The general conventions guide on how authorship order is decided across disciplines notes, in passing, that large multi-institutional collaborations in high-energy physics and astronomy handle order differently from every other field: they do not rank individuals by contribution at all. This guide is the deep dive on that specific case — how the ATLAS, CMS, and LIGO/Virgo collaborations actually decide who is an author, in what order names appear, and how individual credit is tracked when a single paper can carry more names than most conference attendee lists.
Why megacollaboration authorship is a different system, not just a bigger one
A typical multi-author paper still assigns some meaning to position: first author for hands-on work, senior or last author for the lab head, middle authors somewhere in between. That system assumes the number of authors is small enough that contribution can plausibly be read off the list. Large-scale physics experiments broke that assumption decades ago. The 2015 joint ATLAS and CMS measurement of the Higgs boson mass, published in Physical Review Letters (PRL 114, 191803), listed 5,154 authors across a paper whose author and institution listings ran to roughly 24 of its 33 total pages — widely reported at the time as a record for a single research article. With author counts in the thousands, no plausible ranking by individual contribution is attempted. Instead, these collaborations assign authorship collaboration-wide, under membership rules the group itself sets and enforces, and list names alphabetically rather than by any contribution hierarchy.
This is a genuinely separate system from the alphabetical convention in mathematics or economics covered in the discipline-conventions guide above. In those fields, alphabetical order is a norm individual co-authors choose to signal equal intellectual contribution on a paper with a handful of names. In ATLAS, CMS, and LIGO, alphabetical order is collaboration policy applied automatically to a standing author list that exists independently of any single paper — a researcher who joins the collaboration and meets its authorship criteria is added to that list, and every paper the collaboration publishes during their tenure carries their name, regardless of what specific role, if any, they played on that particular analysis.
ATLAS: a standing, collaboration-wide author list
ATLAS, one of the two general-purpose particle-detector experiments at CERN’s Large Hadron Collider, maintains its own Authorship Policy, overseen by a Publications Committee (PubComm) of twelve members appointed by the ATLAS Collaboration Board on two-year terms, with roughly half the seats turning over each year. PubComm recommends authorship and publication guidelines to the Collaboration Board and reviews collaboration documents before public release. Under that policy, the author list on an ATLAS paper is alphabetical by surname and applies to the collaboration as a whole rather than being assembled paper by paper: a physicist who qualifies for authorship under ATLAS’s membership rules is added to a single standing list, and that list, not a bespoke judgment about who contributed to a specific result, is what appears on each publication issued while they remain a qualifying member.
Qualification is not automatic on joining an ATLAS institution. Collaboration authorship policies of this kind typically require a documented period of substantive service to the experiment — commissioning or maintaining detector hardware, developing and operating software and computing infrastructure, or performing data-taking shifts — reviewed by the collaboration itself, not simply institutional affiliation or working on one analysis. The practical effect is that “being an ATLAS author” and “having personally worked on the physics result in a given paper” are frequently different things, which is precisely why megacollaboration papers list authorship separately from the specific analysis credit described further down this guide.
CMS: the same convention, independently arrived at
The Compact Muon Solenoid (CMS) experiment, the LHC’s other general-purpose detector and ATLAS’s counterpart on the 2015 joint Higgs-mass measurement, follows the same broad model: a standing, collaboration-wide author list, alphabetical by surname, governed by membership and qualifying-service rules the collaboration sets for itself rather than negotiated contribution order per paper. That the two largest LHC collaborations independently converged on the identical alphabetical, membership-based model is itself informative — it is not an idiosyncrasy of one experiment but the default that particle-physics collaborations of this scale settle on once contribution-based ranking becomes impractical at four-figure author counts. The joint ATLAS/CMS Higgs-mass paper is the clearest illustration precisely because it merged two independently-alphabetized standing lists into one combined list for a single publication.
LIGO and Virgo: alphabetical order with an explicit effort threshold
The LIGO Scientific Collaboration (LSC) documents its rule in a public policy (the LSC Publication and Presentation Policy, maintained in LIGO’s Document Control Center): the author list on LSC scientific publications includes every collaboration member who has earned that status, including engineers and technicians who contributed importantly to the design, construction, installation, commissioning, or operation of the detectors and major LSC facilities, and the list is alphabetical. Qualification carries an explicit, dated effort threshold rather than a vague service requirement: a member must have joined the LSC before December 1 of the previous year and devoted more than 50% of their research effort to LIGO since that date to be included as an author, with a related provision retaining authorship for members who left the collaboration or dropped below that effort level after a specified date. Some paper types use a tiered structure — a smaller Group 1 of authors most directly responsible for that specific result, ordered by their own agreement, followed by Group 2 listed alphabetically — but the collaboration’s flagship discovery papers use a single alphabetical list. The 2016 detection announcement “Observation of Gravitational Waves from a Binary Black Hole Merger” (Physical Review Letters 116, 061102) carries roughly 1,000 named authors under “B. P. Abbott et al. (LIGO Scientific Collaboration and Virgo Collaboration),” alphabetical across both collaborations combined.
Where individual credit actually lives, if not in author order
If author order carries no contribution signal on these papers, and CRediT-style contribution statements — designed and calibrated for teams of a handful to a few dozen people — are not practically published per-author against a list of thousands, the obvious question is where individual credit is tracked at all. It is tracked, just not on the public byline. Large physics collaborations maintain extensive internal documentation separate from the published paper: individual analysis notes, internal review records naming who performed which piece of an analysis, conveners’ assessments of specific contributions, and service-work records used for authorship qualification itself. Research examining this gap directly — including the arXiv working paper “Large publishing consortia produce higher citation impact research but co-author contributions are hard to evaluate” and CERN-focused work on “formal and informal credit attribution mechanisms in collaborative research” — documents both that this internal machinery exists and that it is a genuinely unresolved evaluation problem: the credit a researcher needs for hiring, tenure, and grant review is real and internally documented, but it is largely invisible to anyone reading the published author list from outside the collaboration.
This is the structural reason megacollaboration authorship cannot be evaluated the way a five-author biomedical paper can. See CASRAI’s overview of the CRediT taxonomy for the model that does work at ordinary team scale, and the group authorship dictionary entry for how journals and indexing databases handle a collaboration name standing in for an author list generally.
What this means in practice for research administrators and hiring committees
A CV or funding biosketch listing “co-author, ATLAS Collaboration” on a paper with several thousand names cannot be read the way a five-author paper’s byline is read, and evaluators who are not themselves particle physicists routinely need to know that. In practice, researchers from these fields typically supplement the raw publication list with a separate, explicit account of their individual contribution — a personal-statement section describing specific analyses led, hardware built, or software developed, distinct from and unverifiable purely from the shared byline. This mirrors the function that a CRediT contribution statement serves on an ordinary paper, but it takes the form of prose in a CV, biosketch, or reference letter rather than a structured per-role tag next to each name, precisely because a structured statement covering thousands of individually-tagged authors is not a realistic format. Administrators reviewing files from these fields should expect, and can reasonably request, that kind of supplementary account rather than trying to infer contribution from position in an alphabetical list of thousands, since that position carries no such signal by design.
Frequently asked questions
Is author order on an ATLAS or CMS paper ever meaningful?
No. Both experiments list their standing, collaboration-wide author roster alphabetically by surname on every paper. A name’s position reflects where it falls in the alphabet, not seniority, role, or how much work that person did on the specific result being published.
How does someone qualify to be an author on an ATLAS, CMS, or LIGO paper?
Each collaboration sets its own membership and service requirements, reviewed internally rather than self-declared. LIGO’s policy is the most precisely documented publicly: a member must have joined before December 1 of the prior year and devoted more than 50% research effort to LIGO since that date. ATLAS and CMS require a comparable documented period of substantive service — hardware, software, computing, or operations work — under each collaboration’s own authorship policy, not simply institutional affiliation.
Does everyone on a 5,000-author paper actually work on that specific result?
Not in the sense of having personally performed that analysis. Authorship in these collaborations attaches to standing collaboration membership, not to demonstrated involvement in one particular measurement. Someone qualifying as an ATLAS author through detector operations work, for instance, is listed on physics-analysis papers they had no direct hand in, and vice versa.
Does CRediT apply to megacollaboration papers?
Not in its usual published form. CRediT’s fourteen contributor roles were designed for author teams realistically countable in the tens, not thousands, and no major HEP or gravitational-wave collaboration publishes a per-author CRediT statement against its full byline. Individual contribution in these fields is instead documented internally (analysis notes, service records, conveners’ reviews) and surfaced externally through CVs, biosketches, and reference letters rather than a structured public statement.
How should a hiring or tenure committee read a megacollaboration author list?
Not as a ranked or contribution-weighted list — alphabetical position carries no signal. The documented practice in these fields is to evaluate a separate, explicit account of the individual’s specific contribution (analyses led, systems built) rather than inferring it from the byline, since the byline reflects collaboration membership rules, not a judgment about that paper specifically.
For the broader landscape of author-order conventions this is one variant of, see How Authorship Order Is Decided: Conventions Across Disciplines. For what qualifies someone as an author at all under the criteria most journals outside physics apply, see ICMJE’s four authorship criteria, and for what specific byline positions signal in conventionally-ordered papers, see Types of Authorship in Research.
