Ars Technica: A textual analysis of the summaries of half a million grant applications has determined that millions of dollars may have been awarded to duplicate proposals by the same researchers. Harold Garner of the Virginia Bioinformatics Institute and his colleagues examined applications, dating back to 1985, to the National Institutes of Health, NSF, the Department of Energy, the Department of Defense, and the Susan G. Komen for the Cure foundation. Their text similarity tool found that, of the half a million grant summaries, 1300 were considered highly similar. After examining those manually, the team noticed that 167 pairs of applications had “suspicious” overlaps in their descriptions. Of the $200 million in total funding, almost $70 million was the result of overlapping grants. However, only one-third of the funding grants ran concurrently, which means that up to two-thirds could have consisted of proposals that were legitimately resubmitted once the original grant had been used up. Nevertheless, Garner believes that the team’s finding concerning project overlap is conservative because of limits on the text comparison tool. The team was able to use only summaries that were freely available and longer than 200 words. For a more comprehensive examination, extensive Freedom of Information Act requests would have to be filed with the government agencies. Garner says that agencies should work to eliminate overlapping grants because any duplication reduces the total number of research proposals that receive funding.