Thursday, December 29, 2005

On Scientific Scandals



The Korean stem cell scandal is still growing:

The scandal surrounding disgraced South Korean stem cell researcher Hwang Woo Suk deepened today as an investigator told reporters in Seoul that none of the 11 tailor-made cell colonies Hwang claimed to have created earlier this year actually exist.
Korean news outlets also reported that the ongoing probe into one of the biggest scientific frauds in memory had broadened to embrace allegations that government officials -- concerned about the shame such revelations could bring upon their country -- may have attempted to bribe scientists who were considered potential whistleblowers.

Scientists are supposed to be ethical. Like clerics, aren't they, in some ways? Those of us who don't believe in religious ideas often have the same kind of blind belief in science. Thus these scandals that crop up once in a while are a good reminder not to take anything as a matter of blind faith.

The scientific system has many built-in checks for problems in someone's research project, but they are not perfect. Having to present papers in conferences and having to offer the research to unknown reviewers are not only fun ways of harassing other researchers; they do have a point in trying to keep them honest. But none of these safeguards is perfect as the Korean story reminds us. Then keep in mind that most political think tanks don't even use these basic safeguards. A good reason to read very critically indeed.

What research gets to be published can be biased even in the absence of any actual fraud. We have a tendency to focus on the unusual findings, on differences, on a new drug being successful as opposed to it being a failure. Published research therefore overstresses findings of a particular sort and understresses other types of findings.

In the field of gender research these biases mean that what we tend to hear about are new findings of differences between sexes, and especially those new findings which can be easily popularized. Research that doesn't find any differences between men and women will not even get printed in the obscure academic journals, let alone discussed all over the popular press.

All of this is good to keep in mind when leafing through scientific publications. Another nice check on our desire to take new findings at face value is to do some historical research. For example, go to the library and see what the popular scientific publications touted in, say, 1975 as absolute truth. You will be very surprised.