The National Science Foundation (NSF) has a new tool for detecting plagiarism in its grant proposals—and in its first application, it has proved alarmingly effective.

On Friday, Jeffrey Mervis reported on ScienceInsider that NSF is investigating nearly 100 cases of suspected plagiarism turned up in an internal audit by NSF's Office of Inspector General (IG) of "NSF's entire portfolio of some 8000 awards" made in 2011. That's a lot of plagiarism cases for a single year: The office has investigated just 120 cases of scientific misconduct since 2003, about 80% of them plagiarism cases. "My group is now swamped," James Kroll, head of administrative investigations within the IG's office, told ScienceInsider.

Commenters on the ScienceInsider post speculated that because the rate of plagiarism is so high, NSF must be including cases of self-plagiarism—taking text from an old paper or another grant proposal written by you or a co-author—which most people consider much less serious than appropriating someone else's words or ideas without credit. But, as Kroll points out in an e-mail to Science Careers, according to the federal definition, self-plagiarism doesn't count as plagiarism.

Kroll tells Science Careers that the software that detected the suspected plagiarism cases was "similar to products being used by many universities" to detect plagiarism in student essays, "although the version we use drill[s] down into many of the science journals … for potential sources of plagiarized materials."

Jim Austin is the editor of Science Careers. @SciCareerEditor on Twitter

10.1126/science.caredit.a1300040