编者按：我们团队分别在2010年和2015年对在公司注册的中国生物医学科研工作者进行了问卷调查，了解他们对我国科研工作中存在的学术不端行为的看法。该研究结果于今年3月发表在《科学与工程伦理》（Science and Engineering Ethics）杂志上，引起了多方关注[1]。应 Retraction Watch 特邀记者 Mark Zastrow 的邀请，该文的共同通讯作者夏华向教授和何华女士接受了Mark Zastrow的电话采访[2]。
Retraction Watch 【撤稿监察】
Four in 10 biomedical papers out of China are tainted by misconduct, says new survey 【一项新的调查估计：中国每10篇生物医学类论文有4篇涉嫌学术不端】
Chinese biomedical researchers estimate that 40% of research in their country has been affected in some way by misconduct, according to a new survey. 【根据一项新的调查，中国生物医学研究人员估计中国有40%的研究在某种程度上受到学术不端行为的影响。】
The authors are quick to caution against putting too much stock in this figure due to the subjective nature of the survey, published in Science and Engineering Ethics. The estimates also spanned a wide range, with a standard deviation of ±24%. But they say that the responses to this question and others on the survey suggest that scientists in the region feel academic misconduct remains a major problem that authorities have failed to adequately address. (Indeed, a recent analysis from Quartz using Retraction Watch data showed that researchers based in China publish more papers retracted for fake peer reviews than all other countries put together.) 【此项调查结果发表在《科学与工程伦理》（Science and Engineering Ethics）杂志上。由于调查类研究的主观性，同时也由于估计的范围比较广（标准差达到±24%），作者们提醒不要过于关注这个数字（40%）。但是，他们同时指出，从对问卷上的这一问题和其他问题的回复来看，生物医学研究人员认为学术不端行为依然是一个尚未被有关管理部门足够重视的重大问题。（事实上，Quartz近期通过调取撤稿监察的数据所进行的一项的研究表明，中国的研究者因为虚假评审而遭到撤稿的论文数比其他所有国家加在一起还要多。）】
The survey was designed by employees at Medjaden, a Hong Kong-based editing company that assists mainland Chinese biomedical researchers publishing in English-language journals. They invited all of their registered users by email to complete two surveys—roughly 10,000 users in 2010 and 15,000 in 2015. Like most online surveys, this one had a low response rate—around 5%, so caveats about sampling bias apply. 【这项研究是由美捷登（Medjaden）的工作人员设计和完成的。美捷登是一家注册在香港、致力于协助中国大陆生物医学研究者将科研论文发表在英文期刊上的编辑公司。他们使用邮件邀请了所有美捷登公司的注册用户，分别于2010年（约10000个用户）和2015年（约15000个用户）完成了两次调查。如同大部分的网上调查一样，这份调查的回复率较低（在5%左右），所以可能存在样本偏倚性的问题。】