Sunday, February 03, 2008

Charcuterie de science

“Gaming the system” is something that inevitably occurs whenever the quantitative outcomes matter (such as impact factors, usage statistics, number of articles on a CV, money, et cetera). Salami-slicing, the subject of a current thread on Liblicense-l, is just one of the ways of gaming the system. I’m not completely convinced that salami-slicing (or even auto-plagiarism, though that goes rather further, of course) is all that unethical. Or rather, that it is more unethical than, say, mutual citation cliques, boosting a journal’s impact factor by publishing review articles, improving usage statistics and impact factors by publishing with open access, et cetera. In the ‘ego-system’ of science, they’re all ways of gaming the system.

The gist of the discussion on Liblicense-l is that salami-slicing is bad. The motives of salami-slicing authors are presented as suspect, and there are strong suggestions that salami-slicing is bad for science.

As always, in discussions like this, the definition of what is salami-slicing is nor clear. In other words, how thin is a slice? Multiple publication of the same article is even brought under topic. But let’s take as a definition that salami-slicing is the practice of publishing a series of articles in each of which just one, or a small number, of a larger array of connected contributions to knowledge are presented, that could have been presented in one, more substantial article. For instance, "a inhibits b" (just one finding of a set that includes "a inhibits c, and f, and n, and p, and enhances the actions of h, of k, and of z"). Is it really bad for science if these findings are salami-sliced for publication?

I’m not sure, but mining the data from such articles with small units of information may conceivably be easier than mining them from articles that present the whole lot. Or it may make no difference. In certain disciplines, where automated analysis of articles is overtaking actual reading, it may even be desirable and should be the future of science publication. Salami-slicing may come close to publishing entries one by one in a database. If peer-reviewed entries in databases were to give their authors the same sort of acknowledgement as journal articles do, and ‘the system’ (those who decide on funding, promotion, tenure, et cetera) would formally recognise such contributions to science, would we still get upset about salami-slicing?

Gaming the system is human, and it happens in all walks of life, all the time. Usually it’s the result of flaws in the system. In science it is among the survival mechanisms, an evolutionary adaptation, if you wish, to the stresses of the ego-system, and it is done in all manner of guises. Isn’t freely disseminating peer-reviewed research results that are published in journals, by depositing in open repositories, while expecting the journals to continue to be paid for via subscriptions (i.e. via mechanisms intended for and dependent on exclusivity of dissemination), also a way of gaming the system? Ideas about correcting the flaw in the system that makes this particular form of gaming it possible range from stricter copyright enforcement (i.e. abolishing ‘green’ and not publishing if copyright isn’t transferred to the publisher), to open access publishing (i.e. securing payment for the services rendered, via article processing charges, subsidies, and the like). Obviously, the second idea has my preference.

Jan Velterop

No comments:

Post a Comment