Diana Crane - Testing the 'Invisible College' Hypothesis

The term 'invisible colleges' was developed in sociology by Diana Crane based on Derek J. de Solla Price's work on citation networks. She sees it as referring to an elite group of highly productive and mutually interacting scientists in a Research Area (RA) who are at the centre of the formal network of communication among all scientists working in that particular Research Area (RA). 

Diana Crane, in the paper 'Social Structure in a Group of Scientists: Testing the 'Invisible College' Hypothesis,' says that a group of scientists working on a particular RA form a social group that has not been paid much attention primarily because: they are geographically scattered, most scientists do not have more than one or two papers published in a particular RA, the boundaries of RAs themselves is hazy at times, and the participation in these groups is highly voluntary thus making it difficult to study the existence of 'invisible colleges' among members of the scientific community. Crane is, in this paper, trying to see if: 

• there exists any sort of social organisation among scientists working in the same Research Area by studying the social ties among scientists who have published in an RA with other scientists who haven't published in the same RA and, 
• scientists who have published in the RA can be differentiated by the degree of social participation within the RA. 

She studies a group of rural sociologists involved in 'agricultural innovations' - senior and junior - by using questionnaires and response sheets for information about their references. She tries to establish her understanding of 'invisible colleges' by studying the citation patterns similar to the study made by Price. 

Scientists are part of different networks of communicating with their peers and superiors including informal communication networks and formal networks such as collaboration with other authors, or thesis directors before, during or after one's research. Intellectual linkages often reflect the influences of one scientist on another. Using the information about references and citations received from the rural sociologists, she sets out to map their influences with the help of a matrix developed by James Coleman with the choices received by the author on one axis and the choices made by the author on the other. Continuous multiplication of this matrix will eventually yield all the indirect relationships between scientists in the RA. 

To analyse the direct and indirect relationships between members of different subgroups, Crane divides the group on the basis of productivity and commitment to the RA. With this division, she comes up with five subgroups, 3 based on productivity: 

• 8 High Producers - those who had published more than 10 papers in the RA.
• 11 Medium Producers - those who had published between 4 and 10 papers in the RA.
• 33 Aspirants - those who had published less than 4 papers in the RA.

and 2 based on their commitment to the RA - those who had not continued their research in the RA (it was found that all those had published more than 10 papers continued research in the RA.):

• 9 Defectors - they had between 4 and 10 papers published in the RA.
• 86 Transients - they had less than 4 papers published in the RA.

Studying both the choices made by scientists favouring other scientists who had published in their RA and with 'Outsiders', Crane found that the scientists chose both on an almost equal level (49:51 for 'insiders':'outsiders'). However, she found that the 'Outsiders' were less likely to be chosen more than twice (not more than 84% were chosen more than twice) and since only one 'outsider' had their name cited more than ten times, it was not possible for a group of outsiders to influence them as much as their own did. 

Characteristics of Members of Subgroups:

1. Selection of Group Members versus Outsiders: The 5 subgroups divided based on productivity and commitment were expected to exhibit varying degrees of linkages both with insiders and outsiders. While members of the highly productive group were closely linked with each other and with outsiders, members of another group which was relatively unproductive were not as closely linked with either group.   
2. Direct and Indirect ties by Subgroups: Choices made by the High Producers and Medium Producers led to them having greater ties with their own members than did the Aspirants. In addition, a high proportion of choices made by members of other subgroups led to the High Producers being placed at the middle of their communication network. The High Producers are also linked through published collaboration to a large number of members in the RA and also influenced many others in their roles as thesis directors. 

Changes in Networks of Social Ties

Over the last few decades, the number of High Producers has increased significantly and so has their proportion in the scientific community. High Producers associate themselves with other High Producers and most of their students become High Producers too. The High Producers place themselves at the centre of the communication network through their high productivity and commitment to developing and making the RA known in the scientific community. 

The type of social organisation one sees in among a group of scientists working in a particular RA can be likened to a social circle. The social circle is not well instituted compared to the bureaucracy or social institutions such as family. Members come together on the basis if their interests rather than their ascribed statuses. Indirect interaction and interaction mediated by common associates is an important part of the social circle and it is not necessary to know a person to be influenced by them. At most, a member of the social circle would know a few other members but never all. The presence of scientists whose high productivity is sufficient for them to attract the attention of those who enter the field, produce a social circle which then plays an important role in the growth of the RA.

The diffusion problem area deals with the diffusion of a sizable number of papers in a RA. A majority of these articles could be published either in a few 'core' journals with the rest scattered among numerous others, or be published in numerous different journals with no relation to each other. Neither scenario would be conducive to the growth of knowledge in that RA for, if there were only a few core researchers producing articles and papers regularly citing each other's work, that would only allow for a restricted view point with each being influenced by the other; it would lead to stagnation of ideas in the field; if scientists and researchers preferred to be isolated and avoided each other's contact, that too would be harmful for the growth of science for many ideas would never come to fruition. 

One of the things that needs to be further studied is the points of intersection between different areas especially if scientists are going to continue moving from one area to another to study related problems. The communication network then has to be traced and the influences form each area have to be analysed further to get a wholesome understanding of the communication among scientists. 

-----------

Diana Crane is a professor emerita of sociology at the University of Pennsylvania. Diana Crane is a specialist in the sociology of culture, arts, media, and globalization. She has also taught at Yale University, Johns Hopkins University, University of Poitiers (France), Erasmus University (The Netherlands) and Columbia University in Paris. She received her Ph.D. from Columbia University. She has been awarded a Guggenheim Fellowship, been a Member of the Institute for Advanced Study (Princeton, NJ), and a visitor at the Bellagio Study and Conference Center (Rockefeller Foundation, Bellagio, Italy). She has held Fulbright Awards in France and the Netherlands. She was chair of the Sociology of Culture Section of the American Sociological Association in 1991-1992. She has been a member of the Advisory Board of Poetics since 1992.

Confronting Merton's Norm of Universalism

Kumar, Neelam (2001): ‘Gender and Stratification – An empirical study in the Indian setting’ in Indian Journal of Gender Studies, 8:51

Available at : http://ijg.sagepub.com/content/8/1/51

This paper is based on a research study, the attempt of which was to understand if gender played a role or was a variable used in the stratification system within Indian scientific institutions.

The author looks at the assignment of ranks to scientists/lecturers in scientific institutions and tries to determine if advancement in rank is guided by universalistic norms. (Merton’s norm of Universalism requires that when a scientist makes a contribution to scientific knowledge, the community's assessment of the validity of that claim should not be influenced by personal or social attributes of the scientist and should be subject to pre-established impersonal criteria. Universalism also requires that a scientist be fairly rewarded for contributions to the body of scientific knowledge.) For comparing academic rank, the major determinants of advancement in academia like achievements/recognitions (measured by awards won and membership in various professional bodies) and research productivity (measured by research publications, research grants received and reviews done for journals*) were taken into account.

The sample for the survey consisted of physical scientists of both sexes in four different Indian cities and eight scientific institutions. The study covered two work contexts: national laboratories and universities. The resulting sample included 117 scientists-56 women and 61 men-matched on the basis of age.

Results

The table below shows a comparison between men and women scientists in terms of academic rank held by them. It reveals the proportion of each sex in the sample in a given academic rank.

It is seen that as the rank gets higher, the number of women holding that rank declines. Only 3.6% of the women held the rank of professor, whereas 60.7% were assistant professors. In the case of male scientists, 18% were professors and 44.3% were assistant professors. In the rank of associate professor the difference is smaller. About 35.7% female scientists and 37.7% male scientists were in the rank of associate professor. Further comparison of men and women scientists (within each particular rank) in terms of age and number of years spent within the organisation revealed larger differences but these differences were evident only at the topmost level of the hierarchy. The mean age of women scientists at the professor (or equivalent) level was 54.5 years while for men it was 46.4 years. An analysis of the career trajectories of a few women scientists also revealed that many women stayed for an unusually longer time in the same rank than their male counterparts.

Gender differences on variables like research productivity and rewards/honours were analysed with t-test. The table below summarises the findings.

It is seen that the two groups differ significantly (at the level of .01 or .05) in terms of academic rank. But there was no evidence of significant gender differences in research performance. The result appears to contradict numerous studies showing that women are less productive. Findings not presented in this table revealed that there were no significant differences between men and women in terms of the time they reported committing to teaching and research (p > .05 in each case). The two groups also did not differ in terms of recognition measures (awards and membership in various professional bodies). However, it was also revealed that there was a significant difference in terms of reviews done by men and women scientists. The authors attribute this to bias and discriminatory practices in the selection of reviewers.

Conclusion

Ideally and in accordance with the norms of universalism, advancement in rank should be governed by research productivity. This study has clearly revealed a lower percentage of women in higher academic positions within Indian scientific institutions. Thus, it might be reasonable to assume that women’s performance levels were lower than that of men’s. The results of this study reveal that while there are significant differences in the academic ranks of women and men scientists, in research performance they do not differ in a statistically significant manner. The finding thus makes it clear that rank disparities between men and women scientists are not attributable to differences in research performance and leaves us with the hypothesis of sex discrimination, indicating the prevalence of particularism.

* As the various types of performance measures are of unequal scientific importance, a weight was assigned to each measure to reflect the importance of the measure in relation to other measures. A composite measure of performance was created by adding the scores for each measure. In the resulting order of importance, books rank first in published written output (and received a weight of 24), articles published in refereed international journals appear next (receiving a weight of 8) and so on. Similarly, single or co-authored articles and books were differentiated.

On Peer Review, Fraud and Plagiarism

The article I presented is titled "What is the future of peer review? Why is there fraud in science? Is plagiarism out of control? Why do scientists do bad things? Is it all a case of:“All that is necessary for the triumph of evil is that good men do nothing?” by Chris R. Triggle and David J. Triggle. It was published in February, 2007 in the journal, Vascular Health and Risk Management. Here is a link for the article: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1994041/

This article deals with Merton's norm of Organised Skepticism according to which all scientific claims must be exposed to critical scrutiny before being accepted. This happens through the peer review process but as the paper suggests, there are inherent problems with the very process.

Here are some of the important arguments/questions in the paper:

Peer Review

• How do you define 'peer?'
• Anonymity in the review process assumes maximal effort and fair judgement. At the same time, it protects reviewers from retribution. But does it lead to laziness?
• Use of vague reasons such as 'my gut feeling is this will not work'
• Why not publish the reviews then?
• "In what direction and should the peer review process actually police scientific fraud and should the peer review itself be subjected to review and potential legal action if scientific fraud by the reviewer is suspected?"
• Tendency to accept only positive data
• Bias - free passage for well-known laboratories, country of origin also influences acceptance
• "If the peer review process is unfair, if the rights of the individual under review are not protected, if the “facts” presented during the review are inaccurate, and the result is a damaged reputation and loss of income then why shouldn’t you sue?"
• Conflict of interest can arise from differences between author, journals and funding sources, vested interests, ideologies, religions etc.

Journal Impact Factor (JIF)

• Means of defining the impact of a scientist's research
• The number of citations for an article in a given year divided by the number of articles and reviews published in the same journal during the past two years
• “80:20” phenomenon - 20% of publications account for 80% of the citations
• But it is essential to evaluate the impact of the individual paper and take into account not only where it was published but, in particular, also how well it has been cited and by whom —a paper in a high impact journal does not necessarily equate with a high impact paper, it is the citation frequency that is more important
• Alternative method - Index of individual productivity, “h”, has been proposed by Hirsch (2005) where h is defined as the number of papers with citation number >h

Fraud

• Difficulty - to determine with complete certainty that malicious intent and not interpretation error, or simply bad laboratory practice, was the cause.
• "Fraud in science, whether initially intended as hoaxes or planned with career and profit-making intentions, not only ruins the careers of the perpetrator, but also, potentially, their innocent colleagues, as well as tarnishing the reputation of the institution where the work was performed and reducing the confidence of the public in the value of scientific research"

Plagiarism

• Self-plagiarism
• Cryptomnesia - unconscious plagiarism?
• Is reuse of descriptions of experimental methods (like A + B = C) plagiarism?
• "Scientists are no different from any other groups in society and, like many other analogous comparisons, a few rotten apples will always be found."

Recommendations

• The establishment of the equivalent of the Office of Research Integrity (ORI) in other countries
• Appropriate safeguards designed to protect both the whistleblower and the accused
• Processes whereby the apparent bias in peer review can be reduced are urgently required and should be evaluated
• Heightened awareness and education at all levels concerning the seriousness of scientific fraud in all of its manifestations

Feel free to comment.

- Dipali.

Understanding Merton's Universalism (Gender)

Merton's passion and interest towards the study of the Sociology of Science, especially through the interactions of social and cultural structures with Science. In this regard, he formulated the Norms of Science, namely: Universalism, Communism, Organized Scepticism and Disinterestedness.
This post deals with the norm of Universalism ('claims to truth' ought to be evaluated fairly and not determined by characteristics such as race, nationality, gender, religion, etc.) especially in the context of Gender.

Stephen J. Ceci and Wendy M. Williams of Cornell University, in a paper titled 'Understanding current causes of women's underrepresentation in science' try to explore the prevalence and causes for gender bias in the academia. The primary causes for under-representation, reports suggested were discriminations in the fields of granting funds, hiring and the journal reviews.

With respect to Journal Reviewing, the authors examine manuscript acceptance rates between males and females while keeping the quality of work as a constant. Analysing works ranging from Budden's study of blind reviews faring better gender-wise in Behavioural Psychology to longitudinal studies spanning two and a half decades, the authors dismiss the argument of gender bias in Journal Reviwing, primarily citing weak justifications and proof of discrimination with respcet to manuscript acceptance rates.In the case of Grant Funding, Ceci and Williams (2010) primarily compare the works of Wennerås and Wold with the Cochrane Methodology Review Group that concluded that apart from the former's study conducted a decade ago, none of the other studies indicate a strong case of discrimination.Hiring, witnessed rather interesting results- one study that looked at positions in R1 universities concluded that women has a better chance of getting interviewed and receiving offers than their male counterparts.

In conclusion, Ceci and Williams are rather sceptical about these conventional claims of discrimination against women. So how do they answer the puzzle of under-representation? The authors attribute it to fertility (life style choices), career preferences (teching over research, for example) and 'work-home' balance issues. While they acknowledge gender differences, they define the differences as 'secondary' and conclude that the real problem lies in the resources attributable to the abovementioned choices and it is those that have to be the focus of the remedy.

Thus, placing it in the context of Merton's Universalism, we see that Ceci and Williams bear a tinge of optimism and infact indicate a move towards Universalism. However, some points to note about the article are that the argument is centered around American contexts and hence cannot be extrapolated to other contexts; the authors primarily look at data for explicit indicators of discrimination, which is rarely the case.

Reference- Stephen J. Ceci and Wendy M. Williams. Understanding current causes of women's underrepresentation in science. PNAS vol. 108 no. 8 3157-3162 (2010)

-Uttara