New Scientific Journal To Publish "Discrete Observations Rather Than Complete Stories" (sciencemag.org)

sciencehabit writes: Is the pressure to publish tempting scientists to improperly tweak their findings in order to create more cohesive stories? If researchers could report just the one finding they felt comfortable with, perhaps there would be no need to be dishonest. That thinking has spurred the creation of a new scientific journal, Matters. The open-access publication aims to boost integrity and speed the communication of science by allowing researchers to publish discrete observations rather than complete stories. "Observations, not stories, are the pillars of good science," the journal's editors write on Matters' website. "Today's journals however, favor story-telling over observations, and congruency over complexity Moreover, incentives associated with publishing in high-impact journals lead to loss of scientifically and ethically sound observations that do not fit the storyline, and in some unfortunate cases also to fraudulence."

Comments

Will this affect perception of research?

[Transist]

I think a concerning matter is that journalists (not science journals necessarily) also destroy the credibility of science by taking these observations ("according to a recent study...") and running with the "results" as news. A recent one that comes to mind is that researchers noticed that the diabetes medication Metformin seemed to have effects on life expectancy. Of course news outlets are currently running with the story that we might have found the miracle anti-aging pill. You can turn up a bunch of articles by googling the drug. It's usually later found that the claims are hugely inflated by the media and further research really goes nowhere. I suspect that the fatigue of constantly hearing these kind of false-hope and misleading reporting articles might hurt the image of legitimate scientific research. I wonder if this will have an effect on this issue. I suspect researchers may be complicit in providing journalists with these stories that they love to run with. Keeping that kind of speculation to a minimum might help.

And what's the impact factor?

[TheRaven64]

(Speaking from the perspective of a UK academic, may vary between countries) There is no pressure to publish, as an abstraction. There is pressure to demonstrate impact. The easiest way to demonstrate impact is to publish in top-tier publications. Publishing in a new journal or conference is always a big gamble - if the journal does well later then you may retroactively benefit from a later assessment of its impact, but typically it's in the noise of all of the spammy journals.

Re:And what's the impact factor?

[chihowa]

Impact factor is determined from the number of citations to a journal's articles, so a journal that hasn't published anything has no definable impact factor. It's not zero or low, it's undefined and referencing it is meaningless.

I would expect that a journal that let people publish observations without requiring an accompanying narrative could acquire a decent number of citations, even if the overall impact factor is low. Demonstrating impact through the proxy of a journal's impact factor is just lazy accounting by management types. It's easy enough to count actual citations to a publication to determine an author's impact (even if it can be gamed). Most of the papers that I actually read and cite are not in the ultra-high-impact journals. Most of a fantastically high impact researcher's publications are going to be in a field's bread and butter journal.

Re:honesty...

[NostalgiaForInfinity]

Scientist speaking here. One finding in no finding. It's luck or mistake. If there's just one "finding" you're "comfortable with", it's not publication you should think about, it's changing what you do and how you do it.

Whether it's one finding supporting your theory or one hundred findings really makes no statistical difference when your approach is to keep doing experiments until you get the results you want. And, sadly, that is what academics generally do: they vary experimental conditions, parameters, samples, and approaches until they get the results they expect and their peers are likely to accept for publication. Statistically, "one finding" supporting a theory can actually be a lot more meaning full than many findings supporting a theory, if you got your "one finding" on the first try without any selection.