Performance, Specialization and International Integration of
Science in Brazil: Changes and Comparisons with Other Latin
America and Israel.
Department of Sociology
University of Pittsburgh
Paper prepared for a science policy study carried on by the Escola de Administracao
de Empresas, Fundacao Getulio Vargas for the Brazilian Ministry of Science and
Technology and the World Bank, within the PADCT II agreement. The opinions
expressed in this text are the sole responsibility of its author.
N ovember, 1993
This analysis is also a by-product of the project "Center-periphery relations and
international diffusion of scientific information" funded by theU.S. National Science
Foundation. The survey of scientists was funded by faculty grants from the Center for
Latin American Studies and the Faculty of Arts and Sciences at the University of
Pittsburgh. I am also grateful to the interviewed scientists in Brazil, other parts of Latin
America, Israel and elsewhere, to my project staff, and to Drs. Rachel Hovneand Simon
Schwartzman for comments on the study.
National scientific development and participation in world science. 1
Brazil in regional and peripheral locations. 5
Historical and comparative perspectives: changes and contrasts. 6
Institutionalization of science in Brazil, other Latin America, and Israel. 7
Concepts and their indicators based on the literature and a survey. 10
Performance and specialization in research. 14
Deference to world science. 22
Influences from and upon world science. 24
Communal attachments: Emulation and recognition. 39
Audience-orientations. 4 1
How have Brazilian research performance, specialization and ties with local,
regional and distant colleagues been shaped by the institutionalization of science?
Changes in Brazilian research are here ascertai ned and compared to research in the rest
of Latin America and Israel which have similar and different, respectively, institutional
arrangements for science.
Science is now institutionalized in Brazil insofar as it is appreciated and granted
autonomy and support. Scientific activities such as research and training of researchers,
however, have been concentrated in a fraction of the universities in Brazil, like
elsewherein Latin America but unlike in Israel where scientific research and training
has been the primary purpose at all the institutions of higher education.
This difference between Brazil, or Latin America more generally, and Israel in
their institutional arrangements for science has enhanced scientific performance in
Israel so that research performance has been higher than in Brazil and in the rest of
Latin America. Israeli performance has been higher despite the smallness of the
country; actually, relative to its population and economy, Israeli performance has been
higher than any other nation. But scientific research has expanded in Brazil, and the
Brazi I ian rate of i ncreasefrom the 1960s up to 1993 was higher than i n the rest of Lati n
America, in Israel, and intheworld asa whole. The current Brazilian crisisand doubts
about the worthiness of science may entail a decline in the creation of scientific
knowledge in the near future.
Brazilian research is specialized i nsofar as certai n disciplines and specialties are
foci of scientific attention in Brazil more than elsewhere. Brazilian specialization
emphasizes the disciplines of physics, biology and mathematics, it is rather typical in
biomedicine and earth and space science, and it deemphasizes clinical medicine,
chemistry and technological science. Within medicine, though, tropical medicine and
parasitology are fields of strong specialization in Brazil, like in the rest of Latin
America, butunlikein Israel where specialization is more concentrated on mathematics
and fertility. Brazilian growth has been highest in technological science, especially in
computing. These directions of Brazilian specialization seem shaped by national needs
and research policies.
Brazilian scientists are tied to col leagues who have influenced their research, and
also are variously collaborators and competitors in research, and often also are
significant recognizers of their work. These ties are not confined within Brazil but are
also rather extensive with foreign colleagues. Brazilian research is somewhat
integrated with science in the rest of Latin America insofar as regional influence and
collaboration are higher than expected. This integration in collegial ties is promoted
by the embeddedness of scientific ties in the I inks between Brazil and the rest of Lati n
America in other spheres of life. Although this regional integration is noticable, it is
rather weak and is overshadowed by the attachment to the world centers of science,
located in North America and Western Europe. Scientists in Brazil, like elsewhere,
defer to science in these centers, travel there, receive influence from science in the
centers, seek collaboration there and value recognition from colleagues in the centers.
The involvement of Brazilian scientists with the centers, though, has been slightly less
than the participation of other Latin American scientists and much less than the
integration of Israeli scientists in the centers. The Israeli ties with the centers, especially
the North American center, is enhanced by their embedded ness in the strong political-
economic links between Israel and the United States and also by institutional
arrangements for science such as comprehensive scientific cooperation agreements with
the centers and travel funds which are results of Israeli research policies for enhancing
research performance through integration with the world centers of science.
National scientific development and participation in world science.
Throughout history people have interpreted sensed phenomena, and their
interpretations have fueled intellectual traditions within many civilizations. Typically,
a tradition has been parochial insofar as its cultivators have not been interested in
extending participation, and even if they occasionally have sought to extend
participation, is has remained a local endeavor because people in other civilizations
aware of the exogenous tradition have ref rai ned from participating in it (Shi Is 1981).
Science, however, is an atypical tradition in its extensive communal ity. Present-
day participation in the scientific tradition is world-wide. Not only is science practiced
in every society but its practitioners are receptive to ideas from any place on earth,
pursue long-distance collegial ties, and disseminate their creations globally. This
global ity has evolved only recently. Until the 19th century, the tradition was
concentrated within one area, Europe, where it had become established just a few
centuries earlier. Global participation emerged in the 20th century. The formation of
a global communal endeavor was a process of institution-building: both
institutionalization, that is, granting legitimacy, appreciation and autonomy, and
establishing institutional arrangements for global communal participation.
In early-seventeenth-century Europe the pursuit of knowledge about natural
phenomena had some legitimacy and some appreciation but little autonomy. It was
controlled by political and ecclestial authorities who respected religious dogma over
natural knowledge. But a social movement had formed in Europe which was inspired
by an Utopian faith in progress by advancing empirical knowledge applicable to
control ing and exploiting nature. Thisscientistic movement advocated legitimation for
the search for truth through empirical inquiry as an activity worthy of cultivation. The
movement was successful first in England in the middle of the seventeenth century.
The movement won legitimation for science as an intellectual tradition cognitively
distinguishable from intellectual traditions such as contemplative philosophy, practical
arts, crafts and religious thought. The shift to a considerable appreciation for empirical
inquiry into God's creations, especially nature, was triggered by Puritan thought (see
the content analysis of Puritan statements by M erton  1970, chapters 4-5). But the
enhanced appreciation was broad as evidenced by the participation in science of not
only Puritans but also many people from diverse other faiths (M erton  1970,
chapter 6, shows the high participation of Puritans and the high participation from
diverse other faiths is shown in many studies; whether Puritans were over rep resented
is debated but not crucial here where the documented great diversity of participation
is the relevant evidence showing the breadth of the appreciation of natural inquiry).
Political and religious authorities granted natural inquiry considerable autonomy and
authority as a source of truth along with the Bible. This institutionalization in England
was imitated in the continental societies that in the culturally well-connected but
politically decentralized Europe were competing for national esteem through
appreciated cultural forms (Ben-David  1984, 75-87; Greenfeld 1987; Wuthnow
1987, 265-298). Empirical inquiry became appreciated as worthy of pursuit and
support, it was granted autonomy from control by other spheres, and it was
differentiated from other activities by establishing organizational arrangements
specifically for this endeavor. These arrangements included the separation of a
dignified social role or vocation devoted to empirical inquiry and organizational
arrangements for formulating and implementing policies shaping the endeavor, for
recruiting and training incumbents of the new social role, for validating claims to
knowledge and disseminating knowledge, for evaluating and rewarding role-
performance, for promoting and regulating communication and other role-relations,
and, more generally, arrangements for integrating the practitioners into a community
with a self-organization spanning across long distances, across nations, religions and
other social divisions. This differentiated tradition became coined 'science' and the
social role became coined 'scientist' . Science became supported by the provision of
social resources such as laboratories, apparatus, libraries, and livelihood of the
incumbents of the new role. Many of the scientists were employed as teachers in
universities and in the nineteenth century scientific research became adopted as another
purpose of many universities, along with teaching, especially in Germany. Although
the scientific role became anchored on universities, hospitals, museums and other
organizations, science continued to be practiced as a communal enterprise. The
establishment of distinctiveness, legitimation and autonomy for the pursuit of science
iscommonly refered to as the institutionalization of science (Ben-David,  1984).
This first institutionalization occurred mainly in Northwestern and Central
Europe but less extensively on the Iberian Peninsula and elsewhere.
Institutionalization was then extended to the European settler-societies around the
world, notably in North America, Australia and in what became Israel, but it did not
extend much beyond the boundaries of Western civilization. There was little
institutionalization of science on the Iberian Peninsula and therefore institutionalization
did not extend to the colonies in Latin America (Schwartzman 1991, chapters 2-3).
Despite the boundedness of the scientists movement within Europe, the
institutionalization of science had a potential for expansion.
The institutionalization of science in Europe codified a conception of nature as
invariant in time and place. Knowledge was conceived to have a truth value
independent of the researcher and therefore have the same truthfulness in all
civilizations. The participants were recognized and considered themselves as
discoverers of truth. Knowledge was considered to be cumulative and to be a
component in the broader vision of human progress as an ongoing advancement of
humanity. The doctrine of progress of humanity as an imagined global civilization
promoted a cosmopolitan orientation among the participants (Anderson 1983;
Robertson 1992; Schott 1993b, forthcoming). Natural inquiry was to be open to
participation from any part of mankind (extended later to include women), and the
resulting knowledge was to be widely disseminated as a collective good of humanity.
This cosmopolitan orientation sustained knowledge diffusion and collegial ties across
social divisions and long distances (Daston 1991). Thefaith in invariance of nature and
i n truthful ness of knowledge across places, together with the cosmopol itan orientation
of the participants, created a potential for adoption of the European tradition in the
The non-Western civilizations encountered European civilization mainly
through its military, political, economic, and religious expansion. European intruders
were not interested in imposing the European scientific tradition on indigenous
peoples, but brought local people into contact with European goods and equipment;
systems of communication, transportation, extraction, and production; and means for
curing diseases and conducting warfare. Some local people considered European
techniques better than their own and therefore judged European knowledge worth
obtaining. Thei r efforts to import that knowledge promoted interest in participating
in the creation of such knowledge, especially when some went to study at universities
in Europe where they acquired a taste for research and upon their return home
advocated establishing scientific organizations. This local legitimation and
appreciation for science led to some autonomy for efforts to assimilate the tradition
from the West (Adas 1989, 1991; Altbach and Selvaratnam 1989; Eisemon and Davis
1991; Petitjeanjami and Moulin 1992; Shils 1976, 1991). The adoption of the scientific
tradition seems to have preceded the adoption of other Western traditions such as the
ideology of human progress and rationalization. The institutionalization proceeded,
albeit with fragility and frequent moves toward deinstitutionalization, mainly around
the 19th century in the non-colonial societies and also in the colonial societies, despite
many colonizers' opposition. Institutionalization was one step of institution-building;
the other step was to establ ish social arrangements for practici ng science.
The institutionalization was accompanied by establishing institutional
arrangements comprizing a specific social roleor vocation for science, organizations for
housing the scientific role, and arrangements for mobilizing needed social resources.
Such arrangements, established in every society, are strikingly similar around the
world. The similarity of institutional arrangements can be explained by thei r common
source in world standards, both model arrangements diffusing through the web of
scientists from center to periphery and doctrines promulgated by a global science
Over the centuries scientific ideas and institutional arrangements have diffused
from a center, located where scientists have recognized the greatest accomplishment
(Ben-David  1984; Shils  1972a,  1972b). Scientists have attri buted the
achievements of the center partly to its institutional arrangements and therefore
considered the center arrangements to be especially effective. In addition, their
deference toward the achievements created in these institutions have infused a sense
of appropriateness i n the center arrangements so they have become standards to which
adherence is expected. The foreign observers' beliefs in the effectiveness and in the
appropriateness of the center arrangements have made them standards for i mitation i n
their home countries (Schott 1980, 1987b). Notably, non-Western participants regarded
Western practitioners of science as role-models and the Western role-definition of a
'scientist' became imitated as a new role in the non-Western societies. This role of
assimilator-creator-disseminator of knowledge also included a cosmopolitan orientation
and an identity as a participant in a communal endeavor transcending traditional
civilizational boundaries. Modified by the local social ecology, organizations for the
practi ce of sci ence were I i kew i se establ i shed i n the non-Western soci eti es as copi es of
Western organizational forms such as the university and the scientific academy.
M obilization of social resources has been regularized, as in the West, by taxes, by grants
and contracts for research, and by combining science with higher education.
Institutional arrangements are established not only by the process of diffusion
of center models through the global scientific community, but also by adherence to
doctrines promulgated by the global science policy regime, that is, the complex of
arrangements shaping science policy around the world (King 1974; on international
regimes in general, see Young 1989). Sci ence policy, including doctrines of sci ence and
policies for and through science, is increasingly formulated and advocated by a regime
sustained by and operating through a network of organizations such as the United
Nations Educational, Scientific and Cultural Organization (UNESCO), the Organization
for Economic Co-operation and Development (OECD), the Rockefeller Foundation, and
the World Bank. These scientists doctrines become standards that are adhered to
around the world partly because organizations such as UNESCO and OECD work
through experts in rational organization and acceleration of progress with authority to
design and recommend policies and strategies, because organizations such as the
Rockefeller Foundation sponsor scientific endeavors selectively around the world, and
because organizations such as the World Bank make loans to a state contingent on its
policy for science. The scientists doctrines of the global sci ence policy regime are also
influential because by adhering to the standards, nation-states and their leaders can
enhance their esteem as participants in the emerging global civilization of modernity
in which the scientists doctrines are part of the prevailing broad ideology of human
progress that is also promulgated by these organizations. Indeed, the global science
policy regime is an element of the world-polity (Meyer 1987; Ramirez 1987). Science
continues to be considered a component of the progress of individuals, nations, and
humankind, and there is a widespread belief that such progress can be accelerated by
rational organization and mobilization of resources for the pursuit of science. The
meaning of human progress, though, is no longer mainly religious, but focuses on
social and economic growth, power, and wealth. In mighty and wealthy societies,
progress is considered in terms of national security and competitiveness (Hill 1989),
and in poor societies it is concretized in terms of catching-up, development and
modernization (Apter 1974; Sagasti 1979; Shahidullah 1991). Thusthe rapid growth of
scientific activity in virtually all societies since mid-century can be explained as a result
of the scientism and the doctrines promulgated by the global science policy regime
(Ben-David 1991, 521-559).
The science policy regime also shapes communal participation. The rational
organization of science for national progress is thought to require integration of the
national research endeavor with world science. The doctrine against intellectual
protectionism, combined with the scientists' cosmopolitanism, counteracts parochialism
and enhances participation in communal formations spanning the globe.
Although each scientist knows very few of the other practitioners, they consider
themselves cultivators of a collective tradition, forming not only local communities but
even a global community (Anderson 1983). Their sense of intellectual and moral
communion is reinforced by their ties across social divisions and long distances and by
institutional arrangements such as the Nobel Prize ritual and international scientific
meetings and organizations, eg. the International Council of Scientific Unions, and the
agreement on free dissemination of scientific literature, an international agreement that
is vital. The global scientific community of all scientists is informally organized into
a global web of collegial ties. Between the macro-level of the global scientific
community and the micro-level of the individual scientist are numerous inter mediate
groupings. The kinds of groupings commonly examined areaKuhnian community of
specialists, a research team, the scientists in an organization, and those in a country
(reviewed by Shrum and M ullins 1988). Specialization, membership in a research team,
organizational affiliation and societal location are groupings that all shape the web of
ties among scientists (eg. Schott 1987a; Shrum 1985). For understanding national
participation in science, an especially significant grouping is the national scientific
community. The national scientific community of the scientists in a country is
informally organized by the web of their ties with one another and with foreign
colleagues. Their foreign ties shape national participation in world science. Individual
participation in world science can be described by the concept of a social circle
consisting of a scientist's web of collegial ties with the local and distant peers who are
significant for the scientist's work. These individual, national, and global webs are
nested: a scientist's collegial circle is nested in the web of the scientist's national
scientific community, which in turn is nested intheglobal scientific community. These
web enable network modeling (Schott 1991a).
Brazil in regional and peripheral locations.
The scientific tradition is being adopted in virtually every society around the
world. Science is becoming legimated and appreciated and practiced with support and
backing from the local population or at least from a powerful stratum of the society.
Science is being valued variously as an endeavor worthy in itself and variously for its
multiple uses in nation-building and national development, and perhaps also as a
symbol of modernization signalling participation in the world system of significant
societies. Institutional arrangements for scientific activity are typically not local
innovations but are usually modeled on arrangements prevai I i ng in the pi ace of highest
performance, the center of world science. Scientists from much of the world congregate
in the center. They consider the institutional arrangements in the center as most
appropriate and also as most efficient insofar as the high performance of the center
plausibly can be attributed to its institutional arrangements.
Despite the widespread modeling of the central institutional arrangements,
however, there are considerable differences among societies in their institutionalization
of science. Institutionalization differsamong societies in degree and manner. Societies
differ in their appreciation of science, in their granting of autonomy, in their provision
of social resources, and in their institutional arrangements for science, and thisshapes
their scientific activities, including their research performance (Schott 1987b). Scientific
activity in a small national scientific community cannot be self-reliant - not even Soviet
research was self-reliant (Schott 1992a, 1992b). Local scientific work is more or less
integrated with scientific traditions elsewhere, especially in the places of greatest
creativity. World science has had its shifting center, the center was in England at time
of institutionalization, then moved to France, then to Germany, and si nee the Second
World War the major center has been in the United States (Ben-David,  1984),
with Western Europe emerging as a secondary center (Schott 1991b). Scientifically
small countries have tended to be peripheries that are attached to and variously
integrated with the center (Shils  1972a,  1972b). Integration of the
periphery with the center shapes the specialization in the periphery and enhances
performance (Schott 1987b, 1992a).
This study focuses on a scientifically rather small country, Brazil. I shall
ascertain Brazilian research performance and specialization and examine the manners
and degrees of integration of Brazilian research into world science. I shall also attempt
to ascertain how these developments have been shaped by the institutionalization of
science in Brazil.
Brazi I i an soci ety i s part of the Lati n A meri can regi on. Brazi I has rather i ntense
links with other Latin American societies in most cultural domains, especially in
religion and lifestyle. In the economic and political spheres, Brazil participates in the
regional life of Latin America and is dependent on trade and political I inks with other
continents, mainly North America and Europe. Scientific ties are partly embedded in
political-economic and other links (Schott 1988). But insofar as scientists are attracted
to loci of creative work, their scientific ties may also transcend political and economic
constraints, so also for this reason wewould expect Brazilian scientists to pursueties
with col leagues in the North A meri can and Western European centers of world science.
In short, wewould expect Brazilian research to be attached to and variously integrated
with sciencedone in elsewhere in Latin America, in North America, and in Western
Historical and comparative perspectives: changes and contrasts.
My purpose is to ascertain contemporary science in Brazil. But contemporary
science can be understood better when contrasted to its own past and to science
elsewhere. Therefore I adopt a historical perspective and shall ascertain changes in
recent decades (as many years as time series data are available, which is back only a
few decades). Furthermore, I adopt a comparative perspective and shall contrast Brazil
to a place that is similar in its institutionalization of science and therefore expected I y
also similar in its research, namely the rest of Latin America, and I shall compare Brazi I
toa place that isdifferent in its institutionalization of science and therefore expected I y
also different in its research, namely Israel. These two comparisons are suitable insofar
as one is highly similar to and the other is highly different from Brazil in
institutionalization of science.
I shall attempt to provide an interpretation that is up to date. However,
although data go up to 1992, they tap outcomes of processes that have occured over
some years. For example, I shall use indicators derived from publications up to 1992.
But, typically, publications appear a year or so after submission of a manuscript and
a manuscript is the result of a project conducted in the preceding year or so which has
been designed earlier on the based of influence received even earlier. I shall use
responses from interview surveys of scientists which I directed around 1991 but the
questions focused on the scientists' activity in the period from 1985 to time of interview.
Futhermore, changes in policy and in funding do not have instant effects but affect
scientific activity, especially results of research, with a lag of some years (for example,
funding reductions in the United Kingdom entailed decline in publications only about
three years later; Irvine et al. 1985). I am emphasizing that the effects are lagged
because the recent acknowledged crisis in Brazilian science (see, e.g. Schwartzman
1991, ch. 10) may not yet have had its impact on the currently available outcome
measures up to 1992.
My study of individual, national, and global communal formations in science
takes an institutional and interactional approach (the institutional approach is reviewed
in Zuckerman 1988, and the interactional approach is reviewed in Shrum and Mull ins
1988). M y conceptual ization of nested communal formations was developed over the
years mainly from the general center and periphery theory (e.g. Greenfeld and Martin
(eds.) 1988, see especially the essay by Shils). Accordingly, my perspective shares much
with world-polity theory (e.g. Meyer 1987; Ramirez 1987) and globalization theory
(Robertson 1992) and complements or supplements the more specific political -economic
world-systems theory (e.g. Chase-Dunn 1989) and organizational field perspective (e.g.
Powell and DiMaggio 1991; Shrum 1985). The general usefulness of my nested-level
approach for disentangling complex local and global dynamics has been emphasized
by Knoke(1990, pp. 201-202), and it is compared and contrasted to the organizational
approach to research endeavors by Shrum and Bankston (forthcoming).
Institutionalization of science in Brazil, other Latin America, and Israel.
Science has become institutionalized in Brazil, in other Latin American societies
and in Israel insofar as science is legitimated, appreciated and practiced with public
backing and support. Science is appreciated in these societies partly as a cultural
endeavor in its own right and partly for its multiple uses, notably for nation-building
and national development in higher education, industry, agriculture, health services,
the military, and as a channel for social mobility. The societies differ, however, in the
firmness and pervasiveness of institutionalization and in their institutional
arrangements for practicing science.
In Brazil, like in other Latin American countries, an emerging local scientific
community has in the course of this century created a space for natural, medical and
technological science around research institutes, some of the hospitals and some of the
universities (on Brazil, see Schwartzman 1978, 1991, esp. Part 2; Velho 1990; Velho and
Krige 1984; Verhine 1991; and also Ben-David 1976, 1987; on other parts of Latin
America, see Albornoz 1991; Frame 1977; Glick forthcoming; Inter-American
Development Bank 1988; Levy 1986; Schwartzman 1985, 1986; Vessuri 1986, 1990). But
most of the up toward one hundred universities and university-like institutions in
Brazil, like the many professional schools and other tertiary schools, have primarily
emphasized teaching. In by far most of the nearly one thousand tertiary institutions in
Brazil, research has not been a significant organizational purpose, and thefaculty has
had little training and experience in research and has not engaged in training
researchers. The small space for the education of researchers implied that when the
educational system expanded rapidly si nee the late 1960s in Brazil, I ike in other Latin
American societies, the supply of researchers was far below the number of available
academic positions, which in turn entailed a perpetuation of the two-tier academic
system of a small fraction of research-oriented universities and a mass of teaching-
oriented universities and professional schools. The financial support for scientific
research, furthermore, has been rather turbulent and the public backing of science as
a value in itself and as a means for national development is not widely rooted. The
institutionalization of science is therefore rather precarious in Brazil and elsewhere in
Earlier studies comparing scientific research in Brazil to that in other Latin
American countries havefound Brazil to be rather typical of Latin America with respect
to social and human resources expended on scientific activity relative to population,
such as stock of scientists and engineers, employment of scientists and engineers in
research and experimental development, and expenditures on research and
experimental development. Brazil has also been rather typical i n terms of outcomes of
scientific activity such as scientific publications relative to the national economic
product and the influence of these publications on world science (Inter-American
Development Bank, 1988). This similarity between Brazil and the rest of Latin America
in both institutional conditions and some outcomes of course entails the hypothesis that
the more detailed analyses undertaken here will show more similarities. Indeed,
insofar as similarities are shown by all indications in the analyses, this will show the
rel iabil ity of the i ndicators.
Israel provides a contrasting case. Israel is already known to be exceptional
among the societies in the world by its high extent of scientific activity. Already in the
1960s were Israeli scientific authors exceptionally numerous relative the the national
economic product (Price 1986). This high scientific activity has continued until today
when Israel is the nation with the higest number of scientific articles relative to
population and also relative to gross national product (Schott 1993a). This exceptional
activity is the results of several conditions.
In Israel science has been highly appreciated, both in thedominant stratum and
increasingly throughout the population (Ben-David 1962, 1964, 1986, 1991, chapter 3
with Aran and chapter 13 with Katz; Katz 1979; Tal and Ezrahi 1973; Zahlan 1970).
There has been a traditionally high appreciation of learning inthejewish culture and
there has been a historically high tendency among Jews in Europe and North America
to pursue higher education and scientific activities. Israeli society has an especially
high demand for new and useful knowledge and for highly educated labor in several
sectors, specifically health services, agriculture, industry, and the military in particular.
The Israeli government and the academic institutions have had a policy of treating
knowledge as a commodity that can be produced in a transnational market for research,
notably by obtaining research grants and contracts from local and foreign funding
sources. I n short, science has i n Israel been treated as a cultural endeavor as wel I as a
social and economic investment, especially somewhat likean export industry.
Israeli science is institutionally integrated with higher education, seemingly
more than anywhere else in the world. By the middle of this century the Israeli
institutions of higher education were all converging on the nineteenth century German
university model combining research and teaching and were also all modeling the
American education of researchers through doctoral programs. This widely
institutionalized training has produced a supply of researchers largely matching and
occasionally even exceeding the demand for academic staff in the expanding Israeli
system of universities (Ben-David 1986; Schott 1987a). By the early 1970s the Israeli
research enterprise had crystallized as an essentially one-tiered set of seven universities
and several university- related hospitals performing by far most of the research in the
natural, mathematical, medical and technological sciences. The Israeli universities have
been similar in so far as they al I emphasize scientific research far more than teaching
or d i rect servi ce to the I ocal commu nity, provi de research faci I iti es and have a moderate
teaching load. The self-governing faculty have been appointed largely on the basis of
research accomplishment and the typical qualification for entry into an academic career
has been a doctoral education and postdoctoral research training, frequently abroad,
and some research experience. Likewise, advancement up through the several levels
in the professoriate has been based chiefly on the criterion of accomplishments in
research. These standards have been promoted by including foreign eminent scientists
in the evaluation of staff, departments and universities, and also on the editorial boards
of the journals published in Israel. Moreover, the universities and their affiliated
hospitals have been providing facilities for research and have been promoting ties with
foreign colleagues in several ways, by annual travel funds to each scientist and
regularized sabbaticals and frequently granted leaves of absence for sojourns in foreign
research institutes, and also by bringing visiting scientists from abroad. Actually, the
policies of the Israeli authorities in government and the academic institutions have
included establishing rather comprehensive bilateral scientific cooperation agreements
with foreign institutions and countries, for example the U.S.-lsrael Bi national Science
Foundation. These comprehensive institutional arrangements have promoted an
uncommonly strong integration of Israeli research with the centers of world science
which in turn has enhanced the research performance (Katz 1978; Schott 1987b).
The differences in institutionalization of science makes Israel a contrast to Brazil,
a contrast that expected ly is informative insofar as it can be used to ascertain and
explain developments in Brazilian research.
Concepts and their indicators based on the literature and a survey.
The concepts for descri bi ng the scientific research enterprise are either attri butes
or relations. Performance is an attribute, a characteristic of a scientist or, in the
aggregate, of a national scientific community. Specialization is also an attribute, here
a characteristic of a national scientific community. The other concepts are kinds of role-
relations between scientists or, in the aggregate, between their national scientific
communities (Burt and Schott 1989). I shall consider six analytically distinct kinds of
role-relations - namely deference, travels, influence, collaboration, emulation, and
concern for recognition. I shall briefly consider each concept and how the concept can
be measured by one or two i ndicators.
Performance i n scientific research refers to the creation of new public scientific
knowledge. Performance in a country can be indicated by the articles in the world's
major journals in science which are written by authors in the country. Another
indicator isthe extent to which the articles written in the country are subsequently cited
in the literature. Two other indicators are the extents to which scientists in the country
are named as principal contributors or as influencers on researchers surveyed in other
parts of the world. These four measures are indicators of scientific performance.
Specialization refers to the substantive foci of scientific attention, the cultivation
of some rather than other fields of research. Specialization in a country can be
indicated by the concentration of publishing in some disciplines and subdisciplines.
Deference of researchers refers to their appreciation of others' performance.
Researchers' deference toward performance can be indicated by asking them to name
principal contributors to the field.
Travel s by researchers can be i nd i cated by aski ng them to report thei r vi sits to
other institutions and their participation in meetings abroad.
Influence upon research refers to the transfer of intellectual material and its
impact on research. Influence upon researchers can be indicated by asking them who
influenced their research. Another indicator of influence upon researchers can be based
on their bibliographic citations to earlier publications.
Collaboration refers to joint research. Collaboration between researchers can be
indicated by asking researchers about thei r work jointly with others. Another indicator
of collaboration is jointly authored publications.
Emulation refers to competition among researchers to excel in the performance
of the scientific role. A researcher's emulation can be indicated by asking about felt
competition with others to be first or best in research.
Recognition refers to the validation and acknowledgment of contributions to
public knowledge, and as a social reward for role performance such recognition is often
of great concern to a researcher. A scientist's desi re for recognition from others can be
indicated by asking about caring about others recognizing the research.
These indicators are based on data from two sources: the literature of world
science and surveys of scientists in Brazil and other countries. Some of the above
concepts have several indicators, based on different approaches. This enables
methodological triangulation by which the imperfection of each approach is countered
by using several approaches. Influence, for example, can be indicated both by survey
data and by literature data. Questionnaire responses to a question asking about
influence are not perfect, eg. because the sample is not perfectly representative or
because the respondent does not recall influences. Also citations in articles are far from
perfect, eg. because not all influence is reflected in citations and because not all
citations reflect influence. But if surveys and citations show similar distributions then
we consider the indicators rather reliable.
The literature-based indicators for this study are derived from the Science
Citation Index which largely covers the world's important journals in the natural,
medical and technological sciences, including journals published in the poorer parts of
the world (Frame 1985; Garfield 1979, 1983a, 1983b; Stevens 1990). The main criterion
for i ncl ud i ng a journal i s its qual ity or i mpact, so the vol ume of i ndexed arti cl es written
by authors in a country is a reasonably good indicator of research performance i n the
country. Since 1973, the number of indexed articles has exceeded 260,000 annually.
Annually, more than 4,400 of the articles were coauthored between people at different
institutions. And the articles indexed in 1980-82 contained more than 1,835,000
citations refering to indexed articles published 1978-80. The citations and
coauthorships are indicators of influence and collaboration occuring in collegial ties.
These indicators derived from the scientific literature are far from perfect, especially
when thefocus is on the areas of the world where there may be considerable research
for local dissemination but few journals are published with significant impact on world
science (Eisemon and Davis 1989; Frame 1985; Moed 1989; Morita-Lou 1985). This is
one reason for supplementi ng the I iteratu re- based i ndicators with survey data. I n fact,
the survey yields indications si mi lar to the i ndications derived from the I iteraure and
thereby validate the indicators based on the literaure.
The other indicators are derived from a survey of collegial ties among scientists
in Brazil, elsewhere in Latin America, and in Israel, using a questionnaire ad ministered
during 1990-91. Focusing on collegial networks, the unit of analysis will not be a
respondent but a relationship or a dyad of researchers within a country or in different
countries. More precisely, the unit of analysis will be variously a travel (a visit to
another institution or a trip to a conference), an expression of deference, a named
influences, a collaborator, a competitor, or another specific kind of role-relation.
Brazilian, other Latin American, and Israeli participation in collegial networks will in
the analyses be indicated by their reported 709 travels, 499 expressions of deference,
and 1,228 influencers, etc., as listed in the tables. These relations were reported by 167
scientists, namely 64 in Brazil, 87 in other parts of Latin America (47 in Chileand 40 in
Uruguay), and 16 in Israel.
Sci enti sts were sel ected for the su rvey as fol I ows. In each cou ntry I sel ected a
geographically circumscribed site for survey. In Brazil, the site was a large city some
distance from Sao Paulo with a large university with considerable research in the
natural, medical and technological sciences. In the other part of Latin America, the sites
were the capitals in Chile and Uruguay. In Israel the site was a city with a
comprehensive university. These sites were selected because of avai labi I ity of contacts
and local interviewers. In each site the respondents were sampled mainly from a list,
namely the list of authors of articles covered by the Science Citation Index which has
a listing of authors in each city in the geographically arranged author index. The most
recent list provided a sample of authors which was classified into disciplines according
to the classification of journals they published in. Thisprovided a classification of the
sampled scientists into the disaggregate of science as eight disciplines: clinical
medicine, biomedicine, biology, chemistry, physics, earth and space science, technology
and engineering science, and mathematics. Each discipline is so broadly defined that
the eight disciplines comprize all sciences. In each site the respondents were thus
randomly sampled. H owever, this does not assure representativeness.
The restrictive selection of sites, one or two cities in each place, implies that the
ties of the selected respondents are not assured to be representative of the ties of all
scientists in Brazil, the other part of Latin America, and Israel. Representativeness can
be gauged by methodological triangulation. It turned out, as will be shown in the
tables, that in each place the respondents' reported collaborations and received
influences are similar to the collaborations and influences indicated by the
coauthorshi ps and the citations i n the articles by authors i n the place. Therefore we can
be reasonably confident that the ties reported by the respondents are rather
representative of the ties of all the scientists.
The participation rate in the survey in each site exceeded 90% of the contacted
scientists (except that I do not know the participation rate in Israel) and the response
rate on any si ngle item i n the questionnai re was also above 90%.
A questionnairewasusedfortappingtiesoftherespondents. The questionnai re
briefly asked for attributes of the respondent such as education, career, research
orientation (pure versus applied research), and research mode (experimental versus
theoretical). Travels were tapped by two questions, "Which institutions haveyou visited
in the last 12 months?" and "To which conferences abroad haveyou gone in the years from
1985 to present?" The respondent's deference was tapped by aski ng about contri butors,
" Who are the people in the world who have performed the best scientific research in your field
since 1985?" noti ng also the place of each named contri butor.
The main part of the questionnaire tapped the collegial circle around the
respondent. The interpersonal circle around a respondent can be tapped by first
identifying the significant others and then asking about their attributes and their
relations, a procedure that has recently gained codification by adoption in theGeneral
Social Survey in the U nited States (Burt 1984).
In my survey, each scientist listed significant colleagues in response to the
question, "Who are the people whose specific ideas have influenced your research sinceabout
1985? ".The tie to each named col league was tapped by the foil owing questions,
"Through which media did their ideas influenceyour research?"
" To what extent has each person influenced your choice of problems for research?"
"To what extent has each person influenced your research through your reading of the person' s
"To what extent has each person influenced your research through personal communication?"
"To what extent has the person been a collaborator on your research?"
" To what extent do you feel that you and the person are competing with one another to be
first or best in research?"
"To what extent do you careabout each person's recognition of your research?"
The attri bute of a named col league which is crucial for this study is of course the
colleague's location, "Wherein the world is each person?"
This questionnaire provides rather detailed information on collegial networks
(for validation, see Schott 1992b). The surveyed scientists were mostly working in
academic settings, in universities or in university-affiliated hospitals. Typically they
had a doctoral level education and were performing research for publication in the
literature gaining international circulation. In each place roughly half reported to be
doing mainly pure research and roughly half reported to be doing mainly applied
research. In each place, several weredoing primarily theoretical work and several were
doing primarily experimental work. The age composition was also similar across the
places, most had obtained their highest degree in the 1970s or first half of the 1980s, a
few were further in their careers and a few had less experience. The groups of
respondents were thus rather similar in these personal background characteristics.
They differed, however, in their participation in the international collegial networks,
as will be shown in the later analyses.
Performance and specialization in research.
How much of the world's scientific research has been performed in Brazil and
how does this compare to the size of the country in terms of population and economy?
How much of the world's research in each discipline has been performed in Brazil?
What have been thedirections of specialization in Brazilian research? How have the
performance and the specialization changed during recent decades?
Performance i n Brazil and in every other place is here indicated by its scientists'
share of scientific articles, the extent to which their articles are cited in other articles,
other researchers' mentions of them as pri nci pal contri butors, and mentions of them as
influencers upon research elsewhere in the world. These four indicators of scientific
performance are juxtaposed to two important conditions of science (Teitel 1987),
namely the share of Brazil in the world's population and Gross National Product, as
listed in Table!
Table 1. Scientific performance, juxtaposed to economy and population.
Percentage distribution of articles, 1986; Percentage distribution of citations, 1980-85;
Percentage distribution of contributors named in a survey; Percentage distribution of
influencers named in a survey; Percentage distribution of Gross National Product, 1986.
Percentage distribution of population, 1986.
Performance in scientific research
Articles Cites Cont r i but or s I nf I uencer s GN P
I srael 1.
North Amer i ca 40,
We stern Europe 30,
Rest of world 26 .
Popul at i on
Notes: The percentages in each column sum to 100% except
for rounding. Sources: Articles (298, 815) are from a
dataset derived from the Science Citation Index for 1986
(Institute for Scientific Information) (dataset compiled by
CHI Research; Stevens, 1990) . Citations (exceeding 5
million) are derived from the Science Citation Index 1980—
85 (Institute for Scientific Information) (data published
by Schubert et al . 1989). Contributors (759) and
influencers (2, 159) are reported in a survey of scientists
in Bangladesh, Czechoslovakia, Greece, India, Indonesia,
Japan and the then Soviet Union (Schott 1992b) . Their
mentions of compatriots are excluded from the counts. Gross
National Product is from World Bank, World Tables 1989-90
(Baltimore: Johns Hopkins University Press), P. S. Shoup,
The East European and Soviet Data Handbook (New York :
Columbia University Press, 1981) and C. L. Taylor and D. A.
Jodice, World Handbook of Political and Social Indicators.
3rd ed. (New Haven: Yale University Press) .
Population is from World Bank, World Tables 1989-90
(Baltimore: Johns Hopkins University Press) and United
Nations, Demographic Yearbook 1987 (New York: United
Nations, 1988) .
Table 1 shows that Brazil is a scientifically small country, performing much less
than 1% of the scientific research in theworld, and thisattracts much less that l%of the
citations in subsequent literature. No Brazilian scientist was among the nearly three
thousands mentioned as principal contributors or significant influential in a survey of
scientists elsewhere. Brazilian research amounted to a little less than half of the
research performed i n the rest of Lati n A meri ca and about a thi rd of that performed i n
Israel where scientific performance was high as indicated by the rather frequent
mentioning of Israelis as great contributors and influentials. In economy and
population, Brazil is roughly half the size of the rest of Latin America, as in science.
But Brazil is a whole order of magnitude larger than Israel i n terms of the economy and
even more in terms of population and yet far less research is performed in Brazil than
in Israel. This showsthat scientific performance i n a country is not a reflection of the
size of the country in terms of population or economy (there is only a very weak
correlation with population and a weak correlation with the economy; see Schott
1991b). These differences in scientific performance seem shaped by differences in
institutionalization of science.
Brazil and other Latin American countries have in recent decades been investing
many human and material resources in expanding their scientific activities. So the
question is not only one of the width of the gap to nations with higher scientific
performance, but the question is also whether the gap is narrowing. The historical
perspective is adopted by exami ning data from different years; Table 2. The avai lable
data covering different decades are unfortunately not compiled by exactly the same
method, but they are from similar sources, so the different counting methods
presumably do not alter the indications. The earliest data are counts of authors
recorded in the Current Contents which is published by the Institute for Scientific
Information that also publishes the Science Citation Index which has been used to
cou nt arti cl es from the 1970s to the 1980s and to esti mate the vol u me of I i sti ngs u p to
Table 2. Changes in scientific performance.
Authors in Brazil, in the rest of Latin America, and in Israel, as percentage of the world's
authors in Who is Publishing in Science, annually 1967 to 1973. Articles by authors in Brazil,
in the rest of Latin America, and in Israel, as percentage of world's articles in a fixed set of
journals in the Science Citation Index, annually 1973 to 1986. Listings of articles by authors
in Brazil, in the rest of Latin America, and in Israel, as percentage of the world's volume of
listings in the Science Citation Index, 1975 to 1993.
Br az i
Other LA Israel
N ote: Total numbers of authors exceeded 125,000 and of indexed articles exceeded 265,000 annually.
Source: Authors are from Who Is Publishing In Science (Institute for Scientific Information; the
numbers of authors arelisted in Price, 1986, pp. 203-205). A rticles are from the Science Citation Index
from 1973 to 1986 (Institute for Scientific Information); they are the articles in a set of journals that has
been constant, namely fixed as those indexed in 1973 and articles are classified by address of author
(dataset compiled by CHI Research; cfr. Stevens 1990). Li stings are from the published Science Citation
Index 1975 through February 1993; the percentage for a country isits percentage of the printed columns
in the geographical section of the author index.
Table2 indicates that Brazilian scientific performance grew considerably in its
share of world science from the 1960s throughout the 1970s and the 1980s to the early
1990s. Research in the rest of Latin America grew slowly from the 1960s to the early
1980s but leveled off around the early 1980s and has been rather constant in the last
decade. I srael i research grew duri ng the early 1970s and was a rather constant share
of world science from the mid-1970s through the mid-1980s but its share has been
declining in recent years. Constant share of a place during a growth of world science
of course means a conti nued growth of research i n the place.
Another way to consider these shift is in terms of rank ordering. According to
the listings, Israel ranked 15th among the nations in scientific activity in 1975-79 and
was in the 1980s surpassed by Spain and the People's Republic of China, so that Israel
moved down to 17th in science by the early 1990s. Brazil ranked 26th in 1975-79,
trailing Austria, Norway, Finland, Czechoslovakia, Hungary and South Africa, which
it surpassed in the 1980s during which it was surpassed by the People's Republic of
China, so that Brazil moved up to be the 21st among the nations in scientific activity by
the early 1990s.
Let me, however, reemphasize that the conti nued growth in Brazilian scientific
activity as measured by this outcome up to 1993, does not fully capture the
developments in the early 1990s because of the timelag between performance of
research and its resulting publication. The precarious institutionalization of science in
Brazil makes science very vulnerableto political and economic crises and the current
turbulence, crisis and decay in Brazilian is quite likely to lead to some withdrawal of
social support for science (Schwartz man, 1991, ch. 8). Following the decades of
enthusiasm and growing support for science, which gave the scientists considerable
self-confidence, the recent lessening of appreciation of science and a leveling of support
is likely to lead, if not to contraction of the national endeavor, to a loss of sense of
purpose and meaning among scientists, actuation of anomie (Ben-David  1984,
N ational research endeavors are not evenly distri buted across fields of science,
but are more or less concentrated in selected fields. The selection of foci of attention in
a country entails a national specialization. Diversity among countries in national
specialization entails a global division of labor in research. Specialization can be
revealed when control ing for the research performance of each place. I shall therefore
indicate the specialization in a particular field in a country not only relative to the
world effort in the field but also relative to the overall research effort of the country.
Specialization in a particular field is high to the extent the measure exceeds land it is
I ow to the extent the measu re i s I ess than 1, as I i sted i n Tabl e 3.
Table 3. Specialization into eight disciplines and their subdisciplines, 1970s and 1980s.
Rati o of the percentage of the arti cl es from Brazi I (from the other Lati n A meri ca
and from I srael ) that are i n the (sub)disci pi i ne to the percentage of the world's
arti cl es that are i n the (sub)disci pi ine.
PI SCI PLI NE S and subdi sci pi i nes 1970s 1980s
CLI Nl CAL MEDI Cl NE . 70 . 58
1. 39 1. 13
1. 02 1. 15
General and internal medicine
Anesthesi ol ogy
Cardi ovascul ar
Dent i s t r y
Der mat ol ogy and
Endocri nol ogy
Ger i a t r i c s
H e ma t o I ogy
I mmunol ogy
Obs t etrics & gynecology
Neurology & neurosurgery
Opht a I mo I ogy
Arthritis & rheumatology
Otorhi nol aryngol ogy
Pedi a t r i c s
Pha r mac ol ogy
Pha r mac y
Ps yc hi a t r y
Radiology & nuclear
Hygiene & public health
Miscellaneous clinical medicine
Bl OMEDI Cl NE
me d i c i n e
1. 33 1. 03
si ol ogy
to my & morphology
r yol ogy
etics & heredity
rition & dietetics
c he mi s t r y & mo I ec ul
phys i cs
r obi ol ogy
as i t ol ogy
a r biology
cytology & histol
Microscopy 1.4 .4 .8 .4
Miscellaneous bio medicine .8 .6 .5 .4
General biomedical research 1.5 1.1 1.4 .8
1. 39 1. 60
1. 39 1. 52
1. 15 1. 12
General biology 2
Mi seel I aneous zoology 1
Marine biology & hydrobi ol ogy
Agricultures food science 1
Dairy & a ni ma I science
Analytical chemistry 1
Or ga ni c c hemi s t r y
Inorganic & nuclear chemistry 1
Gener a I c hemi s t r y
Pol y me r s
1. 60 1. 82
1. 11 1. 105
Che mi c a I phys i cs 1
Sol i d s t at e phys i cs 3
Fluids & pi as mas
Acous t i cs
Opt i cs
Nuclear & particle physics 2
Miscellaneous physics 3
EARTH AND SPACE SCI ENCE
Astronomy & astrophysics 1
Meterorology & atmospheric scien
Earth & planetary science 1
Geogr a phy
Oceanography & limnology
TECHNOLOGI CAL SCI ENCE
1. 08 1. 24
1. 12 1. 13
2 2. 3
Electrical enginee & electronics
Miscellaneous engin & technology
Met a I s & met a I I ur gy
Ma t er i a I s science
Computers .4 .9
Library & information science .3 .9
Operations res & management sci .8 .6
MATHEMATI CS 1. 49 1. 58
Pr obabi I i t y & st at i st i cs .7 .9
Applied ma the ma tics .9 1.0
Gener a I ma t he mat i c s 1.7 1.8
Miscellaneous mathematics 2.7 2.5
N otesJhe 1970s denote 1973-79 and the 1980s denote 1980-86. Industrial engineering was not
reported as a subdiscipline in the 1980s. Source. Science Citation Index from 1973 to 1986
(Institute for Scientific Information), (dataset compiled by CHI R esearch; Stevens 1990).
Table 3 shows that Brazilian research was somewhat specialized. Whereasthe
Brazilian cultivation of the disciplines of biomedi cine and earth and space science was
similar to their cultivation in the world as a whole, Brazilian research emphasized
biology, physics and mathematics, with comparatively less research in clinical
medicine, chemistry and technological science. This indication of specialization is
similar to an indication obtained by a somewhat different classification procedure
which shows a Brazilian specialization in physics and mathematics and adeemphasis
especially in chemistry but also in engineering (but did not separately classify clinical
medicine, biomedi cine, biology, and earth and space science; Schubert et al., 1989, p.
The discipline with most rapid growth from the 1970s to the 1980s was
technological science. The increasing specialization in the discipline of technological
science was largely concentrated in its subdisciplines of computers, information science
and electrical engineering and electronics. These fields are known to have been
expanding in Brazil, as a results of deliberate research policies (Schwartzman, 1991, pp.
232-236). Although these are fields of increasing specialization, they havenot become
the fields of major concentrations. Specialization shows concentrations mostly in
tropical medicine and in parasitology (again, relative to the world distribution across
fields) (also this subfield specialization is consistent with that obtained by another
procedure; Schubert, 1989, p. 448). Conversely, Brazilian research hasdeemphasized
the fields of geriatrics and acoustics.
By comparison, specialization has been somewhat different in the rest of Latin
America where there has been more emphasis on the disciplines of clinical medicine
and less emphasis on physics and mathematics. But, like Brazil, other parts of Latin
America have concentrated on tropical medicine and to some degree also on
parasitology. Israeli research is less specialized into particular disciplines except
mathematics which is known to be exceptionally strong in Israel (Schott 1987b). But
there are subdisciplines with particular concentrations in Israel, notably fertility. These
indications of Israeli specialization is similar to that obtain by a somewhat different
procedure (Schubert et al., 1989, p. 416 and 458).
These specializations can be partly explained as resulting from national needs
and interests. Brazil and other Latin American countries have a considerable need for
knowledge in tropical medicine and parasitology whereas Israel has an interest in
enhancing the fertility of thejewish population (and in reducing the high birthrate of
the Arab population). But the scientists' research is also shaped by their attention to
science done elsewhere.
Deference to world science.
Scientists conti nual ly eval uate work that they are aware of. They assess not only
the truthfulness but also the worth of new knowledge. The appreciated contributions
may become exemplars influencing their own research. Their deference toward
contributions provides a mental map for orientation in the world of science. To indicate
Brazilian and other scientists' deference, the respondents in the survey were asked to
name those who had made the major contributions in their field in recent years and to
report the location of each named contributor; Table 4.
Table 4. Deference to contributors to science.
Percentage distribution across places of people named as outstanding by
respondents in Brazil, in other Latin America and in Israel.
Country of Count r y of respondents
contributors Brazil Other LA Israel
Brazil 1 8 % 2 % %
Other LA 16
Israel 1 .4 19
Nor t h Amer i ca 41 47 60
We s t e r n E u r o p e 3 4 3 6 15
Rest of wor I d 6 9 6
N cont r i but ors 170 282 47
Note: The percentages in each column sum to 100% except for rounding. Source: Survey of
scientists in Brazil, other Latin A merica and Israel.
Table 4 shows that Brazilian scientists have defered most to science in N orth
America, second most to science in Western Europe and third most to sciencedonein
Brazi I.Li kew i se, sci enti sts i n the rest of Lati n A meri ca and i n I srael al so have defered
most to science in North America, second most to science in Western Europe and third
most to local science. Deference is thus toward the easily observed and intimately
understood local work and toward work in the world centers of science. Despite
I srael 's spati al proxi mity to Eu rope, the I srael i sci enti sts defer I ess to Eu ropean sci ence
and more to North American science than the Brazilian and other Latin American
scientists do. Thiscan be explained by theembeddednessof collegial ties, including
deference, in political and economic links which are especially strong between Israel
and the United States.
The deference toward science in various places makes for attraction toward
those places and we should expect scientists to travel mainly to those places.
To what degree do scientists in Brazil travel to interact with colleagues and what
are the destinations of their travels? Scientists in Brazil and the other countries were
asked to report thei r visits to other i nstitutions i n the latest 12 months unti I the ti me of
interview and to report their participation in meetings abroad from 1985 until the
interview; Table 5.
Table 5. Travels to other institutions and to meetings abroad.
Percentage distribution of respondents' visits to other institutions; and
percentage distribution of participation in meetings abroad.
Br az i I Ot her Latin Am. I srael
Visits Meet i ngs Visits Meet i ngs Visits Meet i ngs
Brazil 6 8 % n a 8 % 1 8 % % %
Other LA 19 48 38
Israel .4 31 na
North Ameri ca 11 31 17 16 54 56
Western Europe 10 4 9 2 6 2 7 15 3 4
Rest of world 9 12 11 10
Mea n visits or
me e t i n g s f o r a 1.5 1.9 1.8 3.1 1.7 4.9
Note. The percentages in each column sum to 100% except for rounding. Source:
Survey of scientists in Brazil, other Latin A merica and Israel.
Table 5 shows that Brazilian scientists on average made 1.5 visits to another
institution in the latest year and participated in 1.9 meeti ngs abroad in the latest six or
so years. This travel i ng frequency is somewhat less than that of scientists i n the other
surveyed Latin American countries (which are much smaller than Brazil and thereby
encourage travel) and much less than that of Israeli scientists whose geographical
location provides little opportunity for travel. This major difference in travel has rather
little to do with geographically shaped opportunity but can be explained as a
consequence of different institutional arrangements, notably the Israeli scientists'
annual funds for travel as mentioned in the section on 'Institutionalization of science'.
The Brazilian scientists' visits to other institutions were mainly within Brazil.
Their travels abroad were primarily to Western Europe and secondarily to North
America. But travels were also frequent between Brazil and other Latin American
cou ntri es, i n both d i recti ons. Tabl e 5 al so shows that sci enti sts i n other Lati n A meri can
countries also traveled more to Western Europe than to North America while Israeli
scientists traveled far more frequently to North America than to nearby Western
Europe. These destinations of traveling reflect their deference orientations.
The scientists' travels and deference should expected I y be reflected in the
intellectual influence upon their research.
Influences from and upon world science.
To what extent has research in Brazil been influenced by results from within
Brazil and from other places? And, reciprocally, to what extent has research in each
place been influenced by the results of Brazilian research? Into which environment has
Brazilian research been most cohesively integrated in terms of influence? How have
influences from and upon Brazilian scientists changed in recent decades? Through
which channels have Brazilian scientists been influenced? What are the origins of
influence through publications and through personal communication? What are the
sources of influence on selection of problems for research by Brazilian scientists? These
questions will be answered here, comparing Brazilian influences with influences
i nvol vi ng sci enti sts i n the rest of Lati n A meri ca and i n I srael .
Influences upon scientists in each place are indicated by the percentage
distribution of the citations in their articles, as listed in a column in Table 6.
Table 6. Influence among national scientific communities; 1980.
Percentage distribution of citations in articles by authors in each place (column)
across places of cited author (rows).
I n f I uenc i ng
( ci ted)
I s r ael
North A me r i c a
We stern Europe
Rest of wor I d
Influenced (citing) scientists
LA I s r ael Amer
We s t e r n Rest of
ca Europe wor I
3,960 9,449 18,041 914,857 621,210 267,590
Note: The percentages in each column sum to 100% except for rounding. Source: Science
Citation Index (Institute for Scientific Information); citations in articles 1980-82 refering to
The first column in Table 6 shows that Brazilian research wasinfluenced mostly
by results from North America and then by results from Western Europe and from
within Brazil. Likewise, comparing to the second and third columns, scientists in the
rest of Latin America and in Israel were also mainly influenced by results from North
America and then from Western Europe and from within their place. These origins of
i nf I uence are highly si mi lar across the three compared pi aces. The origi ns of i nfl uence
upon the sci enti sts fol low the patterns of thei r deference more than the patterns of thei r
travels. Notably, although Brazilian and other Latin American scientists have traveled
much more to Western Europe than to North America, they havedefered much more
to North American science and have been much more influenced by science in North
America than in Western Europe.
The first row in Table 6 shows that Brazilian scientific results exerted some
influence on research in the rest of Latin America but very little on research in other
continents. By comparison, the second and third rows show a little more influence
from other Latin American science and much more influence from Israeli scienceupon
research in North America, Western Europe and elsewhere.
Table 6 shows that influence has been rather strong between Brazil and the rest
of Latin America, in both directions, relative to their small influence on research
elsewhere. This suggests some regional integration in influence. Regional integration
can beindicated astheoccuring influence (in Table 6) relative to the influence that we
should expect if influence were not embedded in particular I inks between countries but
influencees were independent of influencers. If influence were not selective, influence
would just be proportional to thei nfl uencer's tendency to exert influence and also to
the influencee's tendency to receive influence. This conception of independent
influence can be formalized (such formalization is considered in Schott, 1986, and used
in studies of scientific influence in Schott 1987a, 1988, 1992a).
The conception of behavioral independence between influencersand influencees
can be formalized by the model of statistical independence in a two-way table and
involves computing expected values. The expected values are I ike those for the usual
chi-square test of i ndependence i n a two-way frequency table except that we have no
diagonal in the table of citations from each scientific community to the other
communities. Under the model of independence, the expected number of citations
from an influenced community r to an influencing community c is the product of two
numbers P r Qc w here P r is the tendency of r to cite others and Q c is the tendency of cto
be cited by others. The expected number can be computed from the observed
frequencies of citations from each community to the other communities. Thediagonal-
less matrix of expected numbers has the same row-sums and column-sums as the
diagonal-less matrix of observed citations (modeling of a diagonal-less table with
formulas for use in computing expected numbers is in Haberman 1979, ch. 7, and is
implemented in publicly available software, Eliason 1990, pp. 16-18).
The ratio of the observed citations to the expected number of citations is a
measu re of i ntegrati on i n the web of i nf I uence. The measu res of i ntegrati on are I i sted
in Table 7. Integration has been weak to the extent the measure is less than 1 and
i ntegrati on has been strong to the extent the measure exceeds 1.
Table 7. Integration in influence among national scientific communities.
I nf I uenc i ng
( ci ted)
I s r ael
North A me r i c a
We stern Europe
Rest of wor I d
Influenced (citing) scientists
LA I s r ael Amer
lorth We stern Rest of
ca Europe wor I
N ote: Influence within each place is here ignored. Source: Same as for Table 6.
Table 7 shows in the first row that Brazilian research has been highly integrated
with science in the rest of Latin America, influence has been more than six times higher
than expected. Reciprocally, as shown in the first column, other Latin American
research has also been highly integrated with Brazilian science, influence has been more
than four times higher than expected. Integration has been stronger between Brazil and
the rest of Latin America, in both directions of influence, than between any other listed
places. Such regional integration also exists among, for example, the Scandinavian
national scientific communities (Schott 1992a). Contrasting the regional integration in
influence, integration has been as expected between Brazil and the other major places
such as North America and Western Europe but comparatively weak between Brazil
and Israel that has been integrated with North American science. The Latin American
regional integration in science can be explained by the embeddedness of scientific ties
in integrative links among Latin American countries in other spheres of life such as
education, religion, language, economy and politics. The scientific integration between
Israel and North America can be explained by its embeddedness in the strong political,
economic and educational links between Israel and the United States. Collegial ties
between scientists in different countries, more generally, tend to be embedded in other
links between their societies, notably in education, language, politics and economy
Influence can be indicated not only by citations but also by a survey asking
scientists to report influence on their research (as discussed in the earlier section on
'Concepts and their indicators'). Scientists in Brazil, in the rest of Latin American, and
in Israel named people who had influenced their research. Sources of influence are
indicated by the distribution of these named influencers across their locations; Table
Table8. Influence upon scientists from local and distant colleagues. Circa 1990.
Percentage distribution of influencers named by respondents in Brazil, in other
Lati n A meri ca and i n I srael .
Br azi I Other LA Israel
Brazil 4 3 % 4 % %
Other LA 1 35
I srael 1 1 26
Nort h Amer i ca 25 32 55
Western Europe 25 25 13
Rest of wor I d 6 4 7
N i nf I uencer s 408 710 110
Note. T he percentages in each column sum to 100% except for rounding. Source Survey of
scientists in Brazil, other Latin A merica and Israel.
Table 8 shows, like the citations in Table 6, that influences upon scientists in
Brazil, other Latin America, and Israel originate primarily from North America and
secondarily from Western Europe, with considerable influence from local colleagues.
These survey responses also indicate, I ike the citations, that there is a small but notable
regional integration in influence among scientists in different parts of Latin America.
Influence seems to have changed from 1980 to 1990 as can be seen by juxtaposing
Tables 6 and 8. Around 1980 North American influence upon research in Latin
America was roughly twice as strong as the influence from Western Europe (Table 6)
but by 1990 North American influence seemed only slightly stronger than influence
from Western Europe upon research in Latin America (Table 8). Opposite changes
seem to have occurred in the case of influence upon research in Israel, namely as
follows. Around 1980 North American influence was roughly twice that from Western
Europe but by 1990 North American influence was roughly four times the influence
from Western Europeupon research in Israel. These changes i n scientific influence may
be explained as consequences of changes in the political -economic links between the
places. During the 1980s, Latin American political-economic links seem to have grown
more with Western Europe than with North America. Conversely, West European
political-economic I inks with Israel havecooled during the 1980s. Explaining the main
differences between Tables 6 and 8 as changes in influence resulting from changing
geopolitical -economic links seems more plausible than attempting to attribute the
differences between the tables to differences in thedata sources (indexed journalsfor
one tabl e and sampl ed sci enti sts for the other tabl e) or i n the i nd i cators (ci tati ons i n one
table and a questionnaire item in the other table).
Influence is channeled through a variety of media, notably publications,
preprints, lectures, discussions, telephone, facsimile messages, postal mail, and perhaps
also through rumor. The interviewed scientists reported which one or more media
channeled influence from each named col league; Table 9.
Table 9. Media channeling influence upon scientists from colleagues.
Percent of the col leagues in each location who influenced through each medium.
Public- Pre- Lect- Disc- Tele- Fax Elec. Post Rumor NUMBER OF
a t i ons pr i nt s ur es us s i on phone ma i I ma i I COLLEAGUES
Brazil 48 19 31 89 10 2 2 12 5 177
Other LA 75 100 50 100 50 4
Israel 100 50 2
North Ameri ca 94 22 16 24 111 6 12 101
Western Europe 97 19 2 6 5 2 3 6 17 9 101
Rest of world 91 29 9 22 9 23
OTHER LATI N AMERI CA
Public- Pre- Lect- Disc- Tele- Fax Elec. Post Rumor NUMBER OF
a t i ons pr i nt s ur es us s i on phone ma i I ma i I COLLEAGUES
Brazil 50 4 31 85 4 15 26
Other LA 42 1 1 25 96 4 2 3 9 244
Israel 75 25 25 25 50 4
North Ameri ca 92 18 31 47 8 10 8 202 226
Western Europe 82 13 2 8 61 5 17 6 3 6 1 180
Rest of world 90 17 31 3 28 29
Public- Pre- Lect- Disc- Tele- Fax Elec. Post Rumor NUMBER OF
a t i ons pr i nt s ur es us s i on phone ma i I ma i I COLLEAGUES
Israel 50 44 43 82 29 7 18 18 4 28
North Ameri ca 52 38 18 70 25 15 28 37 2 60
Western Europe 57 71 21 7 9 21 5 7 4 3 14
Rest of worl d 100 75 38 25 13 25 38 38 38 8
N otes: N umber of colleagues in a location is the number of colleagues in that location who were
reported to have exerted influence through one or more media. The Israeli respondents reported
no colleagues in Brazil and the rest of Latin America. Source. Survey of scientists in Brazil,
other Latin A merica and Israel.
The I eft-most col u mn i n the u pper part of Tabl e 9 shows that Brazi I i an sci enti sts
were influenced by publications by about half their significant colleagues in Brazil,
pu bl i cati ons by more of thei r col I eagues i n the rest of Lati n A meri ca, and pu bl i cati ons
by the vast majority of their significant colleagues elsewhere. Publications were the
most frequent medium of influence from colleagues outside Latin America. But
influence from colleagues in Brazil and in the rest of Latin America was more
frequently channeled through face-to-face discussions. Also other interpersonal media
were channels of influence more frequently from local and Latin American colleagues
than from col I eagues outsi de Lati n A meri ca.
Scientists in the other part of Latin America reported influence through the
various channels to be rather similar to the channels of influence upon Brazilian
scientists. Face-to-face discussion was the most frequent medium of influence from
Latin American colleagues and publications were the far most frequent medium of
i nf I uence from col leagues outsi de Lati n A meri ca.
Brazilian scientists, however, differed from scientists in the other part of Latin
America in that they less frequently were influenced through interpersonal media such
as face-to-face discussion and postal mail. This difference indicates that Brazilian
scientists, compared to scientists in the rest of Latin America, were slightly less
integrated into networks of colleagues in the North American and European centers.
The bottom part of Table 9 shows that Israeli scientists differed considerably
from scientists in Brazil and other parts of Latin America in that publications were a
less significant channel of influence upon the Israelis, whilethe interpersonal media
were highly utilized channel of influence, not only from local colleagues but also from
colleagues in North America and Europe. Through a multiplicity of interpersonal
media, Israeli scientists have been intensely integrated into circles of colleagues in the
world centers, evidently far more integrated than Brazilian and other Latin American
The influence is exerted through the professional mass media, i.e. publications,
and through interpersonal media, i.e. direct contact between scientists. Each
respondent was asked to rate the extent to which each named col league's publ i cations
influenced the research of the respondent. The influence from each colleague's
publications was rated on a scalegoing from Of or 'none' through lfor 'little 1 and 2for
'some' up to 3 for 'great' influence of the publications on the respondent (the
questionnaire was described in the earlier section on 'Concepts and their indicators').
These ratings of influence of each named colleague's publications on the respondents
can be used to indicate influence from the various places, namely as follows. The
influence from a place is indicated by the sum of the rated influences from the
colleagues in that place, asa percentage of the overal I sum of influences from all the
colleagues. A colleague whose publications had no influence, i.e. was rated 0, is
ignored by the computation, and a colleague weighs heavily according to the rated
influence. In other words, the computation is a weighted distribution across the
sources; it is the distribution of the col leagues across places where the colleagues are
weighted by the rated influence of each colleague's publications. This indication of
sources of influence through publications is listed in the left side of Table 10.
The left-most column in Table 10 shows that Brazilian scientists were influenced
through publications written by colleagues mainly in Brazil, North America, and
Europe, and in similar amounts from these three places. Also scientists in the rest of
Latin America were influenced through publications by col leagues from Latin America,
North America, and Europe, and in similar degrees from these three pi aces. The Latin
American scientists differed from the scientists in Israel who received influence from
publications by colleagues in North America much more than in Europe.
Table 10. Influence upon scientists from publications and from personal contacts.
Percentage distribution of named colleagues, weighted by their rated influence
through their publications. Percentage distribution of named colleagues,
weighted by their rated influence through their personal communications.
Br azi I
Brazil 3 %
Other LA 1
I srael 1
North America 28
We stern Europe 33
Rest of world 8
N influencing 324
Influence from publications
Other LA Israel
Influence from contacts
Br azi I
Other LA Israel
Note The percentages in each column sum to 100% except for rounding. Source: Survey of
scientists in Brazil, other Latin A merica and Israel.
The respondent was also asked to rate extent of each named colleague's
influence through personal contact, using the same scale from to 3. The influence
from a place through personal communication isthen indicated by the sum of the rated
influences from the colleagues in that place. In other words, influences from direct
contacts in the various places are indicated by the distribution of the colleagues
weighted by each colleague's direct influence, as listed in the right side of Table 10.
The right side of Table 10 shows that the personal contacts influencing Brazilian
scientists were mainly with col leagues within Brazil whereas they were less influenced
through personal contacts with colleagues in other places such as North America and
Europe. In this respect, the Brazilian scientists differed from scientists in the rest of
Latin America who were more influenced through personal contact with colleagues
outside Latin America. But the Brazilians and other Latin Americans differed more
from the scientists in Israel who were far more influenced by personal contact with
foreign colleagues, especially in North America.
Influence is an encompassing concept. Influence on research in general
encompasses influence on selection of problems for research, influence on conceptual
frameworks, influence on methods of investigation, and influence on yet other aspects
of research. The selection of problems for research is of most i nterest for understandi ng
the direction of scientific research. Brazilian agricultural researchers have earlier
reported to be much influenced by regards for national usefulness in their selection of
problems for research (Velho, 1990). Here, the respondent was asked to rate each
named colleague's influence on selection of problems for research by the respondent,
again using the scale from to 3 as described earlier. The sources of influence on
probl em choi ce are then i nd i cated by the d i stri buti on of the rati ngs across pi aces; Tabl e
Table 11. Influence upon scientists' problem selection.
Percentage distribution of named colleagues, weighted by their rated influence
on the respondents' problem selection.
Brazil Other LA Israel
Brazil 4 2% 4 % %
Other LA 1 34
Israel 1 .4 24
N o r t h A me r i c a 2 4 3 2 5 5
We s t e r n E u r o p e 2 7 2 5 10
Rest of wor I d 5 4 11
N i nf I uenci ng 382 633 99
Note The percentages in each column sum to 100% except for rounding. Source Survey of
scientists in Brazil, other Latin A merica and Israel.
Table 11 shows that the main source of influence on Brazilian scientists' selection
of problems was their local colleagues. Secondarily, their problem choice was
influenced by colleagues in North America and in Europe, with influences in similar
amounts. Their sources were rather similar to the sources influencing selection of
problems for research by other Latin Americans. The scientists in the rest of Latin
America were influenced mainly by Latin American colleagues, secondarily by
colleagues in North America, and third most by colleagues in Europe. The Brazilian
and other Latin American scientists differed from the Israeli scientists who were less
influenced by their local colleagues and more influenced by their foreign colleagues,
especially North Americans, in their choice of problems for research.
The above analyses have disentangled sources and channels of influence on
research in general and problem selection in particular. The analyses show that there
has been a small but notable degree of regional integration in influence among
scientists in Brazil and other parts of Latin America. But Brazilian science, I ike science
done in other Latin America countries, has had little impact on research outside Latin
America, in contrast to the notably larger impact of science done in Israel. Brazilian
research, like research elsewhere in Latin America, has been influenced mainly by
science done in North America and Europe. Earlier, North American influence was
stronger than influence from Western Europe but during the 1980s the European
influence has become equally strong. By contrast, research in Israel has continually
been far most influenced by science in N orth America and this influence has become
even stronger during the 1980s. These influences have been influences on research in
general and on problem selection in particular. Brazilian scientists, likescientistsinthe
rest of Latin America and in Israel, haveselected problems under the influence of local
colleagues and have been even more influenced by their distant colleagues. The
influences have been channeled through a variety of media. Brazilian scientists, like
scientists elsewhere in Latin America, have been influenced by Latin American
colleagues most frequently by face-to-face discussions and other interpersonal media,
while influences from outside Latin America have been mediated mainly by these
distant colleagues' publications. Contrastingly, influences upon scientists in Israel from
their foreign colleagues have been channeled through a multiplicity of media, not
primarily through publications but mainly through multiple interpersonal media,
especi al I y face-to-face d i scussi ons, tel ephone, facsi mi I e messages, el ectron i c mai I , and
postal mail. The Israeli scientists have been more intensely integrated into circles of
distant colleagues, especially in North America. Israeli scientists' strong international
integration is in some degree a consequence of their past high performance, but the
high integration has a separate and considerably enhancing effect on performance (the
enhancing effect of integration on performance was estimated in another study; Schott
Influence is the intellectual substance of collegial ties in science. Another kind
of intellectual role-relation among scientists is collaboration.
To what extents have Brazilian scientists been collaborating with local, Latin
American and distant colleagues? Conversely, to what degrees have they been sought
out for collaboration by scientists in various other places? Has there been a notable
degree of regional integration among Latin American scientists in collaboration? How
has Brazilian collaboration changed in recent decades, in science-as-a-w hole and in
each discipline? How does Brazilian collaboration compare to the collaboration of
scientists in the rest of Latin America and in Israel? These questions will be answered
in this section.
Scientists' collaboration will be indicated first by scientists' own reports about
joint research. In the survey in Brazil and elsewhere, each respondent was asked to rate
the extent of collaboration with each colleague named by the respondent (as discussed
in the earlier section on Concepts and their indicators'). Each colleague's collaboration
was rated on the scale going from Ofor 'none' through 1 for 'little 1 and 2 for 'some' and
up to 3 for 'much' collaboration with the respondent. The respondents' extent of
collaboration with colleagues in a particular pi ace can then be indicated by the sum of
the rated col laboration with col leagues i n that place, computed as a percentage of the
overall sum of collaboration with all colleagues. In other words, the extent of
collaboration with various places is indicated by the distribution of the colleagues
across places, where each colleague is weighted by her or his rated collaboration; Table
Table 12. Collaboration with local and foreign colleagues.
Percentage distribution of named colleagues, weighted by their rated
collaboration with the respondents.
Br azi I Other LA I sr ael
Brazil 6 6 % 5 % %
Other LA 47
Nort h Amer i ca 9 21 55
We s t e r n E u r o p e 2 5 2 4 18
Rest of wor I d 1 3 1
N col I aborat ors 190 425 71
Note. The percentages in each column sum to 100% except for rounding. Source Survey of
scientists in Brazil, other Latin A merica and Israel.
Table 12 shows that Brazilian scientists have been collaborating mainly with
col leagues within the country, secondarily with colleagues in Europe and thirdly with
colleagues in North America. Scientists in the other part of Latin America had about
half of their collaboration with Latin American colleagues, and most of the other half
with colleagues in Europe and North America. Israeli scientists had only about one
fourth of their collaboration with local colleagues, most of their collaboration was with
colleagues in North America, and a small fraction was with colleagues in Europe. A
notabl e d ifference i s that the I srael i s have mai nl y been col I aborati ng w ith the center i n
North America, whereas the Brazilians' foreign collaboration has been mainly in
Scientists' collaboration with others can also be indicated by the extent to which
their publications are coauthored with people at other institutions, either within the
country or in other countries. To detect change over time, the degrees of local and
foreign collaboration are listed for both the earliest and latest periods for which data
are avai I abl e; Tabl e 13.
Table 13. Collaboration among national scientific communities, 1970s and
Percentage of the articles by authors in the scientific community (column) which
have coauthors i n other institutions (rows).
Brazil Other Israel North Western Rest of
Coauthors LA Amer i ca Europe world
L0j_81i_ L0j_81i_ 71i_ 8_0_s_ 7_0_s_ 8_0_s_ 7_0_s_ 8_0_s_ 70s 80s
Brazi I 16.1 .8 .1 .1 .1 .02
19.7 1.1 .2 .1 .2 .05
Other LA 1.9 20.8 .1 .2 .1 . 06
2.4 22.8 .1 .3 .3 .13
Israel .3 .1 25.7 .3 .2 .03
.5 .2 29.9 .5 .4 .07
North Ameri ca 12. 8 12. 8 10. 3 31. 4.2 2.4
13. 6 14. 1 17. 36. 9 6.9 3.9
We s t e r n E u r o p e 8 . 5 6.1 6.0 3.3 26.1 2.7
12. 4 9. 7 10. 6 5. 5 35. 9 4. 5
Rest of wo r I d 2.0 2.1 0.8 1.5 2.1 17. 1
3.2 3.9 1.5 2.5 3.6 22.
N articles 6,149 14,252 20,916 814,350 628,527 500,696
8,452 18,523 24,534 873,649 696,177 561,055
N otes: The percentages in each column sum to less than 100% insofar as many articles
are not coauthored. The 1970s denote 1973-79 and the 1980s denote 1980-86.
Source. Science Citation Index from 1973 to 1986 (Institute for Scientific Information)
(dataset compiled by CHI Research; Stevens 1990).
The coauthorsh i ps i n Tabl e 13 are si mi I ar to sci enti sts' reported col I aborati ons
in Table 12 which show that the indicators have satisfactory reliability. The Brazilians'
most extensive collaboration with colleagues at other institutions has been with
colleagues at other Brazilian institutions. An earlier study observed a rather extensive
in-house publication (i.e. publishing in printed materials issued by the institution of the
author) and that the publications had 13 to 20 percent of their citations to publications
by scientists at other Brazilian institutions and on this basis concluded thattherewas
very little communication between scientists at different Brazilian institutions (V el ho
and Krige, 1984). This percentage, however, is quite high compared to the less than 1
percent of the world literature that is written by Brazilian scientists, so the observed
percentages actually indicate a considerable communication among Brazilian
institutions. That Brazilians collaborate more with colleagues at other Brazilian
institutions than with peers in any other country and that this percentage is not much
lower than the percentage of other Latin American's publications that are with
colleagues at other Latin American institutions (a pool of publications about twice that
in Brazil) or the percentage of Israeli's publications that are with peers at other Israeli
institutions (a pool of publi cations about twice that in Brazil), as shown in Table 13, also
i ndicate considerable col I egial ties among the Brazilian institutions, quite comparable
to those among other Latin American institutions and among Israeli institutions.
Table 13 shows that the foreign collaboration has been concentrated in North
America and Europe. Theforeign collaboration from the 1970s to the mid-1980s was
apparently mainly with North American colleagues (Table 13) but then became mainly
with European colleagues (Table 12). The Brazilians' switch in collaboration from
North America toward Europe is consistent with the switch in Brazilians' received
influence, from mainly American influence toward mainly European influence (as was
discussed in the preceding section).
Table 13 shows that Brazilian scientists' collaboration has increased from the
1970s to the 1980s, both with colleagues at other Brazilian institutions and with
colleagues in every foreign place. But their foreign collaboration has increased at a
faster rate than their local collaboration. Thisfast increase in foreign collaboration is
not unique to Brazil but obtains in every place as shown in the table. The i ncreasi ng
transnational collaboration is part of the globalization of science, which may be partly
explained by the embedded ness of scientific ties in I inks among nations in spheres such
as the economy and the polity which are in a process of globalization (Schott 1991b).
Table 13 shows that Brazilian scientists have had some collaboration with
colleagues in the rest of Latin America. Their Latin American collaboration has been
small relative to their collaboration with colleagues in North America or Europe, but
evidently large compared to, say, Israeli scientists' collaboration with colleagues in
Latin America. This suggests some regional integration in collaboration within Latin
America. To highlight this integration wecan control for the tendency of each country's
sci enti sts to col I aborate.
Integration in collaboration between places can be indicated by their actual
coauthorships relative to their expected frequency if collaboration occured
independently between scientific communities, that is, in proportion to each
community's tendency for outside collaboration. The expected frequency of
coauthorships between two communities can be computed from the diagonal-less
matrix of coauthorships, like earlier described for influence. Integration in
collaboration between two places is then indicated by the ratio of the actual
coauthorships to their expected frequence. The resulting measures of integration in
collaboration are listed in Table 14. Collaboration between two places is dense to the
extent the measure exceeds 1 and collaboration is sparse to the degree the measure is
less than 1.
Table 14. Integration in collaboration among national scientific communities, 1970s and
Brazil Other Israel North Western Rest of
Coauthors LA Amer i ca Europe world
L0j_81i_ L0j_81i_ L0j_81i_ L0j_81i_ L0j_81i_ 70s 80s
Brazi I 4.2 .7 1.3 .9 .4
4.5 .7 1.1 1.0 .6
Other LA 4.2 .3 1.4 .7 .5
4.5 .2 1.3 .8 .7
Israel .7 .3 1.5 .9 .3
.7 .2 1.5 .9 .3
North Ameri ca 1.3 1.4 1.5 1.0 1.0
1.1 1.3 1.5 1.0 1.0
We s t e r n E u r o p e . 9 .7 .9 1.0 1.1
1.0 .8 .9 1.0 1.1
Rest of wor I d . 4 .5 .3 1.0 1.1
.6 .7 .3 1.0 1.1
Note: The 1970s denote 1973-79 and the 1980s denote 1980-86.
Source: Same as Table 13.
Table 14 shows that there has been a considerable regional integration in
collaboration between Brazilians and scientists in the rest of Latin America,
collaboration has been four times more frequent than expected. The regional
integration in Latin America has exceeded the integration in collaboration between
Israelis and scientists in North America which has also been strong. The integration in
Latin America has even increased slightly from the 1970s to the 1980s. In the 1970s
Brazilian researchers collaborated especially much with colleagues in North America,
but that integration has weakened and instead Brazilians have intensified collaboration
with colleagues in Western Europe and other parts of the world.
Col I aborati on vari es from one d i sci pi i ne to another. Col I aborati on may be easi er
in some disciplines than in others, perhaps especially long-distance collaboration.
Research in some disci pi i nes may be so complex that it requi res extensive collaboration.
The extents of local and foreign collaboration in each discipline can be indicated by the
frequency of coauthorships, listed in Table 15.
Table 15. Collaboration with local and foreign colleagues, by discipline, 1970s and
The percentage of the articles involving scientists in Brazil (and in the rest of
Latin America, and in Israel) which had coauthors at other institutions within
the country, and the percentage which had coauthors abroad.
Local collaboration Foreign collaboration
Brazil Other LA Israel Brazil Other LA Israel
L0i_81i_L0i_81i_L0i_81i_ 70s 80s 70s 80s 70s 80s
Sci ence- as- a- whol e 18 23 24 27 28 35 25 32 24 30 17 28
clinical medicine 27 33
biomedicine 21 29
biology 22 25
chemistry 16 21
physics 13 17
earth & space sci 10 10
technological sci 12 20
4 39 55 61 19 31 17 20 10 15
9 26 23 30 21 33 20 28 20 33
6 21 22 26 34 36 35 41 11 22
13 11 16 17 19 19 23 14 27
8 23 16 15 28 31 37 35 27 50
2 13 16 22 48 47 59 74 26 47
6 23 12 13 35 44 29 38 17 28
ma t h e ma t i c s 4 13 8 7 5 9 4 3 4 7 4 2 3 8 3 8 5 3
Notes: The 1970s denote 1973-79 and the 1980s denote 1980-1986. The number of
articles involving Brazilian authors (and authors in each of the other two places) in each
discipline in each period is at least 238. Source: Science Citation Index from 1973 to
1986 (Institute for Scientific Information) (dataset compiled by CHI Research).
Table 15 shows that collaboration varied considerably among the disciplines.
In all three places collaboration in mathematics has been infrequent within the place
but frequent with foreigners, which is undoubtedly due to the ease of long-distance
collaboration in this discipline that does not depend on tinkering with apparatus for
making empirical observations and experiments. The more experimental disciplines
have higher degrees of local collaboration.
The increase in both local and foreign collaboration in science as a whole has
occured in some disciplines more than in others. In Brazilian earth and space science
there has been no increase in local or in foreign collaboration. But there have been
considerable increases in local collaboration in mathematics, technological science and
biomedicine and in foreign collaboration in technological science, biomedicine, and
clinical medicine. The simultaneous increases in local and foreign collaboration suggest
that a discipline may be both local and cosmopolitan in its collaborative orientations.
The above analyses of collaboration have shown that Brazilian scientists have
had a considerable degree of collaboration with one another and with foreign
colleagues, mainly in North America earlier and increasingly in Western Europe. They
have also had a notable collaboration with colleagues in other Latin American
countries, so that there is some regional integration in collaboration, but the regional
collaboration is infrequent compared to the extensive collaboration with colleaguesin
the North American and European centers.
Communal attachments: Emulation and recognition.
The above analyses of influence and collaboration have exami ned the scientists'
intellectual integration into collegial networks. A scientist lives not only from
intellectual material. The intellectual integration is also sustained by a social
integration, by the scientist's communal attachments to her collegial circlewhich in turn
i ntegrates her into wider circles in the world scientific community. A scientist's social
integration will here be exami ned in terms of her emulation with colleagues and her
caring about receiving collegial recognition of her work.
Emulation refers to the desire to excel, the ambition to equal or surpass others
(the word emulation also occasionally denotes imitation but that is not the meaning
used here). A researcher competes with others i n the performance of the scientific role,
and competes with others in the creation of contributions to public knowledge,
specifically for rewards for making contributions. Emulation was tapped in the survey
of scientists in Brazil, in other parts of Latin America, and in Israel by asking each
respondent to rate the degree of felt competition with each named colleagueto be first
or best in research. The respondent rated competition with each col league on the scale
from for 'none' up to 3 for 'much' competition with the colleague. Sources of
emulation are then indicated by the distribution of the col leagues, weighted by their
degree of competition; Table 16.
Table 16. Emulation of local and distant colleagues.
Percentage distribution of named colleagues, weighted by their rated extent of
competition with the respondents.
Brazil Other LA Israel
Brazil 6 4 % 3 % %
Other LA 35
Israel 2 25
North America 5 40 48
We s t e r n E u r o p e 2 7 21 15
Rest of wor I d 3 1 11
N e mu I a t o r s 7 5 15 8 5 3
Note. The percentages in each column sum to 100% except for rounding. Source. Survey of
scientists in Brazil, other Latin A merica and Israel.
Emulation is seen, from Table 16, to originate partly from the local environment
and partly from abroad, but the scientists differ in their feelings of local versusforeign
competition. The Brazilian scientists feel competition mainly from their local
colleagues, some from their European colleagues but little from their N orth American
colleagues. Scientists in other parts of Latin A merica feel less competition from their
local col leagues but feel much more competition from their North American col leagues.
Scientists in Israel feel even less local competition and even more competition from
North American colleagues. Therefore, also in terms of emulation, Brazilian scientists
seem less integrated into the world scientific community than other Latin American
scientists and, especially, than scientists in Israel.
Social integration can also be ascertained in terms of caring for collegial
recognition as a validation and a social reward available for the performance of the
scientific role. Scientists often care about receiving recognition for their work, but they
consider recognition more i mportant from some than from others. Each respondent
rated her/ his caring about receiving recognition of her/ his work from each named
colleague, using the scale from to 3. The distribution of the col leagues, weighted by
the salience of their recognition, indicates the salience of various places as valued
sources of recognition; Table 17.
Table 17. Importance of receiving recognition from local and distant
Percentage distribution of named colleagues, weighted by their rated
importance as recognizers of the respondent's work.
Brazil Ot her LA I s r ael
Brazil 5 1 % 4 % %
Other LA 1 35
Israel .4 1 28
Nor t h Amer i ca 20 32 49
Western Europe 23 26 15
Rest of world 5 3 8
N r ecogni zers 338 629 100
Note. The percentages in each column sum to 100% except for rounding. Source Survey of
scientists in Brazil, other Latin A merica and Israel.
Recognition is valued from local col leagues and from foreign colleagues. But the
salience of the local environment and of foreign places as sources of recognition differ
among the respondents. Brazilian scientists attach about the same i mportance to local
and foreign sources of recognition. Scientists in other parts of Latin America attach less
importance to local recognition and more importance to recognition from colleagues in
North America. Israeli scientists attach even I ess significance to local recognition and
care even more about recognition from their col legues in North America.
In short, the scientists are not only intellectual I ly but also socially integrated in
terms of emulating and desiring recognition from local and distant colleagues.
Brazilian scientists seem a little more integrated with local colleagues and a little less
with distant colleagues than do scientists in other parts of Latin America and,
especially, than do scientists in Israel.
The last kind of role-relation among scientists to be examined here is the
orientation towards collegial audiences. A scientist makes her/ his findings common
property by publishing them and may often want to reach the widest possible
audience, namely the global scientific community in the scientist's specialty, but a
scientist also has a primary audience, the col leagues that are primarily addressed.
Scientists' ad dressing of audiences can be indicated by where they publish their
articles. They publish some works in locally published journals and some in journals
published in various other places. The distribution of articles written byauthorsina
scientific community across places of publication is listed in a column in Table 18. As
emphasized at the beginning of this essay, however, this study does not examine
research for I ocal consu mpti on, but focuses on research contri buti ng to worl d sci ence.
Specifically, for example, the distribution in the first column of Table 18 largely ignores
Brazilian authors' works in Brazilian journals published mainly in Portugese and in
other Latin American journals publishing mainly in Spanish. Rather, the first column
shows the primary audiences that Brazilian authors ad dress when they present their
contributions to world science.
Table 18. Audience-seeking among national scientific communities; 1982.
Percentage distribution of the articles written by authors in the scientific
community (column) across places of publication (rows).
publ i c a t i on
I s r ael
North Ame r i c a 44
We s t e r n Europe 39
Rest of First Wor I d
S e c o n d Wo r I d 2
Thi rd Worl d (ex LA)
Brazil Other Israel North Western Rest of Second Third
N articles by
authors in place
ca Europe F . Wo r I d Wor I d WexLA
1,060 2,233 3,795 187,054 135,028 8,987 33,350 15,689
/Votes: The percentages in each column sum to 100% except for rounding. Rest of First World
denotes Japan, South Africa, Australia, New Zealand; the Second World denotes the then
communist East Bloc of theSoviet U nion and Eastern Europe; and the Third World, excluding
Latin America, denotes most of Asia and Africa. Source: Science Citation Index for 1982
(Institute for Scientific Information).
The first column of Table 18 shows that Brazilian authors had their primary
audiences in North America and Europe. A I so authors in the rest of Latin America and
in Israel addressed these primary audiences. Authors in North America and in
Western Europe addressed their works primarily toward their local colleagues.
Authors in the rest of the so-called First World addressed themselves toward local
audiences and audiences in Europe and North America. Works by authors i n the then
communist countries, the so-called Second World, were addressed mainly to local
audiences, much I ess often were their primary audiences in Western Europe and even
far less frequently in North America. Authors in the so-called Third World (here
excluding Latin America listed in other columns) addressed local audiences and
audiencesin Europe and North America. Nowhere outside Brazil did authors notably
address col leagues in Brazil. And nowhere outside Latin America did authors notably
address colleagues in Latin America. Only to a very small extent did Brazilian authors
add ress col I egi al aud i ences i n the rest of Lati n A meri ca and to an even I esser extent d i d
authors i n the rest of Lati n America address a Brazi I ian audience. The smal I ness of this
orientation is probably due to the difference in language, Portugese in Brazil and
Spanish elsewhere in Latin America.
The phenomenon of audience-orientation can of course also be viewed from the
other end of the tie, the audience. The attraction of each national scientific community
as a primary audience can be indicated by the extent to which their locally published
journals are outlets for works by authors around the world. This is indicated by counts
of articles by various countries' authors published in journals in various countries as for
Tabl e 18, but now percentagi ng across cou ntri es of authors, as I i sted i n a row i n Tabl e
Table 19. Audience-attraction among national scientific communities; 1982.
Percentage distribution of the articles published in each place (row) across
countries of the authors (columns).
publ i cat i on
I s r ael
North A me r i c a
We stern Europe
Rest of First Wi
Second Wo r I d
Third Wo r I d (ex
I s r ae
No r t h
Note The percentages in each row sum to 100% except for rounding. Source: Same as Table
The first row in Table 19 shows that Brazilian scientists have been a primary
audience only for work by local authors. They have not been an audience addressed
by foreign authors. The second row shows that scientists in the rest of Latin America
have been an audience for work by Latin American authors and also somewhat of an
audience for work by authors outside Latin America, notably work by authors in North
America. Scientists in Israel have also been an audience for local work and also rather
frequently a significant audience for work by foreign authors. Likewise, scientists in
each part of the First World have been an attractive audience for local work and also
for work by others. Scientists in the Second World were an audience mainly for local
work and much less for work by authors in the First or Third Worlds. Scientists in the
Third World have been an audience mainly for local work and, like Brazilian scientists,
have hardly been addressed by other authors.
Audience-seeking and audience-attraction result partly from the performance
of authors in the various scientific communities, partly from the volume of mainstream
publications published in the various countries, the preferences of authors in certain
places for publishing in certain places, and of course also the quality of the journals
through which the audience-seeking is expressed.
The last aspect of the research enterprise to be examined here is the journals
through which the scientists present their contributions to the common stock of
knowledge. I shall examine the influence of a journal on subsequent research.
Brazilian journals shall be compared to journals published in the rest of Latin America,
in Israel, and to the mainstream journals in the world (whether or not to include a
jou rnal of cou rse occasi onal I y becomes a conf I i ct; for a Lati n A meri can case, see Vessu ri ,
1987). The journals to be examined are those that have been covered by the Jou rnal
Citation Reports of the Science Citation Indexduring 1975-88. The Brazilian journals
Revista BrasileiradeMedicina in Internal and General Medicine,
Revista Brasileira de Pesquisas Medicas e Biologicas in Internal and General Medicine,
Brazilian Journal of Medical and Biological Research in Experimental and Research Medicine,
Memoriasdo Institute Oswaldo Cruz in Experimental and Research Medicine,
Revista do Institute deMedicina Tropical deSao Paulo inTropical Medicine,
Arquivos da Escola de Veterinaria da Universidade Federal de Minas Gerais in Veterinary
Revista Brasileira deGenetica in Genetics and Heredity,
Anaisda Academia Brasileira deCienci as in Multi disciplinary Sciences, and
Pesq uisa A gropecuari a Brasileira in Multidisciplinary Sciences.
Their six fields are classes constructed by the source of these data which also
provided the classification of the Brazilian journals except two (Revista Brasileira de
Medici na and Pesquisa A gropecuari a Brasileira) which weclassified into one of these
six fields mainly according to the provided clasification of jou rnal s that they refered to.
These Brazilian journals shall be compared to the journals in their field which
were also covered by the Science Citation Index. The index in the period 1975 to 1988
included 433 journals in these six fields, namely 109 journals in Internal and General
Medicine, 55 in Experimental and Research Medicine, 14 in Tropical Medicine, 93 in
Veterinary Medicine, 74 in Genetics and Heredity, and 88 in M ultidisci pi i nary Sciences,
but not all thesejournals were included in a particular year.
These fields in the index also included seven journals published in the rest of
Latin America, namely
ArchivosdeBiologicay Medici na Experi mental es in Internal and General Medicine.
Medicina- Buenos Aires in Internal and General Medicine.
Revistade Investigation Clinica in Internal and General Medicine.
Revista MedicadeChile in Internal and General Medicine,
Archivosde Investigation Medica in Experimental and Research Medicine,
Interciencia in Multidisciplinary Sciences, and
ActaCientifica Venezolana in Multidisciplinary Sciences.
Thesixfieldsin the index also included four journals published in Israel, namely
HaRefuah in Internal and General Medicine,
Israel Journal of Medical Sciences in Internal and General Medicine,
Refuah Veterinarith in Veterinary Sciences, and
Israel Journal of Technology in Multidisciplinary Sciences.
Thesejournals will be compared to the journals published in Brazil with respect
to their influence. The influence of a journal refers to the impact of its articles on
further research, I ike the phenomenon of influence of people examined earlier in this
study. Influence of an article can be indicated by its subsequent citations, citations in
subsequent publications refering to the article (as influence was also operationalized
forTable6). Such frequency of citations to the articles appearing in the journal have
been listed in the Journal Citation Reports (Institute for Scientific Information,
annually) published together with the Science Citation Index. More precisely, this
measure of impact of a journal in a year is the ratio of the number of citations given in
the year referring to articles that appeared in the preceding two-year period to the
number of these articles in that period in the journal. The influence of a journal is
hereby measured as the average impact of its articles and is thus not affected by the
number of articles appearing in the journal (considerations of this impact measure are
provided by Sivertsen 1991, 64-68).
A journal may be compared each year to the other journals in its field which are
reported in thejournal Citation Reports in the year. The reported journals can be
ranked according to their impact each year. The percentile rank of a particular journal
denotes the percentage of the reported journals in its field that have impact at or lower
than the particular journal. Theannual percentile rank of each Brazilian, other Latin
American and Israeli journal, within their fields, are listed in Table 20. A journal is of
course ranked only for those years in which its impact measure was reported in the
Table 20. Influence of journals, by field and year, 1975-1988.
Percentile rank of each Brazilian, other Latin American and Israeli journal.
1975 '76 '77 '78 '79 '80 '81 '82 '83 '84 '85 '86 '87 '88
Internal and General Medicine
Revista Brasileira de M edicina
12 12 2 12 11111
Revista Brasileira de Pesquisas M edicas e Biologicas
4 2 8 6 7
Archivos de Biologica y M edicina Experimental es
. . . . 47 65 69 50 63 60 59 59 58
M edicina - Buenos Aires
43 17 57 47 49 52 53 40 40 27 35 34 42 29
Revista de I nvestigaci on Clinica
14 33 21 27 31 23 38 27 13 19 13 19 3 4
Revista Medicade Chile
11 20 31 26 13 20 24 17 35 20 33 21 22 12
Israel Journal of M edical Sciences
53 57 69 63 60 67 59 53 62 45 42 40 38 41
Experimental and Research Medicine:
Brazilian Journal of Medical and Biological Research
28 47 31 25 29 28 30
M emorias do I nstituto Oswaldo Cruz
Archivos del nvestigaci on Medica
8 7 17 12 6 38 10 21 5 8 2 2 2 6
Revista do I nstituto de M edicina Tropical de Sao Paulo
16 28 21 30
Veteri nary M ed i ci ne:
Arquivos da Escola de Veteri naria da Universidade Federal de M inas G erais
2 2 3 5 2 1....
R ef u ah V eteri n ari th
. 32 25
Genetics and H eredity:
Revista Brasileira de G en etica
4 4 6 7. 11
M ultidiciplinary Sciences:
Anais da Acad emia Brasileira deCienci as
28 5 15 51 34 29 29 41 22 49 23 17 10 12
Pesq u i sa A gropecu ari a B rasi I ei ra
10 3 3 5 6 5
. 29 52 52 26 20 33 45 51 60 37 29
33 22 50 42 42 38 20
Israel Journal of Technology
45 49 25 7 19 8 12 23 1 11
N ote: Years in which influence was not reported are indicated by a dot (.) Source: Journal Citation
Reports for 1975 to 1988 (Institutefor Scientific Information).
Before interpreting the rankings in Table 20, I reemphasize that the journals
included in the Journal Citation Reports (as in the Science Citation Index) are largely
the jou rnal s that are of emi nence. Therefore, the I i steal jou rnal s and thei r compari sons
are largely journals with considerable influence, so the ranking is among influential
journals. A journal listed with a low percentile rank is thus a journal which seemingly
is influential and much more influential than most journals in the world (largely those
that are not included in the data source) but is not among the few most influential
journals in its field. Table 20 shows a tendency for the least influential ones to be
included only occasionally, notably the two Israeli journals with titles in Hebrew for
medicine and veterinary medicine (in HaRefuah the articles are published in Hebrew
and the abstracts are also published in English).
The first row of Table 20 shows that the Revista Brasileirade Medici na has
not been one of the top journals among the covered journals in Internal and General
Medicine which were reported in the Journal Citation Reports. Another Brazilian
journal in the same field, Revista Brasileira de Pesquisas Medicas e Biologicas, has
also not been among the top journals. The four journals published elsewhere in Latin
America have ranked somewhat high while one Israeli journal ranked low and one
Israeli journal ranked rather high among the covered journals in Internal and General
In Experimental and Research Medicine, the Brazilian Journal of Medical and
Biological Research has been quite influential while another Brazilian journal,
Memoriasdo Institute Oswaldo Cruz, was less influential (and was probably found
not to be so influential as to be covered since the 1970s). A journal published elsewhere
in Latin America has been among the rather influential journals in Experimental
In Tropical Medicine, the Brazilian journal Revista do Institute de Medicina
Tropical deSao Paulo, has been quite influential.
In Veterinary Medicine, the Brazilian journal, Arquivos da Escola de
Veterinaria da Universidade Federal de Minas Gerais, has been one of the less
influential ones and also the Israeli journal Refuah Veterinarith has been one of the I ess
influential ones among the covered journals in the field.
The journal Revista Brasileira deGenetica has been among the I ess influential
ones in Genetics and Heredity.
In the Multidici pi i nary Sciences, the A nais da Academia Brasileira deCiencias
has been of changing influence, apparently more influential around the first half of the
1980s than earlier and later. The other Brazilian journal in this field, Pesquisa
Agropecuaria Brasileira, has been among the less influential one. The two journals
published elsewhere in Latin America have been somewhat more influential while the
Israeli journal has apparently become I ess influential overtime.
To summarize these rankings among the influential journals within each field,
about a handful of Brazilian journals have evidently been among the highly influential
journals in the world, and a couple of these Brazilian journals have actually been
especially influential. Another handful have been somewhat influential. Apartfrom
annual fluctuations which may be due to just an occasional influential article, the
influence of each Brazilian journal has apparently been rather stable over the years,
with the exception of onejournal that apparently had an especially influential period.
A mong the other Latin American and Israeli journals in these fields, some were more
and some were I ess influential, and most were stable but several were apparently of
declining influence. Compared to the mixed stableand declining influence of journals
published in the rest of Latin America and Israel, the stability of influence of the
Brazilian journals appears as a slight comparative increase.
The above series of analyses were undertaken to answer the question of how
Brazilian scientific performance, specialization and ties with local, regional and central
colleagues have been shaped by the institutionalization of science in Brazil. Therefore
trends in Brazilian research were ascertained and Brazil was compared to the other
Latin America and to Israel which have had similar and different, respectively,
Science has become institutionalized in Brazil in so far as it has become rather
appreciated and been granted considerable autonomy and some support. Scientific
activities such as research and training of researchers, however, have been concentrated
in a fraction of the universities in Brazil and elsewhere in Latin America, unlike in
Israel where scientific research and training has been the primary purpose at all the
This difference between Brazil, or Latin America more generally, and Israel in
their institutional arrangements for science has enhanced scientific performance in
Israel so that research performance has been higher in Israel than in Brazil and in the
rest of Latin America. Israeli performance has been higher not only relative to
population and the economy but also in total, despite the smal I ness of the country. But
scientific research has expanded in Brazil, and the Brazilian rate of increase was higher
than in the rest of Latin America, in Israel, and in the world as a whole. The Brazilian
performance has been increasing throughout the 1980s. A decline in social and human
i nvestment i n scientific activity takes some years to cause a decl i ne i n the outcomes of
research. The recent Brazilian crisis has not showed up as a decline in the Brazilians'
publications by 1993 but will probably entail a decline within a few years.
Brazilian research has been specialized in so far as certain disciplines and
specialties have been foci of scientists' attention in Brazil more than elsewhere.
Brazilian specialization has emphasized the disciplines of physics, biology and
mathematics, it has been more typical in biomedi cine and earth and space science, and
has deemphasized clinical medicine, chemistry and technological science. Within
medicine, though, tropical medicine and parasitology have been fields of strong
specialization in Brazil, I ike in other parts of Latin America, but unlike in Israel where
specialization has focused on mathematics and fertility. Brazilian growth has been
highest in technological science, especially in fields of computing. These directions of
Brazilian specialization seem shaped by national needs and research policies.
Brazilian scientists have been tied to colleagues who have influenced their
research, and also been variously collaborators and competitors in research, and often
also been significant recognizers of their work. None of these ties have been confined
within Brazil but have also been rather extensive with foreign colleagues. Brazilian
research has been integrated with science in the rest of Latin America insofar as
regional influence and collaboration have been higher than expected. This integration
in collegial ties has been promoted by their embedded ness in the strong links between
Brazil and the rest of Latin America in other spheres of life. The regional integration
has been notable but has been overshadowed by the attachment to the centers of
science, located in North America and Western Europe. Scientists in Brazil, like
elsewhere, havedefered mainly to these centers, have traveled there, been influenced
by science in the centers, occasionally collaborated there and valued recognition from
colleagues in the centers. The involvement of Brazilian scientists with the centers,
though, has been slightly less than the involvement of other Latin American scientists
and much less than the participation of Israeli scientists in the centers. The Israeli ties
with the centers, especially the North American center, have been enhanced by their
embeddedness in the strong political and economic I inks between Israel andtheUnited
States and also by institutional arrangements for science such as comprehensive
scientific cooperation agreements with the centers and travel funds which are results
of Israeli research policies enhancing research performance through integration with
centers of science.
Adas, Michael. 1989. Men, Machines as the Measure of Men: Science, Technology and
Ideologies of Western Domination . Ithaca, New York: Cornell University Press.
Adas, M ichael. 1991. Scientific standards and colonial education in British India and French
Senegal. Pp. 4-35 in Teresa Meade and Mark Walker (eds.), Science, Medicine
and Cultural Imperialism . N ew York: St. M arti n's Press.
Albornoz, Orlando. 1991. Latin America. Pp. 853-867 in Philip G. Altbach (ed.),
International Higher Education: An Encyclopedia . N ew York: Garland.
Altbach, Philip G., and Viswanathan Selvaratnam (eds.). 1989. From Dependence to
Autonomy: The Development of Asian U niversities . Dordrecht: Kluwer.
Anderson, Benedict. 1983. Imagined Communities . London: Verso.
Apter, David E. 1974. The role of the new scientific elite and scientific idelology in
modernization. Pp. 398-408 in Sal P. Restivo and Christopher K. Vanderpool
(eds.), Comparative Studies in Science and Society . Columbus, Ohio: Merrill.
Ben-David, Joseph. 1962. Scientific endeavor in Israel and the United States. The
American Behavioral Scientist , 6, 4 (December): 12-16.
Ben-David, Joseph. 1964. M ada BaEretz Katana [Science in a small country].
H aOni versita (H ebrew magazine published by the Hebrew University of
Ben-David, Joseph. 1976 Report on a visit to Brazil. Rio dejaneiro: Financiadora de
Estudos e Projetos [manuscript; translated and published in 1987 as listed
Ben-David, Joseph.  1984. The Scientist's Role in Society: A Comparative Study .
2 nd ed. Chicago: University of Chicago Press.
Ben-David, Joseph. 1986. Universities in Israel: dilemmas of growth, diversification and
administration. Studies in Higher Education , 11:105-130.
Ben-Davi d, J oseph. 1987. U niversidade e ciencia observadas por Ben-D avid. R elatorio de uma
visita ao Brasil (25 dejulho - 8 de agosto de 1976). Ciencia Hoje , 7 (November
Ben-David, Joseph. 1991. Scientific Growth: Essays on the Social Organization and
Ethos of Science . Berkeley: University of California Press.
Burt, Ronald S. 1984. Network items and the General Social Survey. Social Networks ,
Burt, Ronald S, and Thomas Schott. 1989. Relational contentsin multiple network systems.
Pp. 185-213 in L. C. Freeman, A. K. Romney, and D. R. White (Eds.),
Research Methods in Social Network Analysis . Fairfax, Va.: George Mason
Cabral, Regis. 1992. Latin A merica, Spain and Portugal: Scientific networks and exchange
with Sweden. Science Studies , 5, 2:97-107.
Chase-Dunn, Christopher. 1989. Global Formation: Structures of the World-Economy .
Oxford: Basil Blackwell.
Daston, Lorraine. 1991. Theideal and reality of 'the Republic of Liters in the Enlightenment.
Science in Context 4:367-386.
Eisemon, Thomas O., and Charles H . Davis. 1989. Publication strategies of scientists in
four peripheral Asian scientific communities: Some issues in the measurement and
interpretation of n on -main stream science. Pp. 325-375 in Philip G. Altbach et al.,
Scientific Development and Higher Education: The Case of Newly
Industrializing Nations . New York: Praeger.
Eisemon, Thomas O., and Charles H. Davis. 1991. University research and the
development of scientific capacity in sub-Saharan A frica and A sia. I n International
H igher Education: An Encyclopedia , edited by Philip G. Altbach, 275-295. New
Eliason, Scott R. 1990. The Categorical Data Analysis System Version 3.5 User's
Manual . Iowa City: University of Iowa.
Frame, J. Davidson. 1977. M ain stream research in Latin America and the Carribean.
Frame, J . Davidson. 1985. Problems in the use of literature-based science and engineering
indicators in developing countries. Pp. 117-122 in H . M orita-Lou (ed.), Science and
Technology Indicators for Development . The United Nations Science and
Technology for Development Series. Boulder: Westview.
Garfield, Eugene. 1979. Citation Indexing: Its Theory and Application in Science,
Technology, and Humanities . New York: Wiley.
Garfield, Eugene. 1983a. Third World research. Parti. Whereis it published, and how often
it is cited. Current Contents , 33 (August 15):5-15.
Garfield, Eugene. 1983b. Third World research. Part 2. High impact journals, most-cited
articles, and most active areas of research. Current Contents , 34 (August
Glick, Thomas F. Forthcoming. "Science in the Twentieth Century." Cambridge
History of Latin America. Vol VII .
Greenfeld, Liah. 1987. Science and national greatness in seventeenth-century England.
Greenfeld, Liah, and Michael Martin (eds.). 1988. Center: Ideas and Institutions .
Chicago: University of Chicago Press.
Haberman, Shelby. 1979. Analysis of Qualitative Data . New York: Academic Press.
Hill, Christopher T. 1989. Technology and international competitiveness: methaphor for
progress. In S. L. Goldman (ed.), Science, Technology, and Social Progress .
Bethlehem, Pennsylvania: Lehigh University Press).
Institutefor Scientific Information. Annually, journal Citation Reports . Philadelphia:
Institute for Scientific Information.
Institutefor Scientific Information. Annually. Science Citation Index . Philadelphia:
Institutefor Scientific Information.
Inter-American Development Bank. 1988. Economic and Social Progress in Latin
America. 1988 Report . New York: Inter-American Development Bank.
Irvine, John, Ben Martin, Tim Peacock and Roy Turner. 1985. Charting the decline in
British science. Nature , 316 (August 15)587-590.
Katz, Shaul. 1978. Quantity and quality in Israeli science. Science and Public Policy ,
King, Alexander. 1974. Science and Policy: The International Stimulus . Oxford:
Oxford University Press.
Knoke, David. 1990. Political Networks: The Structural Perspective . Cambridge:
Cambridge University Press.
Levy, Daniel C. 1986. Higher Education and the State in Latin America. Chicago:
University of Chicago Press.
Merton, Robert K.  1970. Science, Technology and Society in Seventeenth-
Century England . N ew York: H arper and Row.
M eyer, J ohn W. 1987. The world polity and the authority of the nation-state. Pp. 41-70 i n
G. M. Thomas, J. W. Meyer, F. O. Ramirez, and J. Boli (eds.), Institutional
Structure: Constituting State, Society, and the Individual . Newbury Park: Sage.
Moed, Henk F. 1989. The Use of Bibliometric Indicators for the Assessment of
Research Performance in the Natural and Life Sciences . Leiden: DSWO Press.
Morita-Lou, H. (ed.). 1985. Science and Technology Indicators for Development . The
United Nations Science and Technology for Development Series. Boulder:
Petitjean, Patrick, Catherinejami and Anne Marie Moulin (eds.). 1992. Science and
Empires: Historical Studies About Scientific Development and European
Expansion . Dordrecht: Kluwer.
Powell, Walter W., and Paul J. DiMaggio (eds.). 1991. The New Institutional ism in
Organizational Analysis . Chicago: University of Chicago Press.
Price, Derek J. de Sol la. 1986. Little Science, Big Science . . . And Beyond . New
York: Columbia University Press.
Ramirez, Francisco O. 1987. Institutional analysis. Pp. 316-328 in G. M . Thomas, J. W.
Meyer, F. O. Ramirez, and J. Boli (eds.), Institutional Structure: Constituting
State, Society, and the Individual . Newbury Park: Sage.
Robertson, Roland. 1992. Globalization: Social Theory and Global Culture . London:
Sagasti, Francisco R. 1979 Technology, Planning, and Self-Reliant Development: A
Latin American View . New York: Praeger.
Schott, Thomas. 1980. Fundamental research in a small country: M athematics in Denmark
1928-1977. Minerva 18:243-283.
Schott, Thomas. 1986. M odels of dyadic and individual components of a social relation:
applications to international trade, journal of Mathematical Sociology , 12:225-249
Schott, Thomas. 1987a. Interpersonal influence in science, mathematicians in Denmark and
Israel. Social Networks , 9:351-374.
Schott, Thomas. 1987b. Scientific productivity and international integration of small
countries: mathematics in D enmark and Israel. M i nerva , 25:3-20.
Schott, Thomas. 1988. International influence in science beyond center and periphery. Social
Science Research , 17:219-238.
Schott, Thomas. 1991a. N etwork models. Pp. 107-206 i n STRU CTU RE Reference M anual .
New York: Columbia University.
Schott, Thomas. 1991b. The world scientific community: globality and globalisation.
Minerva , 29:440-462.
Schott, Thomas. 1992a. Scientific research in Sweden: orientation toward the A merican centre
and embeddedness in N ordic and European environments. Science Studies , 5:13-27.
Schott, Thomas. 1992b. Soviet science in the scientific world system: was it autarchic, self-
reliant, distinctive, isolated, peripheral, central? Knowledge: Creation, Diffusion,
Schott, Thomas. 1993a. Israel's scientific performance. Near East Report , 37, 1 (4
Schott, Thomas. 1993b. World science, globalization of institutions and participation.
Science, Technology, & Human Values , 18:196-208.
Schott, Thomas. Forthcoming. Therise, spread and growth of science-Joseph Ben-David's
contribution and a continuation.
Schott, Thomas. Forthcomi ng. D eclining and emerging centers of technological science The
United States and Japan.
Schubert, A., W. Glanzel, and T. Braun. 1989. Scientometric datafiles. A comprehensive set
of indicators on 2649 journals and 96 countries in all major science fields and subfields. 1981-
1985. Scientometrics , 16:3-478.
Schwartzman, Simon. 1978. Struggling to be born: the scientific comunity in Brazil.
Minerva , 16:545-580.
Schwartzman, Simon. 1985. The quest for university research: policies and research
organization in Latin America. Pp. 101-116 in B. Wittrock and A. Elzinga (eds.),
The University Research System: The Public Policy of the Home of Scientists .
Schwartzman, Simon. 1986. Coming full circle: a reappraisal of university reform in Latin
America. Minerva , 24:456-475.
Schwartzman, Simon. 1991. A Space for Science: The Development of the Scientific
Community in Brazil . College Park: Pennsylvania State University.
Shahidullah, Shahid M. 1991. Capacity-Building in Science and Technology in the
Third World: Problems, Issues, and Strategies . Boulder: Westview Press.
Shils, Edward.  1972a. Center and periphery. Pp. 93-109 in E. Shils, The
of Society . Chicago: University of Chicago Press.
Shils, Edward.  1972b. M etropolis and province in the intellectual community. Pp.
355-371 in E. Shils, The I ntel lectuals and the Powers, and Other Essays . Chicago:
University of Chicago Press.
Shils, Edward. 1976 The cognitive expansion from centreto periphery. Pp. 63-95 in Say
Not the Struggle , 63-95. Delhi: Oxford University Press.
Shils, Edward. 1981. Tradition . Chicago: University of Chicago Press.
Shi Is, Edward. 1988. Center and periphery: an idea and its career, 1985-1987. Pp. 250-282,
in Liah Greenfeld & Michael Martin (Eds.), Center: Ideas and Institutions .
Chicago: University of Chicago Press.
Shi Is, Edward. 1991. Reflections on tradition, centre and periphery, and the universal validity
of science, the significance of the life of S. Ramanujan. M i nerva 29:393-419.
Shrum, Wesley. 1985. Organized Technology: Networks and Innovation in Technical
Systems . West Lafayette: Purdue University Press.
Shrum, Wesley, and Carl Bankston. Forthcoming. Organizational and geopolitical
approaches to international science and technology networks. Knowledge and Policy .
Shrum, Wesley, and Nicholas C. Mullins. 1988. Network analysis in science and
technology. Pp. 107-134 in Anthony F. J. van Raan (ed.), Handbook of
Quantitative Studies of Science and Technology . Amsterdam: North-Holland.
Sivertsen, Gunnar. 1991. Internationalization via journals: Scientific and Scholarly
journals Edited in the Nordic Countries . Nord 1991:49 . Copenhagen: Nordic
Council of Ministers.
Stevens, Kimberly A. 1990. Updating and maintaining fourteen bibliometric data
through 1986 . H addon Heights, N.J. : CHI Research.
Tal, Eliezer, and Yaron Ezrahi (eds.). 1973. Science Pol icy and Development . (London:
Teitel, Simon. 1987. Science and technology indicators, country size and economic
development: A n international comparison. World Development , 15:1225-1235.
Velho, Lea, and John Krige. 1984. Publication and citation practices of Brazilian
scientists. Social Studies of Science , 14:45-62.
Velho, Lea. 1990 Sources of influence on problem choice in Brazilian university agricultural
science. Social Studies of Science , 20:503-517.
Verhine, Robert E. 1991. Brazil. Pp. 885-898 in Philip G. Altbach (ed.), International
Higher Education: An Encyclopedia . New York: Garland.
Vessuri, Hebe(ed.). 1984. Ciencia academica en la Venezuela moderna . Caracas.
Vessuri, Hebe. 1986. The universities, scientific research, and thenational interest in Latin
America. Minerva , 24:1-38.
Vessuri, H ebe. 1987. La revista cientifica periferica: El caso de A eta Cientifica Venezolana .
Interciencia , 12:124-134.
Vessuri, Hebe. 1990. inventamosoerramos: the power of science in Latin America. World
Development , 18:1543-1553.
Wuthnow, Robert. 1987. Meaning and Moral Order: Explorations in Cultural Analysis.
Berkeley: University of California Press.
Young, Oran R. 1989. International Cooperation: Building Regimes for Natural
Resources and the Environment . Ithaca, N.Y.: Cornell University Press.
Zahlan, Antoine. 1970. Science and Higher Education in Israel . Beirut: Institute for
Zuckerman, Harriet A. 1988. The sociology of science. Pp. 511-574 in Neil Smelser (ed.),
Hand book of Soci ol ogy . Newbury Park: Sage.