Saturday, August 13, 2011

Why has there been such a huge increase in retractions of science publications?

Here's a finding in my working paper on retractions of papers by Indian authors: The 1990s (1991-2000) saw just 5 retractions; this number jumped dramatically to 69 in the noughties (2001-10). [If you feel like it, take a look at the list of retracted publications by Indian authors at PubMed. It has 77 papers right now, with two retractions from 1990 and one from 2011.]

Sure, India's scientific output has also grown, but not as fast. It went up from 31,500 in the '90s (1991-2000) to 103,000 in the noughties (2001-10) -- an increase by a factor of 3.25, which pales in comparison with the nearly 14-fold increase in the number of retractions.

According to Erik Hayden at The Atlantic Wire,

A lot more science findings have been found out to be wrong recently, and no one is sure why. Yesterday, The Wall Street Journal quantified the trend: since 2001, there's been a 15-fold increase in retractions in scientific journals.

Hayden goes on to discuss some of the possible reasons for this increase.

Coming back to India, we know that almost all cases of misconduct were due to plagiarism, and that a lot more Indian scientists resorted to this shady practice in the noughties than in the 90s. However, we also know that there has been a steep fall in the number of retractions (green line in the figure below), as well as in misconduct cases (red line) since 2007. I suspect two possible reasons: (a) widespread use of plagiarism detectors by journals has now made it difficult to publish plagiarized papers, and (b) this has (already) had a deterring effect on potential offenders.