Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
12 On Being a S c i e n t i s t Mistakes and Negligence All scientific research is susceptible to error. At the frontiers of knowledge, experimental techniques often are pushed to the limit, the signal can be difficult to separate from the noise, and even the question to be answered may not be well defined. In such an uncertain and fluid situation, identifying reliable data in a mass of confusing and sometimes contradictory observations can be extremely difficult. Furthermore, researchers sometimes have to take risks to explore an innovative idea or observation. They may have to rely on a theo- retical or experimental technique that is not fully developed, or they may have to extend a conjecture into new realms. Such risk taking does not excuse sloppy research, but it should not be condemned as misguided. Finally, all researchers are human. They do not have limitless working time or access to unlimited resources. Even the most re- sponsible researcher can make an honest mistake in the design of an experiment, the calibration of instruments, the recording of data, the interpretation of results, or other aspects of research. Despite these difficulties, researchers have an obligation to the public, to their profession, and to themselves to be as accurate and as careful as possible. Scientific disciplines have developed methods and practices designed to minimize the possibility of mistakes, and failing to observe these methods violates the standards of science. Every scientific result must be carefully prepared, submitted to the peer review process, and scrutinized even after publication. Beyond honest errors are mistakes caused by negligence. Haste, carelessness, inattentionâany of a number of faults can lead to work that does not meet scientific standards or the practices of a discipline. Researchers who are negligent are placing their reputation, the work of their colleagues, and the publicâs confidence in science at risk. Er- rors can do serious damage both within science and in the broader society that relies on scientific results. Though science is built on the
M i s ta k e s a n d N e g l i g e n c e 13 Changing Knowledge In the early part of the 20th century, astronomers engaged in a prolonged debate over what were then known as spiral nebulaeâdiffuse pinwheels of light that powerful telescopes revealed to be common in the night sky. Some astronomers thought that these nebulae were spiral galaxies like the Milky Way at such great distances from the Earth that individual stars could not be distinguished. Others believed that they were clouds of gas within our own galaxy. One astronomer who thought that spiral nebulae were within the Milky Way, Adriaan van Maanen of the Mount Wilson Observatory, sought to resolve the issue by comparing photographs of the nebulae taken several years apart. After making a series of painstaking measure- ments, van Maanen announced that he had found roughly consistent unwinding motions in the nebulae. The detection of such motions indicated that the spirals had to be within the Milky Way, since motions would be impossible to detect in distant objects. Van Maanenâs reputation caused many astronomers to accept a ga- lactic location for the nebulae. A few years later, however, van Maanenâs colleague Edwin Hubble, using a new 100-inch telescope at Mount Wilson, conclusively demonstrated that the nebulae were in fact distant galaxies; van Maanenâs observations had to be wrong. Studies of van Maanenâs procedures have not revealed any inten- tional misrepresentation or sources of systematic error. Rather, he was working at the limits of observational accuracy, and his expectations influenced his measurements. Even cautious researchers sometimes admit, âIf I hadnât believed it, I never would have seen it.â idea that peers will validate results, actual replication is selective. It is not practical (or necessary) to reconstruct all the observations and theoretical constructs made by others. To make progress, researchers must trust that previous investigators performed the work in accor- dance with accepted standards. Some mistakes in the scientific record are quickly corrected by subsequent work. But mistakes that mislead subsequent researchers can waste large amounts of time and resources. When such a mistake appears in a journal article or book, it should be corrected in a note, erratum (for a production error), or corrigendum (for an authorâs
14 On Being a S c i e n t i s t error). Mistakes in other documents that are part of the scientific recordâincluding research proposals, laboratory records, progress reports, abstracts, theses, and internal reportsâshould be corrected in a way that maintains the integrity of the original record and at the same time keeps other researchers from building on the erroneous results reported in the original. Discovering an Error Two young faculty membersâMarie, an epidemiologist in the medi- cal school, and Yuan, a statistician in the mathematics departmentâhave published two well-received papers about the spread of infections in pop- ulations. As Yuan is working on the simulation he has created to model infections, he realizes that a coding error has led to incorrect results that were published in the two papers. He sees, with great relief, that correct- ing the error does not change the average time it takes for an infection to spread. But the correct model exhibits greater uncertainty in its results, making predictions about the spread of an infection less definite. When he discusses the problem with Marie, she argues against sending corrections to the journals where the two earlier articles were published. âBoth papers will be seen as suspect if we do that, and the changes donât affect the main conclusions in the papers anyway,â she says. Their next paper will contain results based on the corrected model, and Yuan can post the corrected model on his Web page. 1. What obligations do the authors owe their professional colleagues to correct the published record? 2. How should their decisions be affected by how the model is being used by others? 3. What other options exist beyond publishing a formal correction?