Research fraud among scientists in the US might be more common than previously thought and the rate at which it is reported "alarmingly" low, a new study suggests.
A report polled more than 2,200 federally financed scientists on instances of research misconduct such as fabricating data and plagiarism that they themselves had witnessed.
The results, published in the journal Nature, give a conservative estimate of 2,325 possible cases of research fraud each year.
The survey, which is said to be the largest and most systematic to date, found that only 1,350 of these incidents - 58 per cent of all the cases - were reported to university officials.
The report's authors included an official from the US Government-financed Office of Research Integrity, which sponsored the study. In a peer-reviewed commentary, the authors called the reporting rate "alarming" and said that it "calls into question the effectiveness of self-regulation".
In the US, institutions that get federal grants must investigate allegations of research misconduct themselves. Few cases, however, are referred to the ORI. Previous studies have suggested that up to four in ten whistleblowers are pressured by their institution to drop allegations.
Institutions were often keen to avoid the expense and embarrassment of a full-scale investigation, said the authors of the new study.
The authors expressed worry about the "failure to foster a culture of integrity" within the research community.
"Nearly one generation after the effort to reduce misconduct in science began, the responses by scientists suggest that falsified and fabricated research records, publications, dissertations and grant applications are much more prevalent than has been suspected to date ... We hope it will lead individuals and institutions to evaluate their commitment to research integrity," they said.
No comparable study has been carried out in this country, but the UK Research Integrity Office said a working group was considering one.
The US is generally thought to have a more entrenched "whistleblower culture" than elsewhere, which makes the findings on low reporting all the more surprising.
Andrew Stainthorpe, director of UKRIO, said: "There may be a greater willingness and confidence to bring forward concerns in the US than in other countries, although I would not single out the UK as being particularly backward.
"I think major cases of misconduct are equally likely to be reported on both sides of the Atlantic, but perhaps with matters of poor practice in research there may be greater awareness of previous investigations in the US, which gives whistleblowers the confidence to report their concerns."