Jointly published by Elsevier Science Ltd, Oxfi)rd and Akaddmiai Kind6, Budal)est
Scientometrics, Fol. 39, No. 3 (1997) 293-314
COLLABORATION PROFILE OF THEORETICAL POPULATION GENETICS SPECIALITY B. M. GUPTA,* SURESH KUMAR,* C. R. KARISIDDAPPA**
* Scientometric~" and hTJbrmetrics Group, National Institute of Science, Technology and Development Studies, Dr I'L. S. Krishnan Marg, New Delhi 110012 (lndi& ** Deparm~ent Of Library and lnfi~JTnation Science, Karnataka Univerxity, Dharwad 580003 (India) (Received March 19, 1997) Traces the growth of collaborated and funded research as reflected in research papers in theoretical population genetics research speciality from 1916 80 through a case study. Analyses the proportion and extent of collaborated papers, averge number of authorship per paper, and collaborative coefficient index of research papers thereby giving an overall perspective of the growth o f professionaiism in the field. Studies the relation between co/iaboration, productivity, and funding of research papers in theoretical population genetics. Classifies the total collaborative papers/authors by type of collaboration and studies the trends and shifts in the nature and type of collaborative research over the years.
Introduction Collaboration can be defined as "a process of functional interdependence between scholars in their attempt to coordinate skills, tools and rewards, t However, the complex nature of human interaction that takes place between collaborators and the magnitude of their collaboration are not easily captured by quantitative tools. For example, the precise relationship between quantifiable activities (e.g. data analysis) and intangible contributions (e.g. ideas) and their weightage in the final product of the collaboration (e.g. a research paper) is extremely difficult to determine. Science indicators, however, provide additional quantitative information of a more direct and objective nature to be used for identification, comparison, and evaluation of the nature, magnitude and geographical patterns of cooperation among scientific institutions. 2 Although collaboration has existed in science since its beginning, however the first collaborative paper was published in 1655. Since then, the number of collaborative papers has increased, initially slowly, and then dramatically after the middle of the eighteenth century. The collaborative linkages further increased towards the end of this
013~9130/97/US $15.00 Copyright 9 1997 Akaddmiai Kiad6, Budapest All rights" reserved
B. M. GUPTA,S. KUMAR,C. R. KARASIDDAPPA: COLLABORATIONOF POPULATIONGENETICS
century. The present century has shown increasing growth of international collaboration as against national collaboration in earlier centuries. The significant increase in collaboration is also matched with simultaneous increase in funding available to research projects. The earlier research studies emphasised on different determinants of scientific collaboration. They employed three sets of factors: economic, cognitive, and social to explain collaboration. The relative importance of these factors varies depending on the type of analysis and the kind of aggregates one is studying. The factors indicated above can be categorised as internal to science or scientific enterprise. There may be other factors which can be considered as external to science but nonetheless enhance scientific collaboration. Presently, many governments in different countries have taken initiatives to enhance contacts among scientists in science through collaborative research programmes, both at the national and the international levels. Such initiatives and programmes do have their impact in increasing the collaboration at national and international levels in scientific research. In a comprehensive study, B e a v e r and R o s e n , have pointed out that the collaboration came forward or advanced in response to professionalism and increased knowledge, by analysing publications data from 17th century through the 20th century. The professionalism refers here to a complex of dynamic processes involving organisations and scientists, individually and collectively, with respect to both the scientific group and society at large. It also sets criterion for the new entrants, and formalizes rules of behaviour among members, and promotes interaction between the specific groups and outsiders. Some indicators of the professional maturation of an individual or research field can take the form of initiating professional societies, specialised journals, innovative applications and institutionalization of prizes and awards. The society, in turn also recognises the contribution made by an individual scholar or profession either through increased funding or research allocation or by taking professional advice from professional scientists in the fields of policy making and advisory capacities. B e a v e r and R o s e n 3-5 maintained that the collaboration functions as a social regulator within a profession, through its influence on recognition and subsequent access to resources, thereby increasing the visibility and productivity of authors.
Existing literature There has been an increasing trend towards collaboration in almost all the fields of science and technology over a period of time, but it was found to vary from one subject to another, one country to another, one branch to another within the same subject. In the
294
Scientometrics 39 (1997)
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA: COLLABORATION OF POPULATION GENETICS
field of mathematics, 6 94% of the papers in 1940 were single authored and their proportion remained at a relatively high level of 79% in 1960. In contrast, in the field of astronomy, the overall decline in the percentage of single authored papers has been fairly gradual. According to Meadow and Connor, 7 the proportion of single authored papers stood at 95% in 1910 and had fallen to 86% in 1940. It declined to 70% in midfifties and to less than 50% by late sixties. Price 6 in a study of Chemical Abstracts, 1910-60 indicated that multi-authored papers were increasing over the years. Among them, three author papers were accelerating more rapidly than two-author papers, fourauthor papers more rapidly than three-author papers, etc. Price s also suggested that the extent of co-authorship in a field was directly linked to the size of financial support available to research in a subject area. He also concun'ed that the collaboration reflects more economic than intellectual dependence. Price 9 later suggested that the number of authors per paper was directly proportional to the size of grant or funding at work, as a grant allows a researcher to here collaborators "who become co-authors as 'an almost fiscal' artifact of funding". The contribution of funding is normally directly proportional to increase in collaboration which has also been reflected and confirmed in some research studies. Hirsch and Singleton 10 found that the papers with financial support, in both the American Sociological Review and Social Forces, had more authors per paper than those without support. Patel I analysed 7908 papers from four sociological journals and found that one in every four multiauthored papers was funded, as compared to one in every 14 single authored paper. Heffner ll investigated four disciplines and found that funding was statistically significant with multiple authorship in chemistry and biology, but not for political science and psychology. Another study on biochemistry and chemical engineering endorsed, Haffner's view, indicating that multi-authored papers in both fields were more supported by grants than single-authored papers. 12 Hart, Carstens, La Croix and May 13 also endorsed the same view in the field of library and information science. They found that the authors of funded research publications are more collaborating in their work than the unfunded authors. One of the major impacts of collaboration on scholarly research is the increase in productivity associated with multiple authorship. Beaver and Rosen 4 have reported that scientific collaboration was a key element in the advancement of knowledge and the productivity of researchers in a field. Price, 14 in one study, found that out of 555 authors in 1966, high producers with 14 or more papers were those who was most collaborative . Those who worked alone or with only one scientist were unable to produce more than four papers in the same period. Zuckerman,15 on a study of Nobel laureates, showed a high correlation between collaboration and productivity. Similarly,
Scientometrics 39 (1997)
295
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA: COLLABORATION OF POPULATION GENETICS
Pao 16 had found that only 38% of sample authors took part in co-authored activities, the most collaborative were also the most productive in computional musicology. Pao, 17 in another study, made a comparison between funded and non-funded coauthored papers in the field of schistosomiosis and found that the average productivity of funded collaborators was much higher than that of non-funded ones. Collaboration in research can take a variety of paths, but necessarily takes the form either of cooperation between near equals, or between senior and junior scientists employed in the same or different organisations. Depending upon the type of participants, their status and location, etc., Subramanvam 12 has identified the following types of collaboration: (a) Teacher-pupil collaboration; (b) Collaboration among colleagues; (c) Supervisor-assistant collaboration; (d) Researcher-consultant collaboration; (e) Collaboration between organisations; (t) International collaboration. Qin 18 further simplified this classification and indicated the following types of collaboration: (a) Collaboration in a department within the institution; (b) Collaboration between two or more departments within the institution; (c) Collaboration between two or more institutions within a country; and (d) International collaboration. Qin also studied the type of collaboration in scientific research papers and the number of authors involved by taking data fi'om Philosophical Transactions of Royal Society of London for the period 1901-1991. Kundra 19 has studied type of collaboration in Indian medical research during 1900-1945. Among the types of collaboration, the international collaboration has been the most widely studied. Many prominent authors, namely Frame and Carpenter, 2~ Tijssen and Moed, 21 Luukkonen, 22 Schubert and Braun, 23 etc. have worked on several aspects of international collaboration. Frame and Carpenter using Science Citation Index data revealed that the more basic the field, the greater the proportion of international coauthorship. They also indicated that the extent of international collaboration is inversely, proportional to the size of a country's scientific enterprise and the geographical, political, and language factors played a stronger role in determining who collaborates with whom in the international scientific community. Keeping the above perspective ih view, the present study aims to: 1. Understand the trends in the growth of collaborative and funded research, as reflected in research papers. 2. Investigate the relation between collaboration, funding productivity of authors. 3. Investigate at what productivity level the authors collaborate among themselves. 4. Analyse the trends and shifts in nature and type of collaborative research.
296
Scientometrics 39 (1997)
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA: COLLABORATION OF POPULATION GENETICS
Database
The paper analyses literature in the field of theoretical population genetics as well as the companion fields of theoretical quantitative genetics and statistical human genetics. The data being considered here for analysis were taken from research papers on theoretical population genetics appearing in the international journal Genetics and listed in a comprehensive bibliography Bibliography of Theoretical Population Genetics, compiled by Felsensteen (Pennsylvania, USA; Dowden, Hutchinson & Ross, 1981). 24 This bibliography is considered as one of the most comprehensive printed source of information on theoretical population genetics. An analysis of this bibliography indicates that the most important core journal in this field is Genetics. This journal is international in coverage and was initiated in as early as 1916. In order to get a representative sample of developments and papers in the field, we have therefore decided to analyse all papers published in Genetics since its inception in the area of theoretical population genetics. The duration for study was from 1916-1980. The articles in a five years block were cumulated beginning in 1916 was taken as sample. In each block, the total number of papers and authors were recorded according to the type of collaboration and the availability of funding. Each paper has been classified under the following types of collaboration: (a) Type 1-Intra-departmental collaboration: collaboration with in a department within an organisation or institution; (b) Type 2-Inter-departmental collaboration: collaboration between two or more departments within an organisation or institution; (c) Type 3-Inter-institutional collaboration: collaboration between two or more organisations or institutions across cities in a country; (d) Type 4 - Inter-institutional collaboration: collaboration between two or more organisations or institutions within a city in a country; (e) Type 5 - International collaboration: collaboration between two or more organisations or institutions across countries. For discussion, we have further simplified the above classification and will now refer to generally three types of collaborative relationships: (1) Internal collaboration (type 1 and 2 taken together), (2) domestic collaboration (type 3 and 4 taken together), and (3) international collaboration. In all 615 research papers were published on theoretical population genetics in the international journal Genetics during 1916-1980. O f these, 367 research papers were single-authored (41.4% of which were funded) and 248 were multi-authored (73.38%
Scientometrics 39 (1997)
297
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA: COLLABORATION OF POPULATION GENETICS
of which were funded). A chronological breakdown of total sample indicates that 28.61% pertain to the period 1976-80. For subsequent periods, the percentage contributions in total sample were: 24.39% (during 1971-75), 19.83% (during 1966-70), 7.96% (during 1961-65), 5.52% (during 1956-60), 2.60% (during 1951-55), 2.43% (during 1946-50), 1.30% each (during 1941-45 and 1936-40), 1.46% (during 1931-35), 0.48% (during 1926-30), 1.62% (during 1921-25) and 2.43 % (during 1916-20). In terms of geographical dispersion, the largest number of papers, 468 came from the USA, constituting 76.09% of the total sample. The next important group of countries in terms of total contribution were: Japan, Australia, the UK, Canada, and Scotland, which when taken together contributed 106 papers, constituting 17.23% of the total sample. The 13 other countries taken together have contributed 41 papers, constituting 6.66% of the total sample. In the last group, FRG and Italy contributed four research papers each; Sweden, France, Denmark and Brazil, three research papers each; New Zealand, India, Puerto Rico and Greece, two research papers each; and finally Israel, Spain, Finland and USSR one research paper each. For simplification of analysis, we have divided the total sample into two groups: A and B. In Group A, we have included only the United States (USA), as it is the major contributor in the total output. In the second Group B, we have kept all the other 18 countries which together have contributed 23.91% to the total research output.
Analysis and results
Funding Table 1 provides data on the growth of total and funded research papers form a block of five years each for the period 1916-80. In the total sample comprising 615 research papers, 53.33% papers were observed to be funded. The funding of research started since 1936 and it systematically increased over the years along with growth of research papers. There exists a strong correlation between the growth of overall research papers and funded research papers, the correlation coefficient being 0.9862. The percentages of funded research papers were 12.50% (during 1936-40) and it increased up to 75% (during 1976-80). The mean average percentage increase in funding of research papers was 12.77% (1936-50), 31.55% (1951-65), and 65.33% (1966-80).
298
Scientometrics 39 (1997)
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA:COLLABORATIONOF POPULATIONGENETICS
Table 1 Growth in total research papers and funded papers Total
Group A USA
Group B Other countries
Period
No. of papers
No. of funded papers
No. of papers
No. of funded papers
No. of papers
No. of funded papers
1916-20 1921-25 1926-30 1931-35 1936-40 1941-45 1946-50 195l 55 1956-60 1961-65 1966-70 1971-75 1976 80
15 10 3 9 8 8 15 16 34 49 122 150 176
0 0 0 0 1 2 2 2 10 26 85 78 123
14 9 2 7 7 8 13 14 24 33 88 110 139
0 0 0 0 l 4 4 2 9 21 68 64 I02
l 1 1 2 1 2 2 2 10 16 34 40 37
0 0 0 0 0 0 0 0 l 5 17 14 21
Total
615
329
468
275
149
58
In Group A (the United States), the financial support to research started earlier (during 1 9 3 6 ~ 0 ) and the rate o f increase and extent o f support to research were observed to be larger than in Group B. In Group A, the percentage increase in funded research papers was from 14.28% (1936-40) to 73.38% (1976-80). The mean average percentage increase in funding for research papers was 31.68% (1936-50), 38.47% (1951-65) and 69.61% (1965-80). In Group B (Other 18 countries), funding o f research started much later (1956-60) than in Group A and it increased from 10.00% (1956-60) to a m a x i m u m o f 56.75% (1976-80). The mean average percentage increase in funding o f research papers was 14.44% (1951-65) and 47.25% (1966-80). For significant countries in Group B, the percentage o f funded papers in total papers was: Japan (52.25%), Australia (28.57%), U K (22.22%), Canada (62.50%) and Scotland (15.38%). There is also a strong correlation between the growth o f overall research papers and funded papers both in Groups A and B, the correlational coefficient being 0.9887 and 0.9631 respectively.
Scientometrics 39 (1997)
299
13. M. GUPTA, S. KUMAR, C. R. K A R A S I D D A P P A : C O L L A B O R A T I O N OF P O P U L A T I O N GENETICS
Proportion of collaborated research p a p e r s The actual proportion of collaborated papers indicates the extent of collaboration in the field at a particular time. The data on the growth o f collaborated research papers is presented in Table 2. Taking the entire sample of 615 research papers, we observe that 243 papers (39.51%) are observed to be collaborative. The percentage of collaborated papers in the total sample has been found to be slowly increasing over the years, and it has doubled from 22.22% (1931-35) to 46.59% (1976-80). The mean average percentage increase in collaborated papers is 25.33% (1936-50), 33.73% (1951-65) and 43.61% (1966-80). A strong correlation is also observed between the growth of total research papers and collaborated papers, the correlation coefficient being 0.9968. Along with increase in the proportional of collaborated papers, the funding support to collaborated research papers also has been found to be increasing. A strong correlation is observed between these two, the correlation coefficient being 0.9927. Out of 242 collaborated research papers, 74.79% were observed to be funded. The increase in funded co-authored research papers has grown from 50.00% (during 1936-40) to 84.14% (during 1976-80). The mean average percentage increase in funded collaborated papers was 50% (during 1936-50), 50.93% (during 1951-65), and 79.83% (during 1966-80). Table 2 Contribution of collaborated research papers Period
Total TP
Group A
TCP
TFCP
TP
Group B
TCP
TFCP
TP
TCP
TFCP
0 0 0 2 0 0 0 1 4 5 18 14
0 0 0 0 0 0 0 0 1 3 7 9 9
55
29
1916 20 1921--25 1926 30 1931-35 1936~,0 1941-45 1946 50 1951-55 1956-60 1961-65 1966 70 1971-75 1976-80
I5 10 3 9 8 8 15 16 34 49 122 150 176
1 0 0 2 2 2 4 5 13 18 54 60 82
0 0 0 0 1 1 2 2 6 12 47 41 69
14 9 2 7 7 8 13 14 24 33 88 110 139
1 0 0 0 2 2 4 4 9 13 43 42 68
0 0 0 0 1 1 2 2 5 9 40 32 60
1 1 1 2 l 0 2 2 10 16 34 40 37
Total
615
243
181
468
188
152
147
11
TP = Total papers; TCP = Total collaborated papers; TFCP = Total funded collaborated papers.
300
Scientometrics 39 (1997)
B. M. GUPTA, S. KUMAR, C. R. K A R A S I D D A P P A : C O L L A B O R A T I O N OF P O P U L A T I O N GENETICS
In Groups A and B, the overall percentage of collaborated papers was observed to be 40.17% (188/468) and 37.41% (55/147), respectively. The increase in percentage of collaborated papers in total papers was observed to be much larger in Group A than in Group B. The increase in percentage of collaborated papers in total papers was from 28.57% (1936-40) to 48.92% (1976-80) in Group A and from 0% (1936-40) to 37.83% (1976--80) in Group B. The mean average percentage increase of collaborated papers in total papers for Group A was 17.66% (1936-40), 34.52% (1951-65) and 45.32% (1966-80) and for Group B 0% (1936-50), 40.41% (1951--65), and 38.39% (1966-80). A strong correlation exists between the growth of total papers and collaborated papers both in Groups A and B, the correlation coefficient being 0.9904 and 0.9835, respectively. The percentage of funded collaborated papers in total collaborated papers in Group A and B was observed to be 80.85% (152/188) and 52.72% (29/55), respectively. The increase in percentage of funded collaborated papers in total collaborated papers was observed to be much faster in Group A than in Group B. In Group A, the increase was from 50% (1936-40) to 88.25% (1976-80), while in Group B, the increase was from 0% (1936-40) to 64.28% (1976-80). The mean average percentage increase in funded collaborated papers of total collaborated papers were: Group A - 50% (1936:-50), 58.26% (1951-65), and 85.81% (1961-65); Group B - 0% (1936-50), 28.33% (1951-65), and 59.30% (1966-80). A strong correlational also exists between the growth of collaborated and funded collaborated papers in both Group A and B, the correlational coefficient being 0.9960 and 0.9806, respectively.
Extent of coauthorship Of the total sample of 615 papers, 372 (60.48%) were single-authored, 165 (26.82%) two-authored, 67 (10.89%) three-authored, 10 (1.62%) four-authored, and 1 (0.16%) five-authored. Table 3 presents data on the extent of single authored and multiauthored papers in three chosen time periods. Table 3 Trends in multi-authored papers Share of papers (%) No. o f authors 1 2 3 4 5
Scientometrics 39 (1997)
1976-80
1956-60
1931-35
53.40 30.78 13.65 1.70 0.56
61.76 23.52 11.76 2.94 -
77.77 11.1 l 11.11 -
301
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA:COLLABORATIONOF POPULATION GENETICS
The share o f single-authored papers has decreased from 77.77% (1931-35) to 53.40% (1976-80). In contrast, the share o f two-authored papers has substantially increased from 11.11% (1931-35) to 30.78% (1976-80) and three-authored papers have marginally increased from 11.11% (1931-35) to 13.63% (1976-80). In order to get a better picture o f the relative growth o f single-authored, twoauthored, three-authored, etc. papers, we have analysed data on the relative frequency o f papers by number o f authors for different block years from 1916-80. I f the total number o f two-authored papers during 1966-70 and 1916--80 is 42 and 165, respectively, then the relative frequency o f two-authored papers during 1966-70 is calculated as 42/165 = 0.25. The analysis clearly indicates that growth in the frequency o f papers by number o f authors increases as we go up from one-authored to four-authored papers. In one-authored papers, the growth in frequency has occurred from a minimum o f 0.08 to a m a x i m u m o f 0.25, while in others it has increased from 0.06 to 0.32 (twoauthored), from 0.15 to 0.36 three-authored), and from 0.10 to 0.30 (four-authored).
A verage number of authorship per paper The average number o f authorship per paper is another indicator throwing light on the intensity o f collaboration and number o f contributors in the field. The data on the average number o f authorship per paper in theoretical population genetics are computed and presented separately for total, funded, and non-funded papers in following Table 4. "Fable4 Average number of authorship per paper Period
Total papers
Average authorship per paper Total papers
1936-40 1941~,5 1946-50 1951-55 1956-60 1961-65
1966-70 1971-75 1976 80
302
8 8 15 16 34 49 122 150 176
Funded papers
Non-funded papers
2.00 2.00 2.00 2.50 1.60
1.14 1,28 1.15 1.28 1.37
1.25 1.37 1.26 1.43 1.55
1.48
1.37
1.42
1.67 1.74 1.75
1.19 1.33 1,50
1.52 1.53 1.65
Scientometrics 39 (1997)
B. M. GUPTA,S. KUMAR,C. R. KARASIDDAPPA:COLLABORATIONOF POPULATIONGENETICS
The average authorship per paper is observed to be 1.86 from 1936-80. In the funded and non-funded subsets the average authorship per paper is 1.4 and 1.29. The average authorship per paper in total sample has been found to be increasing over the years. It has increased from 1.25 (1936-40) to 1.56 (1956-60) and finally to 1.65 (1976-80). Similar trends are also noticed in non-funded set of papers, while in funded set no uniform trend is noticed. Measures o f collaboration Various mathematical measures have been suggested by scholars which permit the expression of collaboration as a numerical value. Subramanyam 12 has measured degree of collaboration (DC) as the proportion of multi-authored publications in the total sample of publications. DC has a limitation in the sense that it does not distinguish between different levels of authorship. Lawani 25 has established a collaboration index (CI) of the mean number of authors per paper for the total sample of publications. In order to overcome some of the problems posed by these two indices, Aj!feruke, Burell and Tague, 26 proposed a comprise measure, the Collaboration Coefficient (CC), that takes a proportional mean of the sum of the papers with each number of authors and set the values between 0 and 1. In order to make a comparison between different periods in a discipline, we have computed here the values of CI, DC and CC for each block of five years for 1916-80 and the results are presented in Table 5. As we can see from Table 5, the mean number of authors per paper as reflected through CI is noticed to be increasing over the years, without major deviations from 1.06 (1916--20) to 1.65 (1975-80). The growth in collaborative papers over the years is reflected in the decreasing value.of DC and increasing value of CC. The trend in the computed values of CI, DC and CC over the years is more or less consistent, reflecting growing collaboration and professionalism in the field. If we take CC as a more reliable indicator, we notice substantial increase in the computed values of CC in the block years from 1946-50 to 1951-55, 1951-55 to 1956-60, and 1971-75 to 1976-80. This has happened because during these periods there had been a substantial increase in the growth of collaborated papers also in the two cases out of the three case mentioned above.
Scientometrics 39 (1997)
303
B. M. GUPTA, S. KUMAR, C. R. KARAS1DDAPPA: COLLABORATION OF POPULATION GENETICS
Table 5 Indices of collaboration Period 1916-20 1921-25 1926 30 1931-35 1936-40 1941---45 1946-50 1951--55 1956-60 1961-65 1966-70 1971-75 1975-80
CI 1.06 0+00 0.00 1.33 1.25 1+37 1.23 1.43 1.55 1.53 1.43 1.57 1.65
DC
CC
0.9334 0.0000 0.0000 0.7778 0.7500 0.7500 0.7334 0.6875 0.6177 0.6327 0.5574 0.6000 0.5341
0.0334 0.0000 0.0000 0.1300 0.1250 0.1463 0.1334 0.1719 0.2183 0.2111 0.2385 0.2267 0.2617
CI = Collaboration Index. DC = Degreeof Collaboration. CC = Collaboration Coefficient.
Productivity and collaboration In this section, we will explore to what extent productivity of authors is enhanced through their collaboration activities. To begin with, we must understand how the productivity of scientists is to be measured. Lindsey 27 has suggested three methods of measuring productivity o f authors in research papers: Normal count, Straight count, and Fractional count. In the normal count method, every author is assigned a weightage one for each of his/her publication irrespective of whether he/she is a first author or a second/third author. Only the first author is assigned a weightage one in the second method of straight counting. Here the second or third author gets only zero weightage. In the last method o f fractional counting, every author is assigned a weight 1/n in the n-authored paper. In our analysis we have restricted to normal and straight count method for measuring the productivity o f authors. For the total sample the average productivity of authors per paper is observed to be 2.34 (n = 404) and 2.08 (n = 285) using normal count and straight count methods. Table 6 presents data on the productivity of authors using normal count method. For collaborated papers the average productivity is reported to be 1.99 (n = 291) as against the non-collaborated papers where it is 1.84 (n = 173).
304
Scientometrics 39 (195)7)
B. M. GUPTA, S. KUMAR, C. R. K A R A S I D D A P P A : C O L L A B O R A T I O N OF P O P U L A T I O N GENETICS
Table 6 Productivity of authors using normal count method No. of papers
Number of authors Total
Collaborated
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 20 22 24
250 64 25 21 9 6 3 6 5 2 3 2 3 0 0 1 0 2 0 1 1
200 37 17 14 4 3 4 3 0 3 1 2 i 0 1 I 0 0 0 0 0
110 35 15 5 0 1 3 0 2 0 0 1 0 0 0 0 1 0 0 0 0
167 45 17 13 7 2 4 0 5 2 0 1 1 1 0 0 0 0 1 0 0
146 26 12 10 5 3 4 0 1 0 0 0 0 0 0 0 0 0 0 0 0
Total
404
291
173
267
208
2.34
1.99
1.84
2.15
1.67
Average authorship per paper
Non-collaborated
Funded
Non-funded
The number of authors who have produced more than 10 papers are 2.06% in collaborated authors subset and 1.15% in non-collaborated authors subset. Similarly, the number of authors who have produced papers in the range o f 6--10 are 4.46% in collaborative authors subset and 3.46% in non-collaborative authors subset. Considering the productivity o f authors in the funded and non-funded subsets, it is observed that the averge productivity o f authors is 2.15 (n = 267) and 1.67 (n = 208) using normal count method and 1.94 (n = 165) and 1.65 (n = 164) using straight count method. It is clear from the data that the productivity o f authors is much higher in the funded subset than in the non-funded subset, in both the methods used. These results were expected because collaborated papers are more likely to be funded than noncollaborated papers.
Scientometrics 39 (1997)
305
B. M. GUPTA, S, KUMAR, C. R. KARASIDDAPPA: COLLABORATION OF POPULATION GENETICS
Table 7 Range of collaborated activity of authors at different productivity levels Number of Authors Productivity (Range of
No. of authors
papers) 11-24 7-10 3-6 1--2
Percentage of Collaborative Papers in the range of 80-100
15 20 53 316
5 5 2l 149
60-79
. 40-59
0--39
7 7 8 0
0 2 5 20
3 6 19 147
Number of Authors
Productivity (Range of papers) 11-24 7-10 3-6 1-2
No. of authors
Percentage of Founded Papers in the range of 80-100
15 20 53 316
8 5 15 145
60-79 4 7 18 0
40-59 2 2 1 22
0-39 1 6 19 149
These results are further substantiated by the data presented in Table 7 on the percentage of papers collaborated and funded (of 404 authors in different productivity range). The table clearly indicates that as one goes up in the productivity range of authors one finds that percentage of the collaborated papers increases in authors. Simultaneously the percentage of funded papers in authors also increases. So we can conclude that highly productivity authors are more collaborative in nature and also receive more financial support from funding agencies.
Average number of collaborators per author There were in all 404 authors who have contributed 615 papers. Six hundred nineteen collaborators participated in the production of these papers. The mean averge number of collaborators per author is 615/404 = 1.53. Table 8 provides data on the authors in theoretical population genetics ranked by number o f collaborators and productivity. From Table 7, we notice that the authors who ranked from 1-10 papers collaborated with 85.46 per cent of the total collaborators and the other 14.56 per cent collaborators with 3.21 per cent of the more prolific authors. Less productive authors
306
Scientom etrics 39 (1997)
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA:COI~LABORATIONOF POPULATIONGENETICS
improved their productivity by joining the teams o f prolific or productive authors and similarly prolific authors improved their productivity by increasing the number o f members o f their team to work on different areas and projects simultaneously. Thus, the average number o f collaborators per author could serve as an important indicator o f author potential productivity. The more collaborators an author
has,
the more
productive he/she would be. The data presented in Table 8 demonstrate that the difference between one paper authors and those at other few ranks are distributed asymmetrically and seem to be quite close to Lotka distribution. If w e compare the relative frequencies of number of collaborators, we may find that the observed data is also distributed asymmetrically though at a much higher extent than observed in a typical Lotka distribution. The correlation between the number o f papers and the average number o f collaborators per author as 0.6944 (p = 0.002) seems to be strong. The negative correlation coefficients, -0.4751 (p = 0.027) for the number o f authors and the number o f papers, and -0.4361 (p = 0.042) for the number o f authors and the average number o f collaborators per author, indicate that the number o f author decreases as the rank (number o f papers) and number o f collaborators per author increases. Table 8 Distribution of number of authors vs the number of papers NP 1 2 3 4 5 6 7 8 9 10 ll 12 13 16 18 22 24
Note:
NA
TNC
250 64 25 21 9 6 3 6 5 2 3 2 3 l 2 1 1 404
238 86 38 64 19 16 12 28 22 6 26 15 24 8 19 2 10 619
ANCPA 0.952 1.344 1.440 3.047 2.111 2.666 4.000 4.666 4.400 3.000 8.666 7.500 8.000 8.000 9.500 2.000 10.000
NP = Number of Papers; NA = Number of Authors; TNC = Total Number of Collaborators ANCPA = Average Number of Collaborators Per Author.
Scientometrics 39 (1997)
307
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA:COLLABORATION OF POPULATION GENETICS
Level o f productivity and collaboration among authors The data already given in Table 8 on the number o f authors/ papers written (columns 1 and 2) have been clubbed together in few close groups and data on the total number o f collaborators (column 3) have also been clubbed together and further expanded and classified by range o f collaborator author productivity in Table 9 to get an idea on the level o f productivity at which authors collaborate among themselves. For example, there are 339 authors in all, who have written I - 3 papers and have 362 collaborators [228 (67.25%) have written 1-3 papers, 49 (14.45%) have 4 - 6 papers, 28 (8.25%) have 7 - 9 papers, 43 (12.68%) have 10-16 papers, and 14 have 17-24 papers]. Similarly, there are 36 authors who have written 4 - 6 papers and have 99 collaborators [49 (49.49%) contributing 1-3 papers, 15 (15.15%) contributing 4 - 6 papers and etc.]. Table 9 clearly indicates that authors tend to collaborate with authors having same productivity level (same rank) or below their productivity level (lower rank). As the level o f productivity o f authors increases, the major percentage o f collaborating authors (having either the same or lower productivity status) also increases. It means that the frequency o f contacts, between partners with the same rank is highest and decrease with the increasing distance between the ranks. This is in accordance with the social psychological theory. Table 9 Level of productivity and range of collaboration NP
1-3 4-6 7-9 10-16 17-24
Number of collaborators by range of papers
NA
339 36 14 11 14
1-3
4-6
7-9
10-16
17-24
TNC
228 49 28 43 14
49 15 18 12 5
28 18 2 9 5
43 12 9 8 7
14 5 5 7 0
362 99 62 79 31
Note: NP = Number of Papers. NA = Number of Authors. TNC = Total Number of Collaborators.
308
Scientometrics 39 (1997)
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA: COLLABORATION OF POPULATION GENETICS
Type of collaboration Table 10 throws light on the type of collaboration, as reflected through addresses in research papers in theoretical population genetics. The research papers are deseggregated under five types explained earlier. Of the total 248 collaborated papers, 47.58% (118 papers) reflect intra-departmental collaboration, 8.06% (20 papers) interdepartmental collaboration, 21.37% (53 papers) inter-institutional collaboration within the country, and 22.98% (57 papers) inter-institutional collaboration across countries (international). In the total sample of 248 collaborated papers, 193 papers (77.82%) fall in Group A and 55 papers (22.17%) in Group B. In Group A (i.e. USA) and Group B (18 other countries), the contribution of Intra-departmental collaborated papers are 94 papers (48.70%) and 24 papers (43.63%); inter-departmental collaborated publications are 18 papers (9.32%) and 2 papers (3.63%); inter-institutional collaborated papers across cities within the country are 46 papers (23.83%) and 7 papers (12.72%); and intercountry collaborated publications are 35 papers (18.13%) and 22 papers (40%). Table 10 Types o f collaborated research No. of collaborated papers Types of collaboration
Total papers Total papers
Type Type Type Type Type
Group A papers
Group B papers
% funded
Total papers
% funded
Total papers
% funded
l 2 3 4 5
I18 20 4 49 57
64 15 3 43 54
94 18 3 43 35
70 13 2 39 32
24 2 1 6 22
9 1 l 4 14
Total papers
248
182
193
153
55
29
Among the total collaborated 248 papers, the percentage of papers funded under various type of collaboration is: 54.23% (64/118 papers) under intra-departmental collaboration, 75% (15/20 papers) under intra-departmental collaboration, 86.79% (46/53 papers) under inter-institutional collaboration within the country, 94.93% (54/57 papers) under inter-institutional collaboration across countries. It is observed that as the number of papers are shifting from Type 1 to Type 5, the percentage of funded papers
Scientometrics 39 (1997)
309
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA:COLLAI3ORATIONOF POPULATION GENETICS
has been found to be increasing. For inter-institutional collaborative research papers across cities within the country (domestic collaboration) and across countries (transnational collaboration) the funding is very high. Types 1 and 2 are found to be traditional form o f collaborations either among colleagues within the same department or across departments within the same institute; that is why the percentage o f funded papers is small. Table 11 throws light on the type o f collaboration among research papers over different time periods for total papers as well as for papers falling under Groups A and Group B. The emphasis on the type o f collaborative research has clearly being changing over the years. A distinct shift is taking place from one type o f collaboration to another. Inter-institutional collaboration (domestic collaboration) within the country and the international collaboration have been slowly gaining ground and started becoming prevalent over the years. Domestic collaboration (Types 3-4) started during 1946-50. From contribution o f just one paper (14.28%) its share has increased to 18 papers (21.95%) during 1976-80. There has been sharp increase in domestic collaboration in Group A (USA) because o f availability o f larger financial support in USA.
Table 1l Types of collaborated research papers over time period Period
1976-80 1971-75 1966-70 1961-65 1956-60 1951-55 1946-50
1941-45 1936-40 1931--35 1916-20
TNCP
82 [67] 59 [32] 54 [43] 18 [13] 13 [9] 5 [4] 7 [6] 4 [4] 3 [3] 2 [0] 1 [1]
Number of collaborative papers under Type 1
Type 2
Type 3
Type 4
Type 5
33 [29] 22 [16] 26 [24] 9 [7] 9 [5] 4 [3] 5 [2] 4 [4] 3 [3] 2 [2] 1 [1]
6 [6] 4 [2] 8 [7] 0 [0]
0 [0] 2 [11 1 [1] 1 [1]
18 [I5] 14 [13] 10 [6] 4 [3]
25 [17] 17 [10] 9 [5] 4 [2]
1 [2] l [1] 1 [1]
2 [2]
1 [11 1 [11
T N C P - Total number of Collaborated Papers. Note: Within brackets outputs of Group A is reflected.
International collaboration among research papers is the most recent phenomenon, and there is a substantial increase in its proportion over the years. The number o f international collaborative publications has grown from just 2 papers (15.38%) during 1956-60 to about 25 papers (30.48%) during 1976-80. In sharp contrast to these two
510
Scientometrics 39 (1997)
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA: COLLABORATION OF POPULATION GENETICS
types of collaboration, there is a sharp fall in the internal collaboration (Types 1 and 2). The contribution of this group has decreased sharply from 100% during 1916-40 to 47.56% during 1976-80. Over the last 30 years the government science policy in many developed countries has stimulated and encouraged research collaboration. 29 These governments are encouraging the researchers located in universities, research institutes, etc., to work more closely through joint research programmes and mission-oriented research. Each of these programmes are being provided with substantial funding and encouraging an active exchange of information, sharing of equipments and facilities, and exchange of scholars among organisations. Of late, inter-governmental collaboration (international) among research institutions is becoming more frequent and dominating, and is playing an important role in the production of scientific knowledge. Various kinds and modes of cooperation between researchers and institutions of different countries normally take place through meeting of experts, exchange of students and scientists, sharing of equipment and facilities, etc. The government initiatives to increase international contacts through provision of fellowships and travel grants to foreign scholars, bilateral and multilateral agreements between countries, and initiatives of the international (e.g. UNESCO) and intergovernmental agencies are also actively promoting transnational collaboration.
Number of authors The total number of authors involved in collaborative research in different years is presented in Table 12. It is observed that number of collaborative authors has increased from just three (1926-30) to a maximum of 166 authors (1976-80). The percentage of collaborating authors in total authors has also been found to he increasing from 18.10% (1916-20) to a maximum of 71.08% (1976-80). Table 13 provides a break-up o f total collaborative authors by type of collaboration. It is observed that percentage of collaborating authors in internal collaboration (Types 1 and 2) is found to be decreasing over the years. There number and percentage share have decreased from just two authors (100%) during 1916-20 to 56 authors (47.46%) during 1976-80. Against this, number and percentage of collaboration authors involved in domestic collaboration (Types 3 and 4) and international collaboration (Type 5) has increased systematically. The number of collaborating authors involved in domestic collaboration has increased from three authors (1956-60) to 47 (1976--80). In international collaboration, the number of collaborating authors has increased from
Scientometrics 39 (1997)
31 l
13. M. GUPTA, S. KUMAR, C. R. KARAS1DDAPPA: COLLABORATION OF POPULATION GENETICS
seven authors (1956-60) to 42 authors (1976-80). The most systematic increase has been in the types 4 and 5 of collaboration. Table 12 Number and percentage of collaborated authors Period
No. of authors
1916-20 1921-25 1926-30 1931-35 1936-40 1941-45 1946-50 1951 55 1956-60 1961-65 1966-70 1971-75 1976-80
11 5 3 8 9 6 11 17 36 56 114 124 166
No. of collaborating authors
% of collaborating authors
2 0 0 4 4 3 5 8 20 34 72 85 118
18.18 00.00 00.00 50.00 44.44 50.00 45.45 45.45 55.55 60,71 63.15 68,55 71.08
Table 13 Number of collaborating authors by type of collaboration Period
1916-20 1931-35 1936-40 1941-45 1946-50 1951-55 1956-60 1961-65 1966-70 1971-75 1976-80
312
Total collaborating authors 2 4 4 3 5 8 20 34 72 85 118
Number of authors Type 1
Type 2
Type 3
Type 4
Type 5
2 4 4 3 2 5 10 18 36 30 56
0 0 0 0 2 0 2 0 12 14 0
0 0 0 0 0 0 0 2 2 2 7
0 0 0 0 3 4 3 9 24 30 40
0 0 0 0 0 0 7 9 14 27 42
Scientometrics 39 (1997)
B. M. GUPTA,S. KUMAR,C. R. KARASIDDAPPA:COLLABORATIONOF POPULATIONGENETICS
Conclusions
The analysis presented in this paper offers evidence to support the hypothesis that the increased collaboration is generally associated with proportional increase in funding of research papers in theoretical population genetics. In all 72% (291 authors) of the total authors are associated with collaborated research, reflecting a high index of professionalism in the field. The subset of research papers associated with funding (333 papers) is observed to be much more collaborated than the non-funding subset of literature (276 papers). The actual percentage of collaborative papers in these two subsets were 54.65% and 21.75%. The proportion of co-authored research papers over time is a good index of the level of professionalism emerging in the field. The extent of authorship in research papers and the averge number of unique names or authorship per paper provide a general idea on authorship patterns in the field. The averge authorship per paper in funded and non-funded subset of papers throws some light on the intensity of collaborative activity associated with funded research. The averge authorship per paper in funded and non- funded subset of papers using normal count method was 1.89 and 1.25, respectively. A strong relationship is observed between collaboration, productivity, and funding of research papers. Authors who were most productive in theoretical population genetics are also found to be most collaborative in nature and they are also able to manage and get more funds from the funding agencies. This argument was further substantiated from the analysis of averge number of collaborators per author at different productivity level of authors in the field. The averge number of collaborators per author was found to be much higher in highly productive authors than in medium or low productive authors. The analysis clearly brings forth the changes taking place in the nature of collaborative research. A distinct shift is found to be taking place over the years from local collaboration to domestic and international collaboration. This trend is encourged by deliberate government policies and supported by funds through concerned funding agencies.
References
1. NARSI PATEL, Collaboration in the professional growth of American Sociology, Social Science Information, 12(6) (1972) 77-92, 2. National Science Foundation, Science and Engineering Indicators 1987. Report of the National Science Board, Washington, D.C. 1987.
Scientometrics 39 (1997)
313
B. M. GUPTA, S. KUMAR, C. R. KARASIDDAPPA: COLLABORATION OF POPULATION GENETICS
3. D. DE B. BEAVER, R. ROSEN, Studies in scientific collaboration: 1. The professional origins of scientific coTauthorship, Scientometrics, 1 (I 978) 65-84. 4. D. DE B. BEAVER, R. ROSEN, Studies in scientific collaboration: 2. Scientific co-authorship, research productivity, and visibility in the French scientific elite, 1799-1830, Scientometrics, 1 (1979) 133-149. 5. D. DE 13. BEAVER, R. ROSEN, Studies in scientific collaboration: 3. Professionalisation and the natural history of scientific co-authorship, Scientometrics, 1 (1979) 231-245. 6. D. J. DE SOLLA PRICE, Little Science, Big, Science, New York; Columbia University, 1965. 7. A. J. MEADOW, Communication in Science, London; Butterworth, 1974. 8. D. J. DE SOLI,A PRICE, Citation measures of hard science, soft science, technology and non-science. In: C. E. NELSON, D. K, POLLAK, (Eds), Communication Among Scientists" and Engineers, Lexington, M. A.; D.C. Heath, 1970, p.3-22. 9. D. J. DE SOLLAPRICE, Multi authorship, Science, 212 (1981) 986. 10. W. HIRSCH, J. F. SINGLETON, Research support, multiple authorship and publications in sociological journals, 1936- 64. Purdue University, 1964 (Unpublished). ll. A. G. tlEI~NER, Funded research, multiple authorship and subauthorship collaboration in four disciplines, Seientometrics, 3 (1981) 5-12. 12. K. SUBRAMANYAM,Bibliometric studies of research collaboration: A review, Journal ofb~fiJrmation Science, 6 (1983) 33-38. 13. R. HART, T. CARSTENS,M. LACROIX, K. R. MAY, Funded and non-funded research: Characteristics o f authorship and patterns of collaboration in the 1986 library and information science literature, Library and hT~rmalion Science Research, 12 (1980) 71-86. 14. D. J. DE SOLLA PRICE, D. DE B. BEAVER, Collaboration in an invisible college, American Psychologist, 21 (1966) 1011-1018. 15. H. A. ZUCKERMAN,Nobel laureates in science: Patterns of productivity, collaboration and authorship, American Sociological Review, 32 (1967) 391-403. 16. M. L. PAO, Collaboration in computational musicology, Journal (~f the American Society .fi~r ln/brmation Science, 33 (1982) 38-43. 17. M. L. PAO, Global and local collaboration: A study of scientific collaboration, Information Processing and Management, 28(1) (1992) 99-t09. 18. J. QIN, An investigation of research collaboration in the sciences through the Philosophical Transactions, Scientometrics, 29 (1994) 219-238. 19. R. KUNDRA, Investigation of collaborative research trends in Indian medical sciences, 1900-1945, Scientometrics, 36 (1996) 69-80. 20. J. D. FRAME, M. P. CARPENTER, International research collaboration, Social Studies of Science, 9(4) (1979) 481-497. 21. R. J. W. TIJSSEN, H. F. MOED, Literature-based statistical analysis o f international scientific cooperation, ln: Science and Technology Indicators, Leiden; DSWO Press, 1989. 22. T. LUUKKONEN, O. PERSSON, G. SIVERTSEN, Understanding patterns of international scientific collaboration, Science, Teehnology andHuman Values, 17(2)(1992) 101-126. 23. A. SCHUBERT, T. BRAUN, International collaboration in the sciences, Scientometrics, 19 (1990) 3-10. 24. J. FALSENTEEN,Bibliography ~?f Theoretical Population Genetics, Pennsylvania; Dowden, Hutshinson and Ross, Inc. 1981. 25. S. M. LAWANI, Some bibliometric correlates of quality in scientific research, Scientometrics, 9 (1986) 19-35. 26. I. AJIFERUKE, Q. BURELL, J. TAGUE, Collaborative coefficient: A single measure of the degree of collaboration in research, Scienmmetrics, 14 (1988) 421-433. 27. D. LINDSEY, Production and citation measures in the sociology of science: The problems of multiple authorship, Social Studies of Science, 1 (I 980) 101-125. 28. J. QIN, Collaboration and publication productivity: An experiment with a new variable in Lotka law. In: M. E. D. KOENIG, A. BOOKSTEIN (Eds), Fifth International Conference of the International Society for Scientometrics and hlfometrics Proceedings-1995, June 7-10, 1995, Medford, N J; Learned Publication, Inc. 1995, pp. 445-454. 29. D. HICKS, J. S. KATZ, Science policy for a highly collaborative science system, Science and Public Policy, 23(1) (1996) 39-44.
314
Scientometrics 39 (1997)