Falagas Matthew E, Karveli Efthymia A, Tritsaroli Vassiliki I
Alfa Institute of Biomedical Sciences (AIBS), Athens, Greece.
Int J Med Inform. 2008 Apr;77(4):280-6. doi: 10.1016/j.ijmedinf.2007.07.001. Epub 2007 Aug 21.
We sought to examine the frequency and the specific problems encountered in accessing Internet references in two leading medical journals during the last 3 years.
Two investigators independently reviewed all publications in the issues of the New England Journal of Medicine and The Lancet during October 2005 to March 2006, November 2004 to January 2005, and November 2003 to January 2004. We calculated the total number of references and the subset referred to an Internet source of each article. Then, we visited the electronic sources to identify the Internet references and noted the problems of accessibility, if any. When we failed to directly access the reference in the electronic address provided by the authors, we visited the referred website; if this was also inadequate, we performed Google searches to retrieve the missing reference(s).
465/18,850 (2.5%) and 952/24,630 (3.9%) of the reviewed references in the New England Journal of Medicine and The Lancet, respectively, referred to Internet sources; from these we could not access 68/465 (14.6%) and 170/952 (17.9%) in the two journals, respectively. There were increasing proportions of lost Internet references as they age. Searching into the website referred by the authors of the reviewed articles could not provide the missing information in a considerable proportion (62.2%). However, the use of an Internet search engine (Google) helped us to identify references in other websites, reducing the proportion of missing Internet references to 17/465 (3.7%) and 17/952 (1.8%) for the two journals, respectively.
The response "page not found" was commonly encountered when we tried to access Internet references in publications of leading medical journals during the last 3 years. A considerable proportion of missing references was identified with the use of Google search engine. Authors of scientific articles should be aware of the problem of missing Internet references and until well-established Internet archiving solutions are in use, they should choose carefully their Internet references from reliable websites whenever it is impossible to avoid using them.
我们试图调查在过去3年中,两份顶尖医学期刊中引用互联网参考文献的频率及遇到的具体问题。
两名研究人员独立审查了《新英格兰医学杂志》和《柳叶刀》在2005年10月至2006年3月、2004年11月至2005年1月以及2003年11月至2004年1月期间各期的所有出版物。我们计算了每篇文章的参考文献总数以及引用互联网来源的参考文献子集。然后访问电子来源以识别互联网参考文献,并记录可访问性问题(如有)。当我们无法直接访问作者提供的电子地址中的参考文献时,我们访问所引用的网站;如果该网站也无法访问,我们进行谷歌搜索以检索缺失的参考文献。
《新英格兰医学杂志》和《柳叶刀》中经审查的参考文献分别有465/18850(2.5%)和952/24630(3.9%)引用了互联网来源;在这两份期刊中,我们分别无法访问其中的68/465(14.6%)和170/952(17.9%)。随着互联网参考文献年代的增加,无法访问的比例也在上升。在相当大比例(62.2%)的情况下,搜索经审查文章作者所引用的网站无法提供缺失的信息。然而,使用互联网搜索引擎(谷歌)帮助我们在其他网站上识别参考文献,使两份期刊中无法访问的互联网参考文献比例分别降至17/465(3.7%)和17/952(1.8%)。
在过去3年中,当我们试图访问顶尖医学期刊出版物中的互联网参考文献时,“页面未找到”的回复很常见。使用谷歌搜索引擎识别出了相当比例缺失的参考文献。科学文章的作者应意识到互联网参考文献缺失的问题,在可靠的互联网存档解决方案投入使用之前,每当无法避免使用互联网参考文献时,应从可靠网站中谨慎选择。