 |
|
| Misconduct.... | | Date Created: Jun 10, 2005, 12:39 PM |
A recent study found that about 33% of American biomedical scientists have engaged in questionable research practices. Considering that the entire scientific enterprise depends on the accurate reporting of research results, the findings of this study are disappointing, to say the least.
Richard Feynman once defined scientific integrity as "utter honesty." Unfortunately, this utter honesty seems to be giving way to utter greed. I don't want to overstate the case, those 33% of scientists were not, for the most part, guilty of the most agregious sins of scientific misconduct, but the statistics certainly point to a problem.
In discussions about misconduct there often seem to be two lines of thought: 1) scientific misconduct occurs, but it is rare and when it happens it's usually discovered, or 2) Scientific misconduct occurs frequently and there's no way to find out. Unfortunately, the merging together of the number of occurances and the discovery of misconduct clouds the issues, I think.
This study, based on self-reporting, basically eliminates the idea that scientific misconduct is rare. One-third of scientists admiting to misconduct isn't a "rare" event. So, let's move to the next question: is it usually discovered?
If you remember being taught the scientific method back in high school, you may have been taught that scientific discoveries must be repeated in order for them to become accepted by the scientific community. That is indeed the case, but it isn't quite as straightforward as it seems. First of all, no one sets out simply to repeat someone elses work, generally. The repetition that does take place is in the context of extending something someone reports in order to further your own research. What happens if a scientists can't repeat a published result? Does that mean that the first scientist lied? Probably not. Given the complexity of most scientific experiments, it is often extremely difficult to reproduce someone else's results exactly. Trust me on that one.
The problem is that most work never gets read. There are over 6000 scientific scholarly journals out there. This huge number of journals means that almost anyone can get almost any study published somewhere. Most people I know keep up with 3 or 4 journals in their field, and usually only the biggest and most prestigious journals. That means that someone could easily publish an article full of crap in some small journal, but the impact of such an aricle would be minimal because its unlikely to get read. Big important studies, however, usually get published in big important journals and likely get read by many people. Some of the results may get repeated by other scientists. If there were falsified data in there, it would likely get discovered eventually.
So my view is that most misconduct is probably not going to get discovered, because it'll get published somewhere small, but it probably didn't matter anyway. But, some certainly slips through. Famous examples would include Charles Dawson, Steven Breuning, and Cyril Burt among others.
So what's the big deal?
The big deal is that the media short circuits this scheme. Any idiot with an ax to grind can go out, find some journal with some statisics that support his/her cause from any crap journal out there. Suddenly armed with statistics, they can try to make their case. (How many times have you heard some nutjob start an argument with, "Studies show that...") I'm reminded of the idiot on CNN who cited bullsh*t "scientific research" purporting to show that children in gay households are 11 times more likely to be sexually abused.
Of course, even good science can get screwed up by people with an ax to grind, as we've recently learned. Heck, why worry about scientists lying when you've got the Bush administration doing it for you? |
|
|
|
|
