CC BY
57
Dominique Ra ynaud
Doctor in Sociology, Habilitation in Epistemology and Sociology of Science Associate Professor, Grenoble University, Grenoble, France; e-mail: dominique.raynaud@upmf-grenoble.fr
Collaborative Networks in Particle Physics: A Sociological Inquiry into the ATLAS and CMS Collaborations
The discovery of the Higgs boson is one of the most significant advances of particle physics in recent years. It led to the award of the Nobel Prize in Physics 2013 to Englert and Higgs for the theory explaining the origin of the particle mass. The Nobel Prize cannot conceal the fact that the results about the new particle have been achieved by the experimental physicists of the ATLAS and CMS collaborations, who are among the largest international collaboration of scientists in the world (2898 and 2932 physicists, respectively). This article is dedicated to the study of the organization and operation of the ATLAS and CMS international collaborations. The disparities between countries, structure of collaborative networks, physicists' cooperative vs. competitive preferences, and emerging properties of research work in large scientific collaborations are reviewed.
Keywords: ATLAS, CMS, Particle Physics, Scientific collaborative networks.
Introduction
As anticipated by Derek de Solla Price with the concept of "Big Science" (Price, 1963), a growing part of today science is taking place within large cosmopolitan collaborations — often composed of young researchers of all nationalities. This is the case in particle physics, where the size of collaborations now results in articles co-authored by thousands people. At first glance, the alphabetical lists of names appended to the papers gives an impression of "mass effect," which is hardly compatible with the very nature of the work done in these collaborations. The larger an organization, the more detailed the division of labor and internal structuring. This effect is not apparent in the lists of authors.
The ATLAS and CMS collaborations in particle physics have been selected as a subject matter because they provide favorable conditions for investigation:
1. The ATLAS and CMS collaborations have done research on the Higgs boson that has rewarded Englert and Higgs with the Nobel Prize in Physics 2013.
2. The findings were released in two twin articles of Physics Letters B (ATLAS, 2012; CMS, 2012).
3. These articles were signed by NA = 2932 and NC = 2898 authors, whose names are recorded in two lists representing 13 in 29 (ATLAS) and 17 in 32 (CMS) pages of text.
4. Until the release of the 2015 joint article by ATLAS and CMS (Aad G. et al., 2015), the ATLAS and CMS papers, with nearly 3,000 authors, held the record for the largest number of authors of a scientific article in the world.
5. The spirit of transparency that prevails at CERN makes it possible to investigate most documents from the CERN site (see Appendix 1). They consist of scientific papers, preprints, notes and conferences reports, as well as technical notes, personnel statistics
:_L
and regulation documents. In particular, the Letters of Intent (CERN/LHCC/92-3; CERN/LHCC/92—4) and Memoranda of Understanding (ATLAS 2002; CMS 2002) define the framework in which the research on the Higgs boson was done.
6. The operating rules are clear and explicit: "All codes are written," and are basically the same for both collaborations.
7. Because the Tevatron at Fermilab closed on September 30, 2011, ATLAS and CMS are now the best instruments for doing research on high-energy particles. The lists of authors appended to the 2012 articles thus establish the comprehensive register of all experimental physicists who discovered the Higgs boson.
Plan. After introducing the subject (1. Recent Advances in Particle Physics), we review the differences between the collaborations (2. Frequency and Rank of the Laboratories). Then we correlate these data with socioeconomic variables of the countries providing the physicists to the twin experiments (3. Size, Population and Wealth). Despite equivalent sizes, it appears that the two international collaborations are structured very differently (4. Organizational Differences). Next, we examine the benefits that scientists can gain through a cooperative vs. competitive attitude (5. Cooperation, Competition, and Strategic Edge). Emergent properties of large collaborations are also described.
1. Recent Advances in Particle Physics
The hypothesis of the Higgs field has been formulated to address a gap of "gauge theory" proposed by Glashow, Weinberg and Salam in the 1960s. This theory predicted the existence of a massless boson, whereas all bosons known at the time were massive bosons: 80 GeV for bosons W± and 91 GeV for the boson Z0. The Higgs field confers mass to gauge particles, among which W and Z bosons, that acquire mass by interacting with the field. The particles that do not interact, such as photon, have a zero mass. Although it has been proposed to name it "BEH boson" (Brout, Englert, Higgs), "BEHHGK boson" (Brout, Englert, Higgs, Hagen, Guralnik, Kibble) or "scalar boson," most physicists continue to speak of the "Higgs boson." The detection of the Higgs boson only became realistic with the commissioning of the LHC (Large Hadron Collider) at CERN, which is to date the most powerful particle collider in the world, with some 1200 electromagnets, each weighting 34 metric tons, a current of 12,000 amperes and a magnetic field of 8.33 Tesla. The collision energy (7 TeV in 2012, 13 TeV in 2015) far exceeds the 100 to 200 GeV of the LEP and 1 TeV of the Tevatron, closed on 30 September 2011. The LHC was inaugurated on 10 September 2008, after fourteen years of construction. CMS (Compact Muon Solenoid) and ATLAS (A Toroidal LHC ApparatuS) are two twin detectors positioned on opposite sides of the LHC beamline. These instruments are now the only ones able to detect the Higgs boson. Signal detection (600 million collisions per second) produces a Po of data per second. Data are filtered (25 Po per year) and archived on the WLCG grid, a network of 200 computer centers distributed worldwide. Computing centers are divided into a cascade of levels: CERN computing center (Tier-0) stores and redirects the data with high speed dedicated connections to eleven sub-centers (Tier-1). These secondary centers perform preprocessing data before redistributing them to the centers responsible for physics analysis (Tier-2).
sociology of science and technology. 2016. Volume 7. No. 1
171
2. Frequency and Rank of the Laboratories
The authors' names of the twin articles on the Higgs boson released in Physics Letters B can be aggregated by laboratories or by countries.
1. Let us first aggregate the co-authors by country. For volumetric reasons, these rankings are given at the end of the article (Appendices 2-3). Data indicate wide disparities in the contribution to the experiments: a few countries are enough to form half of the workforce of the collaborations.
In terms of number of researchers, CMS major contributing countries are the USA (958), Italy (291) and Germany (275), which together provide 52 % of the staff of the collaboration. The first contributors to ATLAS are the USA (593), Germany (415), the UK (292) and Italy (223), which together provide 52 % of the staff of the collaboration.
In terms of number of laboratories involved in the experiments, major CMS contributors are the USA (49) Italy (13) and Russia (7); major contributors to ATLAS are the USA (40), Japan (17), Germany (15), Great Britain (15) and Italy (13).
CERN physicists represent only 5 % of the collaborations. This low number is because scientists only represent 3 % of the staff, far behind the engineers (39 %), technicians (35 %), and administrative staff (17 %). External scientists working at CERN outnumber the internal researchers. They were 9210 against 2544 in 2007; and 11025 against 2427 in 2010 (CERN, 2010: 44). The originality of large equipments such as the LHC detectors is to be placed at the disposal of physicists from around the world. This is why the articles on the Higgs boson are signed by international networks of co-authors.
2. Then aggregate the lists of authors by home institutions, and consider the laboratories with a staff above the upper quartile (Table 1).
Table 1
Laboratories participating in ATLAS and CMS above the upper quartile
Country Laboratories Workforce
CMS Collaboration
USA Fermi National Accelerator Laboratory, Batavia 116
Germany Inst für Experimentelle Kernphysik, Karlsruhe 93
Germany DESY, Deutsche Elektronen-Synchrotron, Hamburg 73
Switzerland Inst Particle Physics, ETH Zürich 68
GB Imperial College, London 59
Italy INFN Sezione di Pisa, Univ Pisa 46
Russia Joint Institute for Nuclear Research, Dubna 42
USA University of Florida, Gainesville 40
USA California Inst Technology, Pasadena 39
Italy INFN Sezione di Padova, Univ Padova 39
USA Univ of Wisconsin, Madison 37
USA Massachusetts Institute of Technology, Cambridge 35
France Lab Leprince-Ringuet, Polytechnique/IN 2P3, Palaiseau 35
GB Rutheford Appleton Laboratory, Didcot 34
ATLAS Collaboration
Germany Physikalisches Institut, Univ Bonn 58
USA RHIC, Physics Dpt, Brookhaven Nat Laboratories 52
Netherlands Nikhef Nat Institute Subatomic Physics, Univ Amsterdam 50
Germany Fak Physik, Albert-Ludwigs-Universität, Freiburg 48
Germany DESY, Deutsche Elektronen-Synchrotron, Hamburg 44
Germany Max-Planck-Institut für Physik, München 42
France LAL, CNRS/IN 2P3, Univ Paris-Sud, Orsay 42
USA Physics Division, Berkeley Nat Laboratories 40
Russia Joint Institute for Nuclear Research, Dubna 40
France DSM/IRFU, CEA Saclay, Gif-sur-Yvette 38
Spain Instituto de Física Corpuscular, Barcelona 38
GB Dept of Physics, Oxford University, Oxford 35
Italy INFN Sezione di Roma I, Univ La Sapienza, Roma 35
Germany Fachbereich C Physik, Bergische Univ, Wuppertal 35
Germany Institut für Physik, Universität Mainz 34
China Inst of High Energy Physics, Acad Sci, Beijing 33
Germany II Physikalisches Inst, Georg-August-Univ, Göttingen 33
Germany Fak Physik, Ludwig-Maximilians-Univ, München 32
Italia INFN Sezione di Bologna, Univ Bologna 31
Fourteen laboratories provide more than a quarter of the total number of researchers of the CMS collaboration. The CMS collaboration is dominated by US laboratories, both by the one who is the leader (the Fermilab, 116 authors) and by the number of laboratories above the upper quartile (5 in 14). The ATLAS collaboration is dominated by German laboratories, both by the one who is the leader (the Physikalisches Institute in Bonn, 58 authors) and by the number of laboratories above the upper quartile (8 in 19). Among them, six laboratories were involved in the two parallel experiments on the Higgs boson.
The ranking of laboratories may be extended beyond the bottom quartile. Then appear all the laboratories that provide only a few researchers on the Higgs boson, such as the Yerevan Physics Institute (1) or the Institute of Single Crystals of Kharkov (1).
Now let us plot the frequency P(k) of a laboratory with rank k against its rank k, that is, its size expressed in number of researchers, for CMS, ATLAS and ATLAS+CMS taken together. The data exhibit a very clear pattern: whatever the series, the j-axis frequency P(k) correlates with the x-axis rank k of the laboratory (Figure 1abc).
Fig. 1. Laboratory distribution by rank: (a) CMS (b) ATLAS (c) ATLAS+CMS
This result stands out for its novelty. This is indeed the first time that authors' data are numerous enough to reveal a link between the frequency and the rank of the laboratory. As a statistical distribution can be studied only if the sample is large, this observation results from growth in the number of authors in particle physics.
Given the propensity to interpret similar distributions as power laws in recent years, a protocol to compare different hypotheses to fit a probability distribution to a given empirical distribution has beenproposed(Clauset,ShaliziandNewman,2009).Theseare:theactualpowerlawP(:r) <x x power law with exponential cutoff P(x) cx x " e Xj\ exponential law P(x) <x e Xx, stretched ex-
(lnx — /Li)2
ponential law in (lie form Pix) x x'3 1 eAr , and lognormal law P(x) <x - exp
X
2a2
In the case at hand, we conducted tests using power-law-test in R (Shalizi, 2007). These tests reject the simple power law model with significant ^-values, ranging from 10-13 to 10-80. Such data, which deviate from the simple power law, are akin to power law with exponential cutoff. However as the tests do no apply for alternative hypotheses (e. g., the exponential vs. log-normal distribution), power law with exponential cutoff cannot be assert with certainty.1
3. Size, Population and Wealth
To find out if the countries participate in the ATLAS and CMS experiments at their own level of development, the number of the authors of the Physics Letters B twin articles have been compared to various socioeconomic variables. Once the countries are coded according to ISO 3166-1 (alpha-3), the numbers provided to the ATLAS and CMS experiments may be compared to the following variables: national population, wealth of the nation, Gini index, PISA results on scales science and mathematics, number of physics prizes obtained, and costs of maintenance and operation of the detectors broken down by country. Bivariate graphs and correlation coefficients were worked out. Similarities between the countries may be specified through a (multivariate) principal component analysis (PCA). We have used res.pca in R (Figures 2ab; for details see Appendix 1: Methodology Note).
Fig. 2. PCA: (a) variables factor map, (b) countries factor map
1 In any case, this result is consistent with the rule of thumb that an empirical distribution is a power law only if the data are aligned with three orders of magnitude, which is not the case here.
The correlation circle graph (Figure 2a) helps interpreting Figure 2b. The lower left sector includes unequal countries that little participate in experiments (e. g. Peru). The upper right sector hosts countries with high scientific performances (e. g. Korea, Finland). Around the mean one discerns a group of rich countries that earn numerous prizes and provide significant numbers to ATLAS and CMS.
This can be clarified by a hierarchical ascending classification (HCA) by using res.hcpc in R. The dendrogram, cut at the threshold a = 0.3, produces seven classes:
Class 1. Rich and scientific countries, participating at the highest level in the ATLAS and CMS experiments comprise a singleton: the USA.
Class 2. Germany, Russia, Italy, Great Britain and France make up a group of rich countries heavily involved in high-energy physics.
Class 3. India and China are actively involved in experiments, at a level that is a little below that which would correspond to their population or wealth.
Class 4. This class includes the rich and scientific countries, such as Japan, Korea, Finland and Australia, which, however, have a lower contribution to high-energy physics.
Class 5. Next come countries, such as Spain, which moderately contribute to physics experiments, while having a scientific and economic level higher than the average.
Class 6. This class includes the countries about average.
Class 7. Finally come the poor or developing countries, such as Peru or Mexico, which contribute little or not at all to the ATLAS and CMS experiments.
Decorrelation of certain variables on the projection circle is very surprising (Figure 2a). The national workforce (EFF) provided to the experiments is unrelated to the Gini index (GINI) or PISA results concerning science and mathematics (SCI, MATH), and the national workforce depends little on the country's population (POP). By contrast, the other variables are highly correlated (Table 2):
Table 2
Correlation of variables
R Variables Interpretation
+0.992 EFF-FUND Workforce highly correlates with M&O expenditures
+0.976 PRIZE-FUND Countries with physics prizes spend more on M&O
+0.967 SCI—MATH Science and mathematics performances are highly correlated
+0.960 EFF-PRIZE Workforce correlates with the number of physics prizes
+0.844 PIB-PRIZE Rich countries get more physics prizes
+0.844 PIB-FUND Rich countries spend more on M&O
+0.840 EFF-PIB Workforce correlates to the wealth of the country
The relationship between the number of physics prizes and the workforce provided to the experiments (+0.960), or between the mathematics and science performances (+0.967) are self-explanatory. Surprisingly, the closest bond is between the quota of physicists provided to ATLAS and CMS and country's M&O expenditures (+0.992). How is this to be interpreted? The Memoranda of Understanding (ATLAS, 2002; CMS, 2002) make it clear that the number of authors interact with the sharing rule of maintenance and operating costs for the central components of the detector (Category A):
"The costs are to be shared amongst the Funding Agencies or Institutes in proportion to the number of their scientific staff holding PhD or equivalent qualifications who are entitled to be named as authors of scientific publications of the Collaboration" (ATLAS, 2002: 6).
Maintenance and operation expenditures are the missing link between the national GDP and number of authors who sign the articles. The more a country is wealthy, the more it can afford costs of maintenance and operation in ATLAS and CMS, and the more it is entitled to provide researchers to the collaborations that sign the articles. This is a payoff in kind, a rule that always operates for the benefit of the wealthiest contributor — i. e., in the case at hand, the USA.
4. Organizational Differences between the Collaborations
The twin articles on the Higgs boson mention the names of all the authors. The structure of the collaborations being known, their structural differences may be specified by network analysis. The network data were worked out using iGraph in R. Closeness and be-tweenness centralities appear to be the most discriminating indices.
Closeness centrality (Wasserman and Faust, 1994: 183-187) is the ability of a researcher to access other network researchers through the minimum number of steps. Mathematically, closeness centrality is the sum of all geodesic distances from i to other vertices — where the geodesic distance is the minimum number of edges to be scanned between vertices i and j. As the maximum value is (g — 1), the normalized index is written as:
Cc, = Jt?g~ ^ 0 < Cci 1
The vertex of the graph linked to the other vertices by the shorter chain has the higher closeness centrality index.
Betweenness centrality (Wasserman and Faust, 1994: 188-191) expresses the ability of a researcher to intercept knowledge or resources flowing between the other researchers of the collaborative network. Mathematically, it is the probability bjk(i) of some vertex / to stand on the geodesic between j and k. As the betweenness centrality maximum is (g - 1) (g -2), the standardized index is:
E un^d) 0 ^ (<?-!)(<?-2)
Below, the results are shown using the same LGL visualization.2 Values above the bottom decile (N = 300) have been coloured to highlight the most salient differences (Figures 3-4).
ATLAS. The institutions with the highest values are the nations sitting at the Collaboration Board and large laboratories (closeness and betweenness). Individuals with the highest score are the members of the Brookhaven RHIC (closeness), CERN in Geneva and JINR in Dubna (betweenness).
2 In both cases, we used the same LGL visualization to default: 150 iterations, 3000 changes by vertex in one iteration, cooling coefficient 1.5 (Csardi et al 2014, 201). As closeness and betweenness results are very similar, we only show the graphs representing the closeness centrality index.
mean closeness = 0.214
Fig. 3. Graph of the ATLAS collaboration network (closeness)
mean closeness = 0.215 Fig. 4. Graph of the CMS collaboration network (closeness)
CMS. Institutions with the highest values are the "regions" serving on the Collaboration Board and large laboratories (closeness and betweenness). Individuals with the highest score are the members of the Fermilab in Batavia (closeness), HEPHY in Wien and HEP in Minsk (betweenness).
The fact that the same members have high closeness and betweenness scores indicates the key role they play in the collaborations. Visual inspection of the graphs shows organizational differences, which may be highlighted by representing only the core set of the two networks (Figures 5-6).
Fig. 5. The core set of ATLAS collaboration (closeness)
Fig. 6. The core set of CMS collaboration (closeness)
Due to the absence of the "regions" middle level in ATLAS, CMS organization appears more fragmented. This may have implications for research: CMS laboratories could enjoy greater autonomy; communication could be slower between the center and periphery; there could be a loss and distortion of information in exchanges. By contrast, ATLAS appears more unitary. Belonging to a common project could override local cultures; decisions could be more in the dialogue; communication between center and periphery could be faster; and there could be less loss of information. Further research should test these assumptions.
Obviously there is no indication of such organizational differences in the 3,000-author lists of each collaboration.
5. Cooperation, Competition and Strategic Edge
The line drawn between ATLAS and CMS experiments is a deliberate will, which stems from the observation that a result is more reliable when established by researchers who use different measuring devices and work separately.3 Apart from methodological meetings (ATLAS, CMS, LHC Higgs Combination Group, 2011), the two experiments run separately. However, the way how the scientific work is progressing depends on the ability to anticipate the substance of future research. The way in which physicists are distributed in one or the other experiment provides further information about this facet of the scientific work. Laboratories involved in both experiments have a strategic edge over those who participate in one experiment. In straddling labs, researchers have direct information on the progress made by other collaboration. They may better anticipate the results and generate new hypotheses. Physicists have a more complete (by accessing data from both experiments) and immediate (by accessing data before publication) view on ongoing research. This causes an asjmmetrj of information between the laboratories invested in one or both collaborations.
Differential involvement of the nations can be studied by the staff of researchers provided to the one and the other experiment. These preferences can be presented on a graph, whose abscissae represent the share of the workforce provided to CMS (-1) and ATLAS (+1) and whose ordinates represent the number of researchers supplied by the country (Figure 7).
Some countries, such as Belgium, India and Korea only provide researchers to CMS (-1), others, such as Canada, Japan and the Netherlands, exclusively participate in ATLAS (+1). Between these extremes, there are 22 countries, some of which provide important staffs to the experiments. The biggest differences in numbers are caused by the USA (-365), very active in CMS; Great Britain (161) and Germany (140), most involved in ATLAS. In a word, CMS has an American color that contrasts with the European tone of ATLAS.
The propensity to select only one collaboration (less communication and thus competition between teams) vs. to distribute researchers between the collaborations (more communication and cooperation between teams) varies greatly according to nationality. If the measurement is limited to the 50 largest countries — because of low statistical significance of small countries choice — there is a sharp contrast between cooperative and competitive nations (Table 3).
3 Edward W. Morley was the first scientist to consciously use this strategy in the determination of the atomic weight of oxygen (Morley, 1895). The concept of robustness was taken up again later (Wimsatt, 1981).
OUSA
O DEU
OITA
OGB R
„ OFRA ORUS OCHE O ESP 3BEL 3IND rOTUR °chn0POL )KOR 0%N OGRC OAUT BRA OPR 5 FIN JhSN OBLR ,HRV OGE0 OSRB list OCOL HEGY 3CYP «CAN WJPN
C : C „CZE \Jr*\ Ml n
: c J SWE ISR
MEXc _PAKF NORr } AUS ROU DNK -,SVK MAR ? SVN
: ESTE ( 3arg ^ CHL 3ZAF
O UKR OARM
OLTU OTHA c , , , , 5AZE 1
-1 -0.5 0 0.5 1
CMS ATLAS
Fig. 7. Preference for the collaborations (x: ATLAS/CMS; y: workforce)
Table 3
Cooperation vs. competition between ATLAS and CMS collaborations
Countries CMS (a) ATLAS (b) CMS n ATLAS (c) (a + b) (c / a + b)
Russia 149 117 219 266 0,823
China 40 33 49 73 0,671
Italy 291 223 293 514 0,570
Czech Rep. 2 64 19 66 0,288
Turkey 44 17 17 61 0,279
Greece 21 31 13 52 0,250
France 117 201 69 318 0,217
Germany 275 415 119 690 0,172
GB 131 292 56 423 0,132
USA 958 593 180 1551 0,116
Spain 63 78 12 141 0,085
Others — — — — 0,000
Among the largest countries, Russia, China and Italy are characterized by their propensity to adopt the cooperative model. By providing people to both collaborations, these countries are better informed about the technical differences between the two experiments. But on the other hand, as their workforces are shared between the two collaborations, they cannot have significant influence on any of the collaborations.
The intensity of ATLAS-CMS exchanges within the same laboratory can be expressed by the number of constructible relationships among physicists. Let n be the number of physicists dedicated to CMS and let m be the number of physicists dedicated to ATLAS in a laboratory. The number of constructible relationships is then R = n • m (Table 4).
Table 4
ATLAS-CMS constructible relationships within the laboratories
Laboratory n • m R
DESY, Deutsches Elektronen-Synchrotron, Hamburg (Germany) 75 • 44 3 300
3rd quartile
Joint Institute for Nuclear Research, Dubna (Russia) 42 • 40 1 680
University of Wisconsin, Madison (USA) 37 • 28 1 036
DSM/IRFU, CEA/Saclay, Gif-sur-Yvette (France) 26 • 38 988
2nd quartile
INFN Sezione di Bologna, Universita di Bologna (Italy) 28 • 31 868
Rutherford Appleton Laboratory, Didcot (GB) 34 • 22 748
INFN Sezione di Milano, Universita di Milano (Italy) 24 • 31 744
INFN Sezione di Roma, Universita La Sapienza (Italy) 19 • 37 703
1st quartile
Institute of High Energy Physics, Beijing (China) 24 • 25 600
INFN Sezione di Pisa, Universita di Pisa (Italy) 46 •ÎO 460
Institute for High Energy Physics, Protvino (Russia) 20 • 18 360
Lab Instrumentado Física Exp de Partículas, Lisboa (Portugal) 12 • 27 324
INFN Sezione di Napoli, Universita Federico II (Italy) 11 • 20 220
The Ohio State University, Columbus (USA) 13 • 12 156
Moscow State University, Moscow (Russia) 25 • 6 150
The University of Iowa, Iowa City (USA) 27 • 5 135
Petersburg Nuclear Physics Institute, Gatchina (Russia) 13 • 9 117
Boston University, Boston (USA) 14 • 8 112
Institute for Theoretical and Experimental Physics, Moscow (Russia) 22 • 5 110
P. N. Lebedev Institute of Physics, Moscow (Russia) 8 • 11 88
INFN Laboratori Nazionali di Frascati, Frascati (Italy) 5 • 14 70
Vinca Institute of Nuclear Sciences, University of Belgrade (Serbia) 7 • 9 63
Bogazici University, Istanbul (Turkey) 7 • 9 63
INFN Sezione di Genova, Universita di Genova (Italy) 5 • 12 60
Institute of High Energy Physics, Tbilisi State University (Georgia) 7 • 6 42
Massachusetts Institute of Technology, Cambridge (USA) 35 • 1 35
Charles University, Prague (Czech Rep.) 2 • 17 34
University of Athens, Athens (Greece) 3 • 10 30
Universidad Autónoma de Madrid (Spain) 3 • 9 27
E. Andronikashvili Institute of Physics, Acad Sci, Tbilisi (Georgia) 2 • 6 12
Nat Centre for High Energy Physics, Minsk (Belarus) 9 • 1 9
Istanbul Technical University, Istanbul (Turkey) 1 • 9 9
Centre de calcul IN 2P3, CNRS/IN 2P3, Villeurbanne (France) 2 • 3 6
Yerevan Physics Institute, Yerevan (Armenia) 4 • 1 4
Some caution is needed in interpreting Table 4. R estimates the constructible relationships in a laboratory where researchers work for both collaborations. This is an oversimplification: relationships are counted on complete graphs, when laboratories are not; and constructible relationships do not necessarily result in actual relations. The schema however remains informative, because the probability of a scientific relation depends on the number of constructible relationships. In a word: attention should be paid to the ordinal ranking that R provides, rather than to the value of this number.
This ranking highlights the DESY as the most conducive laboratory to the development of scientific relationships, since it alone holds one quarter of all the ATLAS-CMS exchanges. The addition of three other laboratories — the JINR of Dubna, the Physics Department of the University of Madison, and the DSM/IRFU of Gif-sur-Yvette — provide together half of all the internal relationships. ATLAS-CMS exchanges are concentrated in these centers, more than in any other place.
Conclusion
This article was intended to look into the structure of the ATLAS and CMS collaborations through the personal data of the twin articles on the Higgs boson released in Physics Letters B in September 2012. After introducing the subject (Section 1), we have investigated the major disparities between countries and laboratories, which are hidden by the "mass effect" imposed by the authors lists (Section 2). The laboratory distribution meets a rank-frequency law (most probably, a power law with exponential cutoff). This is the first time this property is described because a statistical law is detected only if the sample is large — a condition just fulfilled in particle physics. These data were compared with the socioeconomic variables of the countries providing physicists to the experiments (Section 3). Much to our surprise, the high correlation (+0.992) between the number of authors and the country's wealth is due to a rule for sharing the costs of maintenance and operation of the detectors. Other properties were studied, such as the difference between the center and periphery of the collaborative networks, and inner organization of the twin collaborations (Section 4). The propensity to adopt a competitive vs. cooperative model greatly varies across countries. This preference creates a contrast between Canada-Japan-India vs. Russia-Italy-China. However it appears that information asymmetry, created by the fact of having information on one or both collaborations, is beneficial at the scale of laboratories only (Section 5). Although invisible at first glance, all these properties are included in the lists of authors.
The operation of ATLAS and CMS collaborations help us understanding the new ways of thinking scientific work in some fields of science, primarily in particle physics, secondarily in biomedical research with the genome sequencing or double blinded random trials. All these research fields involve increasingly large human collaborations. That is why the ATLAS and CMS collaborations are forerunners of the way many fields of scientific research will operate in tomorrow's world. The increase in the number of authors leads to qualitatively new emerging phenomena, such as the growing organization of research teams and new signing practices, to which sociology of science should devote attention.
Appendix 1. Methodology Note
Apart from the two twin articles on the Higgs boson, which contain the key data (i. e., the two lists of co-authors), we have used framework data, which are mostly hosted on the CERN Document Server and TWiki:
♦ CERN Document Server: cds.cern.ch
♦ CERN Personnel Statistics: cds.cern.ch/collection/CERN%20Annual%20Person-nel%20Statistics
♦ LHC Experiments: cds.cern.ch/collection/LHC%20Experiments
♦ ATLAS Higgs Results: twiki.cern.ch/twiki/bin/view/AtlasPublic/HiggsPublicResults
♦ CMS Higgs Results: twiki.cern.ch/twiki/bin/view/CMSPublic/PhysicsResultsHIG Relevant information is also echoed on organizations and laboratories pages, e. g.:
♦ National Science Foundation: www.noao.edu/nsf
♦ JINR Physics: www.atlas-jinr.ru
The data set that was used for the Principal Component Analysis is as follows: Population. — Data from the Population Reference Bureau: www.prb.org/pdf12/2012-population-data-sheet_eng.pdf. Wealth of the Nations. — Data from the World Bank: data.worldbank.org/indicator/NY.GDP.MKTP.CD.
To find out whether a country contributes to a research program in relation to its wealth, the nominal GDP should be used rather than the domestic expenditures on R&D. Gini Index. — Data from the World Bank: data.worldbank.org/indicator/SI.P OV.GINI.
Mathematics and science are positively correlated with each other (+0.967) and together negatively correlated with the Gini index (-0.702 science; -0.737, mathematics). PISA Results. — Data from OECD:
www.oecd.org/edu/school/programmeforinternationalstudentassessmentpisa/ 33690591.pdf.
As PISA tests are carried out on 15 year old students, we have selected 2000 data, that correspond to a mean age of 15 + 12 = 27 year old researchers, at the date when the CMS and ATLAS twin articles were released. We have used data mathematical literacy (math) and scientific literacy (sci).
Physics Prizes. — No synthetic data exist on the subject. We selected the twenty most known physics prizes, recording, if any, the nationality of the institution awarding the prize. All prize winners in the last twenty years (1983-2012) were counted for the country where they exercised at the time when it was received. National prizes rewarding only citizens of the country have been excluded. The 12 remaining "international prizes" are the Boltzmann Medal, Elliott Cresson Medal, Dirac Medal, Fundamental Physics Prize, Harvey Prize, Lo-rentz Medal, Majorana Prize, Matteucci Medal, Albert A. Michelson Award, Max Planck Medal, Nobel Prize in Physics, and Wolf Prize in Physics. Finally the percentage of prizes received by country was calculated. The results are as follows: USA 154, Germany 33, Italy 27, France 25, Britain 20, Switzerland 13, Russia 18, Japan 7, the Netherlands 6, Canada 4, Austria 4, Israel 4, India 2, Belgium 2, Hong Kong 1, Palestine 1, Australia 1, and Ukraine 1. Costs of Maintenance and Operation. — CERN Data:
cds.cern.ch/record/1537416 CERN-RRB-2013-053: 2012 ATLAS M&O budgets. cds.cern.ch/record/1336102 CERN-RRB-2011-013: 2012 CMS M&O budgets.
Appendix 2. Members of the CMS Collaboration in 20124
Countries Labs CMS Laboratories Workforce %
39 N=167 All laboratories 2898 100,00
USA N=49 Total USA 958 33,06
1 Fermi Nat Accelerator Lab, Batavia 116
2 University of Florida, Gainesville 40
3 California Inst Technology, Pasadena 39
4 Univ of Wisconsin, Madison 37
5Massachusetts Institute of Technology, Cambridge 35
6 University of California, Davis 32
7 Purdue University, West Lafayette 32
8 University of California, San Diego, La Jolla 30
9 University of Illinois at Chicago, Chicago 27
10 The University of Iowa, Iowa City 27
11 Princeton University, Princeton 27
12 University of California, Santa Barbara 26
13 Cornell University, Ithaca 25
14 University of Notre Dame, Notre Dame 24
15 Rutgers, the State University of New Jersey, Piscataway 24
16 University of Maryland, College Park 23
17 University of Rochester, Rochester 22
18 University of California, Los Angeles 21
19 Brown University, Providence 20
20 University of Minnesota, Minneapolis 18
21 Texas A&M University, College Station 18
22 Florida State University, Tallahassee 17
23 University of Nebraska-Lincoln, Lincoln 16
24 Rice University, Houston 16
25 Vanderbilt University, Nashville 16
26 Boston University, Boston 14
27 University of Virginia, Charlottesville 14
28 Carnegie Mellon University, Pittsburgh 13
29 University of Colorado at Boulder, Boulder 13
30 The University of Kansas, Lawrence 13
31 Northeastern University, Boston 13
32 Northwestern University, Evanston 13
33 The Ohio State University, Columbus 13
34 Johns Hopkins University, Baltimore 11
35 Texas Tech University, Lubbock 11
36 State University of New York, Buffalo 9
"Workforce in decreasing numbers. Laboratories taking part in both ATLAS and CMS experiments are written in italics. Since CERN is an international institution, it is not included in Switzerland. The secondary affiliations of the researchers have not been taken into account. Laboratories were members represent half the national contingent are coded (eg 1 Fermilab) in order to identify them on the graphs (Figs. 3-6).
37 Kansas State University, Manhattan 9
38 University of Puerto Rico, Mayaguez 9
39 Wayne State University, Detroit 8
40 Florida International University, Miami 7
41 Florida Institute of Technology, Melbourne 7
42 University of Mississippi, Oxford 7
43 The Rockefeller University, New York 7
44 University of Tennessee, Knoxville 6
45 Lawrence Livermore National Laboratory, Livermore 4
46 Baylor University, Waco 4
47 The University of Alabama, Tuscaloosa 3
48 Purdue University Calumet, Hammond 2
49 Fairfield University, Fairfield 1
ITA 13 Total Italy 291 10,04
50 INFN Sezione di Pisa, Univ di Pisa 46
51 INFN Sezione di Padova, Univ di Padova 39
52 INFN Sezione di Bari, Univ di Bari 32
53 INFN Sezione di Bologna, Univ di Bologna 28
54 INFN Sezione di Milano, Univ di Milano 24
55 INFN Sezione di Torino, Univ di Torino 23
56 INFN Sezione di Roma I, Univ La Sapienza, Roma 19
57 INFN Sezione di Perugia, Univ di Perugia 17
58 INFN Sezione di Firenze, Univ di Firenze 15
59 INFN Sezione di Trieste, Univ di Trieste 12
60 INFN Sezione di Napoli, Univ Federico II, Napoli 11
61 INFN Sezione di Catania, Univ di Catania 9
62 INFN Sezione di Pavia, Univ di Pavia 6
63 INFNLaboratori Nazionali di Frascati 5
64 INFN Sezione di Genova, Univ di Genova 5
DEU 6 Total Germany 275 9,49
65 Inst für Experimentelle Kernphysik, Karlsruhe 93
66 DESY, Deutsche Elektronen-Synchrotron, Hamburg 73
67 University of Hamburg, Hamburg, Germany 32
68 RWTH Aachen Univ, III. Physikalisches Inst A, Aachen 30
69 RWTH Aachen Univ, I. Physikalisches Inst, Aachen 29
70 RWTH Aachen Univ, III. Physikalisches Inst B, Aachen 18
RUS 7 Total Russian Federation 149 5,14
71 Joint Institutefor Nuclear Research, Dubna 42
72 Moscow State University, Moscow 25
73 Inst for Theor and Exp Physics, Moscow 22
74 Institute for High Energy Physics, Protvino 20
75 Institute for Nuclear Research, Moscow 19
76 Petersburg Nuclear Physics Institute, Gatchina 13
77P. N. Lebedev Institute of Physics, Moscow 8
CMS 78 CERN-CMS, Geneva 143 4,93
GBR 5 Total Great Britain 131 4,52
79 Imperial College, London 59
80 Rutheford Appleton Laboratory, Didcot 34
81 University of Bristol, Bristol 20
82 Brunel University, Uxbridge 12
83 CCCS, University of the West of England, Bristol 6
FRA 5 Total France 117 4,04
84 Lab Leprince-Ringuet, Polytechnique/IN 2P3, Palaiseau 35
85 Inst Phys Nucl, Univ Lyon 1, CNRS-IN 2P3, Lyon 31
86 DSM/IRFU, CEA/Saclay, Gif-sur-Yvette 26
87 Institut Hubert Curien, CNRS/IN 2P3, Strasbourg 23
88 Centre de Calcul, CNRS/IN 2P3, Villeurbanne 2
CHE 3 Total Switzerland 96 3,31
89 Institute for Particle Physics, ETH Zurich 68
90 Paul Scherrer Institut, Villigen 16
91 Universität Zürich, Zürich 12
BEL 6 Total Belgium 93 3,21
92 Univ Catholique de Louvain, Louvain-la-Neuve 22
93 Universiteit Antwerpen, Antwerpen 18
94 Universiteit Ghent, Ghent 18
95 Université Libre de Bruxelles, Bruxelles 16
96 Vrije Universiteit Brussel, Brussel 15
97 Université de Mons, Mons 4
IND 6 Total India 71 2,45
98 Tata Institute of Fundamental Research, EHEP, Mumbai 15
99 Bhabha Atomic Research Centre, Mumbai 12
100 Panjab University, Chandigarh 12
101 University of Delhi, Delhi 12
102 Tata Institute of Fundamental Research, HECR, Mumbai 11
103 Saha Institute of Nuclear Physics, Kolkata 9
ESP 4 Total Spain 63 2,17
104 CIEMAT, Madrid 29
105 Inst de Física, CSIC, Univ de Cantabria, Santander 24
106 Universidad de Oviedo, Oviedo 7
107 Universidad Autónoma de Madrid, Madrid 3
KOR 6 Total Korea 54 1,86
108 Kyongpook National University, Daegu 17
109 Korea University, Seoul 13
110 Sungkyunkwan University, Suwon 10
111 University of Seoul, Seoul 9
112 Chonnam Univ, Inst for Elementary Particles, Kwangju 3
113 Kangwon National University, Chunchon 2
TUR 4 Total Turkey 44 1,52
114 Cukurova University, Adana 24
115 Middle East Technical University, Physics Dept, Ankara 12
116 Bogazici University, Istanbul 7
117Istanbul Technical University, Istanbul 1
TWN 2 Total Taiwan 43 1,48
118 National Taiwan University, Taipei 28
119 National Central University, Chung-Li 15
CHN 2 Total China 40 1,38
120 Inst of High Energy Physics, Beijing 24
121 State Key Lab of Nucl Phys, University of Beijing 16
BRA 3 Total Brazil 35 1,21
122 Univ do Estado do Rio de Janeiro, Rio de Janeiro 17
123 Inst Física Teórica, Univ Estadual Paulista, Sao Paulo 12
124 Centro Brasileiro de Pesquisas Físicas, Rio de Janeiro 6
AUT 1 Total Austria 32 1,10
125 Institüt für Hochenenergiephysik der OeAW, Wien 32
POL 3 Total Poland 29 1,00
126 Inst of Experimental Physics, University of Warsaw 14
127 National Centre for Nuclear Research, Swierk 13
128 Inst Electronic Systems, Warsaw Univ of Technology 2
FIN 3 Total Finland 28 0,97
129 Helsinki Institute of Physics, Helsinki 21
130 Lappeenranta University of Technology, Lappeenranta 4
131 Department of Physics, University of Helsinki, Helsinki 3
GRC 3 Total Greece 21 0,72
132 Inst Nuclear Physics Demokritos, Aghia Paraskevi 11
133 University of Ioánnina, Ioánnina 7
134 University of Athens, Athens 3
BGR 3 Total Bulgaria 20 0,69
135 Inst Nuclear Research and Nuclear Energy, Sofia 11
136 University of Sofia, Sofia 6
137 Inst System Engineering and Robotics, Sofia 3
HUN 3 Total Hungary 19 0,65
138 KFKI Inst for Particle and Nuclear Physics, Budapest 8
139 Institute of Nuclear Research ATOMKI, Debrecen 6
140 University of Debrecen, Debrecen 5
MEX 4 Total Mexico 13 0,45
141 Centro Invest y Estudios Avanzados del IPN, México 7
142 Universidad Autónoma de San Luis Potosí, San Luis Potosí 3
143 Universidad Iberoamericana, México 2
144 Benemerita Universidad Autonoma de Puebla, Puebla 1
HRV 3 Total Croatia 13 0,45
145 Institute Rudjer Boskovic, Zagreb 6
146 Technical University of Split, Split 5
147 University of Split, Split 2
BLR 2 Total Belarus 13 0,45
148 Nat Centre for Particle High Energy Physics, Minsk 9
149 Research Institute for Nuclear Problems, Minsk 4
PRT 1 Total Portugal 12 0,41
150 Lab Instrumentagäo e Física de Partículas, Lisboa 12
IRN 1 Total Iran 12 0,41
151 Inst Research in Fundamental Sciences (IPM), Tehran 12
NZL 2 Total New Zeland 11 0,38
152 University of Canterbury, Christchurch 9
153 University of Auckland, Auckland 2
PAK 1 Total Pakistan 11 0,38
154 Nat Centre Physics, Quaid-I-Azam University, Islamabad 11
GEO 2 Total Georgia 9 0,31
155E. Andronikashvili Inst Physics, Acad Sciences, Tbilisi 2
156 Inst High Energy Physics, Tbilisi State University 7
EGY 1 Total Egypt 9 0,31
157 Academy of Sciences and Technology of Egypt, Cairo 9
EST 1 Total Estonia 8 0,28
158 Nat Inst Chemical Physics and Biophysics, Tallinn 8
SRB 1 Total Serbia 7 0,24
159 Vinca Inst of Nuclear Sciences, University of Belgrade 7
CYP 1 Total Cyprus 7 0,24
160 University of Cyprus, Nicosia 7
UKR 2 Total Ukraine 5 0,17
161 Kharkov Institute of Physics and Technology, Kharkov 4
162 Inst Single Crystals of Nat Academy of Science, Kharkov 1
COL 1 Total Colombia 5 0,17
163 Universidad de Los Andes, Bogotá 5
ARM 1 Total Armenia 4 0,14
164 Yerevan Physics Institute, Yerevan 4
LTU 1 Total Lithuania 3 0,10
165 Vilnius University, Vilnius 3
CZE 1 Total Czech Republic 2 €
166 Charles University, Prague 2
THA 1 Total Thailand 2 €
167 Chulalongkorn University, Bangkok 2
Appendix 3. Members of the ATLAS Collaboration in 20125
Countries Labs ATLAS Laboratories Workforce %
38 N=179 All laboratories 2932 100,00
USA N=40 Total USA 593 20,22
1 RHIC, Physics Dept, Brookhaven Lab, Upton 52
5 Same procedure as for CMS (see note 3).
2 Physics Div, Berkeley Nat Laboratory, Berkeley 40
3 Dept of Physics, Univ of Wisconsin, Madison 28
4 SLAC National Accelerator Laboratory, Stanford 28
5 Dept of Physics, University of Michigan, Ann Arbor 28
6 Dept of Physics, Univ of Pennsylvania, Philadelphia 25
7 Enrico Fermi Institute, University of Chicago 24
8 Nevis Laboratory, Columbia University, Irvington 24
9 Dept of Physics, Michigan State Univ, East Lansing 21
10 High Energy Physics Div, Argonne Nat Lab, Argonne 20
11 Lab for Particle Physics, Harvard U, Cambridge 19
12 Dept of Physics & Astronomy, Univ of Stony Brook 19
13 Dept of Physics, University of Texas, Arlington 18
14 Dept of Physics & Astronomy, Univ California, Irvine 17
15 Dept of Physics, Yale University, New Haven 16
16 Dept of Physics, New York University, New York 14
17 Dept of Physics, University of Illinois, Urbana 14
18 Ohio State University, Columbus 12
19 Dept of Physics, Indiana University, Bloomington 12
20 Inst for Particle Physics, Univ California, Santa Cruz 12
21 Dept of Physics, University of Arizona, Tucson 11
22 Physics Dept, Southern Methodist Univ, Dallas 11
23 H. L. Dodge Dept of Physics, Univ Oklahoma, Norman 11
24 Dept of Physics, Iowa State Univ, Ames 10
25 Dept of Physics, University of Massachusetts, Amherst 10
26 Dept of Physics, University of Washington, Seattle 10
27 Center for High Energy Physics, Univ Oregon, Eugene 10
28 Dept of Physics, Brandeis University, Waltham 9
29 Dept of Physics, Duke University, Durham 9
30Dept of Physics, Boston University, Boston 8
31 Dept of Physics & Astronomy, Univ Pittsburgh 8
32 Dept of Physics & Astronomy, Tufts Univ, Medford 8
33 Dept of Physics, Northern Illinois University, DeKalb 7
34 University of Iowa, Iowa City 5
35 Physics Dept, University of Texas at Dallas, Richardson 5
36 Dept of Physics, Hampton University, Hampton 5
37 Dept of Physics, Univ New Mexico, Albuquerque 5
38 Dept of Physics, Oklahoma State Univ, Stillwater 4
39 Physics Dept, SUNY Albany, Albany 3
40 Dept of Physics, MIT, Cambridge 1
DEU 15 Total Germany 415 14,15
41 Physikalisches Institut, University of Bonn 58
42 Fak Physik, Albert-Ludwigs-Universität, Freiburg 48
43 DESY, Deutsche Elektronen-Synchrotron, Hamburg 44
44 Max-Planck-Institut für Physik, München 42
45 Fachb C Physik, Bergische Universität, Wuppertal 35
46 Institut für Physik, Universität Mainz 34
47 II Physikalisches Inst, Georg-August-Univ, Göttingen 33
48 Fak Physik, Ludwig-Maximilians-Univ, München 32
49 Kirchhoff-Institut, Ruprecht-Karls-Univ, Heidelberg 29
50 Institut für Kernphysik, Technical University, Dresden 17
51 Dept of Physics, Humboldt University, Berlin 16
52 Fachb Physik, Universität Siegen, Siegen 12
53 Inst Exp Physik IV, Technische Universität, Dortmund 7
54 Fak Physik, Julius-Maximilians-Univ, Würzburg 6
55 II Physikalisches Inst, Justus-Liebig-Univ, Giessen 2
GBR 15 Total Great Britain 292 9,96
56 Dept of Physics, Oxford University, Oxford 35
57 SUPA, University of Glasgow 28
58 School of Physics, University of Manchester 26
59 School of Physics, University of Birmingham 25
60 Oliver Lodge Laboratory, University of Liverpool 25
61 Particle Physics Dept, Rutherford Appleton Lab, Didcot 22
62 Dept of Physics, University College London 22
63 Cavendish Laboratory, University of Cambridge 20
64 Dept of Physics, Holloway Univ London, Surrey 18
65 Physics Dept, Lancaster University 17
66 Dept of Physics, University of Sheffield 17
67 School of Physics, Queen Mary Univ, London 16
68 SUPA, University of Edinburgh 12
69 Dept of Physics, Univ of Sussex, Brighton 7
70 Dept of Physics, Univ of Warwick, Coventry 2
ITA 13 Total Italy 223 7,69
71INFN Sezione di Roma I, Univ La Sapienza, Roma 35
72 INFN Sezione di Bologna, Univ Bologna 31
73 INFN Sezione di Milano, Univ Milano 31
74 INFN Sezione di Napoli, Univ Napoli 20
75 INFN Sezione di Roma Tre, Univ Roma Tre 15
76 INFN Laboratori Nazionali di Frascati, Frascati 14
77INFN Sezione di Pavia, Univ Pavia 13
78 INFN Sezione di Genova, Univ Genova 12
79 INFN Gruppo Collegato di Cosenza, Arcavata di Rende 12
80 INFN Sezione di Roma Tor Vergata, Univ Roma II 11
81 INFN Sezione di Pisa, Lab E. Fermi, Univ Pisa 10
82 INFN Gruppo Collegato di Udine, Univ Udine 10
83 INFN Sezione di Lecce, Univ del Salento, Lecce 9
FRA 8 Total France 201 6,85
84 LAL, CNRS/IN 2P3, Univ Paris-Sud, Orsay 42
85 DSM/IRFU, CEA Saclay, Gf-sur-Yvette 38
86 LAPP, CNRS/IN 2P3, Univ Savoie, Annecy-le-Vieux 28
87 CPPM, CNRS/IN 2P3, Univ Aix-Marseille 27
88 LPNHE, CNRS/IN 2P3, Univ Paris 6 et 7 25
89 LPSC, CNRS/IN 2P3, INPG, Univ Grenoble 1 21
90 LPC, CNRS/IN2P3, Clermont-Ferrand 17
91 Centre de Calcul de l'IN 2P3, Villeurbanne 3
ATLAS 92 CERN-ATLAS, Geneva 156 5,32
CAN 10 Total Canada 120 4,09
93 Dept of Physics, University of Toronto, Toronto 26
94 Dept of Physics & Astronomy, Univ Victoria 17
95 TRIUMF, Vancouver BC 16
96 Dept of Physics, McGill University, Montreal 15
97 Dept of Physics, Carleton University, Ottawa 12
98 Group of Particle Physics, University of Montreal 9
99 Dept of Physics, Simon Fraser University, Burnaby 9
100 Dept of Physics, University of Alberta, Edmonton 7
101 Dept of Physics, Univ British Columbia, Vancouver 7
102 Physics Dept, University of Regina 2
RUS 8 Total Russian Federation 11l 4,00
103 Joint Institute for Nuclear Research, Dubna 40
104 State Institute for High Energy Physics, Protvino 1S
105 Budker Institute of Nucl Physics, SB RAS, Novosibirsk 16
106 Moscow Engineering and Physics Institute, Moscow 12
107P. N. Lebedev Inst of Physics, Acad Sciences, Moscow 11
10S Petersburg Nuclear Physics Institute, Gatchina 9
109 Skobeltsyn Inst Nuclear Physics, Moscow State Univ б
110 Institute for Theor and Exp Physics, Moscow 5
JPN 1l Total Japan 116 3,95
111 Int Center Elementary Particle Physics, Univ Tokyo 30
112 KEK, High Energy Accelerator, Tsukuba 27
113 Graduate School of Science, Kobe Univ, Kobe 13
114 Graduate School of Science, Univ of Nagoya 8
115 Fac Pure and Applied Sciences, Univ Tsukuba 8
116 Graduate School of Science, Univ of Osaka 7
117 Dept of Physics, Tokyo Inst of Technology, Tokyo 6
118 Faculty of Science, Kyoto University, Kyoto 3
119 Dept of Physics, Kyushu University, Fukuoka 3
120 Graduate School of Science, Metropolitan Univ Tokyo 2
121 Waseda University, Tokyo 2
122 Dept of Physics, Shinshu University, Nagano 2
123 Fac Applied Inf Science, Inst Technology, Hiroshima 1
124 Kyoto University of Education, Kyoto 1
125 Nagasaki Institute of Applied Science, Nagasaki 1
126 Faculty of Science, Okayama University, Okayama 1
127 Ritsumeikan University, Kusatsu, Shiga 1
ESP 4 Total Spain l8 2,69
128 Instituto de Física Corpuscular, Barcelona 38
129 Institut de Física d'Altes Energies, Univ Aut Barcelona 29
130 Dept de Física Teorica C-15, Univ Aut Madrid 9
131 Dept de Física Teorica, CAPFE, Univ Granada 2
CZE 4 Total Czech Republic 64 2,18
132 Institute of Physics, Academy of Sciences, Prague 24
133 Czech Technical Univ, Prague 21
134 Fac Math and Physics, Charles Univ, Prague 17
135 Palacky University, RCPTM, Olomouc 2
NLD 2 Total Netherlands 60 2,05
136 Nikhef Nat Inst Subatomic Physics, Univ Amsterdam 50
137 Inst for Math, Part Physics, Radboud Univ, Nijmegen 10
SWE 4 Total Sweden 49 1,67
138 Dept of Physics, O. Klein Center, Univ Stockholm 25
139 Fysiska institutionen, Lunds Univ, Lund 12
140 Dept of Physics & Astronomy, Univ Uppsala 8
141 Physics Dept, Royal Inst of Technology, Stockholm 4
CHE 2 Total Switzerland 45 1,56
142 Section de Physique, Univ Genève 29
143 A. Einstein Center, Lab High Energy Physics, Bern 16
ISR 3 Total Israel 43 1,47
144 Dept Particle Physics, Weizmann Institute, Rehovot 19
145 R. B. Sackler School of Physics, Univ Tel Aviv 18
146 Dept of Physics, Technion Israel Inst Techn, Haifa 6
CHN 1 Total China 33 1,12
147Inst of High Energy Physics, Acad Sci, Beijing 33
POL 2 Total Poland 33 1,12
148 Niewodniczanski Inst Nucl Physics, Acad Sci, Krakow 23
149 AGH Univ Science and Technology, Krakow 10
GRC 3 Total Greece 31 1,06
150 Physics Dept, Nat Tech Univ Athens, Zografou 14
151 Physics Dept, University of Athens 10
152 Dept of Physics, Aristotle Univ, Thessaloniki 7
AUS 3 Total Australia 2l 0,92
153 School of Physics, University of Melbourne, Victoria 16
154 School of Physics, University of Sydney 9
155 School of Chemistry and Physics, Univ Adelaide 2
NOR 2 Total Norway 2l 0,92
156 Dept of Physics, University of Oslo 15
157 Dept of Physics, University of Bergen 12
PRT 1 Total Portugal 2l 0,92
15SLab Instrumentaçâo Física Exp de Partículas, Lisboa 27
ROU 1 Total Romania 22 0,75
159 Nat Inst of Physics and Nucl Engineering, Bucarest 22
DNK 1 Total Denmark 21 0,72
160 Niels Bohr Institute, University of Copenhagen 21
TUR 2 Total Turkey 1l 0,58
1б1 Dept of Physics, Bogazici University, Istanbul 9
162 Dept of Physics, Ankara University, Ankara 8
TWN 1 Total Taiwan 15 0,51
163 Institute of Physics, Academia Sinica, Taipei 15
SVK 1 Total Slovakia 14 0,48
164 Fac Math Phy Inf, Comenius Univ, Bratislava 14
MAR 1 Total Marocco 12 0,41
165 Fac Sciences, Université Hassan II, Casablanca 12
SVN 1 Total Slovenia 11 0,37
166 Dept of Physics, Jozef Stefan Inst, Univ Ljubljana 11
BRA 1 Total Brazil 11 0,37
167 COPPE, Universidade Federal do Rio de Janeiro 11
SRB 1 Total Serbia 9 0,31
168 Vinca Institute; Inst of Physics, Univ Belgrade 9
ARG 2 Total Argentina 9 0,30
169 CONICET y Inst de Física, Univ de La Plata 5
170 Dept de Física, Univ Buenos Aires 4
CHL 1 Total Chile 8 0,27
171 Dept de Física, Univ Católica de Chile, Santiago 8
BLR 2 Total Belarus 7 0,23
172 B. I. Stepanov Inst of Physics, Acad of Sciences, Minsk 6
173 Nat Centre for Particle High Energy Physics, Minsk 1
AUT 1 Total Austria 6 0,20
174 Inst für Teilchenphysik, Leopold-Franzens-U, Innsbruck 6
GEO 1 Total Georgia 6 0,20
175 E. Andronikashvili Inst Physics, State Univ, Tbilisi 6
ZAF 1 Total South Africa 6 0,20
176 Dept of Physics, University of Johannesburg 6
COL 1 Total Colombia 5 0,17
177 Centro de Investigaciones, Univ A. Narino, Bogotá 5
AZE 1 Total Azerbaijan 2 €
178 Inst of Physics, Azerb Acad of Sciences, Baku 2
ARM 1 Total Armenia 1 €
179 Yerevan Physics Institute, Yerevan 1
References
Aad G. et al. (2015) Combined Measurement ofthe Higgs Boson Mass inpp Collisions at \/s = 7 and 8 TeVwith the ATLAS and CMS Experiments, Physical Review Letters 114, 191803 (2015).
ATLAS (2002) Memorandum of Understanding [MoU] for Maintenance and Operation of the ATLAS Detector, CERN-RRB-2002-035, 28 March 2002.
ATLAS (2012) Observation of a new particle in the search for the Standard Model Higgs boson with the ATLAS detector at the LHC, Physics Letters B716, 1-29.
ATLAS, CMS, LHC Higgs Combination Group (2011) Procedure for the LHC Higgs boson search combination in Summer 2011, ATL-PHYS-PUB-2011-11, CMS N0TE-2011/005.
Boisot M., Nordberg M., Yami S. and Nicquevert B. (2011). Collisions and Collaboration. The Organisation of Learning in the ATLAS Experiment at the LHC, Oxford, Oxford University Press.
CERN (2011) ATLAS and CMS experiments present Higgs search status, http://press.web.cern. ch/press-releases, 13 December 2011.
CERN (2012) CERN experiments observe particle consistent with long-sought Higgs boson, http://press.web.cern.ch/press-releases, 4 July 2012.
CERN (2013) New results indicate that particle discovered at CERN is a Higgs boson, http:// press.web.cern.ch/press-releases, 14 March 2013.
CERN (2014) CERN Personnel Statistics 2012, CERN Human Ressource Department. https:// cds.cern.ch/record/1571169, 23 May 2014.
Clauset A., Shalizi C. R., and Newman M. E.J. (2009) Power-law distributions in empirical data, SIAMReview 51-4, 661-703.
CMS (2002) Memorandum of Understanding [MoU] for Maintenance and Operation of the CMS Detector, CERN-RRB-2002-033, 28 May 2002.
CMS (2012) Observation of a new boson at a mass of 125 GeV with the CMS experiment at the LHC, Physics Letters B, 716, 30-61.
Csardi G. et al. (2014) Package iGraph. Network Analysis and Visualization, Version 0.7.1 (April 22, 2014). http://cran.r-project.org/web/packages/igraph/igraph.pdf
Graßhoff G., Wüthrich A., eds. (2012) MetaATLAS. Studien zur Generierung, Validierung und Kommunikation von Wissen in einer modernen Forschungskollaboration, Bern, SHPS.
Morley E. W. (1895) On the Densities of Hydrogen and Oxygen and on the Ratio of Their Atomic Weights, Washington: Smithsonian Institution (Smithsonian Institution Contributions to Knowledge, no. 980).
Price D. J. de S. (1963) Little Science, Big Science, New York, Columbia University Press.
Schukraft J. (2004) Un processus, pas un événement, Infiniment CERN. Témoins de cinquante ans de recherches, Genève, Éditions Suzanne Hurter.
Shalizi C. (2007) R Code for the estimation of power laws and their comparison to heavy-tailed alternatives. R package, version 0.0.3.
Wasserman S. and Faust K. (1994) Social Network Analysis. Methods and Applications, Cambridge, Cambridge University Press.
Wimsatt W. C. (1981) Robustness, Reliability and Overdetermination, in Brewer M. et Collins B., eds., Scientific Inquiry and the Social Sciences, San Francisco, Jossey-Bass, pp. 124-163.
Acknowledgments
The author gratefully acknowledges Guillaume Beuf (Doctor in Theoretical Physics 2009, UPMC/CEA-Saclay); Emmanuel Laisne (Doctor in experimental physics in 2012, LPSC/University of Grenoble) for advice about the ATLAS and CMS organization and operation; and Jean-Loup Gilis (University of Grenoble) for IT support.
