Research fraud maybe more widespread than thought

[Ace Pace]

Kinda depressing.

One of the things we assume when we talk about scientists and their research is that the work was carried out with integrity. If someone publishes a paper, we assume that they're being honest and accurate when they report their data. Sadly, that isn't always the case, and scientific misconduct is more widespread than anyone wants to admit. That's the conclusions of an article in this week's Nature, and it makes for provocative reading.

The article, from Sandra Titus of the Office of Research Integrity, Lawrence Rhoades, formerly of ORI, and James Wells from the University of Wisconsin, bases its claims on a survey of 4,298 NIH grant holders, conducted across more than 600 institutions. 2,212 of those scientists completed the survey, and they reported 201 instances of misconduct that occurred between 2003 and 2005, a figure far higher than the number of investigations conducted by the ORI, the body responsible for looking into such occurrences.

Instances of research misconduct spanned all levels of the scientific ziggurat, from lowly graduate students and postdocs to tenured faculty and chairs. Postdocs (24.9 percent) and Professors (21.9 percent) actually accounted for the highest percentages, and the most common type of misconduct was fabrication or falsification of data, responsible for nearly 60 percent of incidents, with plagiarism making up the bulk of the rest.

The thing is, if you work in a lab, this figure probably isn't massively surprising. I can certainly think of instances in the past when I've suspected someone's work was a little fishy, and I've seen instances when results have been fudged or images prettied up in Photoshop. High profile cases such as the one involving a South Korean stem cell scientist prompted some journals to begin checking for image manipulation, but that won't catch someone who intentionally excluded inconvenient outliers from a sample, or just made up numbers for an assay.

Titus and her co-authors make several recommendations, including the greater protection of whistleblowers, creating a zero-tolerance environment for research fraud, and better mentoring of young scientists. That last item is a topic I find myself returning to again and again.

Extrapolating out from their data, Titus and colleagues suggest that, out of the 155,000 researchers being funded via the NIH, there could be at least 1,000 unreported cases of research fraud occurring each year in the biomedical field alone, with another 1,500 cases being made public. That's several orders of magnitude more than the 24 cases investigated by ORI. Hopefully as a field, we can start paying more attention to this issue.

Nature, 2008. DOI: 10.1038/453980a

[Mayabird]

High stress environment. Lack of results can mean someone loses his/her funding, which can mean losing their job. The people with the money don't understand that results don't magically come up if you throw money at it, that it's a risk, an investment in something that might not pan out. And so, to keep themselves afloat, they fudge the numbers a bit. Just a bit. It might get caught later, but it'll keep you afloat for a little while longer, and maybe the next thing actually will pan out, and maybe the idiots with the money will forget about the Next Big Thing that didn't work and will believe in your next thing long enough for you to survive until the next one...

[Alyrium Denryle]

High stress environment. Lack of results can mean someone loses his/her funding, which can mean losing their job. The people with the money don't understand that results don't magically come up if you throw money at it, that it's a risk, an investment in something that might not pan out. And so, to keep themselves afloat, they fudge the numbers a bit. Just a bit. It might get caught later, but it'll keep you afloat for a little while longer, and maybe the next thing actually will pan out, and maybe the idiots with the money will forget about the Next Big Thing that didn't work and will believe in your next thing long enough for you to survive until the next one...

Agreed. A big part of it is that the NSF and other funding organizations are themselves under funded, which makes the grants disgustingly competitive... WHich motivates cheating..

In other words, we need more funds