Department of Materials Engineering, IISc Bangalore
In November 2010, Steen [1] caused a stir by asserting that “American scientists are significantly more prone to engage in data fabrication or falsification than scientists from other countries.” His assertions were contested strongly by O’Hara [2] who reanalyzed Steen’s data to show that, when scaled by the overall scientific output, the fraud rate for US scientists -- at 4.5 per 100,000 papers -- is quite close to the world average of about 4 papers per 100,000. More importantly, O’Hara also showed that the fraud rate is much higher at 18 for Indian scientists; it is this result that prompted my study in which I focus on 69 papers by Indian authors published during the 10-year period of 2001-2010, only to be retracted later.
First, plagiarism is overwhelmingly the primary mode of misconduct in India. Of the 69 retracted papers, the retraction of 45 could be traced to some form of scientific misconduct -- plagiarism (of both text and data) and self-plagiarism accounted for 26 and 18, respectively. Only one paper was retracted due to what might amount to falsification. Second, at 44 per 100,000 papers [3], the India’s misconduct-related retraction rate is far higher than the world average for all retractions (due to misconduct as well as genuine errors) of about 17. And, third, this rate could be said to have accelerated during the decade -- while it was 34 during the first half, it rose to 48 in the second half.
In this study, 9 retractions were due to genuine errors -- including three on the editorial side! For the remaining 15 retractions, a reason could not be assigned primarily due to two reasons: (a) the notices provide little information (or, none at all!) about the cause of retraction, and (b) the notices are behind a subscription wall. We must question the ethics of journal publishers and editors who resort to such practices.
[1] R.G. Steen, “Retractions in the scientific literature: do authors deliberately commit research fraud?” J. Med. Ethics (2010). doi:10.1136/jme.2010.038125
[2] R. O’Hara, “Rates of scientific fraud retractions” http://blogs.nature.com/boboh/2010/11/17/rates-of-scientific-fraud
[3] The difference between Steen’s numbers and mine is due to the classification scheme: I count all plagiarism as ‘misconduct’; Steen counts only data plagiarism as ‘fraud’, and classifies self-plagiarism and text plagiarism as ‘error’.