Follow by Email

Thursday, March 15, 2012


Science should be about truth and truth alone; not what is expedient, not turf protection, not money, not politics, not prestige, not career advancement.  Of course, all we'd need to do, if we could, would be to quiz Galileo about his experiences with conducting science to know that other considerations have long entered into scientific publication.

A clue that bias remains and perhaps has increased in scientific publication is in the proliferation of 14 author papers in even ornithological journals.  It's hard to imagine that political factors are not at play for this many individuals to be claiming manuscript authorship. I had long thought that ornithology was immune to such phenomena, there being no money to speak of in ornithological research, but I've been forced to reassess this view because factors other than money can also be potent biasing agents.

Here's a personal example: ten years ago, I sent off for publication a manuscript about endangered species designation.  It concerned a field in which I had worked since the early 1970s, so it was one in which I had significant experience.  I was fully aware that the topic was controversial and would likely elicit vigorous debate among peer reviewers, but when the reviews arrived their content was beyond this.  They bordered on if not crossed the line into hysteria.  The most telling of these was by a reviewer who wished the ms. rejected because his group had the same ideas and should be allowed to publish them first.  That group, I learned, was part of a new and well-connected national initiative on bird conservation.  I had unknowingly tread on their toes.  To be sure, some of my findings were not politically correct- but again, science is about truth and not about political expediency or bureaucracies established about a particular paradigm.  Three years passed while I searched for a publishing venue.  Publication only came when I sent the ms. to a journal abroad. Notably, that paper has gone on to be one of the more frequently viewed in that journal's history.  Curious readers may view it here.

Of course, this is only a single instance from a single individual.  Any journal editor will tell you that things can on occasion go wrong with peer review.  The question is, to what degree does bias affect scientific publication in general?  The evidence from a growing body of literature is that bias is indeed pervasive across scientific disciplines.  To summarize some of the notable findings, papers with negative results, with women authors, with foreign authors, and with authors who are not celebrities or from prestigious institutions are at a significant disadvantage during editorial review.  One study suggests that up to 25% of all peer reviews are biased.  Typing bias in scientific publication or bias in peer review into a search engine will bring up numerous titles on this topic.

A solution to at least certain classes of publication bias is double blind peer reviewing, in which the identity of reviewers, authors and their institutions are hidden.  A number of scientific journals have moved to double blind reviewing although, curiously, American ornithological journals have not.  A substantial literature on the effects of double blind reviewing exists, and it tends to indicate significant improvement in acceptance rates of manuscripts from individuals otherwise discriminated against.

An issue related to bias is that of journal impact factor- a metric that purports to rank the relative importance of particular journals to scholars.  In order to maximize impact factor, journal editors are obliged to publish what they view as the most novel or otherwise important studies.  From a business perspective, such an approach make sense.  However, from a science perspective it is much more dubious.  How many sound papers with hard won data have we not seen simply because they weren't sufficiently sexy?  How might such selectivity bias the entire view of the state of a particular discipline?

Fortunately for us mere mortals working in the field of science, the internet and open access venues show great promise for ending the stranglehold on publishing long enjoyed by traditional scientific journals.  They also offer real possibilities for curbing peer review bias. Search engine visibility is now what really determines what papers get viewed, and there's more than one way to get that sort of visibility.  I'll write more on that last point later...

1 comment:

  1. This is also true for publishing papers on the impacts of investigators on birds and this bleeds over into conference symposiums chosen for the society meetings. We don't want to know there could be any effects of our work.