Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Annalsof Libraryand InformationStudies
Vol.53, September2006, pp. 114-125
Scientometric profile of Indian science as seenthrough Science Citation Index
K C Garg, B Dutt and Suresh KumarNational Institute of Science, Technology and Development Studies (NISTADS),
Pusa Gate, Dr. K.S. Krishnan Marg, New Delhi 110012
Email: [email protected]
An analysis of 11067 papers published by Indian scientists and indexed by Scie/lce Citatio/l!Ili!ex (SCI) CD-ROM for the year
1997 indicates that academic institutions (universities and colleges) are the major contributors to the scientific publications
output. Major contribution came from 29 institutions, which contributed about 45% of the total Indian scientific output. Based
on the values of different impact indicators (normalized impact per paper, publication effective index and relative quality
index) TIFR outperforms all other institutions on different impact indicators. Fifty seven percent of the output is concentrated
in physical sciences, chemical and medical sciences. Indian scientists widely publish their findings in journals published from
the scientifically advanced countries of the West. Based on these values of the Normalized Impact Factor (NIF), it is observed
that about two-third of the total papers have appeared in low and medium NIF journals.
Introduction
Since independence India has developed a large
infrastructure for science and technology with the settingup of several universities and academic institutions,national laboratories, and autonomous research
institutions. Presently there are more than 200universities, 400 national laboratories, 1300 in-house
R&D institutions of the industry and 200 voluntaryorganizations with S&T involvement. In almost everydiscipline, India's map is dotted with institutions thathave capability to carry out research and development'.
However, India's scientific output is constantly on thedecline. According to a study by Arunachalam, Indiamoved down from 8th position in 1980 with 14,983publications to 15th position in 2000 with 12,127publications2• At the same time, China moved upwardsfrom 15th position in 1980 with 924 publications to 9th
position in 2000 with 22,061 publications. However,these included 4307 publications from Hong Kong.Gupta and Garg have pointed out some of the reasons
for this decline3• The present study attempts to gain aninsight as to where, in fact, the decline lies besides
presenting a scientometric profile of Indian scientificoutput based on the papers indexed by SCI CD-ROM
for the year 1997.
Numerous bibliometric studies dealing with assessmentof Indian scientific output in various disciplines and sub-
disciplines of Indian S&T have been published in theliterature4• However, there have been a few bibliometric
studies looking at India's contribution to the scientificliterature based on SC1'6. It is therefore, intended to
assess the contribution of different scientific agenciesto the mainstream scientific literature in different
disciplines of science and technology during 1997 andalso to identify most prolific institutions and theircontribution in five major disciplines, besides studyingthe impact of their research output using different impactindicators.
Data and Methodology
The present assessment of Indian science is based on
the publication data indexed by Science Citation Index(SCI) published by the Institute of Scientific Information
(now part of the Thomson Scientific), USA for the year1997. However, it would have been better to use time
series data to avoid year-to-year fluctuations in thepublication output of indi vidual institutions. The year1997 for the present study has been chosen with a view
to make a comparative assessment of the basic findingswith the earlier study 7. The data for the present studywas downloaded using SCI-CD ROM, while the data
for 1987 study was processed manually resulting in
higher count of publications than actually published asthe collaborative papers were counted as many times asthe number of collaborating institutions. However, for
I II, II '" I I I
GARG K C, DUTT B & KUMAR S : SCIENTOMETRIC PROFILE OF INDIAN SCIENCE 115
type of journal, use of normalized impact factor has been
made. This is so because the papers published by Indian
scientists are scattered in 104 different sub-disciplines
and in 1621 different journals published from different
parts of the world including India.
Indicators of impact
Following indicators of impact suggested by NagpaullO
and used by Garg and Padhi" in their study for laserresearch have been used for inter-field and inter
institution comparisons of quality.
Different authors have suggested different methods to
normalize the impact factor of the journals. According
to Sen, the impact factor of the journal of the sub
discipline is divided by the highest value of the impactfactor in the set of journals of the sub-disciplines.However, in this procedure, one has to exclude reviewjournals as the impact factor of the review journals are
quite high as compared to research journals in the samesub-discipline. Exclusion of review journals on the basis
of relatively very high impact factor is arbitrary,problematic, and has its own methodological fallacies.Therefore this method has not been used to normalize
the impact factor. In view of the limitation of the abovemethod for calculating the normalized impact factor, wehave used another method to normalize the impact factor,
which is a modification of the method suggested byRajagopal and Kumar9• In the present procedure, wecalculated the average impact factor of the journals in
different disciplines and assigned weights as described
below. Using the different weights mentioned below, wecalculated the normalized impact factor for the different
disciplines.
comparison of the data for two periods, the publication
data for the year 1987 has been normalized.
The publication data used in the study included the name
of the journals with their country of origin, impact factor
of the journals as recorded in Journal Citation Reportsfor the year 1997, normalized impact factor of thejournals, disciplines of the journal based on the
classification provided by loan-Iovitz Popescu, Professor
of Physics at Bucharest University and available at http://www.geocities.com/iipopescu/jo_rankingb.htm and thename of the institution where from the paper originated.
Objectives
The data has been analyzed to study the followingaspects:
• Distribution of output according to agency;
• Identification of most prolific institutions and
to study their activity and attractivity profilesin different disciplines namely chemical,
physical, medical, biological, and engineeringSCIences;
• To examine the impact of the research output
of the prolific institutions using different impactindicators as has been described in the
succeeding paragraphs;
• Distribution of output in different disciplinesand to study the impact of their research output;and
• To examine the communication behavior of the
Indian scientists as reflected by the country of
publication of the journals and normalized
impact factor of the journals.
The impact of research has been examined using
Garfield's impact factor as a surrogate measure of qualityas examination of citations of such a huge data is not
only time consuming but also expensive. However, thecitations data would have certainly given a better picture
of impact.
Range of the Impact Factor of the Journals
~ Half the average impact factor
> Half ~ average impact factor
> Average ~ twice the average impact factor
> Twice the average impact factor
Weight
2
3
4
Normalized impact
The impact factor is an indicator of the impact of the
journal and depends upon the average rate of citations
the articles published in it receive. Since impact factor,as suggested by Garfield, varies with the discipline and
Normalized Impact Per Paper (NIMP Ipaper)
Based on the publication pattern and the normalizedimpact factor of the journals in which the research resultswere published, the normalized impact per paper for
prolific institutions and different disciplineSl1ave been,t
116 ANN. LIB. INF. STU., SEPTEMBER 2006
Table 1 - Distribution of scientific output by agency
Agency
1997 (%) Cum. %1987 (%)*Cum. %
AI
4069 (36.8) 36.84075(39.8)39.8
ENGC
1251 (11.3) 48.11229( 12.0)51.8
MEDC
1273 (11.5) 59.61123(11.0)62.8
AGRU
196 (1.8) 61.4339 (3.3)66.1
CSIR
1310 (11.8) 73.2921 (9.0)75.1
ICAR
176 (1.6) 74.8310 (3.0)78.1
ICMR
134(1.2) 76.0187 (1.8)79.9
DRDO
118 (I.l) 77.191 (0.9)80.8
DST
545 (4.9) 82.0316 (3.1)83.9
DOS
163 (1.5) 83.589 (0.9)84.8
DAE
1072 (9.7) 93.2862 (8.4)93.2
Others
760 (6.8) 100697 (6.8)100
Total
11067 10010239100
* Normalized publication data
AI: Universities/deemed universities and colleges, ENGC: Engineering colleges including IITs, MEDC: Medicalcolleges and hospitals, AGRU: Agricultural universities, CSIR: Council of Scientific and Industrial Research,ICAR: Indian Council of Agricultural Research, ICMR: Indian Council of Medical Research, DRDO: DefenceResearch & Development Organization, DST: Department of Science & Technology, DOS: Department ofSpace, DAE: Department of Atomic Energy
calculated. This has been calculated by using the formula
(LPj.Fi ) / N where Pi denotes the number of papers in ith journal, Fj denotes the normalized impact factor ofthe i-th journal, and N denotes the total number of papers.
Number of High Quality Papers (NHQ)
Based on the average normalized impact factor in
different disciplines, those papers have been consideredas high quality papers that have been published in
journals with normalized impact factor more than twice
the average normalized impact factor.
Publication Effective Index (PEl)
The indicator is the ratio of the proportion of the impact
(TNIMP%) to the proportion of the publications(TNP%), where, TNIMP% = (Total normalized impact
for an institution or discipline / Total normalized impact)
x 100 and TNP% = (Total publications output of aninstitution or discipline / Total publication output) x 100.
The measure indicates whether the impact of
publications of an institution or a disciplinecommensurate with the publication effort devoted to it.A value of PEl > 1 indicates that the impact of
publications is more than the research effort and viceversa.
Relative Quality Index (RQI)
This indicator is the ratio of the proportion of high quality
papers (NHQ%) to the proportion of the publications(TNP%), where, NHQ% = (Number of high quality
papers for an institution or a discipline / Total numberof high quality papers) x 100
The measure relates to the incidence of high quality
papers in a field or by an institution. A value ofRQI > 1
ii' II' '''1'11'1' 11111 'I I
GARG K C, DUTT B & KUMAR S : SCIENTOMETRIC PROFILE OF INDIAN SCIENCE 117
indicates higher than average value, whereas the value
of RQI <1 indicates lower than average quality.
Limitations of SCI
The use of SCI to characterize scientific enterprise in
Third World countries has been criticized on the ground
of the inadequate coverage of Third World science
journals in SCI database, as it does not cover a large
proportion of journals published from these countries.
In the present case also, only 11 journals published fromIndia have been covered as source items in the database,
while it included 3453 journals as source item in its
database in the year 1997. Thus, a large proportion ofthe papers published by Indian scientists in other journals
not indexed by SCI database remain uncovered.However, it is imperative to use SCI as no other databaseprovides information on the publication activity on the
entire gamut of science and technology of a country.Though, for studying a single discipline other secondarysources/databases are available.
Results and discussion
Distribution of output according to agency
Bulk of funding to R&D comes from the central
government. Scientific research is mainly performed at
the universities and the institutes of higher learning,besides the government funded laboratories under the
Department of Atomic Energy (DAE), Department ofSpace (DOS), Department of Science & Technology(DST), Defence Research & Development Organization(DRDO), Council of Scientific & Industrial Research
(CSIR), Indian Council of Agriculture Research (ICAR)and Indian Council of Medical Research (ICMR).
Table 1 presents the data on the distribution of scientific
output according to agency. This indicates that academic
institutions (universities and colleges) are the majorcontributors (-37%) to the total output. This is followed
by output by engineering colleges including Indian
Institutes of Technology (IITs), medical colleges /hospitals and CSIR, which have contributed - 11% eachof the total output. DAE and DST contributed -10%
and -5% of the total output respectively. The
contribution by AGRU, ICAR, DOS, ICMR, and DRDOvaried in between 1% to 2% of the total output.
A comparison of the output data for 1997 with that of1987 indicates that there had been a slight increase in
the output in 1997 as compared to 1987. However, there
is a decline in the scientific output in the academic sector
(universities and colleges). Prathapl2 has also
demonstrated this fact in a study wherein he has looked
into the performance of university, industry, and the
government sector in India. The output of universities
and colleges declined from 40% in 1987 to 37% in 1997.
Output in case of AGRU and ICAR has also decreased.
However, output in case of CSIR, DST and DAE have
increased in 1997 as compared to 1987. Decline in caseof universities and colIeges is more pronounced ascompared to AGRU and ICAR.
Profile of prolific institutions
The total output came from 1101 institutions. Of these29 institutions contributed 85 or more papers (Table 2).We have chosen the cut off as 85, because, in 1987 thecut off was 100, which when normalized on the actual
output of publications for 1987 comes out to be 85. The
names of these institutions along with abbreviations usedin the paper are given in Annexure I.
On comparing the ranks of these institutions with highly
productive institutions for 1987, it is observed that thereis no change in the rank for Indian Institute of Science(lIS c), Bhabha Atomic research Centre (BARC) and
Pune University (POONU). However, in case of others
nine have moved up and eight have moved down in therank list. The institutions that have moved up in the rankare denoted by the symbol (i) and those that moveddown are shown by the symbol (J..) in Table 2.
Central Drug Research Institute (Lucknow), RoorkeeUniversity (Roorkee) now Indian Institute ofTechnology), Osmania University (Hyderabad), Punjab
Agriculture University (Ludhiana), Rajasthan University(Jaipur), and Haryana Agricultural University (Hissar)have lost their place in the list of most productiveinstitutes.
New Institutions added in the list are Indian Association
for the Cultivation of Science (lACS), Indian Institute
of Chemical Technology (lICT), Saha Institute of
Nuclear Physics (SINP), Physical Research Laboratory
(PRL), Bombay University (BOMU), Christian Medical
College and Hospital (CMCH), Sanjay GandhiPostgraduate Institute of Medical Sciences (SGPIMS)
and Regional Research Laboratory (RRLT). Following
118 ANN. LID. INF. STU., SEPTEMDER 2006
Table 2 - Publication output (activity index) of premier institutions in different disciplines
* Institutes of national importance.
Figures in bracket give Activity Index (AI). AI rounded off to the nearest whole number.
Institution
IISC*
DARC
TIFR (I)
AIIMS* (I)
l3I-IU 0·)
IITM* (I)
NCL(I)
lACS
IITK* (I)
IITl3* (I)
IITKI-I* (J-)
DU
IITD* 0·)
JADU (I)
I-IYDU Ci)
PG[MER*O·)
MADU (I)
HCT
UCAL(-1.)
PANU (-1.)
PRL
SINP
CMCI-I
l30MU
AMU
SGPIMR*
POONU
RRLT
ISI*
Others
TOTAL
l3iologicalsciences
86(126)
14(32)
33(93 )
33(104)
52(173)
6(22)
24(89)
5( 19)
3(11)
9(360
1(4)
41(176)
12(52)
7(38)
20(1l1)
21(126)
25( 159)
4(26)
18(122)
15(104)
0(0)
4(34)
8(67)
2(18)
15(136)
14(129)
9(84)
4(38)
4(38)
894
1383
Chemical
sCiences
96(82)
65(103)
10(111)
0(0)
26(64)
42(113)
113(310)
94(259)
55(154)
68(201)
23(72)
37(117)
31 (99)
48(191)
55(225)
2(9)
18(84 )
94(447)
11(55)
19(97)
0(0)
11(68)
0(0)
32(210)
24(161)
0(0)
20(137
37(257)
1(7)
846
1878
Engineeringsciences
74(126)
77(208)
6(20)
0(0)
29(112)
74(314)
23(100)
4(17)
64(283)
49(230)
55(275)
5(25)
55(279)
[5(95)
1(6)
0(0)
1(7)
6(45)
13(103)
5(41)
8(78)
8(79)
0(0)
26(270)
12(127)
0(0)
9(98)
10(110)
19(209)
537
1185
Medical
sCiences
10(11)
15(26)
4(90
216(517)
38(96)
0(0)
0(0)
0(0)
0(0)
3(9)
1(3)
11(36)
2(7)
11(45)
7(30)
108(494)
33(159)
1(5)
15(77)
20(106)
0(0)
0(0)
85(544)
0(0)
3(21)
72(503)
0(0)
0(0)
1(7)
1165
1821
PhysicalsCiences
155(119)
131(159)
189(280)
0(0)
52(91)
59(113)
8(16)
86(169)
62(123)
33(70)
35(79)
62(140)
45(103)
58(165)
46(134)
0(0)
21(70)
7(24)
35(125)
43(157)
8 [(354)
68(301)
0(0)
11(51)
11(52)
0(0)
26(127)
15(74)
27(133)
1269
2635
Others
126
44
41
5
44
39
47
25
27
37
72
30
39
9
15
2
28
12
26
13
7
4
2
19
23
22
19
33
1354
2165
Total
547
346
283
254
241
220
215
214
211
199
187
186
184
148
144
133
126
124
118
115
96
95
95
90
88
87
86
85
85
6065
11067
'I 'ii' "'!HH- II I'i 'I I
GARG K C, DUTTB & KUMARS : SCIENTOMETRICPROFILEOF INDIAN SCIENCE II9
paragraphs present details about the activity and the
attractivity profiles of the most prolific institutions.
Activity profile of premier institutions
Activity profile of premier institutions has been
examined by using Activity Index (AI). AI was first
proposed by Framel3 and has been elaborated bySchubert and Braun 14. It characterizes the relative
research effort an institution devotes to a given subject
field. Mathematically,
AI = {(Nij/ Nio) / (Noj / Noo)Ix 100, where
Nij = total numberof publicationsof institutioni in disciplinej;
Nio = total numberof publicationsof institutioni in all disciplines;
Noj = total numberof publicationsof all institutionsin disciplinej;and
Noo = total Indian output in all disciplines.
The value of AI=lOO indicates that the research effort
of an institution in a given discipline correspondsprecisely to the Indian average; AI> 100 reflects higherthan average activity and AI <100 lower than average
effort dedicated to the disciplines.
The major advantage of using activity index overabsolute count of publications is that it takes into accountboth the size of the institution as well as the size of the
discipline.
For calculating the AI for different institutions, the output
for agricultural sciences, environmental sciences,material sciences, mathematical sciences and
multidisciplinary sciences have been clubbed together,as the output in these disciplines is low as compared to
the five major disciplines for which AI has beencalculated.
The absolute number of publications and the values of
AI for different institutions and different disciplines aregiven in Table 2. From the values of AI given in Table 2,it is observed that the distribution of AI is skewed. There
are certain institutions that concentrate their research
effort only in one discipline, while some otherinstitutions distribute their research efforts in more than
one discipline. In the present study 13 institutions namelyTIFR, AIIMS, BHU, IITM, NCL, PGIMER, IICT,
PANU, PRL, SINP, CMCH, SGPIMS, and RRLTmainly
concentrated their research effort only on one discipline,
while the rest 16 concentrated their research effort on
more than one discipline.
As expected, medical colleges, namely AIIMS,PGIMER, CMCH, and SGPIMS concentrated their
research effort mainly on medical sciences. IITs
emphasized their research effort mainly on engineeringand computer sciences. IITK and IITB also emphasizedon research in chemical sciences. In addition to these
institutions, other institutions emphasizing onengineering research are BARC, BOMU and IS!. Theemphasis of IISC is evenly spread in biological,
chemical and physical sciences. The discipline ofemphasis for other institutes can be seen from Table 2.
Attractivity profile of premier institutions
Attractivity profile of premier institutions has beenexamined by using Attractivity Index (AAI). Like the
absolute publication output, the absolute impact is alsoconfounded by the size of the institution and size ofthe field. Hence, Attractivity Index, also suggested bySchubert and Braunl5 has been used. AAI characterizes
the relative impact, the publications of an institution
make in a given discipline as reflected by the citationsthey attract. However, in the present case we have usedthe normalized impact factor for calculating the AA!.
Mathematicall y
AAI = {(Cij / C io) / (C oj / Coo) J x 100 whereC ij: normalizedimpactof the country i in sub-specialtyj;C io: normalizedimpactof the country i in all sub-specialties;C oj: normalizedimpactof all countries in the sub-specialtyj;
.C00; normalizedimpactof all countries in all the sub-specialties.
AAI =100 indicates that country's citation impact in the
given field corresponds precisely to the world's average,AAI > 100 reflects higher than average, and AAI < 100lower than average.
The values of total impact and AAI for differentinstitutions and different disciplines are given in Table
3. From the values of AAI given in Table 3, it is observed
that like the activity index, medical colleges has thehighest values of AAI in medical sciences. However, incase of AIIMS the value of AAI is also high for biological
sciences. Similarly, like the AI, the highest value of AAI
for IITs is in the discipline of engineering and computersciences. Values for AAI for other institutes can be seenfrom Table 3. Based on the values of AI and AAI in
120 ANN. LIB. INF. STD., SEPTEMBER 2006
Table 3: Normalized impact (attractivity index) of premier institutions in different disciplines
Institutions Biologicalsciences
Chemical Engineering..sCiences sCiences
Medical
sciencesPhysicalsCiences
Others Total
261(114) 175(113)
156(113) 179(193)
26(19) 14(15)
0(0) 0(0)
40(54) 56(112)
84(106) 167(312)
306(319) 67(103)
236(273) 11(19)
151(171) 146(244)
184(227) 132(241)
53(81) 115(260)
59(101) 14(35)
74(108) 115(249)
99(191) 40(114)
161(261) 4(10)
2(5) 0(0)
40(113) 1(4)
231 (463) 14( 41)
20(50) 27(99)
37(89) 13(46)
0(0) 22(79)
31(75) 19(68)
0(0) 0(0)
60(168 72(298)
37(133) 22(117)
IISC
BARC
TIFR
AIIMS
BHU
IITM
NCL
lACS
IITK
IITI3
IITKH
DU
IITD
JADU
HYDU
PGIMER
MADU
IICT
UCAL
PANU
PRL
SINP
CMCH
BOMU
AMU
SGPIMR
POONU
RRLT
ISI
Others
TOTAL
222(141)
26(28)
105(112)
80(135)
88(174)
14(26)
46(70)
8(13)
3(5)
17(31)
1(2)
81(202)
22(47)
9(25)
41(97)
46(172)
42(173)
5(15)
33(120)
24(84)
0(0)
10(35)
18(9 I)
2(8)
30(157)
37(178)
18(76)
6(27)
6(32)
1694
2734
0(0)
56(162)
85(264)
1(4)
1492
3982
0(0)
20(86
26(119)
38(208)
1182
2691
24(12)
24(20)
6(5)
412(541)
64(98)
0(0)
0(0)
0(0)
0(0)
9(13)
2(30
17(33)
3(5)
20(44)
16(29)
172(501)
45(1450
1(2)
22(62)
28(76)
0(0)
0(0)
143(565)
0(0)
8(33)
137(513)
0(0)
0(0)
2(8)
2350
3505
349(106)
309(156)
566(287)
0(0)
99(93)
108(95)
14(10)
182(146)
154(121 )
58(50)
63(670
117(1390
95(96)
112(150)
99(1110
0(0)
27(53)
9(13)
69(119)
118(196)
208(352)
172(287)
0(0)
28(54)
20(50)
0(0)
55(111)
28(60)
52(134)
2619
3730
295
99
73
7
79
84
120
62
56
68
143
49
86
19
35
5
49
28
61
21
7
8
5
44
44
50
41
57
2631
4327
1326
793
790
499
426
457
553
499
510
468
377
337
395
299
356
225
204
288
232
241
237
240
166
206
161
175
199
186
156
11968
22969
Figures in bracket give Attractivity Index (AAI). AAI rounded off to the nearest whole number.
,1,1 I itl I II
GARG K C, DUTT B & KUMAR S : SCIENTOMETRIC PROFILE OF INDIAN SCIENCE
Table 4 - Institutions with AI > 100 and AAI > 100 in different disciplines
121
Disciplines
Chemical
Sciences
PhysicalSciences
Medical
Sciences
Institutions with AI > 100 and AAI >100
IISC, BARC, IITM, NCL, lACS, IITK, IITB,DU, JADU, HYDU, IICT, BOMU, AMU,
POONU,RRLT
IISC, BARC, TIFR, lACS, IITK, DU, JADU, HYDU,UCAL, PANU, PRL, SINP, POONU, ISI
AIIMS, PGIMER, MADU, CMCH, and SGPIMR
Remarks
HYDY has less effort, but more impact.BOMU,AMU and POONU have less impact, but more effort.
PANU has less effort, but more impact.
AIIMS and CMCH have more impact and less effort.
Biological IISC, AIIMS, BHU, DU, PGIMER,effort, but more impactSciences MADU, UCAL, AMU, and SGPIMR
AIIMS, DU, PGIMER, AMU and SGPIMR have less
EngineeringSciences
IISC, BARC, BHU, IITM, IlTK, IITB, IlTKH,IITD, BOMU, AMU, RRLT and ISI
IIITK and IITD have more effort, but less impact.BOMU has less effort, but more impact
different disciplines, we identified institutions for whichthe values of AI and AAI are more than 100. The names
of such institutions are given below in Table 4.
Impact of output of prolific institutions
The 29 institutions (Table 5) account for 45% of all
publications, 48% of the total impact, and 46% of all
high quality papers published by India. Majority of the
papers published by these institutions have appeared injournals originating from the scientifically advancedcountries of the west. This indicates that the research
performed at these institutions evoke considerable
interest among the western scientific community, besidesthe Indian scientists preferring to publish in overseasjournals.
Table 5 provides information about various impactindicators, such as normalized impact per paper,
publication effective index and relative quality index.
The average NIMP/paper for Indian publications is 2.1.Of the 29 institutions (Table 5), 13 institutions have
NIMP/paper more than the Indian average and for seven
institutions the NIMP/ paper is less than the Indianaverage. For the remaining nine, the NIMP/Paper is equalto the Indian average. Further analysis of the dataindicates that TIFR had the highest NIMP/ paper (2.8)
followed by NCL (2.6) and HYDU, PRL and SINP eachbeing 2.5. Like the NIMP/paper, the value of PEl is also
highest for TIFR closely followed by IISC, NCL, IlTK,HYDU, PRL, and SINP. Other institutions for whichthe value of PEl> 1 are BARC, IITB, IICT, BOMU,
and POONU. It implies that these institutions earn more
impact than that is commensurate with their publicationeffort. The impact of IITM, IITD, JADU, PANU,SGPIMR, and RRLT is just commensurate with theirpublication effort. The standing of different institutions
on the basis of incidence of high quality papers can be
judged from the values of RQI. Here also TIFR had thehighest value (3.4) for RQI, followed by SINP (1.9),AIIMS, NCL and PRL each being 1.6 and SGPIMS(1.5). HYDU, PANU, CMCH, IISC, and PGIMER have
also RQI > 1. This implies that these institutions havemore than average incidence of high quality papers andthe rest have less than average incidence of high quality
papers.
Distribution of output according to discipline and their impact
SCI includes about 3500 journals in its database as
source items covering a wide range of subjects.However, for the sake of convenience, we have grouped
them into seven major disciplines excluding others. Thescheme suggested by loan-Iovitz Popescu, Professor of
Physics at Bucharest University has been used for
classification of the output d~ta.
Data presented in Tablb'(jj'iridi6'atesthafthe share of~'.¥ _ ' .. ~., I ,.-, " '_.~' r _ " '>:, .'
physical sciences including geosciences constitute
122 ANN. LII3. INF. STU., SEPTEMBER 2006
Tab]e 5 - Impact indicatorsfor prolificinstitutions
Institutions
TNPTNIMPNIMP/ PAPERNHQPElRQ[
IlSC
54713262.4 551.21.2
BARC
3467932.3 19l.l0.7
TlFR
2837902.8 791.33.4
AIlMS
2544992.0 330.91.6
I3HU
241426 1.8 100.80.5
IlTM
2204572.1 51.00.3
NCL
2155532.6 291.21.6
lACS
2144992.3 8l.l0.5
IlTK
2115102.4 181.21.0
IlTB
199468 2.4 12l.l0.7
IITKH
[87377 2.0 30.90.2
DU
[86337 1.8 90.90.6
liTO
184395 2.1 71.00.5
JADU
148299 2.0 91.00.7
HYDU
1443562.5 171.21.4
PGIMER
133225 1.7 120.81.1
MADU
126204 1.6 I0.80.]
IlCT
124288 2.3 41.10.4
UCAL
118232 2.0 20.90.2
PANU
115241 2.1 131.01.4
PRL
962372.5 131.21.6
SINP
952402.5 151.21.9
CMCH
951661.7 II0.81.4
BOMU
902062.3 4l.l0.5
AMU
881611.8 30.90.4
SGPIMS
871752.0 II1.01.5
POONU
861992.3 51.10.7
RRLT
851862.2 51.00.7
ISI
851561.8 40.90.6
OTH
6065119682.0 4870.91.0
TOTAL
11067229692.1 9031.01.0
TNP : Total numberidfpublications; NIMP: Normalised Impact; NHQ : No. of high quality papers;
PEl: Pubijcation effective index; RQI: Relative quality index
'I "" I "'1"11'1' 111" " I Itl; I I!
GARG K C, DUTT B & KUMAR S : SCIENTOMETRIC PROFILE OF INDIAN SCIENCE
Table 6 - Distribution of output and impact according to disciplines
Discipline TNPTNIMPNIMPNHQPElRQI/Paper
Agro-sciences including environmental sciences
61814282.3351.100.70
Biological sciences
138327342.01200.951.10
Chemical sciences
187839822.11241.000.80
Engineering sciences
118526912.3641.100.66
Materials sciences
86020032.3461.100.65
Medical sciences
182135051.92480.931.67
Physical sciences including geosciences
263557302.22541.101.20
Others
6878961.3 120.630.21
123
about one-fourth of total output followed by the outputin the chemical and medical sciences. These three
disciplines together constitute about two-thirds of the
total output. Biosciences and engineering constitute 12% and - 10% of the total output respectively. Rest ofthe 6% output is scattered in multidisciplinary sciencesincluding mathematics and has been named as others.
Analysis of data for different impact indicators indicatesthat NIMP/paper for different disciplines does not differ
significantly except for others, which is quite low ascompared to the national average (2.1). The value of PEl< 1 for biosciences and medical sciences. It indicates
that in these disciplines the impact of the research is not
commensurate with the publication effort. However, thevalue of RQI for these two disciplines as well as forphysical sciences> 1, which implies that these disciplineshave more than average incidence of high quality papers.
Communication pattern of Indian scientists
This has been examined by using two aspects i.e. the
country of origin of the journals where the Indianscientists have published their research results and the
normalized impact factor of the journals.
Country of origin of joumals
Indian researchers publish their work in journalspublished from many countries. An analysis of the dataindicates that about 89% of the papers have been
published in journals originating from countries otherthan In.dia. Of these, 4553 (41 %) papers are in journals
originating from the US and 1986 (18%) papers are in
journals published from the UK. Thus, almost 59% ofthe papers in different disciplines have appeared in thejournals from the US and the UK. A comparison of thedata with the choice of journals used by Indian scientists
with 1987 indicates that proportion of papers in journalsoriginating from foreign countries have increased from80% in 1987 to 89% in 1997. Further, the share of
publications for the US has gone up from 26% in 1987
to 41 % in 1997. However, the share of papers published
in journals from the UK has gone down. One of thepossible reasons for a steep hike in number of paperspublished from USA may be inclusion of new US
journals in the SCI database or a natural preference ofIndian scientists to publish in US journals.
The number of papers in journals published from theNetherlands are 1295 (12%) and from Germany 509(5%). The number of papers in journals originating fromother countries is Switzerland (464), Japan (155),Denmark (134), France (105), Ireland (102), Hungary
(94), Singapore (90), Czechoslovakia (54), Sweden(47), Canada (45), Italy (42), Poland (39), Australia
(34), Israel (22), Austria (16), Spain (12) and other 11. countries, 39. If the journals published from thescientifically advanced countries of the West in various
disciplines are considered as leading science journals,
then the papers published fro~ India in these journalscould be rated as having attained the internationalstandard.
The number of papers that appeared in Ind~an journals
covered by SCI in the year 1997 is 1191 (11%). The
124 ANN. LIB. INF. STU., SEPTEMBER 2006
low number of papers in Indian journals is due to theinadequate coverage ofIndianjournals in SCI database.
In 1997, SCI covered only 11 Indian journals as source
items in its database. Besides, other important reasons
making Indian scientists to publish in foreign journals
are seeking international recognition, wider readership
offered by international journals, existing 'value system'
in Indian science giving more weight to papers published
in overseas journals and non-availability of specialist
journals in many disciplines in the country.
Distribution of output according to NIF of the journals
Based on the average value of the normalized impactfactor (2.1), the NIF has been divided into four
categories. These are> 0 ::; 1.05 (low), > 1.05 ::; 2.1
(medium), >2.1::; 4.2 (above medium) and> 4.2 (high).Based on these values of the NIF, it is observed that
about two-third of the total papers have appeared in
'low' and 'medium' NIF journals and also the proportion
of papers in low and medium NIF journals is almost the
same. Remaining one-third of the total papers have
appeared in 'above medium' and high NIF journals, out
of which, 25% are in above medium NIF journals andthe rest in 'high' NIF journals.
Findings
The study indicates that academic institutions are the
major contributors to the publications output. However,
the output for the universities as weIl as agriculturaluni versities has gone down in 1997 as compared to 1987.The output for CSIR, DST and DAE has increased in
1997 as compared to 1987. Major contributions camefrom 29 institutions that contributed about 45% of all
publications, 46% of all high quality papers publishedby India and 48% of the total impact. Based on the values
of different impact indicators (normalized impact perpaper, publication effective index and relative qualityindex) TIFR outperforms all other institutions ondifferent impact indicators. Physical sciences, chemicalsciences, and medical sciences constitute about 57% of
the total output. The output is low in mathematical andenvironmental sciences. Bioseiences and medical
sciences have more than average incidence of highquality papers, however, based on the values of PEl,
these disciplines have a relatively low impact as
compared to the publication effort. In addition to thesetwo disciplines, the physical sciences also have more
than average incidence of high quality papers.
Indian scientists widely publish their findings in journalsbrought out by the advanced countries of the West and
the proportion of papers published in journals from theWest have increased from 80% in 1987 to 89% in 1997.
However, this does not imply that these papers are also
cited frequently. Two-third of the total output has
appeared in low and medium NIF journals. Only a
minuscule portion (8%) of the papers has appeared inhigh NIF journals.
Acknowledgements
The authors are highly thankful to Dr. Gangan Prathap,Scientist-In-Charge, CSIR Centre for Mathematical
Modelling and Computer Simulation, Bangalore, for his
valuable comments and suggestions on the manuscript
that have been of immense help in improving themanuscript.
References
1. Rama Rao P, Science and technology in independent India:
Retrospect and prospect, Current Science, 74 (1998) 418-432.2. Arunachalam S, Is Science in India on the decline?, Current
Science, 83 (2002) 107-108.
3. Gupta B M and Garg K C, Is science in India on decline? A
rejoinder, Current Science, 83 (2002) 1431-32.4. Garg K C, An overview of cross-national, national, and
institutional assessment as reflected in the international journal
Scientollletrics, Scientollletrics, 56 (2003) 169-199.5. Garg K C and Dutt 13, Bibliometrics of Indian science as
reflected through Science Citation Index, Joumal of ScientificU1ullndustrial Research, 51 (1992) 329-340.
6. Arunachalam S, Srinivasan R and Raman Y, Science in India
a profile based on India's publications as covered by ScienceCitation Index 1989-1992, Current Science, 74 (1998), 433441.
7. Garg K C and Dutt 13, 13ibliometrics of Indian science as
reflected through Science Citation Index, Joumal of Scientificand Industrial Research, 51 (1992) 329-340.
8. Sen B K, Normalized impact factor, Joumal of Doculllentatio/I,48 (1992) 318-325.
9. Raja Gopal Y and Kumar M P R, Standard of scientific research
publications, Current Scil;nce, 88 (2005) 207-208
10. Nagpaul P S, Contribution of Indian universities to themainstream scientific literature: a bibliometrie assessment,
Scientollletrics, 32(1995) 11-36.
II. Garg K C and Padhi P, Scientometric of laser research literature
as viewed through the Joumal of Current Laser Abstracts,Scientollletrics, 45( 1999) 251-268.
12. Prathap G, Indian science slows down - Y: the slack in the
university sector, Current Science, 87 (2004) 732-734.
13. Frame J 0, Mainstream research jn Latin America and the
Caribbean, lnterciencia, 2 (1977) 143-148.14. Schubert A and Braun T, Relative indicators and relational
charts for comparati~e' assessment Of publication output and
citation impact, Scientollletric.\', 9'(1986) 281-291.15. Schubert A and Braun T, ibid.
I I III I il
GARG K C, DUTT B & KUMAR S : SCIENTOMETRIC PROFILE OF INDIAN SCIENCE
LIST OF INSTITUTIONS
1. IISC: Indian Institute of Science, Bangalore
2. BARC: Bhabha Atomic Research Centre (DAE), Mumbai
3. TIFR: Tata Institute of Fundamental Research (DAE), Mumbai
4. AIIMS: All India Institute of Medical Sciences, New Delhi
5. BHU: Banaras Hindu University, Varanasi
6. IITM: Indian Institute of Technology, Chennai
7. NCL: National Chemical laboratory (CSIR), Pune
8. lACS: Indian Association for the Cultivation of Science (DST), Kolkatta
9. IITK: Indian Institute of Technology, Kanpur
10. IITB: Indian Institute of Technology, Mumbai
11. IITKH: Indian Institute of Technology, Kharagpur
12. DU: University of Delhi, Delhi
13. IITD: Indian Institute of Technology, New Delhi
14. JADU: Jadavpur University, Kolkatta
15. HYDU: Hyderabad University, Hyderabad
16. PGIMER: Postgraduate Institute of Medical Education and Research, Chandigarh
17. MADU: Madras University, Chennai
18. IICT: Indian Institute of Chemical Technology (CSIR), Hyderabad
19. UCAL: University of Calcutta, Kolkata
20. PANU: Panjab University, Chandigarh
21. PRL: Physical Research Laboratory (DOS), Ahmedabad
22. SINP: Saha Institute of Nuclear Physics (DAE), Kolkatta
23. CMCH: Christian Medical college and Hospital, Vellore and other centers
24. BOMU: Mumbai University (now Mumbai), Mumbai
25. AMU: Aligarh Muslim University, Aligarh
26. SGPIMS: Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow
27. POONU: Pune University, Pune
28. RRLT: Regional Research Laboratory (CSIR), Trivandrum
29. ISI: Indian Statistical Institute, New Delhi and Kolkata
Annexure I
125