Importance of Replication of Science vs Me-Too Publishing

A cornerstone of science is the importance of having lots of studies that examine the same phenomena, so that on pooling, we can be more confident about what the total evidence shows. We all know that. But editors know something else. We all have a ratio of papers that come in to papers we can ever publish. My journal runs at about 15%. So we make judgements about quality, originality, reader interest and (let's face it) editor interest when we pass papers into review or reject them.

In my area, we are well into an era of papers that go, "We banned smoking in bars in city/country X. We measured particles/salivary cotinine before & after the ban...and wait for this...there was much less of both after the ban!"

Wow! Who would have guessed!?

Early studies like this were important in providing evidence that the bans did what they were supposed to do. They were eagerly awaited in the context of widespread scepticism. Now each city/country that implements smoking bans likes to wave around local studies that mean more to local doubters. I understand that, as I understand the need for a weight of studies.

But what I feel like doing each time these me-too studies come in now, is to reject them hoping they will find a home down the journal food chain somewhere so that they can contribute to the pool of studies. It occurs to me that these are dilemmas that all editors must face in all subject areas. So what do you do? Are you ever frank about this with authors? I suppose it might depend on where you are in the food chain...

Thoughts?

Simon Chapman
________________________
Maybe I am missing something, but it seems straight forward for me, at least in theory. The devil of course is always in the details.

If there is adequate information, ie when the existing studies are aggregated, preferably via a meta-analysis, the key parameters can be estimated with adequate precision over the range of relevant conditions, then it does not seem to make sense to continue to conduct research on the topic. It sounds like this involves human subject research. If so, it would actually be unethical to continue to research the topic because putting subjects at risk (even minimal risk) for no reason is unethical. It would also be a waste of resources.

On the other hand, if more information would be useful, eg more precise estimates of the parameters under relevant conditions that have not been adequately studied, then additional studies are justified. Whether or not you choose to publish them is up to you as the editor, but they seem appropriate. It would also seem that authors in a situation like this, where there already are many studies addressing the topic, have an obligation to adequately research the literature and clearly document why their study is needed given what has already been done.

I don't know what to think about the "political" aspects of local studies.

I've just told the authors the topic in my view has been adequately researched and would unlikely be of interest to our readers.

Anyway, my thoughts.

Dave Solomon
________________________
I agree with all that has been said, especially the comments from Dave Solomon. There is, however, one wrinkle that Simon (and the rest of us) may need to keep in mind: Although I am not sure I can always detect it, I believe there is a subtle but important difference between a truly 'me-too' paper and one that represents justified replication because some key circumstances are sufficiently different. Or indeed, because having replicated the results are unexpected. That said, when I edited Injury Prevention a common dilemma analogous to Simon's 'no surprise' scenario was what to do with papers that were essentially descriptive, eg of injuries involving a particular age group seen in an ED in a country that was not well represented in the literature. More often than not, the findings were highly similar to all other such papers, and we would reject. But I often agonized about doing so... I wonder what others have done.

Barry Pless

Editor Emeritus, Injury Prevention
________________________
Another reason to publish apparently "me-too" studies.

While repetitious publications are "boring", there is a need for many of them. In the field of health economics and health outcomes research, which is becoming more and more necessary for jurisdictional healthcare decision and policymaking, country-specific and region-specific data is essential. Even if the same sort of information is available for several other regions of the world, if it hasn't been published in the biomedical literature for the region of interest it has to be re-done (costly in time and resources), or extrapolations made from similar countries or regions (usually based on a number of assumptions that limit the validity of the study). Please keep that in mind before rejecting something that seems unnecessary because it has been done already in other countries or similar (but not identical) populations.

Donna Rindress
________________________
What is clear, and this may have been said by others already, is that the editor decides what s/he thinks readers would be interested in. It is a judgment that editors are making all the time, but one that authors sometimes do not seem to appreciate. The fact that one journal has rejected your manuscript does not always mean its bad research. Someone else may be interested. Some papers are better in national journals than in internationals journals of course.

Adamson Muula
________________________
In theory, it would be possible for journals to send such papers out for peer review and publish them in a special 'replications' section of the journal that is on-line only. Often, I would expect them to be short reports. The titles, authors, and topics could be listed in the main print journal. The rest of the publication would be on-line to reduce cost.

The hard part for me as an editor is still—how much replication is too much? There is a difference in a piece that provides a study of a specific population that extends generalization, but not our understanding of a particular phenomenon (as described below) and one that is a solid, but more or less trivial replication with just one more replication of a population we know a lot about.

Nancy Darling