How many authors does it take to write a letter about LHC leptons?

[Deetee]

http://www.ncbi.nlm.nih.gov/pubmed/21517374

 

[Myriad]

Obviously, one to hold the lepton, and several thousand to spin the LHC around really fast.

Serious answer: I've had my name included on scientific papers for being the "instrumentation guy" -- the one who actually developed (built/purchased/integrated), maintained, and operated the data acquisition equipment and collected the data during the study. If that's not part of the actual science being done, then nothing is. (In my case I actually also did help with the writing of the prose of the papers -- but despite the conventional meaning of the word used, being an author of a scientific paper does not require or imply that.)

The LHC just happens to require thousands of elite "instrumentation guys" to develop, maintain, and operate.

Respectfully,
Myriad

 

[Dinwar]

You're complaining that science, which has become an inherently interdisciplinary process, gives credit to too many people?

I can see this being a serious problem in terms of journal space--I believe Nature has a 4 page limit (text, figures, citations, everything). I think Science has a similar limit. But this is not a problem in terms of the science being done, but rather a book-keeping issue.

Perhaps we could contract the authorship by having a set team that we can cite? For example, "LHC Scientific Staff" or something. The problem with THAT is that when searching for an author you don't necessarily pull up those papers. But again, that's a book-keeping issue.

 

[ben m]

Perhaps we could contract the authorship by having a set team that we can cite? For example, "LHC Scientific Staff" or something. The problem with THAT is that when searching for an author you don't necessarily pull up those papers. But again, that's a book-keeping issue.

And it's not necessarily a problem. When an LHC postdoc is applying for faculty jobs, they don't want to say "I have 800 publications, which you can find by searching for my name on SPIRES, so you can see I'm qualified." ("Sorry, our four top applicants *each* have 800 papers. The same 800 papers, in fact.") Nor do they want to say "You won't find anything published under my name."

Authorship is just a set of conventions. For contrast: for both (a) the Hubble Telescope and (b) the LHC, it's true that none of the results would exist if it hadn't been for the hard work of a team of instrument-builders/calibrators/data-pipeliners/engineers. Yet Hubble Telescope data is collected, analyzed, and published by teams of two or three authors, in papers that do NOT cite the Hubble instrument people. (They wrote their own papers, on the instrument-as-an-instrument.) By contrast, even if an LHC dataset is "analyzed" by a small group ("we downloaded some data that had been collected and reconstructed already, and used it to make plots") the published paper conventionally lists everyone on the team.

Is that justified? It makes sense due to the extremely tight integration between all parts of an LHC experiment. The grad student analyzing the final-reconstructed-data may very well phone up the engineer who wrote the slow-control-data-bus specifications, and say, "Let's discuss whether we can speed up that data bus for next run." And vice versa---the engineers and instrument people will propose improvements and changes, that get discussed with an eye to their impact on analysis. That doesn't happen so directly with telescopes.

But that doesn't mean it's the only way to do things. We have our convention; it has its oddities, but most people who care are aware of them and have figured out other ways to understand who contributed what.

(One of the counterbalances is in non-refereed papers. You get your own name on conference talks and posters: "by Rooster Cogburn for the CMS collaboration")

My concern with the particle physics convention is this: it used to be that publications were a *moderately* objective standard for (e.g.) hiring decisions. "This candidate has 12 papers and 60 citations, this one in the same field has 4 papers and 10 citations." Removing that, we're going back to an entirely reputational system---"This candidate was strongly recommended by the Ph.D. advisor, this one's advisor had reservations but the lab director had high praise." That sort of system has very few checks against discrimination on basis of gender, nationality, etc.. There's a lot of research showing that supervisors tend to subjectively under-rate accomplishments by women---the main idea being that women who don't hew to gender stereotype are viewed as bad colleagues ("pushy", "bitchy", etc.) and those who do are viewed as bad scientists ("not assertive enough", etc.) and that such subjective judgements can overshadow and sort of objective report of how much one has contributed.

Publications and impacts and h-factors don't remove that effect, but they have the potential of being a counterbalance. I don't like the idea of a zero-counterbalances system of recognizing work.

 

[geni]

You're complaining that science, which has become an inherently interdisciplinary process, gives credit to too many people?

There are several problems. One is the problem you got with the Andrew Wakefield paper. Bunch of authors there but most don't have a clue if the paper is valid or not.

Another is the outright fraud problem. Two Phd students working next to each other agree to list each other as authors on their papers. People start to trade authorship credits.

The lab bigshot getting his name on every paper that leaves the building regardless of whether he has actualy had any involvement with it is a long standing problem.

Then you run into people falsely listing people as co-authors on the basis that no one actualy cares any more.

 

[Deetee]

I can see this being a serious problem in terms of journal space

In medical articles, each co-author is expected to say what his/her contribution to the paper was.

I expect for papers like these, that would require a few hundred pages of the journal.

BTW, at least Brian Cox is in there, to give it some oomph.

 

[Mr.D]

Obviously, one to hold the lepton, and several thousand to spin the LHC around really fast.

Best physics joke ever.

Giving credit is tricky with these large multi-institutional, multi-national collaborations, particularly ones that can go years between the proposal, design, testing, groundbreaking, turn-on and data acquisition phases, nevermind actually analyzing data.

I was involved in a project that graduated several Ph.D's before the experiment was ever turned on; surely their work designing, prototyping, testing and building equipment needed to be acknowledged, especially since some of it had to set new boundaries on detection levels.

On the other hand, you do get glorified button-pushers on the author list; postdocs who are full-time on a different project, but put in ten days of shift on-site a year so they can get free airfare for a weekend of skiiing - that sort of thing.

Hardly a perfect system, but I think I'd rather see a few too many authors than a few too few. Besides, it was my experience that if you _really_ wanted to know who did most of the hard work, just take a bunch of the senior postdocs to a bar for a couple of rounds of drinks and ask them for funny stories - there will always be funny stories about the goof-offs and the hard workers (for different reasons of course)

 

[Blue Mountain]

How many authors does it take to write a letter about LHC leptons?

http://www.ncbi.nlm.nih.gov/pubmed/21517374

Answer: 3,050

 

[Mojo]

The 1993 Ig Nobel Prize for literature was won By the 976 authors of a paper in the New England Journal of Medicine "for publishing a medical research paper which has one hundred times as many authors as pages". I think this probably beats that (unless it is a very long letter).

 

[Deetee]

Obviously the thing to do if you get into particle physics is to change your surname to "Aad".
Then all the papers will be Aad et al.

Here is another really heavily authored one, I assume the author list is similar but I note some differences.
http://www.ncbi.nlm.nih.gov/pubmed/21230962

 

[paiute]

http://www.ncbi.nlm.nih.gov/pubmed/21517374

How many people does it take to make a fairly simple 90 minute movie? Watch all the end credits sometime.

 

[Blue Mountain]

Here is another really heavily authored one, I assume the author list is similar but I note some differences.
http://www.ncbi.nlm.nih.gov/pubmed/21230962

That paper has 3,172 authors listed, 112 more than the first one. A few names are different, some have been removed and others added. There are also orthographical differences: 'Henss T and 'Mass M' appear in one list as 'Henß T' and 'Maß M'.

I found a copy of this paper on-line at http://prl.aps.org/abstract/PRL/v105/i16/e161801. The paper is 19 pages: 4 pages of scientific text and 15 pages of footnotes and authors.

I wonder how they decide which names to include on the list? Assuming these experiments and the subsequent analysis don't happen in a single day, I'm certain there would be some turnover in the personnel during that time. (Indeed, several people in the paper are listed as deceased.) Maybe they just take a look at a computerized record of everyone who clocked into work on the day(s) they ran the experiment.

 

[ben m]

Hi folks,

Don't get too excited, this is going to be the condition of every paper coming out of the ATLAS, CMS, ALICE, and LHC-B experiments for the next 15 years. It's been the case for big experiments (CDF, D0, BaBar, STAR, SuperKamiokande, Pierre Auger, etc.) for decades.

The author list is basically "the entire collaboration" as of the time the paper was submitted. For CMS, the standards are: incoming people get added to the general author list after a year's activity. Departing persons are retained as authors for some time after they leave, conventionally one year for junior members (grad students etc.) and longer for senior members. (Note, of course, that that doesn't distinguish *whose* work was used *when*.)

Anyone on the author list can withdraw from any individual paper ("I disagree with those error bars, I won't have my name on it.") and individual one-off contributors can request to be added ("Genevieve, the theorist down the hall, helped a lot with this analysis and should be a coauthor"); each institution has a team leader responsible for validating such requests, and the experiment has a committee that collates everything.

Each institution is required to contribute (out of their own grant money) to the operating cost of the experiment; this contribution is proportional to the size of your institution's author list (counting Ph.D.s and up), so there's an incentive not to inflate the author list. ("Heck, let's toss Bob in there, his tenure review packet needs a boost.")

 

[Mr.D]

I wonder how they decide which names to include on the list? Assuming these experiments and the subsequent analysis don't happen in a single day, I'm certain there would be some turnover in the personnel during that time.

As ben m wrote above there are general rules and guidelines which vary a bit per experiment (and probably by university, department and discipline too.) Unfortunately, politics sometimes come in as well.