New Head of China's NSF Speaks Out by Hao Xin on 11 March 2013, 3:15 PM 见: http://news.sciencemag.org/scienceinsider/2013/03/new-head-of-chinas-nsf-speaks-ou.html Hao Xin博客: http://blog.sciencenet.cn/u/ 郝炘 现摘录一段访谈: Q: What measures are NSFC putting into place to promote research integrity and reduce misconduct (LL:加强科研诚信 抑制不端行为)in grant application and review of proposals? Y.W.: NSFC authorizes an independent committee to supervise research integrity. The committee received about 200 misconduct allegations each year in the last 3 years. Considering that over the same period there was a 16% increase of research proposals submitted, the number of allegations vis-à-vis proposals was slightly reduced. Last year, the committee found 19 out of about 200 allegations to be misconduct cases, about 10% of all allegations. The committee has identified five categories of misconduct. The first is inflation of credentials(资历弄虚作假). For example, an investigator expecting his doctorate in June uses the designation Ph.D. on a grant proposal submitted in March. Also, sometimes, signatures of collaborators are forged to inflate the scope of the project. This is the largest category. We have developed training material to warn people of this kind of misconduct. We also hold institutions responsible because they have the obligations to educate new researchers on good conduct when writing proposals. The second category is plagiarism(抄袭剽窃). Some copy from their own past proposals, and some copy others' proposals. We are using software to crosscheck newly submitted proposals against our archive of 5 to 10 years of old proposals. The third category is program managers choosing reviewers in favor of applicants with whom they have good relations(项目官员为申请人挑选与其关系好的评审人). Some reviewers are known to be hard; some are relatively soft. Program managers may choose soft reviewers for their friends. We are thinking about how to address this issue. For instance, we are establishing a database of reviewers with their average review scores. A program manager may select seven to 10 reviewers for one particular proposal. The system will calculate the average score from the reviewers. If the score is very high, then perhaps it indicates that a group of very soft reviewers were chosen. If the average score is very low, then perhaps a group of very hard reviewers are chosen. We can establish some margins and if the reviewer score is outside the margins, we will pay more attention to those proposals. The fourth category is leakage of review information(泄漏评审信息). My colleagues have told me about frequent visits to our database, which contains all proposal-related information. We are thinking about an alert mechanism, which will flag unusually frequent database visits and keep track of logons and accesses to the information system. The last category is fabricated progress reports for NSFC grants.(伪造科学基金项目进展报告) 该访谈的中文摘录,见: http://news.sciencenet.cn/htmlnews/2013/3/275568.shtm