ÁñÁ«ÊÓƵ

The rise of retractions in science – more misconduct or stiffer scrutiny%3F
Year	Retracted articles	Total articles
2009 through July	106	892,2
2008	95	1,437,339
2007	65	1,0,656
2006	59	1,224,567
2005	21	1,315,722
2004	15	1,075,402
2003	23	1,115,193
2002	13	975,480
2001	3	1,000,074
2000	9	956,8
1999	23	974,644
1998	15	960,138
1997	18	9,833
1996	8	903,841
1995	16	854,909
1994	12	798,498
1993	14	754,594
1992	8	741,836
1991	3	695,858
1990	5	689,752

The editor of Stem Cells and Development recently announced the retraction of a paper concerning the production of human sperm from embryonic stem cells after he learnt that the introduction had been largely plagiarised. This event prompted the question of the extent of retractions in scientific literature and whether their rate has changed over the past 20 years. The search was limited to science journals only. As can be seen, the size of the database roughly doubled over the past two decades, whereas the number of retractions increased by roughly tenfold. The notorious case of German physicist J.?H. Schon does not alone account for the increase: 22 of his papers, many published in top journals such as Science, were retracted from 2002 to 2004. The increase is evident for a variety of reasons: misconduct, mistakes of reasoning or interpretation, and the irre?pro?duc?ibility of results. Striking discoveries may be under closer scrutiny, which is a positive sign of science¡¯s self-correcting machinery. More negative interpretations include the suspicion that peer review is not as rigorous as it was and that the pressure to publish is leading to more errors and even fraud.

Next week: Asia Pacific institutes in materials sciences