Sunday 21 April 2013

Coelacanths and Confirmation Bias

I recently read an interesting blog by P Z Myers entitled Coelacanths are unexceptional products of evolution - which discussed why it was inappropriate to call them "living fossils" which were "slowly evolving". It included the following example showing confirmation bias in the scientist researching this interesting fish:

So why is this claim persisting in the literature? The authors of the BioEssays article made an interesting, and troubling analysis: it depends on the authors’ theoretical priors. They examined 12 relevant papers on coelacanth genes published since 2010, and discovered a correlation: if the paper uncritically assumed the “living fossil” hypothesis (which I’ve told you is bunk), the results in 4 out of 5 cases concluded that the genome was “slowly evolving”; in 7 out of 7 cases in which the work was critical of the “living fossil” hypothesis or did not even acknowledge it, they found that coelacanth genes were evolving at a perfectly ordinary rate. 
Research does not occur in a theoretical vacuum. Still, it’s disturbing that somehow authors with an ill-formed hypothetical framework were able to do their research without noting data that contradicted their ideas. 

No comments:

Post a Comment