The Role of Scientific JudgmentThe various problems with collecting and processing data and with writing and reviewing manuscripts have been presented above as if there were no gray areas, no areas ripe for personal interpretations, for scientific judgment. Most scientists would be quick to point out that this is not the case. There are many instances when a scientist must exercise good judgment regarding whether a given experimental animal should be included in the study when results from that animal are at odds with those from the remaining animals in the study. Strong reason to believe that the animal is not representative of the population (e.g., normal, healthy rats) which the research is designed to study may be grounds for such exclusion. Who is to say? Let's take a concrete example. In a molecular biological project, a student does a series of Northern blots. The first few blots in the series have a "high" background level and a lot of "artifact." (These terms are put in quotation marks to emphasize their interpretive nature.) The adviser deduces that these resulted from the inexperience of the student, and he advises the student to try again. With practice, the student manages to produce blots without the "artifacts" and with "low" background. When she or he writes the manuscript describing these experiments the first few blots are omitted from the data set. Is this justifiable? This becomes a thorny ethical issue and a matter of scientific judgment. The adviser may have seen many students go through the same evolutionary process in learning to do blots; he or she may "know" that these are artifacts. Can we be certain that we haven't gradually eliminated real data in doing what we have to do to get rid of artifacts? Remember the oft-used aphorism about throwing the baby out with the bath water. How far is it from the "trivial" example of the last paragraph to the recent case of a student who used a single blot obtained from one experimental condition to illustrate data from several different experimental conditions because it was pretty and he didn't have a pretty blot for the other conditions? This student embarrassed his adviser and may have ruined his own scientific career. Molecular biology is not the only field in which scientific judgment plays a role. The electrophysiologist frequently has to contend with electrical "noise" in recordings. This noise is usually filtered out electronically to reveal the "signal" uncontaminated. But can the electrophysiologist be certain that the filters have not also removed a real signal from the recordings? Experienced researchers in all areas of science must make judgments about the nature of noise in records, about the appropriateness of experimental designs, controls and statistical tests. In all cases, it is a good idea to fully disclose everything that was done in an experiment. A simple statement that a few blots were not included because the student was learning the technique would suffice. A simple statement that some animals were excluded from the discussion because they were obese would be appropriate. A simple statement that the illustrations were "cleaned up" or "edited" might well save embarrassment and ruin! What can a student do to prevent himself from committing any of these forms of unethical behavior? The most important thing is to be certain that he is properly prepared to begin a research career. As Branscomb (1985) puts it,
Young scientists should understand all the subtle ways in which they can delude themselves in the design of observations and the interpretation of data and statistics. They should understand metrology and should know what tendencies to manipulate information are built into their digital signal processors. They should also get to know the algorithms used in their favorite computers, which may under certain circumstances give strange results. Above all they should be trained in detection and control of systematic errors. This is good advice. In addition to these accomplishments, students should develop the same healthy skepticism about their own work that they apply to others. It is natural for people to be skeptical of the work of others. Scientists should develop even more skepticism than the nonscientific public--be a critic. It is healthy for science when the scientist does not take anything as given until there is sufficient evidence that he has no choice; he should not be too quick to jump onto any scientific bandwagon. On the other hand, skepticism should not be carried to extremes such that the scientist does not believe anything anyone else says. A student should learn to apply the same skepticism to his own work. He should not take the obvious conclusion as the only possibility, but work to eliminate the other possibilities through experiments, always looking for sources of systematic error. It is absolutely essential that students learn about statistical tests, their assumptions and their proper applications. Even if they never use the tests themselves (hard to believe, but it's possible), they will be working in a statistical science; reading and evaluating the work of others requires a thorough knowledge of statistics. It is also imperative that students learn about proper design of scientific experiments. They should master the various kinds of experimental designs. Local psychologists or biostatisticians can be very helpful in pointing students to pertinent literature on experimental design. Students should also be very conversant with the proper use of controls, matched controls, and use of the subject as its own control. Psychologists can also help students learn about experimenter bias and where it can occur in experiments. There is already a large literature on this subject. In this regard, students should be aware of their own assumptions and how they may influence what they see. Hopefully, a student will learn to report all his results, even the negative ones when possible. Faculty should help students learn to make appropriate decisions about when and to whom to give credit in manuscripts, grant requests and other places where credit may be due. They should learn that it is in their best interest not to allow others to put their names on work they didn't do. It may look good on their CVs today, but it may come back to haunt them tomorrow. Finally, students should try to do whatever they can to reduce the stresses inherent in research. One thing they can do is learn as much about the subject of their investigations as they can to build up their self confidence as much as possible. This may help them in dealing with more senior investigators. They should also try to get a realistic perspective of what science is about and what its appropriate rewards should be. They must learn that it's ok to do science for the enjoyment of doing science, without any other motive. Students must also know that they can reduce stress by getting out of untenable situations as quickly and gracefully as possible. The lab where they are working, no matter what its world-wide reputation, is not the only one in the world. If they can't work comfortably in it, it's not even the best one for them! Association of American Medical Colleges. Framework for Institutional Policies and Procedures to Deal with Misconduct in Research. Washington, D.C., 1990. Babbage, C. Reflections on the Decline of Science in England and on Some of its Causes. Westmead, England: Gregg International Publishers, 1969 printing of 1830 book. Bacon, F. Novum Organum. Translated and edited by Peter Urbach and John Gibson, Chicago: Open Court, 1993. Branscomb, L.M. Integrity in science. American Scientist 73:421-423, 1985. Broad, W.J. Harvard delays in reporting fraud. Science 215:478-482, 1982 Broad, W. & N. Wade. Betrayers of the Truth. New York: Simon & Shuster, Inc. Publishers, 1982. Committee on the Conduct of Science, National Academy of Science. On Being a Scientist. Washington, D.C.: National Academy Press, 1989. Fahmy, R.N. & J.L. Fahmy. The order of authorship. JAMA 265:865, 1991. Feild, H.S. & A.A. Armenakis. On use of multiple tests of significance in psychological research. Psychol. Reports 35:427-431, 1974. Freiman, J.A., T.C. Chalmers, H. Smith et al. The importance of beta, the type II error and sample size in the design and interpretation of the randomized control trial. New England Journal of Medicine 299:690-694, 1978. Fye, W.B. Medical authorship: Traditions, trends and tribulations. Ann. Intern. Med. 113: 317-324, 1990. Garfield, E. More on the ethics of scientific publication: Abuses of authorship attribution and citation amnesia undermine the reward system of science. Current Contents 30:5-10, 1982. Jackson, C.I. Honor in Science. New Haven: Sigma Xi, The Scientific Research Society, 1984. Lock, S. Repetitive publication: a waste that must stop. Brit. Med. J. 288: 661-662, 1984. Mann, M.D., D.A. Crouse and E.D. Prentice. Appropriate animal numbers in biomedical research in light of animal welfare considerations. Laboratory Animal Science 41:6-14, 1991. Mishkin, B. Responding to scientific misconduct: due process and prevention. JAMA 260:1932-1936, 1988. Neher, A. Probability pyramiding, research error and the need for independent replication. Psychol. Rec. 17:257-262, 1967. Riessenberg, D. & G.D. Lundberg. The order of authorship: Who's on first? JAMA 264:1857, 1990. Selye, H. Can we cope with the 'literature explosion?' J. Med. 1:3-8, 1970. Snow, C.P. The Search, New York: Charles Scribner's Sons, 1959. Towe, A.L., T.T. Kennedy & H.D. Patton. Response properties of neurons in the pericruciate cortex of the cat following electrical stimulation of the appendages. Experimental Neurology 10:325-344, 1964. Wolins, L. Responsibility for raw data. Amer. Psychologist 178:657-658, 1962. Where to Go From Here:Introduction |