[轉錄] Living by numbers

看板PhD (博士班)作者lovenainai (lovenainai)時間16年前 (2010/04/18 13:32)推噓4(4推 0噓 4→)

留言8則, 4人參與討論串1/1

http://stke.sciencemag.org/cgi/content/full/sigtrans;2/99/eg15 http://stke.sciencemag.org/cgi/reprint/sigtrans;2/99/eg15.pdf (PDF 版) 這是一篇 Science Signaling Chief editor 的文章其中我看完最有感觸期刊亂象的是這兩段 "Twenty-five years ago, when I was a graduate student, it seemed like everyone had a pretty clear sense of which journals published what kind of papers. If the work was just overflowing with cool findings, it ended up in Science or Nature. Cell was also known for publishing spectacular findings, depending on geography. PNAS was where you went to get great ideas, because the papers were fascinating but short and often presented only the beginning of the story. If you wanted to get all of the gory details about an experimental system, you read a society journal like Journal of Biological Chemistry (JBC), Biochemical Journal, or Journal of Cell Biology. Now, as journals compete for the highest impact factor, it seems that most journals are trying to publish the identical types of papers, overflowing with cool results, completely self-consistent data, and mind-boggling findings." 很多新 journal 開始想辦法衝高 IF 一些老 journal 就開始掉然後大家就越來越不願意投最後就變成惡性循環 "Perhaps institutions feeling financial pressure only subscribe to the "high impact" journals, like the legislators and insurers who only want to pay the top doctors and hospitals. Then, just like those who refuse to care for the sickest patients, no one will want to work on the really hard problems that can’t guarantee a fast paper in a high-profile journal. The pressure to maintain "high impact" status will further lead to selective data presentation, self-deception, or worse, because truth, in the process of discovery, is often confusingly complex and not always completely consistent" 這也還蠻中肯的 journal 越來越多經費有限能買的種類就相對越少... 以下是全文 ================== Living by the Numbers Michael B. Yaffe1,2* 1 Chief Scientific Editor of Science Signaling, American Association for the Advancement of Science, 1200 New York Avenue, N.W., Washington, DC 20005, USA. 2 Koch Institute for Integrative Cancer Research, Departments of Biology and Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA. Abstract: Quantitation of an article's worth by the impact of the journal in which it appears is a path to undermining scientific enterprise. Through a process analogous to rating medical care, rating journals can lead to loss of research effort in the most challenging questions. A paper should be evaluated on its own impact, not by some arbitrary score for the journal as a whole. Michael B. Yaffe is Chief Scientific Editor of Science Signaling and associate professor of biology and biological engineering at the David H. Koch Institute for Integrative Cancer Research at the Massachusetts Institute of Technology. As scientists, we tend to measure and quantify everything. Until recently, I would have argued that there was probably nothing that couldn’t be, or shouldn’t be, measured. Two examples of "numbers gone crazy" in medicine and science, have changed my mind. First is the mismeasure of medicine. There is an increasing tendency of legislators, third-party payers, and patients to assign some sort of quality ranking numbers to their doctor, surgeon, or local hospital. What could be wrong with that? After all, isn’t everyone entitled to know how good someone or some place is? Isn’t everyone entitled to the best care from the best people? Unfortunately, what looks so right at first glance turns out to be horribly wrong over the long term. This process is fundamentally destabilizing. As everyone rushes to the "best" doctor and the "best" hospital, the system bifurcates into a tiny group of highly skilled but heavily overworked and overburdened professionals, and a much larger group of largely competent folks who, deprived of patients, now have essentially no opportunity to improve. As legislators and insurers continue to reward "the best" and penalizing "the rest," there becomes increasing pressure to maintain or attain elite status. Consequently, patients likely to do poorly are refused care, and there becomes an increasing reluctance to report bad outcomes. In the end, the whole system collapses, taking good patient care down with it. This same process, which I argue is on its way to destroying health care, has also been happening in science. The mis-measure here is the "Impact Factor," a "score" established by Thomson ISI for reporting the frequency with which the "average article" in a journal has been cited in a given period of time. Because Science Signaling is less than 2 years old, the journal does yet have an "Impact Factor." Consequently, I have the freedom to complain about these types of measurements without being accused of "sour grapes." Twenty-five years ago, when I was a graduate student, it seemed like everyone had a pretty clear sense of which journals published what kind of papers. If the work was just overflowing with cool findings, it ended up in Science or Nature. Cell was also known for publishing spectacular findings, depending on geography. PNAS was where you went to get great ideas, because the papers were fascinating but short and often presented only the beginning of the story. If you wanted to get all of the gory details about an experimental system, you read a society journal like Journal of Biological Chemistry (JBC), Biochemical Journal, or Journal of Cell Biology. Now, as journals compete for the highest impact factor, it seems that most journals are trying to publish the identical types of papers, overflowing with cool results, completely self-consistent data, and mind-boggling findings. Even worse, it seems that some universities and research institutions are now using the "Impact Factor" of the journals where a scientist publishes to decide on the size of start-up packages or whether someone will be promoted. These institutions would do well to remember that the original discovery of B cells as the source of antibodies was published in Poultry Science, and that the discovery that PTEN (one of the most important tumor suppressor genes) was a phosphatase for phosphoinositol-phosphate lipids was published in JBC. Why not just judge a paper by the net impact of the work itself, instead of by some arbitrary score for the journal as a whole? Perhaps the silliest aspect of this whole "Impact Factor–centric" approach lies with the potentially wide availability of scientific papers via the Internet. Work that might previously have been missed if one had to read print journals in the library should no longer be at the same disadvantage. Perhaps institutions feeling financial pressure only subscribe to the "high impact" journals, like the legislators and insurers who only want to pay the top doctors and hospitals. Then, just like those who refuse to care for the sickest patients, no one will want to work on the really hard problems that can’t guarantee a fast paper in a high-profile journal. The pressure to maintain "high impact" status will further lead to selective data presentation, self-deception, or worse, because truth, in the process of discovery, is often confusingly complex and not always completely consistent. In the end, the real loser here will be science itself, which despite its high impact factor, will have little, if any, real impact. If we continue to live by the numbers, then we had better be prepared to fall by them. -------------------------------------------------------------------------------- * Corresponding author. E-mail: myaffe@mit.edu Citation: M. B. Yaffe, Living by the Numbers. Sci. Signal. 2, eg15 (2009). -- ※ 發信站: 批踢踢實業坊(ptt.cc) ◆ From: 143.48.119.2