Yeebok Shu'in points us to a report claiming that US scientists are "significantly more likely to publish fake research." Of course, from the writeup, it's not actually clear if that's true. The study involved going through PubMed and looking at every paper that had been withdrawn between 2000 and 2010. There are two reasons why such papers are withdrawn: due to an error or due to fraud. The study did find that the largest number of retracted papers had someone from the US as their first author... but nowhere does it say what the percentage of the overall papers in PubMed are published by US authors. So it's hard to say, just from what's been reported, if US researchers are really more likely to withdraw papers. Honestly, for a scientific publication, the article is a bit weak in leaving out the details. It's entirely possible that the rest of the data is in the actual report, but Science Daily's writeup doesn't provide enough info.

The one stat it provides that is interesting is that 53% of the research withdrawn for fraud came from repeat offenders, while only 18% of the papers withdrawn for errors came from repeat offenders. Given the overall numbers, this actually suggests that fraud really isn't all that prevalent. A total of 243 papers overall were found as fraudulent over a ten year period, which represents about two per month. Perhaps that seems like a lot but given the number of scientific papers published that actually seems relatively low. Perhaps too low to read too much into the details.