It is not easy for universities to investigate their own researchers. But as a case at Tohoku University shows, it is especially difficult in Japan — particularly when the scientist concerned is the university's president.

When concerns were raised in 2007 about irregularities in the research of Akihisa Inoue, who specializes in making metallic glasses, the university formed a committee to look into the matter. The committee determined that there was no case to answer and that the matter need not be investigated formally. But the decision satisfied few of his doubters, who criticized the committee for being heavily populated by university administrators whom Inoue had appointed to their posts (but not to the committee). The university says that there was no conflict of interest.

Four years later, and with claims of problems in Inoue's papers accumulating, the university formed another committee. This time it got an eminent outsider, Akito Arima, former president of Tokyo University, parliamentarian and minister of education, to lead. The committee focused on retractions of six papers on which Inoue had been an author.

The committee concluded that the duplicate submissions that led to the retractions were an example of poor, but somewhat accepted, practice in the field of materials science. Duplicate publications, the report said, were bad, but “different from practices such as FFP [falsification, fabrication and plagiarism], which clearly have a negative impact on research.” In a formal statement, Inoue apologized and said that the university would quickly implement the report's recommendations.

The judgement was surprisingly light. Six retractions is a serious matter, although they may simply indicate that Inoue — who has more than 2,500 papers to his name — was trying to move too fast. As Inoue has told Nature, his laboratory did not always check its papers to ensure that its research was cited properly.

Then, in April 2011, the Japan Science and Technology Agency (JST) formed a committee to find out whether the retractions, corrections and other allegations had undermined the results of a project it had funded between 1997 and 2002. It found no reason to question the main findings, but concluded that 5 of 27 high-impact papers were duplicated in other papers by Inoue's laboratory. But neither of these latest committees, as both acknowledge, investigated the more serious allegations against Inoue concerning irreproducibility of results (see Nature 470, 446–447; 2011). Inoue denies any manipulation of data, and there is no evidence that he has committed any scientific misconduct.

The JST, in accordance with its policies, has now requested that the university look into the allegations. The university says that it will do so, but will give no details about the composition of the committee, or about the investigation's mission, scope or expected date of completion.

Unlike countries such as the United States, Japan has no final arbiter for such issues. The Office of Research Integrity in Rockville, Maryland, for instance, as well as many funding agencies, can aggressively investigate allegations. Even if they don't satisfy everyone, their decisions are final, and for the most part fair.

The primary responsibility for investigating should always remain with the university, research institute or relevant scientific society. But Japan needs a procedure to oversee the committees that make these judgements. The authors of the JST report suggest that the Science Council of Japan (if given the necessary funding to repeat experiments) could fit this role. Such a body seems worthwhile, and it could very well have closed the arguments at Tohoku University that refuse to die even after nearly five years. Japan's policy-makers should use the situation at Tohoku to create a mechanism to handle situations such as these.