as well - dating back to the late 80's. A principle investigator is basically a grant-writing machine who has built enough of a reputation to get his grants funded. He also has to come up with new areas of inquiry and prod his students and post-docs into getting enough data to write a grant for that line of investigation.
They visit the lab, but pretty much never get to do the bench work. And the distance from the lab means that they cannot remember how long things take, so they are always wondering why everyone is so much slower to get the result than they expect.
Being a P.I. is even harder these days, because the funding for the NSF and NIH, etc. have not kept up with the demand. They used to write 2-3 grants expecting to get one funded. These days it is more like 5-8. That is a lot of wheel spinning - writing a grant is not in and of itself productive work.
I publish journal articles and reviews fairly regularly, and we only rarely include authors who did literally nothing other than be in the lab or consult. It happens occasionally when a young researcher has been working hard on a related project, but still does not have nearly enough data for a whole new paper. There are some cases where a researcher in another lab provides a reagent you can't get anywhere else, but even then they often review the final manuscript and make suggestions. However, I think that in larger labs this may occur much more often. Our lab is small (3 scientists and 2 student researchers). On one recent paper several collaborating scientists at NIST worked extensively on a part of the project that did not pan out, so they are in the author list even though their work does not appear in the paper. But you can't say they didn't do any work. There are always issues like this when deciding on authorship that are unique to each situation.
However, the physics and genetic articles that have thousands of authors are much harder to justify, and absurd to even think that anyone would go through the list. All but the first few authors are lost in any citation list when the paper is cited. No one will ever see the other names. Plus, it really messes with citation software like Reference Manager (newer versions of Endnote seem to handle it well).
The bigger journals now require author contributions to be listed, which is a good thing (and would be very difficult on a paper with 1500 "authors")
A brain is a terrible thing to waste... Mind? That's debatable.
That was quite a brief departure for a concept that's nearly 3000 years old. Universities for almost the entirety of that period were utterly divorced from the private sector.
University education was only job training for those jobs that required research skills like Doctors and lawyers. For everybody else there were training colleges (often created and run by unions to help workers advance into management ) but universities were about producing knowledge. Everything else was secondary at best.
It was only in the latter half of the 20th century that this was changed and it was a mistake to change it. We've changed it a dozen different ways since but they've all been disastrous.
Let's not lose tge last bit of research focus left. Product research is for the private sector. University research should be about studying the universe in it's entirety not just the parts we can see profit in.
Unicode killed the ASCII-art *
Collecting the data is the actual work. Any idiot with a computer can make the analysis. And draw the wrong conclusions from that.
It's a shame that so many people seem to agree with this. I would put it exactly the opposite: any idiot can be trained to collect data; knowing what data to collect, why to collect it, how it fits in with 100 years of pre-existing data, and how to condense all of that into a concise but readable story is the difficult (and creative) part.
Maybe it's learned from student science labs, where you spend a lot of time getting the mechanics of an experiment to work, then plug the numbers into a pre-set template for analysis. Of course, those labs are exactly about training students to collect data, and not so much about doing science. Maybe it's learned from the media, where an uninformed reporter picks a bit of data out of a paper and concludes the green coffee extract is a fat-cure-all. Maybe it's just that it takes a lot of work to be able to distinguish good work that advances our state of knowledge from a mountain of data that doesn't really say anything.
Lol, Terabytes.
I think you mean Petabytes.
Terabytes is what we did in the nineties.
- Radioastronomy (and probably other disciplines).
RogerWilco the Adventurous Janitor