Library Philosophy and Practice (e-journal)
Libraries at University of Nebraska-Lincoln
9-13-2019
Visualizing Research Collaboration in Statistical
Science: A Scientometric Perspective
Prabir Kumar Das
Library, Documentation & Information Science Division, Indian Statistical Institute, 203 B T Road, Kolkata - 700108, [email protected]
Follow this and additional works at:
https://digitalcommons.unl.edu/libphilprac
Part of the
Library and Information Science Commons
Das, Prabir Kumar, "Visualizing Research Collaboration in Statistical Science: A Scientometric Perspective" (2019).Library Philosophy and Practice (e-journal). 3039.
V
V
i
i
s
s
u
u
a
a
l
l
i
i
z
z
i
i
n
n
g
g
R
R
e
e
s
s
e
e
a
a
r
r
c
c
h
h
C
C
o
o
l
l
l
l
a
a
b
b
o
o
r
r
a
a
t
t
i
i
o
o
n
n
i
i
n
n
S
S
t
t
a
a
t
t
i
i
s
s
t
t
i
i
c
c
a
a
l
l
S
S
c
c
i
i
e
e
n
n
c
c
e
e
:
:
A
A
S
S
c
c
i
i
e
e
n
n
t
t
o
o
m
m
e
e
t
t
r
r
i
i
c
c
P
P
e
e
r
r
s
s
p
p
e
e
c
c
t
t
i
i
v
v
e
e
P
P
r
r
a
a
b
b
i
i
r
r
K
K
u
u
m
m
a
a
r
r
D
D
a
a
s
s
S ScciieennttiiffiiccAAssssiissttaanntt––AA L Liibbrraarryy,,DDooccuummeennttaattiioonn&&IInnffoorrmmaattiioonnSScciieenncceeDDiivviissiioonn,,IInnddiiaannSSttaattiissttiiccaallIInnssttiittuuttee,,220033,,BB..TT..RRooaadd,, K Koollkkaattaa––770000110088,,IInnddiiaa,,EEmmaaiill::pprraabbiirrddaass@@iissiiccaall..aacc..iinnAbstract
Using Sankhyā – The Indian Journal of Statistics as a case, present study aims to identify scholarly collaboration pattern of statistical science based on research articles appeared during 2008 to 2017. This is an attempt to visualize and quantify statistical science research collaboration in multiple dimensions by exploring the co-authorship data. It investigates chronological variations of col l a b o ra t io n pattern, nodes and links established among the affiliated institutions and countries of all contributing authors. The study also examines the impact of research collaboration on citation scores. Findings reve a l steady influx of statistical publications with clear tendency towards collaborative ventures, of which double-authored publications dominate. Small team of 2 to 3 authors is responsible for production of majority of collaborative research, whereas mega-authored communications are quite low. Country - w i se mapping of research contributions revels that, top five countries have contributed about 66% of the total authors and about 55% of the total affiliated institutions. Indicates few numbers of countries has substantial participation to statistical science research, while large majority has nominal contribution s. Of which, USA contributes the most (31%) followed by India, Canada, France and Japan. Result therefore indicates presence of ‘sort of ‘clique’ with dominant foreign coauthors. Further analysis reveals that, unilateral collaboration dominates at the country level whereas at the institution level bilateral collaboration dominates - implies authors from two different institutions of same country are key contributors of this specialty. Indian Statistical Institute (native institute of the source journal) foun d t o b e the most productive institution. Study therefore signifies skewed distribution of co-authorship with l i mi t e d evidence of cross-country collaboration. Furthermore, Google Scholar citation analysis showed that collaboration has significant positive influence on the article impact.
Keywords:
Bibliometrics, Resea rch colla bora tion, Colla bora tion ma pping, Colla bora tion Density, Sta tistica l science resea rch, Sankhyā, India n Journa l of Sta tistics1.
Introduction
Phenomena l development of ICT (especia lly Internet, web2.0 a nd socia l networking technology) ha s dra stica lly reduced the role of spa tia l bounda ries for a ccessing informa tion. This in turn ha ve fa cilita tes resea rchers of dista nt loca tions to intera ct ea sier tha n before. Nowa da ys, resea rchers of dista nt pla ces intera ct frequently to bra instorm diverse problem with grea ter proba bility of success. Countries, a ll over the world ha ve developed policies to foster cross-fertiliza tion of intellectua l a ssets. Resea rch policy a cross the globe thus ha s become more focused on stimula ting colla bora tive resea rch. Therefore, colla bora tion ha s become the norm of current resea rch towa rds knowledge economy. Reportedly, resea rch colla bora tion being a multi-fa ceted a bstra ction, often been defined loosely in the litera ture. Gene ra lly, it ca n be describ e d
a s a process of working together throughout a la rge pa rt of the project, a nd colla bora tors a re those who either initia te the project, or lea d the project, or a re responsible for the ma in pa rt(s) of the project (Ka tz & Ma rtin, 1997). Therefore, colla bora tion is a n intense form of intera ction tha t fa cilita tes sha ring of competence, skills a nd other resources. Resea rchers colla bora te for va riety of motiva tions a nd considera tions like pooling of intellectua l resources, sha ring of overhea d for expensive experiments, cross-fertiliza tion of thoughts, increa sed visibility a nd recognition (Fra me & Ca rpenter, 1979). Though va rious fields of knowledge ma y exhibit different a ffinity towa rds colla bora tive a ctivity, increa sing colla bora tion is the norms of current resea rch fra ternity. Numerous studies ha ve been underta ken to visua lize a nd a na lyze the fea tures of resea rch colla bora tion a mong the a ffilia ted countries a nd institutions. However, severa l fa ctors like mutua lity of interest, subject specia lty, da ta resources, geogra phica l proximity, cultura l a nd linguistic diversity, politica l interference, socio-economic conditions a nd foreign policy a re some of the ba sic determina nts sha ping the pa ttern of colla bora tion (Ha rirchi et a l., 2007).
Conventiona lly, in a n a ca demic setup resea rch publica tion is rega rded a s perceptible output of resea rch project, a nd co-a uthorship is used a s a proxy indica tor for estima tion of resea rch colla bora tion. Moreover, ever-increa sing trend of resea rch colla bora tion a cross discipline ha s ra ised few pertinent concerns, for which there is ha rdly a ny consensus of opinions. In fa ct, there ha s been mixed opinion a mong resea rchers of va rious disciplines rega rding the rela tionship between pa ttern of colla bora tion a nd its impa ct on qua lity of resea rch (Nude, 2016). Studies ha ve reported significa nt positive impa ct of colla bora tion on the cita tion in the disciplines of Economics (Levitt a nd Thelwa ll, 2010); Biology a nd Biochemistry (Didega h a nd Thelwa ll, 2013); Computer Science (Ibá ñez et a l., 2013); Libra ry & Informa tion Science (Pa tra , 2014) a nd (Abt, 1984). Conversely, good number of studies in the disciplines of Ecology (Leimu a nd Koricheva , 2005); Chemistry (Bornma nn et a l., 2012); Geogra phy a nd Forestry (Slyder et a l., 2011); Socia l Psychology (Ha sla m et a l., 2008) identifies no significa nt correla tion exists between these two va ria bles. Therefore, there ha s been little consensus a bout the proposition tha t colla bora tion boosts or fosters cita tion a cross doma ins. However, in spite of numerous studies, no such effort ha d been ventured in the a rea of sta tistics to a ddress the issue.
2.
Research Collaboration in Statistical Science
There ha s been considera ble a mbiguity a mong world sta tisticia n viewpoints rega rding precise interpreta tion of the term ‘Sta tistics’. According to ASA PresidentMa rie Da vidia n, “Statistics is the science of learning
from data, and of measuring, controlling and communicating uncertainty.” (Davidian and Louis, 2012). Genera lly, “Sta tistics a t its best provides methodology for dea ling empirica lly with complica ted a nd uncerta in informa tion, in a wa y tha t is both useful a nd scientifica lly va lid ” (Cha mbers, 1993). However, Feinberg (2014) opined tha t sta tistica l methods could be used not just within their fields of interest but a cross the spectrum of disciplines. Therefore, sta tistics is a meta -discipline in the sense, a s principles a nd methodologies a re a bstra cted from pa rticula r experiment could be ea sily incorpora ted into innova tive sca ffold tha t fa cilita te them to be used in numerous problems of ma ny other pla ces.
Reportedly, most of the disciplines ha ve distinctive epistemologica l fea ture which regula tes its resea rch pra ctices. Moreover, disciplines show verity of pra ctices towa rds colla bora tion. Sta tistics, being a doma in of multidisciplina ry na ture, often intermingles expositions from both socia l science (soft discipline) a nd ma thema tica l sciences (ha rd discipline). Thereby , significa nt contribution between the resea rchers from both the disciplines is quite norma l a ffa ir. Moreover, Sta tistics is genera lly perceived a s the universa l tool of inductive inferences. Therefore, it is little surprising tha t sta tistics journa l litera ture (sa y Sankhyā ) could be ca ndida te for study of colla bora tion trend. Furthermore, scientific colla bora tion is genera lly presumed t o enha nce qua lity a nd impa ct of resea rch, a s individua ls with network environment ha ve better a ccess to la rger pool of idea s a nd resources. While positive role of resea rch colla bora tion ha s been confirmed a cross disciplines of na tura l a nd socia l sciences, there ha ve been stra y a ttempts (Stigler, 1994, & Visa khi a nd Sriva sta va 2002)to study the colla bora tion pa ttern of Sta tistica l Science resea rchers. Our study intends to identify colla bora tion trend a long with the cita tion scores thereof by exa mining the institutiona l a nd country a ffilia tions of the a uthors published in one of the fla gship journa l of Sta tistica l science.
3.
Scope and Objectives
This study is confined to the schola rly a rticles a ppea red in journa l Sankhyā (both series A a nd B) during the period 2008 to 2017. However, rela tively lesser communica tions like editoria l notes, biogra phy, corrigendum, obitua ry, book reviews, etc. a re excluded from the purview of this study. So the bibliometric a na lysis of peer-reviewed resea rch a rticles a ppea red during la st ten yea rs period (vol. 70 to 79, covers twenty issues) of Sankhyā would certa inly be a n indica tive of current trends of sta tistica l resea rch. However, present study is structured a round the following specific resea rch objectives:
❑ To identify the qua ntum of colla bora tive a nd non -colla bora tive publica tions in this specia lty ❑ To determine the extent of resea rch colla bora tion a mong the a ffilia ted a uthors, institutions a nd
tra nsna tional countries of sta tistica l science resea rch.
❑ To deduce colla bora tive metrics so a s to ma p extent of resea rch colla bora tion ❑ To a na lyze colla bora tion pa tterns a t va rious levels of a ggrega tions
❑ To enumera te prolific a uthors a nd their a ffilia ted institutions of sta tistica l resea rch
❑ To identify the correla tion between colla bora tion a nd cita tion impa ct of sta tistica l science resea rch
4.
Origin and genesis of Sankhyā - The Indian Journal of Statistics
Sankhyā – The Indian Journal of Statistics ha s been chosen a s the da ta source to identify the sta tistica l science, prima rily beca use - it is the fla gship India n journa l in the field of sta tistics. Pra ctica lly it is considered the first India n journa l in the a rea of sta tistics. (Rudra , 1996). However, dedica ted publica tion outlet to showca se contempora ry resea rch communica tion in the a rea of sta tistics a nd rela ted a rea s wa s a ptly perceived by multifa ceted Rena issa nce ma n Prof. P. C. Ma ha la nobis (Ra o, 2006). He founded a nd edited Sankhyā – The Indian Journal of Statistics, a n interna tiona l schola rly journa l from India n Sta tist ic a l Institute. Genetica lly, the journa l Sankhyā ha s close bea ring with India n epic Mahabharata (Ghosh et a l. ,1999), objectively, journa l Sankhyā has evolved to unfold twin a spects of sta tistics, both theoretica l a nd a pplied towa rds his perception ‘sta tistics a s key-technology for socia l welfa re a nd na tiona l development’. In pursua nce of a bove philosophy, the journa l provides a convenient communica tion cha nnel for excha nging innova tive idea s in the different sub-doma ins of sta tistics, which ma ke Sankhyā a n effective a nd relia ble representa tive of contempora ry sta tistica l resea rch. Elsewhere, we ha ve discussed a bout significa nt ebbs a nd flows undergone by the journa l since its inception (Da s a nd Pa l, 2013).
5.
Data Source and Collection
The study ha s been conducted in successive pha ses: da ta extra ction, verifica tion a nd scientometric a na lysis. First of a ll, prima ry da ta of sta tistica l resea rch ha s been extra cted from MATHSCINET. Its sea rch interfa ce provides diverse combina tion of fields like - MR No, a uthor, institutiona l a ffilia tions, country code, MSC cla ssifica tion code, source journa l a nd cita tions thereof, which could be useful to identify the publica tions of journa l. Bibliogra phica l da ta elements of individua l a rticles ha ving source Sankhyā published during 2008 to 2017 were retrieved from the MATHSCINET website by fra ming sea rch query:
Journal = Sankhya AND publication type = journal AND year range = 2008 to 2017. Sea rching ha s yielded 340 hits (a s on 28 April 2018, Fig - 1), were considered a s working da ta set of our study. Further scrutiny of retrieved da ta set (340 in numbers) ha s revea led tha t, a s ma ny a s five item records were not ‘research articles’, they are actually erratum to earlier publications. So, these five records – viz. MR3575747 (erra tum to MR3317478), MR3400122 (erra tum to MR3061848), MR3302278 (erra tum to MR3302274), MR3302277 (erra tum to MR3082814) a nd MR3167778 (erra tum to MR3010292) a re excluded from the purview of our study. Aga in, MR2658160 [conta ins biogra phy of A P Mitra (1938 – 2008)] a nd MR3749268 [conta ins obitua ry of J K Ghosh] a lso excluded for the sa me logic.
Further va lida tion with the physica l volumes of source journa l, we found the record for la st a rticle of
Sankhyā [Vol. 71 pa rt B (2009), p. 264 – 289], entitle “ An application of Martingale Convergence Theorem to Analyse Company Data” by VC Augustine & M G Nadkarni, is somehow missing in the
MATHSCINET, this ha s been included to our study sa mple. Therefore, the fina l counting of Sankhyā a rticles during the study period (2008 to 2017) becomes (340 -7+1) = 334. This ha s been rea ffirmed by visiting the officia l website of the source journa l. Fina lly, releva nt scientometric techniques a re a pplied to determine the bibliogra phic a ttributes of the resea rch publica tions. So, necessa ry bibliogra phic elements of ea ch a rticle like yea r of publica tion, a uthor/s na me, number of a uthors, a ffilia tions, colla bora tion types, ma thema tics subject cla ssifica tion, etc. a re recorded a nd subsequently a na lyzed for ma king insightful interpreta tions.
Fig – 1: Search results of MathSciNet
6.
Methodology
To investiga te the resea rch colla bora tion a nd cita tion impa ct of the sta tistica l publica tions, ba sica lly a descriptive Scientometric scrutiny wa s employed. Schola rly a rticles published in Sankhyā over 10 yea r period (2008 - 2017) ha s been a ssessed retrospectively. Colla bora tion ha s been a ttributed on the ba sis of number of a uthors, where single a uthored pa per implies no-colla bora tion a nd two or more a uthored pa per imply colla bora tive resea rch. Furthermore, colla bora tion ha s been studies in two broa d perspectives – viz. country-wise colla bora tion a nd institution-wise colla bora tion. However, colla bora tion typology used for this study wa s mostly ba sed on the topology used by Soorya moorthy(2009). Here, na tiona l colla bora tion is considered for the ca se in which co-a uthors a re exclusively from different India n institutions; a t the sa me line - foreign colla bora tion is supposed to ta ke pla ce in ca se where a ll the co -a uthors a re from different foreign countries (except India ). Wherea s Interna tiona l colla bora tions is termed - when colla bora tions ta ke pla ce between a uthors from foreign countries a nd India (a t lea st one a uthor). Furthermore, these ha s been subdivided to Intra (within sa me link), Inter (a mong links) modes. Va rious colla bora tion metrics viz ‘degree of collaboration’, ‘collaboration coefficient; and ‘collaboration index’ for the sampled data were
deduced by a pplying esta blished methods (Ajiferuke, 1988; Subra ma nyam, 1983; La wa ni, 1980). Ra nk list of prolific a uthors is enumera ted using fra ctiona l counting method (Va n-Hooydonk, 1997). Cita tion counts of a ll sa mpled a rticles were retrieved from Google Schola r. Individua l a rticle wa s sea rched to find the number of cita tion (including self cita tions) received up to Dec’18. All bibliogra phic a nd cita tion da ta for 334 a rticles were tra nscribed in MS Excel sprea dsheet for a pplying va rious ma thematical a nd sta tistic a l techniques. Fina lly, necessa ry da ta sheets were presented for perceptive interpreta tions.
7. Data Analysis and Interpretation
A deta iled a na lysis of collected da ta revea led lots of informa tion to a nswer va rious interesting questions, which a re presented in the following sections.
7.1 Year wise distribution of contributions
Ta ble-1 shows descriptive sta tistics of bibliometric a ttributes of Sankhyā over the la st ten yea rs period (2008 – 2017). Result reflects tha t, during this period a tota l of 334 a rticles were published by tota l 682 a uthors in the 40 issues of Sankhyā Series A a nd B. These a rticles tota lly consumed 6984 pa ges tha t collectively a ppended 1734 keywords a nd 7275 references respectively. Therefore, a vera ge a uthorship p e r a rticle mea sures to slightly grea ter tha n 2 (2.04), indica tes the domina nce of joint a uthorship. Results a lso reflect stea dy influx of Sankhyā a rticles (a vera ge 33.4 a rticles/ yea r). More significa ntly, up to December 2018, these publica tions receive fa irly good numbers of cita tion (1804) with a n a vera ge of 5.4 cita tions per pa per. Of which, 2011 received highest a vera ge cita tion a nd 2016 ha s received the lea st (2.16). (Expectedly this would increa se over course of time). Result therefore reflects the visibility a nd popula rity of the journa l a mong the peers.
T
Taabbllee– –11::Descriptive Statistics of Sankhyā articles
Y Yeeaarr AArrttiiccllee AAuutthhoorr AAuutthhoorr// A Arrttiiccllee T Toottaall C Ciittaattiioonnss C Ciittaattiioonn// A Arrttiiccllee R Reeff.. a appppeennddeedd T Toottaall P Paaggee K Keeyywwoorrdd a appppeennddeedd 2008 30 55 1.83 282 9.40 614 661 142 2009 33 65 1.97 257 7.79 647 624 185 2010 36 75 2.08 262 7.28 653 650 180 2011 36 70 1.94 353 9.80 759 687 188 2012 31 70 2.26 154 4.97 678 605 167 2013 33 66 2.00 97 2.93 810 724 181 2014 34 79 2.32 140 4.12 773 734 177 2015 38 71 1.87 110 2.89 818 808 175 2016 31 61 1.97 67 2.16 769 728 170 2017 32 70 2.19 82 2.56 754 763 169 T Toottaall 333344 686822 2.2.0044 18180044 5.5.4040 72727755 69698844 17173344 7.2 Authorship Pattern
Ta ble-2 represents ‘pa ttern of a uthorship’ of the published a rticles of Sankhyā during 2008-2017. The study identified a tota l of 682 occurrences of a uthors counted in 334 a rticles produced during the period, thus a vera ge a uthorship obta ined 2.04 for ea ch publica tion. It is a lso observed tha t double-a uthored pa pers a re predomina te (142, 20.82%), followed by single a uthored (98, 14.37 %) a nd triple-a uthored (82, 12 %) respectively. Wherea s mega -a uthored pa pers (four, five & six-a uthored) a re quite low (12, 1.76%). Mea n co-a uthor for colla bora tive a rticles a re 2.47. Therefore, result signifies tha t contributors of Sankhyā prefer to work in colla bora tive fa shion. Observed trend of colla bora tive publica tions ha ve been the norms in ma ny other disciplines a s well (Ba ndyopadhyay, 2001). However, Visa khi a nd Sriva sta va (2002)in their study on resea rch colla bora tion of sta tistica l science a lso endorsed the simila r trend .
T Taabbllee––22::DDiissttrriibbuuttiioonnooffaarrttiicclleessbbyyAAuutthhoorrsshhiipp Y Yeeaarr NNuummbbeerrooff A Arrttiicclleess A Auutthhoorrsshhiippvvaalluuee TToottaall o occccuurrrreennccee a auutthhoorrss A Avveerraaggee a auutthhoorrsshhiipp Solo Two Three Four Five Six
2008 30 12 12 5 1 55 1.83 2009 33 10 14 9 65 1.97 2010 36 9 15 12 75 2.08 2011 36 8 22 6 70 1.94 2012 31 7 12 10 1 1 70 2.26 2013 33 11 12 9 1 66 2.00 2014 34 9 11 10 3 1 79 2.32 2015 38 14 16 7 1 71 1.87 2016 31 12 11 7 1 1 61 1.97 2017 32 6 17 7 1 70 2.19 T Toottaall 333344 9898 141422 8282 8 8 2 2 2 2 686822 2.2.0404
7. 3 Ranking of Prolific Authors
Ta ble-3 enumera tes the ra nking of prolific a uthors ba sed on tota l weigh of a uthorship in a ll publica tions of
Sankhyā during the study period. Tota l Weight of pa rticula r a uthor ha s been summa rized by cumula ting individua l weights in different a uthorship contributions using fra ctiona l counting method. Ta ble reflects tha t 334 a rticles of Sankhyā ha ve been contributed by 565 individua l a uthors in different a uthorship positions resulting tota l of 682a uthorships. Of which, la rge ma jority of a uthors (453, 80.18%) ha ve contributed nomina lly (ha ving cumula tive weight < 1); only 30 a uthors (5.1%) ha ve cumula tive weight > 1 for their entire contribution. Out of which, Bra jendra C. Sutra dha r (Memoria l University of New Foundla nd, Ca na da ) is found to be the most prolific a uthor followed by Deba sis Kundu (India n Sta tistica l Institute, New Delhi); Sa ra less Na da ra ja h (University of Ma nchester, UK); Mike Ja croux (Wa shington Sta te University, USA); M. C. Bha tta cha rjee (New Jersey Institute of Technology, USA); Fuxia Cheng (Illinois Sta te University, USA); Denys Pommeret (CREST-ENSAI, Fra nce); Soumendra N. La hiri (Iowa Sta te University, USA), etc. However to a void long listing, na mes of the contributors who ha ve received nomina l cumula tive weighted (≤ 1) is not ta ken into a ccount.
Table - 3 : Ranking of Prolific authors ( based on Fractional counting method)
Rank Author name
Authorship contribution
Weight Total
Wt. Cu Wt.
One Two Three Four Five Six
1 Sutradhar, Brajendra C. 1 5 4 4.833 4.833 4.833
3 Nadarajah, Saralees 5 2 3.167 3.167 12.667 4 Jacroux, Mike 2 2 3.00 x 2 6.000 18.667 Sengupta, S. 3 5 Kruglov, Victor M. 2 1 2.500 2.500 2.500 6 Sen, Pranab K. 2 1 2.333 2.333 4.833 7 Ghosh, Jayanta K. 1 1 2 2.167 2.167 7.000 8 Balakrishnan, N 2 3 2.00 x 8 16.000 41.667 Withers, Christopher S. 4 Cassese, Gianluca 2 Kosorok, Michael R. 2 Kunita, Hiroshi 2 Ogawa, Shigeyoshi 2 Sarkar, Sanat K. 2 Wellne, Jon A. 2 9 Sinha, Bimal 2 1 1 1.533 1.533 43.200 10 Bhattacharya, Rabi 1 3 1.50 x 6 9 52.200 Pradhan, Biswabrata 3 Nkurunziza, Sévérien 1 1 Patriota, Alexandre G. 1 1 Rajeev, B 1 1 Ramasubramanian, S 1 1 11 Bhattacharya, A 1 1 1.333x4 5.332 57.532 Bhattacharya, Sourabh 1 1 Chesneau, Christophe 1 1 Maitra, Ranjan 1 1 12 Caro-Lopera, Francisco J. 1 2 1.167x2 2.334 59.866 Lin, Lizhen 1 2 13 Pardo, Leandro 1 1 1 1.083 1.083 60.949
15 Basu, Ayanendranath 2 1 0.917 0.917 143.866
16 10 authors each having 1 1 0.833x10 8.330 152.196
17 Klein, Martin 1 1 0.700 0.700 152.896 18 13 authors 1 1 0.667x13 8.671 161.567 19 3 authors having 1 1 0.583x3 1.749 163.316 20 Sengupta, Arindam 1 1 0.533 0.533 163.849 21 203 authors having 1 0.500 101.500 265.349 22 176 authors having 1 0.333 58.608 323.957 23 27 authors having 1 0.25 x27 6.750 330.707 24 7 authors having 1 0.2 x7 1.400 332.107 25 11 authors having 1 0.167x11 1.837 333.944
Total 565 unique authors 98 284 246 32 10 12 334.000
7.4 Research Collaboration
Globa lly, colla bora tive resea rch is becoming more perva sive a cross disciplines (Georghiou, 1998). Here a lso, it is observed (Ta ble- 4) tha t, out of tota l 334 resea rch pa pers, 98 (29.34%) pa pers a re single a uthored, therefore devoid of a ny a uthorship colla bora tion. Of which, 32 non -colla bora tive a rticles a re from India a uthors a nd 66 a re contributed by foreign a uthors. Collected da ta is ma pped for level of colla bora tion underta ken by the a uthors of Sankhyā. Affilia tions of the contributing a uthors were scrutinized to a scert a in levels of colla bora tion a s Domestic (colla bora tion occurs a mong a uthors of different institutions of India ); Foreign (colla bora tion occurs between a uthors of different foreign institutions outside India . Interna tiona l colla bora tion is a ssigned to publica tions in which colla bora tion occurs between a uthors from India n a nd other foreign countries. It is found tha t, out of 236 colla bora tive pa pers, ma jority (186, 79%) a re of ‘Foreign’ category. followed by national (12.71%). Noteworthy is the fact that in both for collaborative and non-colla bora tive publica tions foreign contributions a re significa ntly domina nt. Out of 236 colla bora tive communica tions, domestic colla bora tion constituting only 8% a nd foreign colla bora tion constituting 58%, while the sha re of 34% of multi-a uthored contributions a re colla bora ted a mong sta tisticia ns a cross the countries. Clea rly it brings out the preva lence of colla bora tive resea rch (69.85%) over the solo resea rch (30.15%), which needs to be explored to promote interna tiona l colla bora tion.
Table 4 – Collaboration Scenario
Y Yeeaarr N Noonn--ccoollllaabboorraattiivvee((NNss)) % % C Coollllaabboorraattiivvee((NNmm)) % % D Doommeessttiicc FFoorreeiiggnn TToottaall DDoommeessttiicc FFoorreeiiggnn IInntteerrnnaattiioonnaall TToottaall 2008 3 9 12 12.24 1 15 2 18 7.63 2009 10 0 10 10.20 3 18 2 23 9.75 2010 4 5 9 9.18 5 21 1 27 11.44
2011 1 7 8 8.16 4 22 2 28 11.86 2012 0 7 7 7.14 3 18 3 24 10.17 2013 2 9 11 11.22 3 17 2 22 9.32 2014 2 7 9 9.18 3 20 2 25 10.59 2015 6 8 14 14.29 2 19 3 24 10.17 2016 3 9 12 12.24 2 15 2 19 8.05 2017 1 5 6 6.12 4 21 1 26 11.02 T Tototalal 3232 6666 9898 2929..3344 3030 181866 2020 232366 7070..6666
7.5 Collaboration Topology
Here la tera l rela tionship a mong co-a uthors of colla bora tive contributions is studied under three different levels of a ggrega tion - viz. unilateral, bilateral, a nd multilateral. Unilateral collaboration is described when co-a uthorship of a publica tion occurs within a link, wherea s bilateral collaboration implies the co-a uthorship occurs between two different links co-a nd multilateral collaboration indica tes the pa rticipa tion of co-a uthors from more tha n two different links for producing a n a rticle. Ta ble-5 revea ls the distribution of colla bora tive contributions in order to ma p the la tera l rela tionship a mong co -a uthors.
Y
Yeeaarr
Table 5 – Lateral relations among collaborative contributions
C
Coouunnttrryy--wwiisseeccoollllaabboorraattiioonn IInnssttiittuuttiioonn--wwiisseeccoollllaabboorraattiioonn
Unilateral Bilateral Multilateral Total Unilateral Bilateral Multilateral Total
2008 12 6 0 18 6 9 3 18 2009 17 6 0 23 6 13 4 23 2010 23 4 0 27 8 14 5 27 2011 22 4 2 28 13 12 3 28 2012 13 11 0 24 6 13 5 24 2013 17 5 0 22 11 7 4 22 2014 13 8 4 25 5 13 7 25 2015 17 7 0 24 9 13 2 24 2016 9 10 0 19 5 10 4 19 2017 21 4 1 26 9 16 1 26 T Toottaall 116644 6655 77 223366 7788 112200 3388 223366
Here colla bora tive a ttributes a re ma pped in two levels of stra tum viz. country-wise colla bora tion a nd institution-wise colla bora tion. Institution-wise colla bora tion of a publica tion occurs a mong the (co -) a uthor/s a ffilia ted to sa me institution/s (Unila tera l), or two different institutions (Bila tera l), or more tha n two different institutions (Multila tera l). Simila rly, country -wise Multi -la tera l colla bora tion indica tes tha t the a uthors of pa rticula r a rticle a re a ffilia ted to the institutes loca ted in three or more different countries. Result reflects tha t a t the country level, unila tera l colla bora tion domina tes (164, 70%) wherea s, institution level bila tera l colla bora tion domina tes. However, in both the stra tum, multila tera l colla bora tion is the lea st preference. Results, therefore signifies a uthors from two different institutions of sa me country predomina tes in sta tistica l science resea rch colla bora tion.
Table - 6 : Mapping of Collaboration density
Collaboration Type Freq %
FORN-UNI (authors from same foreign countries) 130 55.08 FORN-BI ( authors from two different foreign countries) 46 19.49 NATL-UNI ( among the Indian authors) 30 12.71 INT-BI ( Indian authors and foreign countries) 20 8.47 FORN-MULT ( authors from two or more foreign countries) 8 3.39 INT-MULT(authors from India and two or more foreign countries) 2 0.85
Total 236 100.00
Further ma pping of colla bora tive a ttribute of Sankhyā a rticles reflects tha t out of 236 colla bora tive a rtic le s majority (130, 55%) are in the form of ‘FORN-UNI‘ (foreign unila tera l) i.e. colla bora tion within the a uthors of sa me foreign countries. Followed by contributions of ‘FORN-BI’ (foreign bilateral) (46) i.e.,
colla bora tion between two foreign a uthors; a nd 8 pa pers a re ‘FORN-MULT’ (foreign multila tera l) i.e, colla bora tion a mong a uthors from more tha n two foreign countries. These foreign contributions (184 in tota l) a re devoid of a ny domestic a uthors. Conversely, 30 pa pers a re in the na ture of ‘NATL-UNI‘(domestic
unila tera l) i.e., these pa pers a re colla bora ted by two or more India n a uthors, therefore devoid of a ny foreign a ffilia tions. However, colla bora tion occurs a mong foreign countries a nd India occurs in the ca se of 22 pa pers. Of which in ca se of 20 pa pers (‘INT-BI’), colla bora tion occurs between one India n a uthor a nd one foreign author (International bilateral) and rest 2 papers are ‘international multilateral’ type i.e, among India n a uthors a nd two or more foreign a uthors. Therefore, country-wise colla bora tion ma pping revea ls clea r domina nce of foreign colla bora tion. Proba bly, tremendous prolifera tion ICT a nd internet with a nd improved resea rch infra structure ha s pa ved the scope for increa sed foreign contributions in this specia lty.
7.8 Collaboration Indices
To deduce the extent of colla bora tion towa rds multi-a uthorship of given discipline, a rra y of metrics ha ve been proposed in the form of Colla bora tive Index (CI); Degree of Colla bora tion (DC) a nd Colla bora tion Coefficient (CI). Here, these three indices a re deduced using the working da ta set to ha ve a n estima te of colla bora tion scena rio.
Table - 7 Yearly variations of Collaborative indices Year
Authorship Total
articles
Total
authors CI CC DC
Single Double Triple Four Five Six
2008 12 12 5 1 30 55 1.833 0.336 0.600 2009 10 14 9 33 65 1.970 0.394 0.697 2010 9 15 12 36 75 2.083 0.431 0.750 2011 8 22 6 36 70 1.944 0.417 0.778
2012 7 12 10 1 1 31 70 2.258 0.459 0.774 2013 11 12 9 1 33 66 2.000 0.386 0.667 2014 9 11 10 3 1 34 79 2.324 0.449 0.735 2015 14 16 7 1 38 71 1.868 0.353 0.632 2016 12 11 7 1 31 61 1.968 0.355 0.613 2017 6 17 7 1 1 32 70 2.188 0.460 0.813 Total 98 142 82 8 2 2 334 682 2.042 0.404 0.707
Subra ma nya m (1983) ma thema tica lly deduced degree of colla bora tive (DC) a s the ra tio of number of colla bora tive publica tions to tota l number of publica tions a s follows:
s m m N N N DC + = 98 236 236 + = = 0.707
Where Nm denotes tota l multi-a uthored contributions a nd Ns denotes the number of non-colla bora tive (single-a uthored) publica tions of the journa l during study period. Thus, a vera ge degree of colla bora tion is found to be significa nt (0.707). Ta ble shows the va ria tions of degree of colla bora tion from 0.6 (in 2008) t o 0.81(in 2017) with a n a vera ge of 0.707. Therefore, result indica tes the preva lence of colla bora tive resea rc h over solo resea rch in the specia lty which is in a greement with previous studies (Da s & Pa l, 2012).
Another significa nt metric, Colla bora tive Index (a proxy mea sure for resea rch colla bora tion) derived by La wa ni, a s the mea n number of a uthors publica tions. This ca n be ca lcula ted by the formula :
CI = [(f1) 1 + (f2) 2 + (f3) 3 + … (fk) k] / N. For, for the working da ta set, it becomes, CI = [(98x1) + (124x2) + (82x3) + (8x4) + (5 x2) + (6x2)] / 334. So, here CI = (682/334) = 2.042.
Result therefore indica tes rea sona ble va lue of colla bora tion index a mong the resea rcher of the specia lty. According to Ajiferuke (1988), Colla bora tion Co-efficient (CC) implies the mea n number of pa pers per joint a uthored publica tions. For the present da ta set, it is ea sily obta ined using the formula suggested by Ajiferuke (1988) a s follows:
N
Fj
j
CC
k j j/
)
1
(
1
1
1
= =−
=
Where, Fj = number of pa per ha ving j a uthors in the collection of k; N = tota l number of pa per published
K = highest number of a uthors per pa per; therefore, N =
iFjCC = 1 – [f1 + (1/2) f2 + (1/3) f3 +… + (1/k) fk] / N. For the given da ta set it becomes CC = 1- [98 + (1/2) x142 + (1/3) x82 + (1/4) x8 + (1/5) x2 + (1/6) x6] / 334
CC = 1- [0.599] = 0.401. From the deductions of three colla bora tive indices it is numerica lly evident tha t resea rchers working in this field prefer to conduct resea rch in colla bora tive fa shion.
7.9 Country Affiliations of the Contributors
Ta ble-8 reflects country-wise distribution of a rticles a s a ppea red in Sankhyā during the study period. Country na mes ha ve been a scerta ined from a uthor a ffilia tions a s reflected in their respective publica tions further verified from their ‘institution code’ da ta -field of Ma thSciNet. Ta bula ted da ta shows tha t a ll 682 contributing a uthors of Sankhyā ha s ema na ted from 51 different countries worldwide. Therefore, indica tion of globa l connectedness of the countries a nd a uthors of the source-journa l is clea rly observed. Ra nk list of a ffilia ted countries of contributing a uthors ha s been prepa red on the ba sis of sha re va lue o f t h e contributions from respective countries. Ana lysis shows a uthors belong to 7 countries of Africa ; 20
countries of Asia ; 15 countries of Europe; 3 countries of North America ; 4 countries of South America a nd 2 countries from the continent of Ocea nia . So there ha ve been fa irly va st distributions of a uthors a cross t h e globe. Of which, USA contributes the ma ximum (31%) of 211 a uthors from 78 institutions; followed by India (18%), Ca na da (8.5 %), Fra nce (4.55%), Ja pa n (3.5%) a nd others. It ha s been found tha t top five countries a re contributing a bout 66% of the tota l a uthors a nd a bout 55% of the tota l a ffilia ted institutions. Result indica tes high concentra tions of sta tistica l resea rch a ctivities in this region. This indica tes tha t a uthor distribution is very much skewed with respect to geogra phica l loca tions of contributing a uthors a nd their a ffilia ted institutions.
Table 8 - Geographical distributions of Authors
Rank Country Continent Authors % Institutions %
1 USA North America 211 30.94 78 26.62 2 India Asia 123 18.04 35 11.95 3 Canada North America 58 8.50 16 5.46 4 France Europe 31 4.55 18 6.14 5 Japan Asia 24 3.52 15 5.12 6 Italy Europe 23 3.37 15 5.12 7 England Europe 19 each 2.79 6 2.05 Germany Europe 10 3.41 Iran Asia 9 3.07 8 PRC China Asia 18 2.64 11 3.75 9 Brazil South America 13 1.91 4 1.37 10 Australia Oceania 10 each 1.47 6 2.05 Greece Europe 7 2.39 Taiwan(R.O.C.) Asia 5 1.71 11 Belgium Europe 6 each 0.88 4 1.37 Spain Europe 3 1.02 12 Poland Europe 5 each 0.73 2 0.68 Sweden Europe 2 0.68 13 Egypt Africa 4 each 0.59 3 1.02 New Zealand Oceania 1 0.34 Russia Asia 2 0.68 Switzerland Europe 3 1.02 14 Algeria Africa 3 each 0.44 2 0.68 Argentina South America 1 0.34 Colombia South America 1 0.34 Israel Asia 1 0.34 Mexico North America 2 0.68 Singapore Asia 1 0.34 Thailand Asia 2 0.68 The Netherlands Europe 2 0.68
Vietnam Asia 2 0.68 15 Cyprus Europe 2 each 0.29 1 0.34 Ethiopia Africa 0.29 1 0.34 Georgia Europe 0.29 2 0.68 Jordan Asia 0.29 2 0.68 Kuwait Asia 0.29 1 0.34 Nigeria Africa 0.29 1 0.34 Saudi Arabia Asia 0.29 2 0.68 UAE Asia 0.29 2 0.68 Ukraine Europe 0.29 1 0.34 16 Bangladesh Asia 1 each 0.15 1 0.34 Botswana Africa 0.15 1 0.34 Denmark Europe 0.15 1 0.34 Ecuador South America 0.15 1 0.34 Gambia Africa 0.15 1 0.34 Iraq Asia 0.15 1 0.34 Malaysia Asia 0.15 1 0.34 Mauritius Asia 0.15 1 0.34 Morocco Africa 0.15 1 0.34 Pakistan Asia 0.15 1 0.34 Uzbekistan Asia 0.15 1 0.34
Total 51 countries from 6 continent 682 293
Table 9: Collaborative behavior of Countries N Naattuurree NNoo..ooffCCoouunnttrryy((CCoollllbb..)) PPaappeerr %% AAuutthhoorr CCiittaattiioonn CCiitt//PPaappeerr Collaborative 1 166 70.33 401 824 4.96 2 63 26.69 160 336 5.33 3 6 2.54 18 150 25.00 4 1 0.43 5 3 3.00 Total 236 70.66 584 1313 5.56 Non collaborative 0 98 29.34 98 491 5.01 T Toottaall 333344 668822 11880044 55..4400
Ta ble - 9 represents the country of a ffilia ted a uthors who ha ve colla bora ted Sankhyā during the study period. Results shows tha t m a jority of the colla bora tive a rticles (70%) a re contributed by a uthors from single country. Wherea s, two countries pa rticipa ted in the a uthoring of 63(27%) a rticles a nd three countries pa rticipa ted in the a uthoring of 6(2.54%) a rticles respectively. In single ca se, colla bora tion occurs a mong the a uthors from four different countries.
Ta ble-10 enumera tes prolific institutions of sta tistica l science resea rch ba sed on the da ta retrieved from t h e ‘author affiliations’ of the contributing authors of Sankhyā. Ana lysis of a uthorship da ta shows tha t 682 contributors of Sankhyā a re a ffilia ted to 293 different institutions a cross the globe. A ra nk list of pa rticipa ting institutions ha s been prepa red ba sed on the frequency of institutions. It is observed from the ta ble tha t India n Sta tistica l Institute (6-ISI) ha s a ppea red on the top; followed by Memoria l University of Newfoundla nd, USA (3-NF), University of Ca lifornia , a nd University of Ma nchester, India n Institute of Technology, Ka npur a nd others. It is a lso evident from the ta ble, tha t though there ha s fa irly la rge number of institutions (293) a re involved in sta tistica l science resea rch only 15 institutions ha ve contributed more tha n one a uthor in the listing. La rge ma jority of institutions (155) contribute minim a lly (1a uthor).
Table 10 - : Prolific Institutions of Statistical Research
Sl no Title of the Institution Country Freq. of
authors %
1 Indian Statistical Institute (6-ISI) India 41 6.01 2 Memorial University of Newfoundland (3-NF) Canada 14 2.05 3 University of California, Dept of Statistics (1-CA-S) USA 13 1.91 University of Manchester, Dept of Mathematics (4-MANC) UK 13 1.91 4 Indian Institute of Technology, Kanpur (6-IITK) India 12 1.76
5
University of North Carolina (1-NC) USA 10 1.47 University of Minnesota Twin Cities (1-MN) USA 10 1.47 University of Manitoba (3-MB) Canada 10 1.47 University of São Paulo (BR-SPL) Brazil 10 1.47 6 Iowa State University (1-IASU) USA 8 1.17
7
Washington State University (1-WAS) USA 7 1.03 University of Washington, Dept. of Statistics (1-WA-S) USA 7 1.03 University of Maryland Baltimore County (1-MD4-MS) USA 7 1.03 Cornell University (1-CRNL) USA 7 1.03 Universität Konstanz (D-KNST-NB) Germany 7 1.03
8
U.S. Bureau of Census, Statistical Res. Div. (1-USBC-SMT) USA 6 0.88 Texas A & M University (1-TXAM-S ) USA 6 0.88 Shahid Beheshti University (IR-SHBH) Iran 6 0.88 University of Calcutta (6-CALC) India 6 0.88
9
University of Rochester Medical Center (1-RCT-BCB) USA 5 0.73 University of Arizona (1-AZ) USA 5 0.73 Purdue University, Dept of Statistics (1-PURD-S) USA 5 0.73 Duke University, Dept of Statistics (1-DUKE-S) USA 5 0.73 Sri Sathya Sai University (6-SSSU-NDM) India 5 0.73 Indian Institute of Management, Ahmadabad (6-IIM3-QM) India 5 0.73 Université de Paris VI (Pierre et Marie Curie) France 5 0.73 University of Windsor (3-WIND-NDM) Canada 5 0.73 McMaster University (3-MMAS-MS) Canada 5 0.73 10 20 Institutions having 4 each 80 11.80 11 22 Institutions having 3 each 66 9.68 12 68 Institutions having 2 each 136 19.72
13 155 Institutions each having 1 each 155 23.25
Total 293 unique institutions 682 100.00
7.11 Research Collaboration and Impact
Rela tion between a uthorship a nd cita tion frequency of Sankhyā a rticles is shown in the Ta ble - 11. Result indica tes tha t, 334 a rticles of Sankhyā ha ve received tota l 1804 cita tions in the Google Schola r (a s on Dec’18). Of which a s ma ny a s 77 (23%) a rticles ha ven’t received a ny cita tion, wherea s only two pa pers ha ve got more tha n 50 cita tions. Highest cita tion (121) received by a triple-a uthored pa per (MR2924199) in 2011. Yea r-wise distribution of cita tion shows tha t the yea r 2011 ha s received highest tota l cita tions (353) wherea s the yea r 2017 ha s received the lea st (67) tota l cita tions. Irrespective of a uthorship, ma jority of a rticles (177, 46.4%) received cita tions within 1 to 5 ra nges. As fa r a s colla bora tion is concerned, 98 non-colla bora ted pa pers ha ve received 491(27.22%) cita tions with a n a vera ge of 5.03 per pa per. Wherea s 236 colla bora ted pa pers ha s received 1313 (72.78%) cita tions with the a vera ge of 5.56 cita tions per pa per. Investiga ting the a vera ge number of cita tions received by the colla bora ted a nd non-colla bora ted pa pers of this journa l, it could be concluded tha t colla bora ted pa pers a re cited more often tha n the non-colla bora t e d pa pers. This is proba bly beca use, while colla bora ting, resea rchers beca me more fa milia r with co -a uthors ‘research network’ that exposes the ‘research’ to a bigger audience that eventually enhance the chance to be cited sooner or la ter. Result shows resembla nce with previous studies (Bordonset a l. 1993; Va n Ra a n, 1997 a nd Soorya moorthy, 2009).
As found in ta ble - 11, Sta tistica l science resea rchers publish in different a uthorship pa tterns ra nging from one to six a uthors per pa per. Approxima tely 71% of publica tions were colla bora tive, of which double -a uthored public-a tions (142, 42.5%) constitute the l-a rgest proportion followed by three --a uthored publica tions. As fa r a s cita tion is concerned, it is found a mong the colla bora tive publica tions six –a uthored publica tions ha ve received highest a vera ge cita tions, followed by three -a uthored publica tions a nd two-a uthored publictwo-a tions. Wheretwo-a s non-colltwo-a bortwo-a tive ptwo-a pers htwo-a ve received comptwo-a rtwo-a tively lower two-a vertwo-a ge cita tions (5.40) tha n colla bora tive pa pers (5.56) ha ve. However, publica tions ha ving grea ter number of a uthors not necessa rily received higher number of cita tions in this study. Therefore, n o significa nt correla tion between a uthorship a nd a vera ge cita tion wa s found in this study.
Table - 11 : Collaborative behavior of Authorship
Type NNoo..ooff A Auutthhoorrss PPaappeerr %% TToottaallCCiittaattiioonn %% CCiitt//PPaappeerr C Coollllaabboorraattiivvee Six 22 00..66 3366 22..0000 1188..0000 Five 2 0.6 3 0.16 1.50 Four 8 2.4 34 1.88 4.25 Three 82 24.55 558 30.93 6.80 Two 142 42.51 682 37.80 4.80 Total Collb. 236 70.66 1313 72.78 5.56
Non collaborative Single 98 29.34 491 27.22 5.01
T
Toottaall 333344 100 18180044 101000 5.5.4040
Table - 12 : Authorship collaboration vs Citation frequency
Authorship Citations distributions
Nil 1 - 5 6 - 10 11 - 20 21 - 30 31 - 40 41 - 50 51 - 100 100 - 200 Total
29.60% 48.98% 10.20% 6.10% 1.02% 1.02% 2.04% 1.02% Double 30 67 27 13 4 1 142 21.12% 47.18 19.01% 9.15% 2.82% 0.07% Three 15 35 20 9 2 1 82 18.29% 42.68% 24.39% 10.98 2.43% 1.22% Four 2 3 2 1 8 25% 37.50% 25% 12.50% Five 1 1 2 50% 50% Six 0 1 1 2 0% 50% 50% Total 77 155 59 29 7 3 2 1 1 334
8.
Major Findings and their Implications:Ba sed on scientometric scrutiny, present study reflects a holistic perspective of colla bora tive la ndsca pe of sta tistica l science resea rch. These a re a s follows:
• Result shows consistent influx of sta tistica l science litera ture (33.4 a rticles/yea r) contributed by tota l 682 a uthors, of which 565 a re unique a uthors. Ma ximum contribution of a ny a uthor (Bra jendra C Sutra dha r) is 10 pa pers, wherea s ma jority (489, 56.55%) of a uthors contributes nomina lly. Of which, ‘double-a uthored’ pa pers domina te but mega -a uthored pa pers a re significa ntly low.
• This study shows positive trends towa rds colla bora tive resea rch a s multi-a uthorship (236, 70.66%) clea rly domina te over non-colla bora tive pa pers (98, 29.34%). Mea n co-a uthor for colla bora tive resea rch is found to be 2.47. Furthermore, estima tion of colla bora tive indices (DC, CC a nd CI) indica tes clea r inclina tion towa rds colla bora tive resea rch with a vera ge a uthorship 2.04 per pa per.
• Though, the source journa l origina tes from Asia n country, there ha ve been fa irly va st distributio n s of a uthors a cross the globe (51 countries). Of which, USA contributes the ma ximum (31%) followed by India (18%), Ca na da (8.5 %), Fra nce (4.55%), Ja pa n (3.5%) a nd others. Top five countries a re contributing a bout 66% of the tota l a uthors a nd a bout 55% of the tota l a ffilia ted institutions. Indica tes high concentra tions of sta tistica l resea rch a ctivities in this region. Therefore, wide geogra phic representa tion of a uthor implies globa l a ccepta nce of the journa l without a ny bia s towa rds na tive a uthors wha t so ever. However, a uthor distribution is very much skewed with respect to geogra phica l loca tions of contributing a uthors a nd their a ffilia ted institutions.
• There ha s been fa irly la rge number of institutions (293) involved in sta tistica l science resea rch , but only 15 institutions ha ve contributed more tha n one a uthor. La rge ma jority of institutions (155) contribute minima lly (one a uthor). Of which na tive Institute of the journa l, India n Sta tistica l Institute is the lea ding producer of a uthors. However, in the top five positions (conta ining 9 institutes), 7 a re foreign institutions - indica tes domina nce of foreign institutions.
• Foreign unila tera l colla bora tion (i.e. Colla bora tion within a uthors of sa me foreign countries) is found to be the most prominent colla bora tion type a nd a ccounts for 55% (130 a rticles) of colla bora ted communica tions, followed by Foreign unila tera l (46 a rticles); NATL-UNI (within two or a uthors of India , 30a rticles); INT-BI (within one a uthor form India a nd other foreign countries, 20 a rticles). However, the most diversified proposition - interna tiona l multila tera l colla bora tion (within two or more a uthors from foreign countries one of which if from India ) is the lea st option pra cticed by the resea rchers.
• At the country level, unila tera l colla bora tion domina tes (70%) wherea s, institute level bila tera l colla bora tion domina tes. Two different countries ha ve pa rticipa ted in the a uthoring of 63(27%) a rticles a nd 3 different countries ha ve pa rticipa ted in the a uthoring of 6(2.54%) a rticles respectively. Only single ca se, colla bora tion occurs a mong the a uthors from four different countries. However in both the stra tum, multila tera l colla bora tion is in the lea st preference. Results, therefore signifies a uthors from two different institutions of sa me country predomina tes in sta tistica l science resea rch colla bora tion.
• Moreover, there were 121 (32.01%) joint a uthored a rticles by a uthors from the sa me institution (domestic intra -institutiona l colla bora tion) a nd 52 (13.76%) a rticles produced by a uthors from different institutions within the sa me country (inter- or externa l institutiona l colla bora tion).The significa nce of this observa tion is the sca rcity of resea rchers involving more tha n two na tions . Preva lence of sma ller tea ms (2/3 links) responsible for the publica tion of colla bora tive pa pers could perha ps be justified by the presence of sort of ‘clique’ in this resea rch community.
•
Study of cita tion performa nce shows tha t a s on da te, 77 (23%) a rticles ha ven’t received a nycita tion, wherea s ma jority of a rticles 262 (78.44%) received cita tions ra nging from 1 to 121. Of which, ma jority (155, 46%) of a rticles received nomina l cita tions (1 to 5). But, proportion of highly cited pa per is very low (only 2). Avera ge cita tion per pa per consistently va ries from 3 to 5.56, only in ca se of three country colla bora tion, it jumped unusua lly to 25. However, this skewed cita tion da ta influenced the a vera ge cita tion/pa per va lue by the few highly cited pa pers. Significa ntly, colla bora ted a rticles a re better cited a s a vera ge cita tion per pa per for colla bora tive publica tions (5.56) is slightly higher tha n non-colla bora tive publica tions (5.40). Therefore, resea rch colla bora tion ha s ma rgina l significa nt impa ct on the cita tion performa nce of the resea rch.
9.
Conclusion
In view of the substa ntia l benefit of colla bora tive resea rch endea vors, countries a ll over the world ha ve reengineered policies a nd stra tegies to fa cilita te resea rch colla bora tion. Present study investiga tes current colla bora tion scena rio of schola rly litera ture of Sta tistica l science in terms of co-a uthorship in va rious perspectives. Ba sed on the a na lysis a nd resea rch findings, it is a ppa rent from the study tha t there ha s been stea dy influx of litera ture produced by the fla gship journa l. However, though a uthor-ba se is va stly distributed a cross the globe, ma jority of individua l a uthors contributed nomina lly. As the study revea led, resea rchers of the specia lty ha ve clea r tendency towa rds colla bora tive ventures. However, both the a vera ge a uthorship for a ll publica tions a nd a vera ge co-a uthorship for colla bora tive pa pers a re modera te (2.04 a nd 2.47 respectively). Moreover, though a tota l 51 countries ha d pa rticipa ted in sta tistica l science resea rch output, only few countries viz. USA, India Ca na da, Fra nce a nd Ja pa n ha ve contributed substa ntia lly; wh ile ma jority of the rest countries ha ve token contributions. Furthermore, ma jority of pa pers in our study sa mple ha d one or two pa rtner countries. Only in single ca se, colla bora tion occurs a mong the a uthors from four different countries. Result indica tes va ria ble distribution of sta tistica l science resea rch a ctivities a cross different pa rts of the world. Little surprisingly, India n Sta tistica l Institute (na tive institution of the source journa l) is the lea ding producer of the sta tistica l science resea rcher. About 46 % of publica tions were under ‘domestic’ category and 20 % of publications were as a result of international collaboration , of which ‘domestic intra-institutional’ collaboration predominates (32%). Therefore, authors from two different institutions of sa me country predomina te in sta tistica l science resea rch colla bora tion. Results, therefore indica tes the presence of ‘sort of clique’ in the sta tistica l science resea rch community. However, country-wise colla bora tion ma pping revea ls clea r domina nce of foreign colla bora tion . On the whole, a t the country level, unila tera l colla bora tion domina tes (70%) wherea s, institute level bila tera l colla bora tion domina tes. As the result shows, resea rch colla bora tion especia lly a t the interna tiona l level is low, which needs to be encoura ged not only for better visibility of ‘loca l resea rch’ but a lso to bring diversity in the resea rch network, which eventua lly promotes the impa ct a nd qua lity of the resea rch. Furthermore, colla bora ted resea rch a rticles tend to ha ve better impa ct (in terms of cita tions) tha n non-colla bora tive counterpa rts, proba bly beca use of a ggrega ted efforts a nd synergies of idea s. Consistent with previous studies (Bordonset a l, 1993; Va n Ra a n, 1997) present study a lso rea ffirms tha t colla bora tion boosts cita tion impa ct of resea rch. However, our study provides useful informa tion for policy ma kers seeking to intensify scientific
colla bora tion, pa rticula rly a t the globa l level. La unching of ‘resea rcher excha nge progra m ’, ‘resea rch forums with interna tiona l institutes’; orga nizing conferences, symposia a nd resea rch curricula with interna tiona l peers a re some of the possible mea sures tha t could provide congenia l environment to promot e interna tiona l resea rch colla bora tion. So, the implica tion of this a na lysis to science policy ma kers a nd sta keholders is tha t - colla bora tion a mong the resea rchers ha s obviously a positive influence on the resea rch endea vors in this specia lty. However, findings a re ba sed on bibliogra phica l da ta limited to a single journa l with focused a uthorship; therefore, a ny multi-journa l study ha ving extensive da ta set could ha ve produced different result. Hence, these might instiga te further resea rch endea vors in this direction.
References:
Abt, HA (1984). Cita tions to single a nd multi-a uthored pa pers. Publications of the Astronomical Society o f the Pacific, 96, 746-749.
Ajiferuke, I et a l. (1998). Colla bora tive coefficient: A single mea sure of the colla bora tion in resea rch.
Scientometrics, 14, 421‑33.
Ba ndyopa dhya y, AK (2001). Authorship pa ttern in different disciplines . Annals of Library and Information Studies, 48 (4), 139-147.
Bordons, M et a l. (1993). Is colla bora tion improving resea rch visibility? Research Evaluation, 3(1), 19–24.
Bornma nn, L et a l. (2012). Wha t fa ctors determine cita tion counts of publica tions in Chemistry besides their qua lity? Journal of Informetrics, 6 (1), 11–18.
Cha mbers, JM (1993). Grea ter or lesser sta tistics: a choice for future resea rch. Statistics and Computing, 3, 182-184.
Da s, PK a nd Pa l, JK (2012). Scientometric eva lua tion of Sankhya- the India n Journa l of Sta tistics.
Malaysian Journal of Library & Information Science, 17(2), 83 – 100.
Da s, PK a nd Pa l, JK (2013). An eva lua tive study on cita tion pa tterns of Sankhyā. International Journal of Library Science, 9(3), 23 – 38.
Da vidia n, M a nd Louis, TA (2012). Why sta tistics? Science, 336 (6077), 12. Ava ila ble a t the: science.sciencema g.org/content/336/6077/12.abstract.
Didega h, F a nd Thelwa ll, M (2013). Which fa ctors help a uthors produce the highest impa ct resea rch? colla bora tion, journa l a nd document properties. Journal of Informetrics, 7 (4), 861–873.
Egghe, L et a l. (2000). Methods for a ccrediting publica tions to a uthors or countries: consequences for eva lua tion studies. Journal of the American Society for Information Science, 52 (2), 145-157.
Fienberg, SE (2014). Wha t is sta tistics? Annual Review of Statistics and Its Applications, 1, 1–9.
Fra me, JD a nd Ca rpenter, MP (1979). Interna tiona l Resea rch Colla bora tion. Social Studies of Science, 9, 481-497.
Georghiou, L. (1998). Globa l co-opera tion in resea rch. Research Policy, 27, 611–626.
Ghosh, JK et a l. (1999). Evolution of sta tistics in India . International Statistical Review, 67, 13-34. Google Schola r a va ila ble a t https://scholar.google.co.in/
Ha rirchi, G et a l. (2007). An explora tory study of the fea ture of Ira nia n co -a uthorship in Biology, Chemistry a nd Physics. Scientometrics, 72(1), 11-24.
Ha sla m, N et a l. (2008). Wha t ma kes a n a rticle influentia l? Predicting impa ct in socia l a nd persona lity psychology. Scientometrics, 76(1), 169.
Ibá ñez, A et a l. (2013). Rela tionship a mong resea rch colla bora tion, number of documents a nd number of cita tions: a ca se study in Spa nish Computer Science production in 2000 –2009. Scientometrics, 95(2), 1–28.
Ka tz, JS a nd Ma rtin, BR (1997). Wha t is resea rch colla bora tion? Research Policy, 26(1), 1-18.
La wa ni, SM (1980). Qua lity, Colla bora tion a nd Cita tions in Ca ncer Resea rch: A 268 Bibliometric Study. Ph.D, Disserta tion, Florida Sta te University; 1980.
Leimu, R a nd Koricheva , J (2005). Does scientific colla bora tion increa ses the impa ct of Ecologica l a rticles? BioScience, 55(5), 438.
Levitt, JM a nd Thelwa ll, M (2010). Does the higher cita tion of colla bora tive resea rch differ from region to region? A ca se study of Economics, Scientometrics, 85(1), 171–183.
MATHSCINET(AMS). Ava ila ble a t: https://mathscinet.ams.org/mathscinet/.
Na udé, F (2016). Country trends a nd schola rly colla bora tion in the ICT4D resea rch community 2000 -2013: A single journa l study. Electronic Journal of Information Systems in Developing Countries, 72(4), 1-26.
Pa tra , SK (2014). Google Schola r-ba sed cita tion a na lysis of India n Libra ry a nd Informa tion Science journa ls. Annals of Library and Information Studies, 61(3), 227–234.
Ra o, BLSP (2006). Sta tistics a s a discipline in India . Electronic journal for History of probability and Statistics, 2(1), 7.
Rudra , A. Pra sa nta Cha ndra Mahala nobis: a biogra phy, Oxford University Press; Delhi, 1996, p.268-270.
Sankhyā journa l website: http://Sa nkhyā.isica l.a c.in
Slyder, JB et a l. (2011). Cita tion pa ttern a nd lifespa n: a compa rison of discipline, institution a nd individua l.
Scientometrics, 89(3), 955–966.
Soorya moorthy, R (2009). Do types of colla bora tion cha nge cita tion? Colla bora tion a nd cita tion pa tterns of South Africa n science publica tions. Scientometrics, 81(1), 177–193.
Stigler, SS (1994). Cita tion pa tterns in the journa ls of sta tistics a nd proba bility. Statistical Science, 9(1), 94-108.
Subra ma nya m, K (1983). Bibliometric studies of resea rch in colla bora tion: a review. Journal of Information Science, 6 (1), 33-38.
Theoha ra kis, V a nd Skordia , M (2003). How do Sta tisticia ns perceive sta tistics journa ls? The American Statistician, 57(2), 115-123.
Va n-Hooydonk, G (1997). Fra ctiona l counting of multi-a uthored publica tions: consequences for the impa ct of a uthors, Journal of the American Society for Information Science, 48 (10), 944-945.
Visa khi, P a nd Sriva sta va , SS (2002). Current trend of resea rch colla bora tion in the field of sta tistica l science - a ca se study. IASLIC Bulletin, 47 (4), 210-215.
