Trendy science findings more likely to be wrong

Should fashionable scientific findings be subject to a higher burden of proof? Yes, says a study that found links between the research popularity of certain proteins and the accuracy of reports about their behaviour.

Researchers have suggested two reasons why trendy research fields may attract spurious results. First, there are greater rewards for getting positive results, so there is a stronger incentive to massage data or ignore "outliers" – a value that varies greatly from the rest of the data. Second, because more groups test trendy hypotheses. This would lead to more negative results too, but the positive ones tend to get reported more.

Now biologist Thomas Pfeiffer at Harvard University has found that "popular" results are indeed less reliable – at least those regarding protein interactions in yeast. Such interactions are of huge interest because they identify links between genes and their function.

Pfeiffer scoured the literature for reports that one yeast protein interacts with another, and compared these to systematic measurements of these interactions. Claims involving extremely popular proteins were only half as likely to be confirmed as ones involving less glamorous ones.
"For some research fields, a higher burden of proof would certainly be appropriate," says Pfeiffer.

http://www.newscientist.com/article/dn17398-trendy-science-findings-