Contemporary Research Environment
THE U.S. SCIENTIFIC RESEARCH ENTERPRISE
Brief Historical Perspective
The U.S. academic research enterprise of the 1990s differs in kind and scale from that of earlier decades. Once an informal, intimate, and paternalistic endeavor, research today is conducted as part of a more formal, complex, highly diversified enterprise that engages the talents of a broad spectrum of individuals and institutions. The organizational structures within which research is supported and performed, the climate within which research is conducted, and the criteria that define scientific achievement today are quite different from those that were in place previously.
The uniquely American multipurpose university was firmly established in the United States by 1890 and thereafter was gradually augmented by professional schools and institutes of technology (Geiger, 1990). Because institutional support for scientific work was scarce in the early part of the nineteenth century, research was usually an avocation rather than a profession. Later in the century, as the university system evolved and the idea of the pursuit of science for its own sake gained support, research was pursued as a full-time vocation (Daniels, 1967).
Throughout the first half of the twentieth century, universities
retained the tradition of a community of independent scholars characterized by autonomy, individuality, and a diversity of research interests. Some faculty research was commercially or industrially oriented, particularly in the engineering schools and in chemistry departments. Some faculty followed government research interests in agriculture. Still others pursued independent research interests with small amounts of philanthropic support.
In response to the vital contributions of science and technology to U.S. victories in World War II, Bush (1945) and Steelman (1947) called for increased government support of research. The Bush report inspired a postwar relationship between government and the scientific community that sought to extend the successes of both government-organized projects such as the Manhattan Project and university-based research such as that performed at the Radiation Laboratory at the Massachusetts Institute of Technology. Both models of scientific work were eventually implemented, and the Bush report provided the blueprint for continued federal support of academic science through a decentralized process driven by investigator-initiated research proposals, eventually institutionalized with the establishment of the National Science Foundation in 1950.
The post-World War II years were thus the formative period for a more intimate relationship between the U.S. government and the scientific research community. The development of the contemporary system of federal support for university-based basic research and the education of new researchers provided the platform for the current preeminence of U.S. research (GUIRR, 1989). This system grew rapidly through the late 1960s (Brooks, 1989).
In the 1970s economic stagnation and concern about the cost of research and the social impact of science-based technologies led to a reexamination of the basic rationale governing federal investments in scientific research (GUIRR, 1989; Brooks, 1989). This reexamination led in turn to increased oversight and involvement of public officials with both science and technology. New regulatory requirements and new standards of accountability were imposed (OTA, 1986a).
In the 1980s renewed growth in federal funding for scientific research stimulated changes in the academic research environment. Support increased for research and development centers, large projects, and single-disease or single-technology programs, often called directed or mission-oriented research. But the accompanying increases in the size of academic administrative staffs and the amount of research overhead costs created concerns among sponsors and faculty.1 In the face of increasing federal budget deficits in the late 1980s and decreasing economic and educational performance by the nation,
motivations for funding research have focused increasingly on technological innovation, economic competitiveness, and education. This shifting rationale for federal support has been accompanied by demands for tighter management and oversight of research.
How the contemporary research environment affects the integrity of the research process and the incidence of misconduct in science is poorly understood. But individual scientists and public officials have expressed concern about several factors that may foster dishonest behavior, which can range from subtle exaggeration of the value of research results to actual fabrication or falsification of research findings. One such factor is the pressure associated with producing research results to attract and maintain stable funding in a research system that cannot support all meritorious research proposals. Such pressure could erode the high standards of honesty and open collaboration that have traditionally characterized the scientific community. This and other concerns, coupled with heightened public awareness of waste, fraud, and abuse in other publicly supported activities, suggest that government oversight of the conduct of scientific research is likely to continue, if not increase. Such scrutiny has profound implications for the system of internal checks and balances in the research enterprise, which were designed for a research environment far removed from the forces of the political process.
THE CHANGING RESEARCH SCENE
Many factors have contributed to the evolving research scene, including the increasing complexity of contemporary research problems and instrumentation, the increasing costs of scientific research, changes in the rationale for supporting and monitoring government-funded research, and increased regulation of federal research. Other principal factors affecting the research environment include the scale, scope, and organization of research centers and groups; the changing character of collaborative efforts; the growing number of contenders for research funds; the reward system; and increasing emphasis on commercialization of research results. Combined, these factors exacerbate conflicts that have always been apparent to some extent in scientific research.
One example of the recent and profound changes characterizing the contemporary research environment is the changing nature of its basic organizational unit. Twenty years ago, a hypothetical laborato-
ry group consisted of less than a half-dozen members. The group was small, closely knit, and composed of individuals who generally shared a common cultural heritage. The group accepted, often without conflict, a hierarchical structure of relationships and shared a common set of craft skills and moral standards, and its members followed well-understood lines of communication.
Today, although many research groups still consist of less than a dozen members, larger and more diverse research groups are becoming more common. The group members in large research teams differ in status; they include research investigators, undergraduate students, postdoctoral researchers, visiting faculty, and technicians. These individuals report, sometimes in an ill-defined manner, to a research director who frequently has many more professional and institutional obligations than his or her counterpart of 20 years ago.
Interpersonal conflicts and professional rivalries have always been part of the scientific culture. Yet good communication, good mentoring, and research supervision may be more difficult to achieve and to sustain in a large, complex, and democratic group environment (Phillip, 1991). Most research supervisors recognize the importance of good manners, civility, professional support, and personal interaction in their laboratories. However, the diverse social environment and the conflicting expectations of researchers offer increased opportunities for misunderstandings and unresolved disputes. If such disputes are not responsibly addressed, they sometimes can lead to allegations of misconduct in science, perhaps accompanied by an accusation that there has been a threat of reprisal. In the current environment, what has traditionally been regarded as an internal concern of a research laboratory or university can be escalated, sometimes rapidly, to a problem involving complex relationships and formal procedures between government agencies and research institutions. Questionable behavior in the research environment today is being publicized and publicly criticized.
Misconduct in science can occur, and allegations of misconduct must be treated seriously. But some complaints may simply reflect a poor research environment rather than actual misconduct in science. The best way to avoid or minimize research disputes is to establish a proper research environment. Research supervisors must devote attention to maintaining an atmosphere of open communication and cooperation in their research groups, with opportunity for appropriate participation by and recognition of all parties. Considering human relationships and interactions is an important aspect of good research practice.
Increased Size and Scope of the Research Enterprise
The U.S. research enterprise is larger today than it has ever been, whether measured in terms of numbers of institutions, research groups, investigators, postdoctoral fellows, technicians, graduate students, proposals, funding, research findings, articles, or knowledge produced. Scientific discoveries, patents, and publications in the post-World War II era all demonstrate the remarkable growth that has occurred in every field and discipline.
Federal funding of academic research and development has grown dramatically over the past 30 years, from less than $2 billion in 1958 to more than $8 billion in 1989, in constant 1988 dollars (GUIRR, 1989). The total number of scientists and engineers employed by universities increased from 120,000 in 1958 to 330,000 in 1988, while the number of Ph.D. degrees awarded annually increased from 6,000 to around 19,000 in the same period (GUIRR, 1989). Also during this time period, the annual growth rate in postdoctoral positions was 5 percent for science and 8 percent for engineering (NSB, 1989). While the number of investigators has been increasing, the number of investigator-initiated research proposals has been increasing at an even faster rate. The increase in proposals per investigator is probably related to the strategy of submitting multiple proposals to increase the likelihood of funding.2
The number of science and engineering articles published by academic scientists and engineers has doubled since 1965, with continued rapid growth through 1980 (NSF, 1990c). As with the number of proposals, the increase in articles has resulted from an increase both in the number of researchers and the number of articles produced per researcher per year. Similarly, the number of patents issued to universities, a different measure of research activity, has grown rapidly during the past decade.
Not only have the numbers of journals and articles increased, but the number of authors per article has also increased in fields such as high-energy physics, molecular biology, and clinical medicine. Huth, for example, reported that the mean number of authors per paper for the journals Annals of Internal Medicine and the New England Journal of Medicine rose from 1 in 1925 to 6 in 1985 (Huth, 1988).3 He added that “in some papers the number of authors is clearly excessive for the intellectual activities represented” and that “the climbing number of authors per paper is tending to cheapen the value of authorship” (p. 40).
With more manuscripts submitted for publication and more pro-
posals submitted for funding, the overall work load associated with critical evaluation has increased. There are concerns that peer review no longer operates as well as in earlier times, although the effects of increased volume on the operation of the system are not known (Chubin and Hackett, 1990).
Complexity of Collaboration
The increased emphasis on collaborative research is another indicator of change in the research environment. Before World War II, for example, scientific papers signed by more than four authors were practically nonexistent. Also extremely rare were papers that reported the results of collaborative efforts involving more than one laboratory or research team. But modern advances in the speed of travel and communication and in research instrumentation have changed the nature of scientific collaboration. Today, many important research papers involve collaboration among three or more laboratories, with a dozen or more authors in all. It is not unusual for authors or contributing laboratories to reside in more than one country. Although the senior investigators in these efforts may know each other personally, it is unlikely that the junior collaborators have ever met.
Different research groups may have different kinds of specialized skills, and complementary expert skills are likely to be the basis of the scientific collaboration. This type of interaction is very different, however, from earlier scientific exchanges in which all members of a research team shared the same laboratory environment and saw each other constantly during their work together.
Many of the achievements of modern science—of molecular biology, for example—have resulted from complex collaborative exchanges. Scientific advances in this field and others show that specialized collaborations can work effectively and are often indispensable to advancing knowledge. Nevertheless, the complexity of such operations, and the fact that many of the participants have limited personal interactions as well as limited abilities to evaluate the qualifications of others with different kinds of expertise, can give rise on occasion to conflicts and serious misunderstandings and can limit the effective operation of internal checks and balances.
Organization, Goals, and Management of Research Groups
Universities are characterized by decentralized organizational structures. The faculty traditionally govern academic programs. The fac-
ulty, in turn, are governed by a broad set of administrative and regulatory policies that affect the scientific research environment, increasingly so today. These policies reflect broad social concerns (e.g., about sexual harassment and equal opportunity) as well as matters explicitly related to the conduct of research (e.g., protection of human and animal subjects, regulation of toxic materials, and handling of hazardous equipment). In addition, many academic research institutions have now adopted policies regarding conflict of interest and the intellectual property developed by their employees.
Research in disciplinary specialties has traditionally been organized in a specific academic department. But research in many fields is now characterized by interdisciplinary approaches and is frequently carried out by individual academic investigators who, though they may have a departmental affiliation, are attached to independent, research centers. Centers may be organized around common research interests (e.g., poverty, energy, the environment) or research styles and resource needs (e.g., surveys, computer modeling, synchrotron light sources). Center directors often assume responsibility for generating support, including ongoing support for facilities and core staff.
Research goals are increasingly linked, by sponsors and investigators, to specific social needs. Indeed, economic development has received explicit emphasis in recent years in some federal and most state-supported research. Research projects aimed at environmental, health, and other particular social problems have, since the 1960s, increasingly been carried out by interdisciplinary academic groups and research centers. Industry has often participated in and sponsored such activities and has provided a diversified source of funding. Research investigators in such organizations include tenured and junior faculty members, visiting scientists, nonfaculty research scientists, postdoctoral research fellows, graduate students, and technicians.
As a result of these trends, scientific research organizations today need an unaccustomed level of structure and efficient management to perform effectively. Many large research groups do not have organizational procedures to support the necessary level of management and oversight. Such circumstances can inhibit the effective resolution of disputes and even incidents of misconduct.
Issues related to authorship, allocation of credit, and data management practices often arise in large research groups. Teams of 100 Ph.D.s are common where research is dependent on major instruments. As instrumentation becomes more specialized, the team size, too, will grow, to 600 or more Ph.D.s in some instances. Some research team efforts are tightly coordinated, whereas other “big sci-
ence” projects have a highly decentralized research culture. For example, the War on Cancer and the Human Genome Project have been described as combinations of “little science” initiatives. However, they typically follow a structured plan to achieve selected research objectives.
Research groups are governed by various management practices. Some groups operate in a collaborative style, choosing research problems through consultation among senior and junior investigators about the appropriate course to follow in pursuing interesting observations. Other groups adopt a more hierarchical style, whereby the principal investigator establishes a course of action for the research team as a whole and encourages efforts that contribute to the central mission of the director. In a few laboratories, research directors may discourage collegial discussion of new results or interpretation of findings or may foster competitive practices by assigning junior researchers to identical research problems.
Regulation and Accountability
Scientific research is increasingly subject to government regulations and guidelines that impose financial and administrative requirements and affect specific elements of the research process as well. Among the subjects of current research regulations are the assurance of a drug-free workplace, laboratory safety, proper use of human subjects and care of animal subjects, and care in the use of recombinant DNA and in the use of toxic and radioactive materials (OTA, 1986a). Regulatory requirements of the Public Health Service, the National Science Foundation, and the Department of Veterans' Affairs have also prompted, and in some cases required, research institutions to adopt policies and formal procedures to handle allegations of misconduct in science.
To assure the full compliance of investigators and institutions with these regulatory requirements, universities have expanded administrative and oversight functions. The associated costs in time and money have escalated tensions between administrators and faculties that would prefer to see the funds going into research. This is one of several issues that has caused schisms in the academic community.
The criteria used to appoint, evaluate, and promote individual faculty members deeply influence the research enterprise. The rewards of a successful academic career traditionally include the per-
sonal gratification derived from scholarship and discovery, recognition by peers, and academic promotion and tenure, as well as enhanced responsibility and outside financial opportunities. The successful researcher can attract continuing research support and can enjoy a reputation that opens new opportunities for prestigious appointments.
The academic reward system today is influenced largely by research performance and productivity, sometimes measured by the number of publications or total amount of research support acquired by individual faculty. Intellectual contributions, teaching, and service to the university and the public are considered in varying degrees depending on the institution and discipline. However, there appears to be an imbalance, with emphasis on publication output and research support as the basis for promotion and tenure (Boyer, 1990).4
Quantitative measures of productivity have occasionally become substitutes for the critical evaluation of scientific work. This reliance on numbers arises in part because departmental peers are less able to evaluate the contribution of an individual researcher to large scientific projects or to interdisciplinary teams with an applied research approach. Attribution of credit among individuals on multiauthored publications is also difficult. Even when the form of an individual's contribution is clear, the significance of the contribution is often arguable.
The “publish-or-perish” dictum can lead to overspecialization, overemphasis on short-term projects, and the organization of research communication around the “least publishable unit.” Theoretical approaches, including computer simulations, that yield especially rapid results can be favored over tedious programs of fundamental experiments. An excessive emphasis on quantitative measures of scientific productivity can penalize scientists who make responsible attempts to protect the quality of science (i.e., by delaying publication until they have completed a series of experiments instead of publishing each experiment). As Jackson and Prados (1983) have observed (p. 464):
Good scientists may publish a lot or a little. But there is a very definite evil in a university that allows or encourages tenure committees to set standards of, say twenty published papers or abstracts in four years as a minimum requirement for consideration, or to discard as irrelevant any paper in a branch of science other than the tenure candidate's principal field of specialization.
Some institutions have responded to the emphasis on large numbers by limiting the number of publications reviewed for promotion
and by rewarding nonresearch scholarship such as teaching and communicating science to a general audience (Kennedy, 1991; Harvard University Faculty of Medicine, 1988).
The importance of contributing to economic development as a national research goal during the last decade has led to an emphasis on prompt transfer of fundamental research findings into commercial use. U.S. universities have often produced discoveries with practical significance, an achievement that has attracted the interest of both U.S. and foreign firms. In many areas of technological significance—microelectronics, biotechnology, materials science, instrumentation, and catalysis, for example—the interval between laboratory discovery and practical application has decreased. Rapid commercialization has provided increased incentives for joint industry-university research programs. Public desires to strengthen the competitive performance of U.S. industry have fostered academic research programs aimed at improving U.S. manufacturing.
A number of federal and state programs now encourage or require cooperation between universities, industry, and national laboratories. University-industry partnerships are implemented by a variety of mechanisms, including long-term agreements with one or more university research groups to pursue a subject of mutual interest, participation in research consortia, research contracts with specific program objectives, and informal collaborations. Consortia efforts, in which several companies combine with research groups at one or more universities to pursue a common research program, are another mechanism. Federal technology transfer programs, such as those in the Department of Defense and several Department of Energy programs at national laboratories, are other examples.
University–industry partnerships stimulate new ideas and innovation by both communities and motivate research teams to achieve important innovations of commercial value. But commercial relationships may introduce conflicts for academic investigators and the university. Some conflicts result from tension between the traditions of openness in the university, where prompt publication and free access to research results is required, and desires to restrict access to research results of proprietary value. Other conflicts can arise because personal profit and commercial interests can become explicit goals of individuals and institutions.
Despite efforts to minimize conflicts, there is growing concern in the scientific research community about the consequences of academ-
ic-commercial collaboration, especially in the area of clinical research. Many universities are adopting new and more stringent rules to govern conflict of interest and ownership of intellectual property, including categories of activities with differing requirements for disclosure as well as prohibited activities and relationships. An instructive case is the debate accompanying the adoption in 1990 of conflict-of-interest rules at the Harvard University Medical School (Harvard University Faculty of Medicine, 1990).
Conflicts of interest have the potential to affect peer review, publication and data management practices, training and mentorship, and other practices and behavior. For this reason, some scientific journals require authors to disclose sources of support and potential sources of bias when submitting their research papers.5 Such conflicts also can influence the investigation of allegations of misconduct in science, especially if biases are not detected in the formation of investigatory panels that review and adjudicate misconduct complaints.
Although the panel does not believe that industry-university research arrangements present unique risks for misconduct in science, the self-serving interests associated with such arrangements pose issues that require institutional attention and oversight to ensure the integrity of the research process.
FINDINGS AND CONCLUSIONS
The contemporary research enterprise is far removed from that of the pre-World War II era. In particular, the academic research community, governed by traditions derived from an earlier model of a community of independent scholars who participated equally in academic governance, is challenged by the complexity of today's issues and of the environment in which research is conducted. Still, basic research continues to flourish, and faculty, postdoctoral fellows, and graduate students continue to contribute extraordinary research capability to science.
Concerns are apparent, however, and it is clear that key environmental factors require attention to protect the high standards of research integrity traditionally associated with scientists and their institutions. In reviewing changes within the scientific research enterprise, the panel reached the following conclusions:
Scientific research is part of a larger and more complicated enterprise today, creating a greater need for individual and institutional attention to matters that affect the integrity of the research process. Scientists themselves and research institutions will be ex-
pected to play a more active role in ensuring that the activities performed by researchers are within the governance mechanisms of their institutions. The need for more explicit forms of institutional accountability and oversight is one price of the growth and success of the academic research enterprise.
The growth and diversity of modern research call for institutions to accept explicit responsibility for fostering the integrity of the research process and for handling allegations of misconduct. In encouraging this acceptance, the panel is not suggesting that institutions assume responsibility for the correctness and accuracy of research results reported by their scientists or students. However, in recognizing that their faculty and research staff are responsible for maintaining the integrity of the research process, institutions should retain and accept certain explicit obligations. Principal among these is providing a research environment that fosters honesty, integrity, and a sense of community. Institutions should strive to attain a research enterprise that emphasizes and rewards excellence in science, quality rather than quantity, openness rather than secrecy, and collegial obligations rather than opportunistic behavior in appointment, promotion, tenure, and other career decisions. Research institutions should also recognize the risks that are inherent in self-regulation and strive to involve outside parties, when appropriate, in investigating or evaluating the conduct of their own members. Steps toward achieving these goals are discussed in Chapter 6.
The increased size, specialization, and diversity of research groups, and other changes in the social relationships of their members, have stimulated personal conflicts and misunderstandings, including disputes within and between research groups about fairness and allocation of credit. These disputes may be prevented by positive efforts to foster responsible research practices and by taking preemptive actions, such as prior discussion and agreement on allocation of credit, to promote a harmonious work environment that encourages collegiality, collaboration, and productivity. Frank discussions, both formal and informal, possibly aided by outside mediators, are additional tools to use in addressing these disputes.
The issues associated with conflict of interest in the academic research environment are sufficiently problematic that they deserve thorough study and analysis by major academic and scientific organizations, including the National Academy of Sciences. Disclosure, either public or institutional, is essential to controlling conflict of interest, and some universities and scientific journals prohibit certain forms of commercial contractual arrangements by their members or
authors. But the responsibility for such disclosure rests with scientists themselves.
The research environment is stressful and yet conducive to the remarkable productivity of researchers. The rewards for successful research are greater now than in the past, but today's rapid pace of development may undermine critical internal checks and balances and may increase opportunities for misrepresentation or distortion of research results. Thus the scientific community must organize to reinforce its standards and to ensure the responsible conduct of research.
1. See, for example, Association of American Universities (1988).
2. The Office of Technology Assessment suggests that a “kind of lottery mentality appears to have taken hold in the 1980s: the more grant proposals submitted, the greater the probability that one would be funded” (OTA, 1990, p. 10).
3. The mean is represented by rounding off to one significant figure.
4. See also, for example, Angell (1986).
5. See, for example, the editorial policies of the New England Journal of Medicine (1992).